Site-specific Loading of an MCM Protein Complex in a DNA Replication Initiation Zone Upstream of the c-MYC Gene in the HeLa Cell Cycle*

The MCM proteins participate in an orderly association, beginning with the origin recognition complex, that culminates in the initiation of chromosomal DNA replication. Among these, MCM proteins 4, 6, and 7 constitute a subcomplex that reportedly possesses DNA helicase activity. Little is known about DNA sequences initially bound by these MCM proteins or about their cell cycle distribution in the chromatin. We have determined the locations of certain MCM and associated proteins by chromatin immunoprecipitation (ChIP) in a zone of initiation of DNA replication upstream of the c-MYC gene in the HeLa cell cycle. MCM7 and its clamp-loading partner Cdc6 are highly specifically colocalized by ChIP and re-ChIP in G1 and early S on a 198-bp segment located near the center of the initiation zone. ChIP and Re-ChIP colocalizes MCM7 and ORC1 to the same segment specifically in late G1. MCM proteins 6 and 7 can be coimmunoprecipitated throughout the cell cycle, whereas MCM4 is reduced in the complex in late S and G2, reappearing upon mitosis. MCM7 is not visualized by immunohistochemistry on metaphase chromosomes. MCM7 is recruited to multiple sites in chromatin in S and G2, at which time it is not detected with ORC1. The rate of dissemination is surprisingly slow and is unlikely to be simply attributed to progression with replication forks. Results indicate sequence-specific loading of MCM proteins onto DNA in late G1 followed by a recruitment to multiple sites in chromatin subsequent to replication.

Members of the minichromosome maintenance (MCM) 1 protein family were first characterized in Saccharomyces cerevisiae as essential for plasmid maintenance during the cell cycle (reviewed in Refs. [1][2][3]. Each of the six proteins MCM2, MCM3, MCM4, MCM5, MCM6, and MCM7 has a counterpart in yeast, and each is highly conserved throughout eukaryotes. Deletion of genes encoding any one of these six are lethal in S. cerevisiae or Schizosaccharomyces pombe (3,4), and these proteins have been ascribed as essential for initiation of DNA replication (5,6). These proteins have been co-isolated as a single complex, and complexes of various combinations of these proteins have been implicated in licensing DNA for initiation of replication in Xenopus egg extracts (7,8). Another MCM protein, MCM10, is reportedly required for phosphorylation of components of the MCM2-7 complex prior to initiation (9) and has also been implicated in the transition from initiation to replication (10). Preceding initiation of replication, and at a point near the end of mitosis (11), assembly of the MCM proteins into a prereplication complex (pre-RC) occurs dependent on coordinated function of the origin recognition complex (ORC) (12) and proteins Cdc6 (5) and Cdt1 (13). Initiation then depends upon activation by cyclin-dependent kinases (14) and the Dbf4-dependent kinase Cdc7 (15). Evidence derived from temperaturesensitive MCM protein degradation during S-phase (16) and from measurements of MCM protein distribution in chromatin (5) suggests that in S. cerevisiae, upon transition from initiation to elongation, the MCM2-7 complex distributes with replication forks. MCM presence at replication forks remains controversial, however, because no data directly colocalizes an MCM protein with fork proteins.
Although there is evidence that MCM proteins play an important role in DNA replication in mammalian cells, evidence regarding an essential role for any individual MCM protein or a functional role for any potential MCM complex remains elusive. Studies of distribution of MCM proteins in chromatin during the cell cycle indicate certain similarities with the yeast model regarding formation of a potential pre-replication complex. MCM proteins 2, 3, 4, and 5 have been reported to enter chromatin late in mitosis, and the MCM proteins alternate between soluble and chromatin-bound forms (17). A subset of the MCM2-7 complex, consisting of MCMs 4, 6, and 7 (hereafter referred to as "MCM4,6,7"), has been reported to possess DNA helicase activity (18,19). As isolated from HeLa cells, helicase activity of the MCM4,6,7 complex could be inhibited by Cdk-dependent phosphorylation of MCM4 (20,21). The MCM4,6,7 complex can be visualized by electron microscopy as a double heterotrimeric ring structure, similar to ring structures of other known helicases, including one of an MCM-like protein from archaebacteria (22). The helicase activity of an MCM4,6,7 complex, reconstituted from recombinant S. pombe proteins, reportedly requires the presence of forked DNA structures (18). Although results taken together suggest that the MCM4,6,7 complex could represent a major replicative helicase, there is little direct evidence for this. In particular, it is not known whether this complex selectively assembles at initiation zones of DNA replication in mammalian cells. Recent reports have dealt with cell cycle distribution of MCM proteins in chromatin in the hamster DHFR locus (23,24), but these did not specifically address MCM4,6,7 localization or requirements for pre-RC assembly proteins. Another recent paper has examined the cell cycle levels of ORC proteins during the HeLa cell cycle, finding that ORC1 undergoes dynamic changes while the other ORC proteins remain at constant levels (25). This study did not examine MCM protein locations. Recent reports indicate that certain MCM proteins (26,27), including MCM7 (28), may play a role in transcription. In this study we assess the localization and distribution of components of the MCM4,6,7 complex, their clamp-loading partner Cdc6, and their partner in the pre-RC complex, ORC1, in a well characterized initiation zone centered upstream of the c-MYC gene during the HeLa cell cycle. This zone was originally mapped in this laboratory (29) and subsequently refined in a variety of studies (30 -33). We report a highly selective localization of MCM proteins on chromosomal DNA in a particular segment of this zone together with Cdc6 and ORC1 in late G 1 -phase. During ensuing S-and G 2 -phases MCM7, an integral component of this complex, remains at the initial site while Cdc6 together with MCM7, but not ORC1, distribute to distant sites, indicating a recruitment of additional MCM proteins to chromatin during and after replication.

