Portal de Periódicos da CAPES

Putting It All Together: Building a Prereplicative Complex

1997; Cell Press; Volume: 91; Issue: 6 Linguagem: Inglês

10.1016/s0092-8674(00)80459-6

1097-4172

Carol S. Newlon,

Fungal and yeast genetics research

The initiation of DNA replication in eukaryotic chromosomes is best understood in the budding yeast Saccharomyces cerevisiae. In this organism, DNA replication initiates at replication origins spaced at intervals of approximately 40 kb whose activity depends upon cis-acting replicator sequences called autonomously replicating sequence (ARS) elements. Replicator activity depends upon a short DNA sequence, the ARS consensus sequence, which acts as the core of the recognition sequence for the origin recognition complex (ORC) (reviewed by15Newlon, C.S. (1996). In DNA Replication in Eukaryotic Cells, M.L. DePamphilis (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), 873–914.Google Scholar). ORC, a six-polypeptide complex, is essential for the initiation of DNA replication and appears to be the yeast replication initiator protein, the analog of E. coli dnaA and SV40 T antigen (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). The binding of ORC to ARS elements depends upon ATP, and it has recently been shown that ORC is an ATPase whose activity is modulated by sequence-dependent binding to DNA (12Klemm R.D Austin R.J Bell S.P Cell. 1997; 88: 493-502Abstract Full Text Full Text PDF PubMed Scopus (204) Google Scholar). cis-acting replicators have not been identified in metazoans, where replication appears to initiate in broad zones. However, ORC homologs have been identified in several metazoans and shown to be essential for DNA replication (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar), suggesting that the mechanism of eukaryotic DNA replication initiation is conserved. The chromatin structure of DNA replication origins, as revealed by DNaseI footprinting, changes during the cell cycle, alternating between the prereplicative complex (pre-RC) during G1 and the postreplicative complex (post-RC) during S, G2, and M. The post-RC pattern is very similar to the footprint generated by ORC binding in vitro, while the pre-RC is characterized by a more extensive region of protection and loss of a prominent DNaseI hypersensitive site characteristic of the ORC footprint (reviewed by2Diffley J.F Genes Dev. 1996; 10: 2819-2830Crossref PubMed Scopus (209) Google Scholar). Initiation of DNA replication requires the prior assembly of pre-RCs. In fact, the regulation of pre-RC assembly appears to play a key role in restricting DNA replication to one round per cell cycle. Activation of the cyclin-dependent kinase (CDK) that triggers DNA replication at pre-RCs blocks assembly of new pre-RCs (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Although a detailed discussion is beyond the scope of this review, the activation of the S phase–promoting CDK (Cdc28p together with a B-type cyclin) is regulated by ubiquitin-dependent proteolysis of an inhibitory protein, Sic1p, that binds to the cyclin B/CDK complex (reviewed by9Hoyt M.A Cell. 1997; 91: 149-151Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). Thus, a key issue in the initiation of DNA replication is to identify the proteins that are components of the pre-RC and to understand their functions. Both ORC and the Cdc6 protein (Cdc6p), which is essential for DNA replication and is required for the assembly and maintenance of the pre-RC, are excellent candidates for components of the pre-RC. Other candidates include products of genes known to be required for the initiation of DNA replication, including the minichromosome maintenance (MCM) proteins, Cdc45p, and the Cdc7p protein kinase and its regulatory subunit Dbf4p. The MCM proteins were first identified in a screen for mutations that caused defects in the maintenance of plasmids in yeast, i.e., mutants in which plasmids were unstable (reviewed by15Newlon, C.S. (1996). In DNA Replication in Eukaryotic Cells, M.L. DePamphilis (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), 873–914.Google Scholar). Six genes identified in this screen, MCM2–MCM7, encode a family of related proteins that are conserved among eukaryotes (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Members of this family share a 240–amino acid domain that contains the motifs characteristic of DNA-dependent ATPases. Adding to the interest in Cdc6p and the MCM family of proteins as components of the pre-RC are experiments in Xenopus egg extracts which demonstrated that association of MCM proteins with sperm chromatin, in a reaction that requires Xenopus ORC and Cdc6p, is necessary for the initiation of DNA replication (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Several recent papers report important progress in teasing apart the pre-RC and provide tantalizing suggestions for the roles of its components (1Aparicio O.M Weinstein D.M Bell S.P Cell. 1997; 91: 59-69Abstract Full Text Full Text PDF PubMed Scopus (630) Google Scholar, 3Donovan S Harwood J Drury L.S Diffley J.F.X Proc. Natl. Acad. Sci. USA. 1997; 94: 5611-5616Crossref PubMed Scopus (422) Google Scholar, 14Liang C Stillman B Genes Dev. 1997; 11: 3375-3386Crossref PubMed Scopus (315) Google Scholar; and19Tanaka T Knapp D Nasmyth K Cell. 1997; 90: 649-660Abstract Full Text Full Text PDF PubMed Scopus (430) Google Scholar). Two approaches were used to examine the association of candidate pre-RC components with DNA. The first approach, used by Aparicio et al. and Tanaka et al., was a chromatin immunoprecipitation (CHIP) assay developed by Grunstein and colleagues (8Hecht A Strahl B.S Grunstein M Nature. 