DNA Breaks and the Cell Cycle:
Checkpoints and Cell Cycle Arrest

Checkpoints and Cancer. Eukaryotic cells go to great lengths to ensure that progeny cells receive accurate copies of parental genomes. Cells with defective checkpoints have progeny with damaged chromosomes, or lost chromosomes which can result in loss of viability or abnormal growth. Checkpoint defects cause a substantial problem in multi-cellular organisms because damaged chromosomes and genetic instability can result in tumors. The mammalian genes P53 and Ataxia-telangiectasia Mutated (ATM) encode checkpoint proteins that recognize damaged DNA and regulate apoptosis (cell suicide). The P53 gene is mutated in greater than 50% of all human cancers, which indicates a strong link between functional checkpoint genes and the predisposition to cancer. ATM is a kinase thought to play an important role in regulating p53 function during the damage response. Mutations in the ATM gene are associated with immunodeficiency, neurodegeneration, radiosensitivity and cancer predisposition. The human homologues of several S. cerevisiae checkpoint genes map to chromosomal regions implicated in the etiology of a wide variety cancers including small cell lung carcinoma, non-small-cell lung carcinoma, duodenal adenocarcinoma, head and neck squamous cell carcinoma, bladder cancer, and colon cancer. An understanding of checkpoint function will shed light on the mechanism of tumor formation and cancer predisposition and may provide insights into new therapeutic targets for cancer treatment. Yeast is an ideal organism for studying checkpoints because of the high level of homology between human and yeast checkpoint proteins. In addition, the genetic malleability of yeast makes it a powerful organism to define the mechanisms of cellular checkpoints.

Checkpoint genes in S. cerevisiae. In Saccharomyces cerevisiae there are three defined checkpoint pathways that recognize the presence of damaged DNA and halt progression at different cell cycle transitions. These include the G1/S, Intra-S, and G2/M checkpoints. The G1/S checkpoint recognizes the presence of damaged DNA during G1 and will delay the onset of S phase and DNA replication. The intra-S checkpoint will stall ongoing replication and initiation of late firing replication origins if DNA damage is detected during S-phase and finally the G2/M checkpoint will delay mitosis to prevent the segregation of chromosomes damaged during G2 or S-phase.

Many of the DNA damage checkpoint genes have been identified from a several genetic screens that identified mutants that were either: sensitive to radiation (rad alleles); defective in the mitotic entry checkpoint (mec alleles); suppressed mutations in other checkpoint genes (ddc1), or bore sequence homology to known S. pombe checkpoint genes (chk1). The checkpoint genes are divided into three basic classes: (A) Sensor genes that actively recognize or modify damaged DNA and activate the (B) Signal transducers which are cellular kinases required for transmitting the checkpoint signal to (C) Target genes involved in regulating the cell-cycle progression machinery and damage response genes (Fig.1). In S. cerevisiae, RAD9, RAD17, RAD24, MEC3, and DDC1 sensor genes are required to recognize the presence of damaged DNA and activate the checkpoint. The sensors elicit the checkpoint response by activating the signal transducer kinase MEC1which is a homologue of human ATM. MEC1 in turn activates several kinases including RAD53 and CHK1. Targets of these signal transducers are not well defined. RAD53 and CHK1 require MEC1 for activation and clearly form parallel pathways that regulate multiple cell-cycle transitions. Activated RAD53 regulates the transcription of G1 cyclins required for G1/S transition, controls replication fork progression and replication origin firing during S phase, and may regulate mitotic exit (G2/M transition) through the target genes, DUN1 and CDC5. CHK1 has been shown to regulate the degradation of Pds1p which an important determinant of the metaphase to anaphase transition. The transducers may also activate the transcription of DNA repair and checkpoint proteins.