EXPERIMENTAL PROCEDURES
Cell Culture and Cell Cycle Synchrony-HeLa cells were maintained in spinner culture in Eagle's Minimal Essential Medium for suspension cultures (with L-glutamine and without calcium and magnesium, Cellgro Mediatech) supplemented with 10% (v/v) fetal bovine serum (Invitrogen) at 37°C in a humidified atmosphere of 5% CO 2 , 95% air. Cell cycle synchrony was achieved using a double thymidine block, exactly as described previously (34). By this method cells are blocked at the beginning of S-phase. Cell cycle progression was verified by FACS analysis (Fig. 1A), and timing of mitosis was verified by microscopic determination of the mitotic index. To specifically examine time points at mitosis the double thymidine block was employed in conjunction with the microtubule disrupter, nocodazole. Eight hours after release from the double thymidine block, the HeLa cells were treated with nocodazole (methyl-[5-(2-thienylcarbonyl)-1H-benzimidazol-2-yl] carbamate, 100 ng/ml, Sigma-Aldrich) for 4 h. This treatment blocks the already synchronized cells at the G 2 -M boundary (35). At that time the cells were released from nocodazole into new medium, and time points were taken for analysis at 1-h intervals.
Chromatin Immunoprecipitation (ChIP) and Re-precipitation (re-ChIP) to Detect Proteins Bound to DNA Sequences Upstream of the c-MYC Gene-Approximately 2 ϫ 10 7 HeLa cells were collected every 2 or 2.5 h after release from double thymidine block (34). Chromatin immunoprecipitation was done using a modification of the procedure of Kuo and Allis (36). Cells were incubated for 15 min in medium containing 1% formaldehyde at room temperature, and cross-linking was quenched by adding glycine to 125 mM. Cells were washed with ice-cold Tris-buffered saline (150 mM NaCl, 20 mM Tris-HCl, pH 7.6) three times. Cell pellets were resuspended in 500 l of lysis buffer (0.1% deoxycholic acid, 1 mM EDTA, 50 mM HEPES, pH 7.5, 140 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml aprotinin), and cells were disrupted on ice with 20 strokes of a Dounce homogenizer. Sonication was performed by pulsing three times for 15 s, incubating on ice for 2 min between each pulse, followed by centrifugation at 15,000 ϫ g for 5 min to remove cell debris. Gel electrophoresis indicates that a substantial fraction of DNA fragments at this stage are from 0.3 to 3.0 kb in length. Centrifugation was then performed for another 15 min to obtain cleared cell lysate. ChIP was performed on the cell lysate by overnight incubation at 4°C with 4 g of primary antibody followed by incubation with protein G Plus-agarose (Santa Cruz Biotechnology) for 2 h. The beads were rinsed four times sequentially with lysis buffer 500 (0.1% deoxycholic acid, 1 mM EDTA, 50 mM HEPES, pH 7.5, 500 mM NaCl, 1% Triton X-100), LiCl/detergent solution (0.5% deoxycholic acid, 1 mM EDTA, 250 mM LiCl, 0.5% Nonidet P-40, 10 mM Tris-HCl, pH 8.0), and TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, pH 8.0). The beads were then incubated for 10 min at 65°C with elution buffer (10 mM EDTA, 1% SDS, 50 mM Tris-HCl, pH 8.0) to elute the precipitates. For re-ChIP, re-ChIP elution buffer (10 mM EDTA, 50 mM Tris-HCl, pH 8.0, 0.7 M NaCl) was used to elute. The elution buffer was diluted to 1/5, and re-ChIP was performed using a second primary antibody as for the first ChIP. To reverse the cross-linking and purify the DNA, precipitates were incubated in a 65°C incubator overnight and then incubated with proteinase K solution (40 g/ml glycogen and 0.4 mg/ml proteinase K stock containing 50 mM Tris-HCl, pH 8.0, and 1 mM CaCl 2 in TE buffer, pH 7.6) for 2 h at 37°C. DNA samples were then purified using LiCl and phenol/chloroform/ isoamyl alcohol. DNA was precipitated by adding ethanol to 70%, and precipitates were washed with 75% ethanol, air-dried, and resuspended in TE buffer.
PCR Amplification of DNA Sequences Obtained by ChIP--The following primers were employed to amplify genomic sequences upstream of the human c-MYC gene. Locations of the primers are specified by distances from the HindIII site at 2325 nt upstream of exon I specified by P1 promoter. myc5F 198, 5Ј-AAGCTGAATTGTGCAGTGCATC-3Ј PCRs were performed as previously described (37). Reactions were carried out with 0.2 mM of each dNTP, 0.3 M primer/each, 10ϫ PCR buffer with Mg 2ϩ , and 1.75 unit of Enzyme mix of the High Fidelity PCR System (Roche Applied Science). Hot-start PCR reactions were carried out for either 34 or 42 cycles as indicated. PCR reactions were analyzed by electrophoresis on 2% agarose gels stained with SYBR Gold nucleic acid gel stain (Molecular Probes).
Immunohistochemistry of MCM7 in Tissue Sections-Paraffin sections were deparaffinized in xylene and rehydrated in 100% ethanol, 90% ethanol, and water. To quench endogenous peroxidase activity, slides were incubated for 20 min with 3% hydrogen peroxide. Antigen retrieval was performed by heating the sections in 0.1 M citrate buffer, pH 6.0, for 3 min. Sections were briefly rinsed with PBS. Slides were then incubated with primary antibody, mouse monoclonal anti-MCM7, in blocking solution (milk and goat serum in PBS) overnight at room temperature. Prior to addition of biotinylated secondary antibody and streptavidin peroxidase (BioGenex) complex for 5 min at 37°C, sections were rinsed in PBS three times and blocked in milk/PBS. Sections were then rinsed in 0.5% Triton X-100/PBS for 10 min. The peroxidase activity was visualized using diaminobenzidine tetrahydrochloride (DAB kit, Vector). Slides were counterstained with hematoxylin, dehydrated, and mounted.