1996; 383: 92-96Crossref PubMed Scopus (437) Google Scholar). Cells expressing epitope-tagged proteins of interest were treated with formaldehyde to cross-link protein–DNA and protein–protein complexes, and then chromatin was isolated. After the chromatin was sheared to fragments containing an average of 500 bp of DNA, antibody was added to precipitate the protein of interest together with material cross-linked to it. Protein–DNA cross-links were then reversed, and the released DNA was subjected to PCR analysis to monitor coprecipitation of specific DNA sequences. While only a small fraction (<0.5%) of any specific DNA sequence was coprecipitated and recovered by this procedure, this approach has the important advantage that interaction of the tagged protein with specific DNA sequences can be detected. The second approach, used by Donovan et al. and Liang and Stillman, was a chromatin binding assay in which spheroplasts were lysed with nonionic detergent and chromatin was partially purified by sedimentation through sorbitol gradients. The association of particular proteins with chromatin was then assessed by Western analysis. The advantage of this approach is that it can be used to examine the total distribution of a protein between soluble and chromatin-bound phases. The disadvantage is that it provides information only about chromatin association, with no indication of the DNA sequences involved. The first important finding, reported by all four groups, is that ORC is associated with chromatin throughout the cell cycle (see Figure 1). While the similarity of the post-RC footprint to the footprint generated by ORC in vitro has strongly suggested that ORC is bound to chromatin during S, G2, and M phases of the cell cycle, the absence from the pre-RC of the characteristic DNaseI hypersensitive sites caused by ORC binding has left open the question of whether ORC remains bound to replicators in the pre-RC. Liang and Stillman found that all six ORC subunits were associated with chromatin throughout the cell cycle. The groups using the CHIP assay found that Orc1p and Orc2p associated with replicators (ARS1, ARS305, and ARS306) but not with sequences that do not contain replicators (within the URA3 gene and at several positions between ARS elements on chromosome III). Moreover, the ORC interaction with replicators required an intact ORC-binding site and was inhibited by mutations in other ORC subunits. Interestingly, Aparicio et al. found that the efficiency of immunoprecipitation of replicators was lower during G1 than during other phases of the cell cycle, because the assembly of pre-RCs partially blocks the epitope used for immunoprecipitation. Incubation of a temperature-sensitive cdc6 mutant strain at the nonpermissive temperature, which causes disassembly of pre-RCs, restored the efficiency of immunoprecipitation by anti-ORC antibodies. These results provide strong support for the idea that ORC is a component of pre-RCs. However, the diminution of signal as a result of pre-RC complex assembly during G1 emphasizes a potential difficulty with the assay. If the structure of complexes containing the tagged protein varies along chromatin, then what appears to be association with specific sequences could reflect instead the relative accessibility of the epitope targeted for immunoprecipitation. In the case of ORC, both its low abundance and the conditions that unmask the epitope in G1 phase cells mitigate this concern. Given its critical role in the assembly and maintenance of the pre-RC, it was of interest to examine Cdc6p. Cdc6p is an unstable protein, present only in the G1 phase of the cell cycle, whose synthesis and degradation is tightly regulated (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Donovan et al. found that approximately one-third of the Cdc6p present in cells arrested in G1 was associated with chromatin. Using the CHIP assay, Tanaka et al. found an enrichment of replicator sequences in Cdc6p immunoprecipitates from G1 cells, as expected if Cdc6p is a component of pre-RCs. Importantly, the interaction of Cdc6p with ARS1 was abrogated by mutations in ARS1 that prevented ORC binding, suggesting that the association of Cdc6p with chromatin was through ARS elements and perhaps through ORC, with which Cdc6p is known to interact. In