The recent completion of the Saccharomyces pombe genome revealed an impressive number of fission yeast genes with human homologs implicated in cancers. Interestingly, many of these genes have known or implied functions in DNA replication in yeast. Current evidence suggests that multiple pathways of control of eukaryotic DNA replication can be disrupted to result in genome instability and predisposition to cancer. Thus, deregulation of CDK activity, impaired origin firing, changes in the timing of firing, loss of control in the order of S phase and M phase, and inability to limit replication to once per cell cycle are all mechanisms that may lead to changes in chromosome structure and gene function. In addition, defects in the checkpoint response to replication blocks, and the inability to respond appropriately to stalled replication forks, also contribute to genome instability. Ultimately, the gain or loss of genetic information may lead to inappropriate expression of proto-oncogenes or loss of tumor-suppressor function.

When the normal timing of origin firing is disrupted, cells are susceptible to deregulated cell cycle progression. This could result either through refiring of origins in a single cell cycle, or through firing late origins of replication under conditions where they are normally prevented from firing. Treatment of Saccharomyces cerevisiae cells with the antitumor drug adozelesin changes the normal pattern of replication such that active replication forks are cations of mammalian chromosomes also may alter replication timing of a particular sequence. Conversely, uncontrolled cell proliferation may result in deregulation of replication timing. This is observed both in checkpoint mutants in S. pombe and in human cancers. Thus, disruption of the timing and coordination of replication is one pathwaytoward genome instability.

Cells extend multiple, overlapping control mechanisms to restrict DNA replication to once per cell cycle. In S. pombe, this is accomplished by regulation of CDK kinase activity, phosphorylation, and destruction of Cdc18, and phosphorylation of STET. In human cells, the Cdc18 equivalent Cdc6 is also negatively regulated by CDK phosphorylation, suggesting that regulation of human Cdc6 likewise contributes to prevention of re-replication. The MCM proteins are another CDK target, at least in some organisms. There are several examples of deregulated CDK activity associated with cancers.

Overexpression of certain replication proteins, such as Cdt1, can promote tumor formation in mammals. In addition, many replication proteins are specifically upregulated in cancer cells. Human Cdc7 (the homolog of the S. pombe Hsk1 kinase) is overexpressed in certain tumor cells. Furthermore, human MCM proteins are specifically expressed (or overexpressed) in cycling cells and are not detectable in quiescent cells. An important consequence of these findings is that the presence of the MCM proteins in cells provides a sensitive diagnostic marker for proliferating cells. MCM proteins are detected in cells that have exited quiescence and reentered the cell cycle; thus, MCM proteins are detected in precancerous cells as well as in tumor cells. MCM transcription is further upregulated by activated oncogenes. Interestingly, human BM28/CDCL1 (the homolog of Mcm2), maps to a chromosomal locus associated with acute myeloid leukemia, suggesting BM28/CDCL1 as a candidate oncogene. Damage tolerance and repair mechanisms are also essential to prevent genome instability. In S. pombe, Rqh1 is needed for recovery from replication blocks. Human cells have at least five Rqh1 homologs, three of which are linked with cancer susceptibility syndromes. Mutations in BLM are associated with Bloom’s syndrome, mutations in WRN lead to Werner’s syndrome, and mutation of RecQL4 results in Rothmund-Thomson syndrome. Hyperrecombination and cancer susceptibility are characteristic of both Bloom’s and Werner’s syndromes. Inappropriate recombination due to the loss of other S-phase functions may generate deletions or expansions in the genetic information, as has been demonstrated in S. cerevisiae. Polymerase slippage may contribute to the formation of triplet repeat sequences, which are associated with several disorders including Huntington’s disease.