Timing of Specific Localization of MCM7 in c-MYC Chromatin
During the HeLa Cell Cycle-We sought a method by which we could compare simultaneous localization of a given protein at multiple points over an extended genomic sequence region. Realtime PCR was not considered practical here, because, although it provides excellent quantitation of sequence representation in a chromatin sample obtained by ChIP, the method becomes cumbersome when applied to multiple sequence locations in multiple cell cycle time points. The high degree of synchrony achieved by the double thymidine block method is shown in Fig. 1A. The adaptation of chromatin immunoprecipitation (ChIP) detailed in Fig. 1B allows assessment of the presence of a given MCM protein at two locations upstream of the c-MYC gene and, simultaneously, on an extended region between the two locations throughout the cell cycle. Fig. 1C shows the positions of two primer sets used to amplify the three indicated PCR segments. The 198-bp segment is near the center of the replication initiation zone originally mapped upstream of c-MYC (29). It contains a prominent site of DNA bending and a purine-rich recognition element for the single-stranded DNA-binding protein, Pur␣ (40, A, kinetics of the synchronous cell cycle as measured by fluorescence-activated cell sorting. HeLa cells were synchronized using a double thymidine block (34). Samples were taken at the indicated hourly times after release from the second block. The y-axis indicates the cell number. The x-axis indicates DNA content as measured by propidium iodide fluorescence (PI Flu). The dotted lines indicate 1ϫ (G 1 ) and 2ϫ (G2) DNA contents. B, ChIP of genomic DNA sequences upstream of the c-MYC gene in the HeLa cell cycle using anti-MCM7 antibody. HeLa cells were synchronized using a double thymidine block as described under "Experimental Procedures." Cells are blocked at the beginning of S-phase and are released by transferring to thymidine-free medium. In this and all subsequent figures the 0 time point was taken immediately after removal of thymidine. Chromatin immunoprecipitation (ChIP) was performed on sonicated, cross-linked chromatin as described using anti-MCM7 antibody. After precipitation and reversal of cross-links, precipitated DNA was subjected to PCR using two primer sets to simultaneously generate three bands of 876, 350, and 198 bp. Mitosis occurs at ϳ11.5 h. M denotes DNA size markers. Four control lanes are included at the right. PCR reactions were performed using both primer sets as for all reactions. B denotes a precipitate obtained using protein G-agarose beads alone with no primary antibody. L denotes a HeLa lysate subjected to precipitation. The last two lanes are ChIP samples obtained using irrelevant antibodies. The precipitate for a-GFP was obtained with an antibody to the green fluorescent protein from Aequorea victoria. That for a-GST was obtained with an antibody to glutathione S-transferase from Schistosoma japonicum. C, positions of primer sets (arrows) relative to a HindIII site 2325 bp upstream of the c-MYC P1 transcriptional start site. Bend, site of demonstrated, constitutive DNA bending; PUR, site of known recognition element for protein Pur␣. A-T, site of an extensive A-T-rich element; FUSE, site of recognition element for far upstream element binding protein FBP. MycPE, various c-MYC gene promoter elements. 41). The 350-bp segment is located nearer to the c-MYC transcription start site and contains various promoter elements. The sequence between these two segments contains an extensive A-T-rich region with multiple potential DNA-unwinding sites and an element for binding another single-stranded DNA-binding protein, FBP (42,43). The 876-bp segment includes the other two segments. It should be noted, however, that due to the presence of annealed primers in the amplifying region, the 876-bp segment is not amplified as efficiently as either the 198 or 350-bp segments.
We hypothesized that, because the 198-bp segment is very near the center of the reported zone of initiation, proteins known to be involved in formation of the pre-initiation complex could be located there after commitment to replication, in late G 1 , and prior to the onset of replication, in S-phase. The cell cycle kinetics of the double thymidine block-synchronized HeLa cells has been thoroughly documented in previous publications (34). We initially focused on MCM7, because it is an essential initiation component in yeast and a component of a potential helicase complex in mammalian cells. It can be seen in Fig. 1B that the presence of MCM7 differs among the three indicated PCR segments during the cell cycle. Anti-MCM7 highly selectively precipitates sequences from the 198-bp segment at 2 h (early S-phase) and 16 h (mid-to-late G 1 -phase) after release from the double thymidine block. At all time points tested MCM7 can be detected on this 198-bp segment. The peak of mitosis in the cell cycle of Fig. 1B occurs at ϳ11.5 h. Mitosis, however, occurs very quickly relative to the scale at which time points are harvested (34). Therefore, the presence of MCM7 on the 198-bp segment at the 10-and 12-h points may reflect the presence of cells both pre-and post-mitosis. The detection of the 350-and 876-bp segments at 12 and 14 h suggests that some MCM7 may be present after mitosis, although the presence of a small percentage of G 2 cells may contribute to this. Data to be presented later in this report indicate that there is actually a short window of time at mitosis when MCM7 is not present on the chromatin.