contrast to the results of Tanaka et al., Aparicio et al. failed to find enrichment of replicator sequences in immunoprecipitates using an HA-tagged Cdc6p. The most likely explanation of the discrepancy between the two reports is that the longer epitope tag used in most of the experiments by Tanaka et al. was differentially accessible. An alternative explanation is that the HA-tagged Cdc6p used by Aparicio et al. is less stable than the derivative used by Tanaka et al. Since initiation of replication at only a fraction of potential origins is sufficient for viability (16Newlon C.S Collins I Dershowitz A Deshpande A.M Greenfeder S.A Ong L.Y Theis J.F Cold Spring Harbor Symp.Quant. Biol. 1993; 58: 415-423Crossref PubMed Scopus (82) Google Scholar), it is possible that pre-RCs were assembled at a fraction of replicators too small to detect in the experiments of Aparicio et al. Taken together, these experiments suggest that Cdc6p is present in a substantial fraction of pre-RCs and is likely to be a component of pre-RCs, although the possibility that it is associated with pre-RCs only while it is loading other components cannot be excluded. Unlike the other proteins implicated to be components of the pre-RC, the abundance of Cdc6p is tightly controlled, and it is targeted for degradation at the G1/S boundary. The same machinery that recognizes and ubiquitinates the CDK inhibitor, Sic1p, to promote its degradation and activate the S phase CDK also targets Cdc6p for degradation (4Drury L.S Perkins G Diffley J.F.X EMBO J. 1997; 16: 5966-5976Crossref PubMed Scopus (239) Google Scholar). In Schizosaccharomyces pombe the degradation of the Cdc6p homolog, Cdc18p, is required to prevent multiple rounds of DNA replication in a single cell cycle, demonstrating the key role of this protein in the regulation of DNA replication (11Jallepalli P.V Brown G.W Muzi-Falconi M Tien D Kelly T.S Genes Dev. 1997; 11: 2767-2779Crossref PubMed Scopus (147) Google Scholar). Additional controls on the re-replication of chromosomal DNA must be present in S. cerevisiae since stable derivatives of Cdc6p do not cause overreplication (4Drury L.S Perkins G Diffley J.F.X EMBO J. 1997; 16: 5966-5976Crossref PubMed Scopus (239) Google Scholarreferences therein). The MCM proteins, which are present at relatively constant levels throughout the cell cycle (14Liang C Stillman B Genes Dev. 1997; 11: 3375-3386Crossref PubMed Scopus (315) Google Scholar, 20Young M Tye B.-K Mol. Cell. Biol. 1997; 8: 1587-1601Crossref Scopus (67) Google Scholar), associate with chromatin in a cell cycle–specific manner. Studies using the chromatin binding assay revealed that the MCM proteins associate with chromatin during G1 and are released during S phase (14Liang C Stillman B Genes Dev. 1997; 11: 3375-3386Crossref PubMed Scopus (315) Google Scholar, 20Young M Tye B.-K Mol. Cell. Biol. 1997; 8: 1587-1601Crossref Scopus (67) Google Scholar). This behavior reflects the behavior of MCM homologs in S. pombe, Xenopus, and mammalian cells, which also associate with chromatin in a cell cycle–specific manner (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar) and emphasizes again the conservation of the DNA replication initiation mechanism among eukaryotes. Using the CHIP assay, it was found that during G1 the two MCM proteins tested, Mcm4p and Mcm7p, associated specifically with replicators and not with other DNA sequences tested. The association was abrogated by mutations in ARS1 and by temperature-sensitive mutations in ORC1 and ORC5, the genes encoding two ORC subunits (1Aparicio O.M Weinstein D.M Bell S.P Cell. 1997; 91: 59-69Abstract Full Text Full Text PDF PubMed Scopus (630) Google Scholar, 19Tanaka T Knapp D Nasmyth K Cell. 1997; 90: 649-660Abstract Full Text Full Text PDF PubMed Scopus (430) Google Scholar). The dependence of loading on ORC and replicator sequences and the specific association with replicators suggest that the MCM proteins are components of the pre-RC. A clear role for Cdc6p in the assembly of pre-RCs was revealed by the finding that the association of MCM proteins with chromatin and specifically with replicators required the activity of Cdc6p (1Aparicio O.M Weinstein D.M Bell S.P Cell. 1997; 91: 59-69Abstract Full Text Full Text PDF PubMed Scopus (630) Google Scholar, 3Donovan S Harwood J Drury L.S Diffley J.F.X Proc. Natl. Acad. Sci. USA. 1997; 94: 5611-5616Crossref PubMed Scopus (422) Google Scholar, 19Tanaka T Knapp D Nasmyth K Cell. 1997; 90: 649-660Abstract Full Text Full Text PDF PubMed Scopus (430) Google Scholar). In contrast to the dependence on ORC of maintaining the association of MCM proteins with replicators seen in vivo, it was found that ORC and Cdc6p could be eluted from chromatin with salt while most of the bound MCM proteins remained associated (3Donovan S Harwood J Drury L.S Diffley J.F.X Proc. Natl. Acad. Sci. USA. 1997; 94: 5611-5616Crossref PubMed Scopus (422) Google Scholar). This observation is consistent with the idea that loading MCM