What is the checkpoint signal? Several types of damage including damage caused by UV irradiation, gamma irradiation, and alkylating agents activate the DNA damage checkpoint genes. In addition, a single double strand break (DSB) generated by the site-specific HO endonuclease is sufficient to establish the checkpoint response. It is not understood how one set of sensor genes is able to recognize these very different types of DNA damage. Sensor proteins could bind to the different types of damage, recognize a specific protein-DNA structure such as a repair enzyme complex, or a DNA structure that results during the repair process. Possible signals include single-stranded DNA (ssDNA) or junctions of ssDNA with double-stranded DNA (ssDNA/dsDNA). Consistent with this idea is mutations in several checkpoint genes are synthetically lethal with the cdc13-1 mutation. At the non-permissive temperature, cdc13-1 mutants accumulate ssDNA and this is thought to elicit the checkpoint response. Cells with defective checkpoint genes may be unable to sense the cdc13-1-induced ssDNA and segregate damaged chromosomes. Our lab uses an HO-induced double strand break (DSB) to induce the checkpoint. Analyses of proteins that regulate the amount of ssDNA generated after a indicate that mutants that accumulate ssDNA activate the G2/M checkpoint, whereas mutants that accumulate less ssDNA do not establish the checkpoint. Several proteins are thought to play a critical role in controlling the amount of ssDNA including KU70, KU80, XRS2, RAD50, and MRE11 (Figure 2). A mutant of the ssDNA-binding protein RPA is also involved in the checkpoint process and may play a role in signaling the presence ssDNA to the checkpoint sensor proteins. These studies implicate ssDNA or the resulting ssDNA/dsDNA as important signals for establishing the checkpoint, however, more studies are required for confirmation.

Checkpoint sensor genes and recognition of the signal. One set of sensor proteins recognizes the presence of damaged DNA, regardless of the cell-cycle transition i.e. G1/S, Intra-S, or G2/M. The DNA damage sensor genes are classified into two parallel epistasis groups with RAD17, RAD24, MEC3, and DDC1 (RAD24 class) in one group and RAD9 in another (Figure 2). During the checkpoint response, mutations in RAD17, RAD24, and MEC3 have decreased exonuclease activity, while mutations in RAD9 have increased exonuclease activity. This suggests that sensor genes may be directly involved in regulating the amount of ssDNA in the cell with inherent exonuclease activity or by modulating the activity of an unidentified exonuclease. RAD17 homologues in Ustilago maydis and humans have been purified and exhibit 3’ to 5’ exonuclease activity although it is not clear whether this activity is required for establishing the checkpoint.

Both genetic and biochemical data indicate functional and physical interactions among several of the proteins in the RAD24 class of proteins. Pair-wise interactions and coimmunoprecipitations have been observed between Rad17p, Mec3p, and Ddc1p, suggesting the presence of a heterotrimeric complex that may be important for establishing the checkpoint signal. Stability of this complex is not dependent upon an activated checkpoint. This implies that the heterotrimeric sensor complex may be present throughout the cell cycle and would recognize damaged DNA and activate the checkpoint. Similar heterotrimeric complexes have been identified in S. pombe and humans, indicating a general functional and structural conservation in the sensor proteins.

Mutations in RAD24 gene locus are associated with a large number of human cancers which indicates a central role in maintaining genomic stability. Mutant forms of RAD24 can be suppressed by overexpression of the other sensor proteins (RAD17, MEC3, AND DDC1) which suggests that RAD24 function may be closer to the damage recognition step. Rad24p does not appear to associate with the heterotrimeric sensor complex, which may indicate a distinct function in establishing the checkpoint. RAD24 has sequence homology to several subunits of Replication Factor C (RF-C) including a nucleotide binding motif (Walker A motif), and physically interacts with several RF-C subunits. It is unclear, how these interactions modulate the checkpoint activity of RAD24 or the replication activity of RF-C. During the course of replication, RF-C specifically binds to ssDNA/dsDNA junctions, which suggests that Rad24p, alone or in concert with RF-C, may also bind to these junctions to elicit the checkpoint response.

Activation of MEC1 kinase activity is dependent upon RAD9 and the RAD24 class of proteins, which indicates a function downstream of the sensor proteins. However, Ddc1p is phosphorylated in a MEC1-dependent manner the presence of damaged DNA. Rad9p is also phosphorylated in a RAD53, MEC1, and RAD24-class dependent fashion and phosphorylation of Rad9p modulates its interactions with Rad53. It is not clear whether these phosphorylation events are important for checkpoint or are simply the result of a feedback pathway, however these results indicate that the functional interactions between the sensor proteins and MEC1 may not fit in a simple linear pathway. Longhese and coworkers have suggested that Mec1p may be activated by directly binding to DNA or DNA-protein structures and that this association may be mediated by the sensor proteins. This would be consistent with the MEC1-related kinase in humans, DNA Dependent Protein Kinase, which binds to ssDNA-dsDNA junctions with the assistance of the Ku proteins.

The identity of the checkpoint signal and the mechanism by which sensor proteins and MEC1 recognize this signal remains unclear. Studies are needed to address these questions in order to gain an understanding of how cellular checkpoints are established. An understanding of the mechanism of checkpoint function will shed light on several human diseases associated with checkpoint defects including ATM and cancer.