Several aspects of the data in Fig. 1B require comment. The 0-time point was taken immediately after removal of thymidine and, thus, represents the status of the cells during blockage. Cells taken in the presence of thymidine yield similar results. At 2 and 16 h after release, when localization of MCM7 on chromatin is highly specific for the 198-bp segment, there is only slight presence of the 876-bp band. Because the larger segment includes the sequence of the smaller segment, one might expect the larger segment to be represented in the PCR products at least equally to that of the smaller one. Its lower representation reflects in part the fact that the 876-bp segment is less effectively amplified than are the two smaller segments due to steric hindrance from primers for the smaller segments. Clearly these data rely on certain controls that would attest to the quantitation of the immunoprecipitation of DNA sequences and the PCR method used to detect them. Certain of these controls, presented at the right in Fig. 1B, show that no chromatin segments are precipitated with beads alone (B), from lysate with no additions (L), or with irrelevant antibodies (a-GFP, a-GST). Additional controls are required to rule out the possibility that the size of the PCR-amplified segments could affect their amplification efficiency due to the DNA shearing during preparation. These controls are presented later in this report. The data of Fig. 1B indicate a dynamically changing distribution of MCM7 during the cell cycle, beginning with a selective localization in late G 1 and early S. Note that even as MCM7 distributes to distant sites in late S and G 2 , it remains at its initial site on the 198-bp segment. Details of this observation will be considered under "Discussion." ChIP of DNA Segments at Different Distances from the Initiation Zone Reveal a Dissemination of MCM7 throughout Sand G 2 -phases-If the presence of MCM7 is expanding outward from an initial assembly point during ongoing replication, then PCR segments that are the same size as the initial segment, but distally located, should yield results similar to those of the 350-and 876-bp segments when analyzed over the cell cycle, and the appearance of MCM7 in the more proximal segments should occur earlier than in the more distal segments. Fig. 2 shows that there is a dissemination of MCM7 to distal chromatin sites, but that dissemination is not a simple progression from a central point. Five segments have been used (a-e), all of which are at different sites from the segments used for Fig. 1. The positions of these segments are indicated in the two maps at the bottom of Fig. 2. It appears that there is increasing distribution to the segments a-d during S and G 2 , similar to those seen with the 350-and 876-bp segments of Fig. 1. MCM7 appears, however, in segments b and d earlier than it does in segment a, which is much closer to the 198-bp site of initial appearance. The segments closest to the initiation zone center, a-c, show the sharpest peaks of MCM7 location, from 6 to 10 h. Segments b and c are each about 4 kb from the 198-bp segment of Fig. 1, and their patterns are quite similar to that of the 350-bp segment of Fig. 1. This similar peaking may be anticipated if there is outward progression from a central segment, but the timing of MCM7 appearance is much slower than expected. Mammalian replication forks are believed to progress at Ͼ100 nt/s. Aspects of the rates of distribution of MCM7 to these segments will be considered in the discussion. Segment e shows a pattern similar to that of the 198-bp segment in Fig. 1. This segment is nearly 60 kb away from the center of the 198-bp segment and may be in or near another replicon. The patterns observed with segments a-d indicate that there is dissemination of MCM7 to multiple chromatin sites during DNA replication. If this dissemination involves progression, however, that must be coupled with an alternate, and slower, form of recruitment.
Double ChIP Reveals That MCM7 and Cdc6 Are Present Together on the 198-bp c-MYC DNA Segment Specifically at Cell Cycle Times of Initiation of Replication-The notion that MCM7 is specifically localized on DNA at initiation and is then distributed during ensuing replication was further examined in an experiment employing ChIP and re-ChIP. Cdc6 has been implicated in the loading of MCM proteins onto DNA to form the pre-initiation complex (5,11). We hypothesized that if Cdc6 is involved in loading MCM7 onto chromatin, then these two proteins would be specifically localized together at times of initiation. ChIP and re-ChIP offer an ideal means to test this. First ChIP was performed using anti-Cdc6 antibody. Then, prior to reversal of protein-DNA cross-linking, the chromatin fragments were subjected to re-precipitation using anti-MCM7 antibody. During subsequent PCR only those DNA sequences would be amplified that are simultaneously bound to both proteins. It can be seen in Fig. 3A that Cdc6 and MCM7 are highly specifically colocalized to the same DNA sequences in the 198-bp c-MYC segment at early S-phase (2.5 h) and midto-late G 1 -phase (17.5 h). Subsequently, throughout S and G 2 , these proteins are distributed to sequences in the 350-bp segment. Distribution to the 876-bp segment (not shown) essentially parallels that to the 350-bp segment. Although Cdc6 and MCM7 can be detected together on DNA at all cell cycle points tested, their highly specific colocalization to the 198-bp band occurs before and just after the onset of S-phase.
A number of controls are required to address questions arising from the data presented. For example, is it possible that sequences representing one of the segments amplified are over-represented in the sonicated chromatin starting material and are thus preferentially immunoprecipitated? Furthermore, is it possible that, in the multiple-segment PCR reaction performed, one of the segments is preferentially amplified over the others depending on the amount of template DNA present? Controls in Fig. 3B address these questions. The left panel of Fig. 3B shows PCR bands from a sample representing primarily G 1phase. It can be seen that in chromatin immunoprecipitated by Cdc6 alone (IP Cdc6), sequences of the 350-bp segment are highly represented. After re-ChIP the fraction unbound by the anti-MCM7 antibody (ub MCM7) still contains a high representation of these 350-bp segment sequences. In contrast, the fraction immunoprecipitated by anti-MCM7 (re-IP MCM7) is highly enriched in the 198-bp segment sequences. Thus, the specific immunoprecipitation of the 198-bp segment sequences is not simply due to their excess representation in the total chromatin starting material. The right panel of Fig. 3B presents a concentration course of input template DNA over a dilution range from 10 Ϫ1 -through 10 Ϫ6 -fold. It can be seen that at no concentration is the amplification of the 198-bp band preferential. Therefore, the specific detection of the 198-bp segment sequences in immunoprecipitated chromatin is not due artifactually to its preferential amplification.
Following the observation that Cdc6 and MCM7 are highly specifically colocalized on DNA in G 1 and early S, and that MCM7 is itself specifically localized, we proceeded to examine Cdc6 alone. Previous studies have indicated that levels of Cdc6 are particularly low in G 1 of the HeLa cell cycle (17). We also found that to be true (Fig. 4, 2.5-and 15-h lanes), an intriguing result given that Cdc6 and MCM7 are maximally colocalized when Cdc6 levels are low. Compare, for example, the 2.5-h lanes in Figs. 3A and 4. Whereas Cdc6 and MCM7 are highly colocalized together on the 198-bp band at 2.5 h (Fig. 3A), Cdc6 levels alone on this band are not particularly high (Fig. 4). The cell cycle timing of Cdc6 in chromatin is somewhat different from that of MCM7. From 12.5 to 15 h Cdc6 alone is at a low ebb on the 350-bp segment, while at these times MCM7 alone (Fig. 1B) and MCM7 together with Cdc6 (Fig. 3A) are relatively high on that segment. One may ask why the double ChIP experiment of Fig. 3 looks so different from the single ChIP experiments of Figs. 1 and 4. The answer is that the double-ChIP actually detects the simultaneous presence of two proteins on a single segment of DNA. A band may appear intense when ChIP is done with either MCM7 (Fig. 1) or with Cdc6 (Fig. 4), but the two proteins may not be on that segment simultaneously. Despite the wide distribution of Cdc6 in chromatin over much of the cell cycle, its copresence with MCM7 at the same DNA site is an event triggered with great precision in G 1 and early S.