proteins onto chromatin requires both ORC and Cdc6p, but that once bound, the MCM proteins interact with some other component of chromatin. One way to resolve the apparent conflict is to suggest that different populations of MCM proteins are detected by the two assays. Both genetic and physical evidence supports the idea that MCM proteins interact among themselves and with other proteins, suggesting that they may execute their essential function(s) in complexes large enough to accommodate two to six MCM protein monomers (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Consistent with this idea, the association of a given MCM protein with chromatin requires the activity of other members of the MCM family. For example, a temperature-sensitive mutation in MCM5 prevented the association of Mcm7p with ARS1 when cells were incubated at the nonpermissive temperature (19Tanaka T Knapp D Nasmyth K Cell. 1997; 90: 649-660Abstract Full Text Full Text PDF PubMed Scopus (430) Google Scholar). A tantalizing and exciting aspect of the association of MCM proteins with specific DNA sequences was revealed in the study by Aparicio et al. While the MCM proteins examined were found specifically associated with replicators during G1, their association with specific sequences changed during S phase, moving from replicators to other sequences. Consistent with the chromatin binding studies, the association of MCM proteins with all DNAs tested decreased to background levels by the end of S phase. In further studies that examined changes in distribution of MCM proteins in a region of chromosome III during a synchronous cell cycle, it was found that their distribution closely paralleled that of DNA polymerase (Pol ε), a known component of the replication fork. Pol ε was first detected in association with replicators at the G1/S boundary, well after the MCM proteins were bound, and then Pol ε and the MCM proteins appeared to move together toward the middle of the region analyzed. With the important assumption that these CHIP assays reflect the actual distribution of proteins on chromatin and not just accessibility of epitopes, these findings suggest that the MCM proteins are part of the replisome, the complex of proteins that participates in the elongation phase of DNA replication. A component of the eukaryotic replisome that remains elusive is the replicative helicase, which functions to unwind the DNA helix at the replication fork. The replicative helicases of E. coli (DnaB) and SV-40 (T antigen) both function as hexamers and use the energy of ATP hydrolysis to unwind DNA (reviewed by18Stillman B Cell. 1994; 78: 725-728Abstract Full Text PDF PubMed Scopus (219) Google Scholar). The DNA-dependent ATPase motifs conserved in MCM proteins, together with the recent report of a DNA helicase activity associated with a complex of Mcm4p, Mcm6p, and Mcm7p from HeLa cells (10Ishimi Y J. Biol. Chem. 1997; 272: 24508-24513Crossref PubMed Scopus (452) Google Scholar), suggest the possibility that at least a subset of the MCM proteins are components of the replicative helicase. An attractive feature of this idea is that it extends the parallel between prokaryotic and eukaryotic replication systems (18Stillman B Cell. 1994; 78: 725-728Abstract Full Text PDF PubMed Scopus (219) Google Scholar) to the initiation steps. For example, in E. coli, the DnaB helicase is brought to the replication origin by a chaperone that interacts with the initiator protein, DnaA, and loads the helicase. By analogy, Cdc6p, which exhibits genetic interactions with ORC, might be responsible for loading the MCM-associated helicase. As attractive as this model is, several issues suggest that it must be entertained with some caution. One problem is that the helicase activity measured by 10Ishimi Y J. Biol. Chem. 1997; 272: 24508-24513Crossref PubMed Scopus (452) Google Scholar is weak compared to known replicative helicases. The weak activity may reflect the properties of the enzyme or it could result from the absence of protein cofactors that stimulate the activity in vivo. Another question is why do MCM proteins, which are relatively abundant, appear to bind specifically to replicators. The levels of MCM proteins are estimated to be 50- to 500-fold higher than ORC, whose levels are estimated to be similar to the number of replicators, suggesting that there could be several hundred copies of MCM proteins per origin (13Lei M Kawasaki Y Tye B.