Checkpoint genes are important gatekeepers of genome stability. Mutations in the ATM checkpoint kinase are linked to ataxia telangiectasia, and mutations in the checkpoint kinase Cds1 (also called CHK2) are found in a subset of patients with Li-Fraumeni syndrome. In addition, Rad17 (one of the checkpoint rad proteins) is overexpressed in certain types of human cancers. The corresponding S. pombe proteins (Rad3, Cds1, and Rad17) are all involved in the cellular response to replication blocks. The S. pombe Rad4/Cut5 protein, which also has a role in cellular checkpoints, contains a BRCT motif that is also present in the human BRCA1 tumor suppressor and the XRCC1 DNA repair protein. Thus, mutations that disrupt function of the replication checkpoint are also implicated in predisposition to cancer. Genomic instability leading to cancers may also result from chromosome structure defects caused by errors in S-phase processes linked to DNA replication. In S. pombe, the Eso1 protein is needed to activate cohesion so that sisterchromatids are held together until mitosis. Part of the Eso1 protein is homologous to DNA polymerase η (Rad30), which is defective in the xeroderma pigmentosum variant syndrome characterized by predisposition to skin cancers. In addition, human securin, normally prevents premature sister-chromatid separation, can induce cell transformation and tumorigenesis when overexpressed. Expression of the Rad21 cohesin is downregulated in certain tumors. In addition, HP1 is downregulated in breast cancer cells that are metastatic or invasive. The S. pombe homolog of HP1, Swi6, recruits Rad21 to centromeres and other regions of heterochromatin. Recently, phosphorylation of another cohesin subunit, Smc1, has been shown to be required for the S-phase checkpoint in human cells. Taken together, these connections suggest a direct role for chromatin structure in maintenance of genome stability.

Meiosis is a specialized cell cycle that generates recombinant, haploid progeny cells from a diploid cell. The meiotic cell cycle differs from the vegetative cell cycle in two outstanding respects. First, the S phase that occurs prior to meiosis (premeiotic S) is followed by two successive rounds of chromosome segregation, rather than alternating between S phase and mitosis. Second, the first meiotic division is reductional, resulting in the maintenance of cohesion between sister chromatids but the separation and segregation of homologous chromosomes. The preparation for this modified cell division cycle involves lengthy interaction between homologous chromosomes during the prophase stage of meiosis I but is likely to initiate as early as premeiotic S phase. Premeiotic S phase is longer than S phase in vegetative cells in most organisms. The cause of this difference is unclear, since experiments in budding yeast suggest that the same replication origins are active in vegetative and meiotic cells. However, other experiments in S. cerevisiae suggest that meiosis-specific chromosomal factors required during prophase might assemble during premitotic STET.

Recent studies in Saccharomyces pombe have addressed the question of whether the replication machinery that functions during premeiotic S phase is the same as that utilized in the vegetative cell cycle. S. pombe proteins required for the actual synthesis of the DNA in vegetative cells, such as DNA polymerase α and ribonucleotide reductase, also are essential for premeiotic S phase. In contrast, mutants defective in initiation of DNA replication, such as the mcm mutants and cdc18, display different phenotypes in meiosis and mitosis. In mitosis, conditional alleles of these mutants allow bulk DNA replication but cause cells to arrest in late S phase. In contrast, in similar conditions these mutants can proceed through the meiotic divisions and sporulate. With more extreme conditions, these mutants delay replication and the subsequent meiotic divisions. This may reflect a quantitative difference: meiotic cells may tolerate a lower amount of certain replication proteins than that needed during the vegetative cell cycle. It is also possible that other meiotic factors, perhaps recombination proteins, can contribute to premeiotic replication.

The S. pombe MCM protein complex is associated with chromatin during premeiotic S phase, consistent with the hypothesis that these proteins function in premeiotic DNA replication. However, MCM proteins are not localized to chromatin in between the meiotic divisions, when an additional round of DNA replication is suppressed. Similar to the vegetative cell cycle, the meiotic cell cycle is subject to checkpoint controls. Fission yeast cells that have been induced to enter meiosis block the cell cycle when treated with HU. As in the vegetative cell cycle, HU-induced arrest during meiosis is likely to be checkpoint-dependent. Future work should resolve the components of this response, which are likely to be essential for the viability of gametes. Importantly, premeiotic S phase is closely coupled to the downstream events of meiosis such as recombination. This has been best demonstrated in budding yeast, where blocks to premeiotic DNA synthesis prevent meiotic recombination and changes in the timing of premeiotic replication result in corresponding changes in the timing of initiation of meiotic recombination. The molecular mechanism by which this occurs is still unclear.