Cell Cycle Disposition of an MCM4,6,7 Complex- Fig. 5A shows that anti-MCM7 coimmunoprecipitates MCM4, and Fig.  5B shows that anti-MCM4 and anti-MCM6 each coimmunoprecipitates MCM7. If these proteins are complexed after synthesis, then one might expect them to be loaded onto chromatin together. If that is the case, the complex should have the same site specificity of location in late G 1 as the complex between MCM7 and Cdc6. We tested this hypothesis using ChIP, and the result is shown in Fig. 5C. Lane HL shows the three c-MYC DNA segments amplified from a sonicated G 1 -phase HeLa lysate prepared as for ChIP. Following precipitation with agarose-anti-MCM7 beads, DNA from the unbound fraction is labeled ub MCM7, and DNA from the immunoprecipitated fraction is labeled IP MCM7. It can be seen that the input HL and ub MCM7 fractions are similar in DNA content. In contrast, the IP MCM7 fraction shows considerable specificity for the 198-bp segment, as was previously documented in Fig. 1B. Finally, the IP MCM7 chromatin fraction was subjected to re-ChIP with anti-MCM4. It can be seen in the re-IP MCM4 lane that copresence of MCM proteins 4 and 7 in G 1 is highly specific for the 198-bp segment. Because this is nearly identical to the double ChIP of MCM7 and Cdc6 seen at 2.5 and 17.5 h in Fig. 3A, it is likely that the complex containing these proteins is loaded onto chromatin in G 1 -phase.
Presence of MCM4 in the MCM4,6,7 Complex Is Diminished in S and G 2 and Re-established during Cell Division-Nocodazole synchrony was employed to examine the nature of the MCM4,6,7 complex during mitosis. For these studies HeLa cells were synchronized by double thymidine block as usual and allowed to progress to G 2 (8 h), at which time nocodazole was added to the medium. This microtubule disruptor blocks the cells prior to chromosome condensation and alignment of the metaphase plate. Upon removal of nocodazole from the medium, at 12 h, microtubule assembly is re-established, and mitosis can commence. The graph at the bottom of Fig. 6 shows that cell division takes place from 2 to 4 h after release from nocodazole. Immunoprecipitation with anti-MCM6 was used to coimmunoprecipitate MCM proteins 4 and 7 from extracts of asynchronous cells and from cells at all time points shown in the graph. It can be seen that MCM6 and MCM7 remain constantly in a complex at all times. As S-phase commences upon release from the double thymidine block, MCM4 is diminished in the complex. As cell division proceeds, MCM4 levels increase in the complex. If the MCM protein complex is loaded onto specific sites of DNA in late G 1 -early S, as suggested by the double ChIP experiment of Fig. 3A, then the MCM4,6,7 complex is formed in cells prior to its actual loading.

. ChIP with anti-Cdc6 and re-ChIP with anti-MCM7 of sequences upstream of the c-MYC gene in the HeLa cell cycle.
ChIP was performed using anti-Cdc6 antibody as described under "Experimental Procedures." Prior to reversal of formaldehyde cross-linking, precipitates were washed, resuspended, and subjected to re-ChIP using anti-MCM7 antibody as described. PCR was then performed as described to analyze presence of three segments upstream of the c-MYC gene as described for Fig. 1B  HeLa cells, and at the peak of this process 29% of cells can actually be visualized as mitotic, a value generally considered high. Thus the cell cycle points taken around this time would have many cells in late G 2 or early G 1 . The nocodazole experiment of Fig. 6 shows that MCM proteins 6 and 7 are together throughout mitosis, but such experiments do not provide information regarding the presence of MCM proteins on the chromatin. Therefore, to determine whether an MCM protein complex is present on chromatin during mitosis, we have examined MCM7 staining directly in cells that can be visually identified as in metaphase. Tissues were obtained from patients with cervical (Fig. 7A) or ovarian (Fig. 7B) cancer, treated with anti-MCM7 antibody and with secondary antibody coupled to horseradish peroxidase for DAB staining (red-brown). Numerous tumor cells can be seen darkly staining for MCM7, which is largely nuclear. Cells indicated by arrows are in metaphase. The blue hematoxylin counterstain can be seen clearly on the condensed chromosomes of the mitotic cells. These chromosomes are nearly devoid of the MCM7 stain. In addition, the mitotic cells possess only slight, cytoplasmic brown staining, suggesting that MCM7 levels are lower overall in these cells MCM Proteins 4, 6, and 7 Each Decrease in Intracellular Level in Mitosis-The same nocodazole block used for Fig. 6 was also used for studies of intracellular levels of several MCM proteins, Cdc6 and ORC1, presented in Fig. 7C. For these studies the amounts of extract loaded in gel lanes were normalized to the number of cells. It can be seen that up until the nocodazole block MCM proteins 4, 6, and 7 are each present at significant levels. Upon release from blockage, however, each of these proteins is diminished in level. MCM4 is seen as several bands ranging upward from about 97 kDa, in keeping with observations that it is post-synthetically modified (21). It is notable that overall levels of MCM4 are high throughout S and G 2 (Fig. 7), times when MCM4 does not effectively coimmunoprecipitate with MCM6 (Fig. 6). The graph in Fig. 6 reveals that cell division occurs 2-4 h after release from nocodazole. This is in part due to the block by nocodazole early in mitosis, prior to chromosome condensation. After release from nocodazole, the cells must re-establish microtubule structure before proceeding with mitosis. MCM7 levels decline significantly 2-4 h after release, reaching a minimum as cell number becomes nearly double. This is consistent with the lower levels of MCM7 visualized in metaphase cells in Fig. 7 (A and B), because the time between metaphase and cell division is generally quite short. Taken together, the data of Fig. 7 (A-C) indicate that MCM7 is released from chromatin in the later stages of mitosis. At the times when levels of MCM proteins 4, 6, and 7 are lowest (Fig. 7C), the MCM4,6,7 complex is beginning to reconstitute (Fig. 6). Taken together with the chromatin localizations of Figs. 1B and 3A, the data of Fig. 7 indicate that there is a short window of time at mitosis during which MCM7 is absent from chromatin. Before and after this time, as at 10 and 12 h in Fig.  1B, MCM7 is on chromatin and broadly distributed.