-K Mol. Cell. Biol. 1996; 16: 5081-5090Crossref PubMed Scopus (154) Google Scholar, 3Donovan S Harwood J Drury L.S Diffley J.F.X Proc. Natl. Acad. Sci. USA. 1997; 94: 5611-5616Crossref PubMed Scopus (422) Google Scholar). The chromatin binding studies of Donovan et al. and Liang and Stillman suggest that a substantial fraction of the MCM proteins are associated with chromatin during G1, so the ratio of chromatin-bound MCM complexes to ORC is considerably higher than the required two replicative helicases per origin. It is possible, of course, that the putative MCM helicase is required for other aspects of DNA metabolism like repair, and that these other functions increase the amount of protein required. An alternative explanation for the apparent movement of the MCM proteins with the replication fork is that MCM proteins are actually distributed relatively uniformly along chromatin but are inaccessible during G1 except at replicator sequences, and that the movement of the replication fork exposes the MCM proteins as it approaches. This view is consistent with observations in mammalian and Xenopus cells that MCM proteins do not colocalize with "replication foci," which are sites of deoxynucleotide triphosphate incorporation into newly synthesized DNA (see Aparacio et al., 1997, for references). However, other proteins involved in DNA replication initiation and fork movement such as Orc2p, Cdc6p, and the single-strand binding protein, RPA, also fail to colocalize with replication foci, so the relationship between replication foci and other components of the replication complex requires further study. The finding that a complex of only three of the six MCM proteins possesses helicase activity is also somewhat surprising. What might be the essential function(s) of the remaining three members of the family? One possibility is that other subunits are actually part of the helicase, perhaps functioning as regulatory subunits or linking the helicase to the polymerase. An alternative is that the other family members have a different enzymatic or regulatory roles. The differences in copy number seen for different members of the family are consistent with the idea that they do not function in a single heterohexameric complex. DNA-dependent ATPases are also known to function in chromatin remodeling as well as origin unwinding. Whatever their precise role, it is clear that they are part of the pre-RC. Moreover, the observation that a mutation in MCM5 allows the survival of strains lacking a protein kinase normally essential for the initiation of DNA replication (Cdc7p) or its regulatory subunit (Dbf4p) indicates that the MCM proteins are also regulatory targets (6Hardy C.F.J Dryga O Pahl P.M.P Sclafani R.A Proc. Natl. Acad. Sci. USA. 1997; 94: 3151-3155Crossref PubMed Scopus (206) Google Scholar). An additional protein that appears to be part of the pre-RC is Cdc45p. CDC45 is an essential gene required for DNA replication, and mutations in CDC45 have been shown to interact with mutations in several components of the pre-RC, including members of the MCM family and ORC. Analysis of replication intermediates in cdc45 mutants revealed that the efficiency of firing at individual replication origins is reduced, a property shared with strains carrying mutations in the genes encoding other components of the pre-RC (reviewed by5Dutta A Bell S.P Annu. Rev. Cell Biol. 1997; 13: 293-332Crossref Scopus (333) Google Scholar). Moreover, cdc45 strains show a defect in the conversion of pre-RCs to post-RCs (17Owens J.C Detweiler C.S Li J.J Proc. Natl. Acad. Sci. USA in press. 1997; Google Scholar). Aparicio et al. found that the behavior of Cdc45p was indistinguishable from the behavior of the MCM proteins. During G1, it associated specifically with replicators and then appeared to move with the replication fork. Although its predicted protein sequence is not informative with respect to its function, Cdc45p may be another component of the replisome that assembles at replicators during G1. The identification of components of the pre-RC is an important step in the elucidation of the mechanism of eukaryotic replication initiation. The SV40 in vitro replication system faithfully recapitulates the complete replication of the virus and has been of crucial importance in dissecting aspects of the elongation reactions, which depend on cellular enzymes (reviewed by7Hassell, J.A., and Brinton, B.T. (1996). In DNA Replication in Eukaryotic Cells, M.L. DePamphilis, ed. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), 639–677.Google Scholar). However, origin recognition, replication initiation, and the replicative helicase activities in the SV40 system are all dependent on a single viral protein, large T antigen. In contrast, the initiation of chromosomal DNA replication is much more complicated than in viral systems, requiring the coordination of multiple replication origins and responding to several layers of cell cycle control designed to assure that the genome is replicated in a timely fashion and with high fidelity. The large number of gene products required for the proper initiation of chromosomal DNA replication reflects these additional layers of control. The information about the components required for pre-RC formation and their order of function, together with tantalizing clues of biochemical function provided by these recent studies, may make it possible to develop partial in vitro reactions that will provide insights into the mechanism of pre-RC assembly and DNA replication initiation.