It is interesting to compare intracellular levels of MCM proteins 4, 6, and 7 with those of MCM proteins 2 and 3 and with those of Cdc6 and ORC1. MCM3 levels are slightly diminished upon release from nocodazole and remain lower through cell division. In contrast, MCM2 levels are diminished in the presence of nocodazole, but they recover during cell division. Cdc6 levels are high during S-phase and decrease prior to nocodazole addition. They appear to recover slightly after release from nocodazole, and they are diminished again upon cell division. ORC1 varies dramatically during the cell cycle. Its levels are high in early S-phase but decrease to be virtually undetectable in late S and G 2 . They remain low in the presence of nocodazole, but they recover after release during cell division.
ORC1 Specifically Colocalizes with MCM7 to the 198-bp Segment in G 1 -Phase-To help assess the functional significance of specific loading of MCM proteins onto DNA at the G 1 /S boundary, a double ChIP experiment was performed using anti-MCM7 and anti-ORC1 antibodies. In various non-mammalian systems thus far analyzed, it has generally been found that the presence of an ORC complex on DNA precedes the loading of MCM proteins (3). It has recently been reported that, whereas other ORC proteins remain at constant levels during the cell cycle, ORC1 levels vary (25). Fig. 8 presents results of a cell cycle analysis in which ChIP was performed first with anti-MCM7 antibody and then with anti-ORC1 antibody. The two primer sets of Fig. 1 were used for PCR in this study. Fig. 8 shows great specificity of ORC1 colocalization with MCM7 in two categories: the two proteins are present together only in late G 1 -phase, and they are present together only on the 198-bp segment and not on the 350-bp segment. The two proteins begin to co-ChIP on the segment at 17.5 h after release from the double thymidine block, and they peak together at 20 h. Interestingly, the two proteins are not detected together at t ϭ 0, which represents cells blocked at the beginning of S-phase. Thus, the co-ChIP of ORC1 and MCM7 at 20 h most likely represents late G 1 -phase cells rather than S-phase cells, even though the synchrony of the HeLa cells is beginning to break down at this point. A comparison of Fig. 8 with Fig. 7 is informative. ORC1 is present in early-S-phase cells (2 h, Fig.  7), but it is not specifically associated with MCM7 on DNA at this time (Fig. 8). ORC1 begins to appear in the cell cycle upon cell division (Fig. 7), but it is only specifically associated with MCM7 on DNA several hours later (Fig. 8). Cdc6 and MCM7 co-ChIP together earlier in G 1 (Fig. 3) than do ORC1 and MCM7 (Fig. 8), but Cdc6 remains specifically associated with MCM7 in early S-phase (Fig. 3). These results indicate that Cdc6 does not completely displace ORC1 from the DNA prior to the loading of MCM7, but they also indicate that Cdc6 and MCM7 remain specifically associated with DNA after ORC1 has departed the complex.

Alterations in MCM Protein Associations in the HeLa Cell
Cycle-MCM proteins 4, 6, and 7 are part of a complex consisting of MCM proteins 2-7, which is essential for initiation of DNA replication in a variety of eukaryotes (2)(3)(4)6). The subcomplex of MCM proteins 4, 6, and 7 reportedly possesses DNA helicase activity (18 -20, 44, 45). Although this complex may represent the primary replicative helicase during S-phase, further evidence must be obtained to indicate that it acts exclusively so or that additional proteins are not required for such activity. In any case, MCMs 4, 6, and 7 act on DNA as a hexameric ring structure consisting of two heterotrimers (46), a structure characteristic of DNA helicases throughout evolution. In Fig. 5 we show that MCM7 is coimmunoprecipitated very effectively with anti-MCM4, and in Fig. 5C we show that double ChIP with antibodies to MCM7 and MCM4 reveals specific binding of these two proteins to the 198-bp DNA segment similar to that observed with double ChIP using antibodies to MCM7 and Cdc6. Fig. 6 reveals that MCM6 and MCM7 coimmunoprecipitate with each other at a near-constant level throughout the cell cycle. In contrast, MCM4 does not coimmunoprecipitate with these proteins in late S-and G 2 -phases ( Fig.  6), times when MCM4 intracellular levels remain high (Fig. 7). At these times MCM4 may participate in processes distinct from those of MCM6 and MCM7. Presently, however, we cannot rule out the possibility that extensive modifications of MCM4 may render it less accessible to antibody detection in late S-and G 2 -phases. We did not detect significant MCM7 on chromatin in metaphase (Fig. 7, A and B). Intracellular levels of MCMs 4, 6, and 7 remain significant after release from nocodazole through the early stages of mitosis, i.e. at times T1Ј and T2Ј in Fig. 7. Levels of each decrease upon cell division. Taken together, the immunohistochemical and immunoblotting data indicate that there is a time late in mitosis when MCM 4, 6, and 7 levels are greatly diminished and when MCM7 is not present on chromosomes. Levels of MCM proteins 2 and 3 are altered differently from MCM proteins 4, 6, and 7 during mitosis. MCM3 levels are reduced slightly as the cell number doubles. MCM2 levels are high in S-phase, remain so in G 2 , and are greatly reduced in the presence of nocodazole. MCM2 levels, in contrast to those of MCM proteins 4, 6, and 7, recover at times T ϭ 3Ј and T ϭ 4Ј.
MCM7 Localizes Specifically within an Initiation Zone Upstream of the c-MYC Gene in G 1 -Phase-It can be seen in Fig. 1 that at 2 and 16 h after release from the double thymidine block, times that represent early S-phase and G 1 -phase, respec-FIG. 6. Coimmunoprecipitation of MCM proteins 4, 6, and 7 in cells subject to nocodazole block. HeLa cells were synchronized by release from double thymidine block (DTB) at T ϭ 0 as for previous figures and as described under "Experimental Procedures." At 8 h after release (T8), the microtubule-disrupting agent nocodazole was added as described under "Experimental Procedures" to block chromosome condensation and thus ensuing mitosis. Four h later, at T12, cells were washed to remove nocodazole and placed in new medium. Times denoted with a prime symbol indicate hours after release from nocodazole block. Filled circles in the graph at the bottom indicate cell numbers corresponding to the gel lanes located directly above them. For the immunoblots shown, 3 g of anti-MCM6 antibody was used for all immunoprecipitations from cell extracts, and the amounts of protein loaded onto gels were not normalized for cell number. AS denotes asynchronous HeLa cells just prior to the thymidine block. T ϭ 0 DTB Rel denotes a point taken immediately after release from the double thymidine block. T8 Add Noc denotes the point at which nocodazole was added. T12 Rel Noc T ϭ 0Ј denotes the point at which nocodazole was removed.
tively, MCM7 is localized primarily to the 198-bp band. This high degree of specific segment isolation via ChIP probably reflects to a certain extent an efficiency of immunoprecipitation based on MCM protein to DNA ratio, the ratio being highest for the smaller, specific segment containing the site of localization. This would in part explain why the 876-bp band is less well represented at 2 and 16 h even though it includes the 198-bp band. The cell cycle experiment of Fig. 1B is internally controlled due to the simultaneous amplification of all three segments. At later time points in late S, G 2 , and early G 1 the 350-and 876-bp bands are very well represented relative to the 198-bp band. The following conclusion can thus be drawn: in G 1 and early S-phases MCM7 is specifically localized to a segment centered ϳ1.7 kb upstream of the c-MYC gene. In late S and G 2 it distributes outward from this segment to localize broadly in chromatin. An important feature of Fig. 1B is that, as MCM7 disseminates from its initial localization site, it does not actually leave this site. As can be seen in Fig. 1B, MCM7 is present throughout the cell cycle on its initial 198-bp localization site. Two possibilities can account for this observation. One is that  A and B). A, staining of a cervical cancer specimen with mouse monoclonal anti-MCM7, seen as red-brown, employed a diaminobenzidine system as described under "Experimental Procedures." Counterstaining (blue) was with hematoxylin. The micrograph was originally taken at ϫ400. B, specimen from an ovarian tumor. The micrograph was originally taken at ϫ200. Lines represent 10 m. C, levels of MCM proteins 2, 3, 4, 6, and 7 as well as Cdc6 and ORC1 in the HeLa cell cycle at time points before and throughout mitosis. HeLa cells in spinner culture were synchronized using a double thymidine block and blocked after 8 h at metaphase with nocodazole as described under "Experimental Procedures" and the legend to Fig. 6. Cells were from the same experiment as that in Fig. 6, and the graph at the bottom of Fig.  6 is equally applicable to this figure. Total cell lysate was loaded onto an SDS-polyacrylamide gel after normalizing for cell number, 0.4 ϫ 10 5 cells per lane. Immunoblot analysis was then performed using primary antibodies to the indicated proteins as shown.  Fig. 3. ChIP was performed using anti-MCM7 antibody as described under "Experimental Procedures." Prior to reversal of formaldehyde cross-linking, precipitates were washed, resuspended, and subjected to re-ChIP using anti-ORC1 antibody. PCR was then performed as described for Fig.  3 to analyze presence of 198-and 350-bp segments upstream of the c-MYC gene as described in Fig. 1C. as MCM7 distributes outward, new MCM7 molecules replace the ones previously located at the initial site. The second possibility is that MCM7 molecules do not migrate from their initial site and that new MCM7 molecules are recruited to effect distribution to multiple sites in chromatin. Our experiments help distinguish between these possibilities.
Timing of Dissemination of MCM7 Throughout the c-MYC Locus Is Consistent with a Slower Recruitment to Replicated DNA Rather than with Replication Fork Progression-In assessing the potential distribution of MCM proteins with replication forks, the overall rate of fork progression must be considered. In Fig. 1B it can be seen that MCM7 is present on the 198-bp segment within 2 h of release from the double thymidine block. In Fig. 2 it can be seen that MCM7 peaks on the 199-bp segment (segment a) about 500 bp upstream of the 198-bp segment, at 8 -10 h, whereas MCM7 peaks on the 219and 203-bp segments (segments b and c) about 4 kb upstream or downstream, at 4 -8 h. MCM7 is similarly disseminated to segment d. MCM7 is present on segment e throughout the cell cycle, although at 60 kb away, this segment may be closer to another replicon. Any potential fork pausing, which could explain this asymmetric timing of localization, is considered unlikely due to the great length of time needed to progress through these relatively short DNA lengths. By 8 h nearly all the HeLa cell DNA is replicated (Fig. 1A). Because eukaryotic DNA polymerases generally progress at Ͼ100 bp/s (47), these observed peaking times would not be consistent with simple progression with a replication fork. In addition, although MCM7 is distributed widely onto chromatin during S-phase, the rates of ChIP progression observed here are not consistent with progression during active DNA replication. Instead, the rates observed indicate an orderly but much slower recruitment to multiple sites on replicated DNA. These data do not rule out distribution of MCM7 with replication forks, which may occupy a percentage of MCM7 molecules. The data, however, suggest that the bulk of MCM7 molecules are recruited to multiple sites to participate in processes subsequent to DNA replication. These processes may involve functions of MCM proteins similar to those known to be involved in DNA synthesis, functions that could include unwinding at stalled forks or aspects of recombination/repair. MCM proteins may also participate in gene transcription (27,28). There is evidence that MCM7 plays a role in autoregulation of its own gene expression (28). A new member of the MCM protein family has now been discovered, MCM8, which exists exclusively in metazoan organisms (37,48). MCM8 evidently plays a protective role against development of neoplasia (37). It is interesting to speculate that demonstrated interaction of MCM8 with MCM4,6,7 (37) alters the function of the latter complex and may play a role in any function of the MCM complex late in the cell cycle.
A Fundamental Difference between ChIP of MCM7 and Double ChIP of MCM7 with Cdc6 during S-phase Sheds Light on Recruitment of MCM7 to Chromatin-A comparison of Fig. 1B with Fig. 3A is highly illuminating. In Fig. 1B MCM7 is specifically localized to the 198-bp band in G 1 and early S. In Fig.  3A MCM7 and Cdc6 together are also highly localized to the 198-bp band in G 1 (17.5 h) and early S (2.5 h). The difference is that MCM7 alone (Fig. 1B) remains on the 198-bp segment during late S and G 2 , whereas Cdc6 together with MCM7 (Fig.  3A) is minimally present on the 198-bp segment as the complex is distributed to the 350-bp segment in late S and G 2 . A logical explanation of the data is that MCM7 is recruited during Sphase to sites distal from the initial late-G 1 /early-S localization site. As MCM7 remains on the 198-bp segment during S and G 2 (Fig. 1B), Cdc6 dissociates from MCM7 at this location (Fig. 3A) and is together with MCM7 primarily at the distal 350-bp site (Fig. 3A). The role of Cdc6 is to load the MCM7, and presumably other MCM proteins, onto the chromatin at the distal sites. The data taken together suggest that new MCM molecules are recruited to sites distal from the initial, pre-replication loading site, i.e. the 198-bp segment, during S-phase and that Cdc6 is involved in loading the new MCM proteins onto the chromatin.
MCM7 Is Released from Chromosomes at a Late Stage in Mitosis- Fig. 1 does not reveal a large change in level of MCM7 on DNA from 10 to 12 h after release from the double thymidine block, a period during which mitosis occurs. Fig. 7 (A and B), however, visually show mitotic chromosomes with little or no MCM7 staining in cells with what appear to be lower levels of MCM7. Fig. 7C indicates that intracellular MCM7 levels remain high during nocodazole block and the early stages of mitosis and decline as cells divide. Taken together, the data indicate that MCM7 is removed from chromatin, and declines in level in cells, during a late stage of mitosis. Either late in mitosis or just after cell division an MCM4,6,7 complex can be detected by coimmunoprecipitation (Fig. 6). There is evidently a short window of time when MCM7 is diminished in chromatin, because it appears on DNA again in early G 1 (Fig. 1). It has been reported that in human cells an MCM protein complex forms on DNA late in mitosis (17). That could be consistent with the present findings if the short window of time during which MCM7 is absent does not extend to the end of mitosis. The appearance of MCM proteins on chromatin appears to precede the time at which MCM7 is specifically loaded onto DNA with Cdc6, which occurs later in G 1 and early S (Fig. 3). It would be of interest, therefore, to determine whether a complex, containing MCM proteins 4, 6, and 7, forms, with possibly other MCM proteins, prior to their loading onto DNA. Comparison of Figs. 1 and 3 suggests that MCM7 is present on DNA in a more diffuse fashion after mitosis and prior to its site-specific loading with Cdc6. Further work must be done to fully elucidate the potentially complicated loading process.
Specific Association of ORC1 and MCM7 with DNA in Late G 1 -phase-We report here dynamic changes in ORC1 levels in the HeLa cell cycle. Furthermore, we report that ORC1 and MCM7 simultaneously associate with a specific DNA segment in late G 1 -phase of the cell cycle. Our results may be compared with recent results of others working with mammalian systems. One group has reported that anti-MCM protein antibodies immunoprecipitate HeLa cell G 1 -phase chromatin enriched for DNA containing a replication origin. As we do, this group further reports that MCM proteins are distributed to multiple sites during S-phase (24). Our results are similar to those of a group that employed T16 to impose mitotic block in HeLa cells (25). In both cases ORC1 was at minimal levels in cells during such blockage. That group found that ORC1 was associated with ORC proteins 1-4 at the G 1 /S boundary, a time at which we find ORC1 associated with MCM7. Both studies are consistent with a key role for ORC1 in regulating the status of the pre-replication complex in the cell cycle.
Characteristics of the c-MYC Region to Which MCM Proteins Specifically Localize Prior to Replication-The 198-bp segment to which MCM7 specifically colocalizes, via ChIP, with ORC1, MCM4, and Cdc6 has certain distinguishing characteristics. It is near the center of an initiation zone of DNA replication originally mapped by PCR-based methods (29). Although the c-MYC locus may contain additional centers of bidirectional DNA replication, this particular zone center has been well characterized among mammalian systems. Segments centered either on or near this region reportedly support initiation in chromosomal DNA (29,31), in plasmids in transfected cells (30) and in a model in vitro system (32). The 198-bp segment con-tains multiple A-T-rich elements similar to yeast ARS consensus sequences. It also contains a region of potential z-DNA and a prominent site of intrinsic DNA bending (40). In addition it contains a PUR element, i.e. a recognition element for binding the single-stranded DNA-binding and DNA-unwinding protein Pur␣ (40). It is not known whether any DNA helical distortions would facilitate the loading of MCM proteins onto chromatin. Intriguingly, MCM7 has been found, using a yeast two-hybrid system, to associate with the retinoblastoma protein, Rb, an interaction that inhibits DNA replication (49). The hypophosphorylated form of Rb, the form predominant in early G 1 , also associates strongly with Pur␣ (50). Genetic inactivation in the mouse reveals that Pur␣ is essential for developmentally timed DNA replication in many tissues (51). The 198-bp segment is adjacent to an extensive A-T-rich region that could facilitate DNA unwinding. Deletions in the 198-bp region and in the adjacent A-T tract have recently been reported to inhibit chromosomal replicator activity in HeLa cells (33). The presence of ORC1, Cdc6, and MCM7 on the 198-bp segment specifically at times near the G 1 /S boundary is highly consistent with an important role for this DNA region, and for these proteins, in the initiation of chromosomal DNA replication.