Mutagenesis Advance Access originally published online on April 20, 2005
Mutagenesis 2005 20(3):153-163; doi:10.1093/mutage/gei031
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW |
Tails of histones in DNA double-strand break repair
Department of Biochemistry, Cambridge University, 80 Tennis Court Road, Cambridge CB2 1GA, UK
 |
Abstract |
|---|
 Top Abstract IntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
DNA double-strand breaks (DSBs) are, arguably, the most deleterious form of DNA damage. An increasing body of evidence points to the inaccurate or inefficient repair of DSBs as a key step in tumorigenesis. Therefore, it is of great importance to understand the processes by which DSBs are detected and repaired. Clearly, these events must take place in the context of chromatin in vivo, and recently, a great deal of progress has been made in understanding the dynamic and active role that histone proteins and chromatin modifying activities play in DNA DSB repair. Here, we briefly review some of the most common techniques in studying DNA DSB responses in vivo, and focus on the contributions of covalent modifications of core histone proteins to these DNA DSB responses.
|
Introduction |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
Within the eukaryotic cell nucleus, genetic information is organized in a highly conserved structural polymer, the chromatin, which supports and controls crucial functions of the genome. The basic unit of chromatin is the nucleosome, which comprises 146 bp of DNA wrapped around eight histones, two of each of the four core histone families—H2A, H2B, H3 and H4. Linker histones, termed H1 or H5, associate with the DNA between individual nucleosomes establishing a higher level of organization.
This method of packaging DNA allows very long negatively charged molecules to exist in relatively small 3D space. For example, the length of DNA contained in a human cell is 
2 m, yet it occupies a spherical space i.e. in some cases, only 3 µm in diameter. In addition to these powerful properties of condensation, the composition of chromatin must allow rapid and precise access to particular regions of the genome during transcriptional regulation and accurate partitioning to daughter cells during mitosis. It is not surprising, therefore, that chromatin is extremely malleable, and that chromatin undergoes dynamic changes, including massive structural reorganization, during many genetic processes, such as DNA replication and cell division, transcription, as well as DNA repair and recombination. Furthermore, nucleosomes can transmit epigenetic information from one generation to another, having the potential to act as a memory bank of the cell.
The chromatin structure has the potential to influence all genetic processes and is frequently exploited as means of gene regulation. The term ‘chromatin remodeling’ has been used to describe transitions in chromatin structure that include histone post-translational modifications (described in more detail below), alterations to the histone variant composition of nucleosomes, changes in the non-histone protein content of chromatin and alterations to chromatin structure due to the action of ATP-dependent chromatin remodeling enzymes [e.g. yeast Swi2, see (1
) for review]. In many cases, these different classes of chromatin alteration are interrelated. For example, recent reports indicate that histone dimers can be removed or exchanged between nucleosomes during the course of a chromatin remodeling reaction.
Core histone proteins are evolutionarily conserved and consist mainly of flexible N-terminal tails protruding outward from the nucleosome, and histone fold-containing domains that make up the nucleosome scaffold. In addition, histones H2A and H2B have C-terminal tails that also protrude outward from the nucleosome structure. These core histone proteins, and in particular, their N-terminal and C-terminal tails, function as acceptors for a variety of post-translational modifications, including acetylation and ubiquitination of lysine (K) residues, phosphorylation of serine (S) and threonine (T) residues and methylation of lysine and arginine (R) residues (Figure 1). Combinations of specific patterns of post-translational modifications provide the basis for the ‘histone code’ hypothesis (2), in which the interplay between different covalent modifications determines chromatin functions, by either altering nucleosome conformation or the binding affinities of the modified histones for various proteins. A major challenge in chromatin biology is connecting particular modifications with distinct biological functions and vice versa.
|
Although a great deal is known about the role of post-translational modifications, particularly acetylation in transcriptional regulation, it is clear that there are many unanswered questions. Moreover, the role of post-translational modifications in DNA repair and recombination is far less understood. In this review, we will focus on recent findings regarding histone modifications that might be involved in DNA double-strand break (DSB) repair in eukaryotes and the tools commonly used to investigate these histone modifications.
DNA double strand breaks
We are exposed daily to a vast range of agents capable of inflicting various types of DNA damage. Of the DNA lesions created, probably the most dangerous is the double-strand break, as it generates loose ends, which have great potential to inappropriately recombine with other parts of the genome. Furthermore, inaccurate repair of the lesion could introduce mutations in tumor suppressor genes. Clearly, there is significant tumorigenic potential if DNA DSB activities are impaired.
DNA DSB repair in eukaryotes is mediated primarily by two pathways: homologous recombination [HR; (3)] or non-homologous end-joining [NHEJ; (4)]. Both processes are likely to require chromatin alterations in order to efficiently undergo processes, such as strand invasion, branch migration, DNA synthesis, ligation and recruitment of checkpoint proteins (5).
In addition to the repair activities, cells respond to DNA damage by instigating a signal transduction cascade that results in the arrest of cell-cycle progression and the transcriptional activation of DNA damage responsive genes. Members of the PI-3 kinase-like family, including mammalian DNA-PK (DNA-dependent protein kinase), ATM (ataxia telangiectasia mutated protein) and ATR (ataxia telangiectasia related protein), have been implicated as important components of these DNA damage responses. While there are no homologs of DNA-PK in lower eukaryotes, clear homologs of ATM and ATR exist and are of principal importance to DNA damage-responsive signal transduction pathways. These are Tel1 and Mec1 in Saccharomyces cerevisiae and Tel1 and Rad3 in Schizosaccharomyces pombe, respectively.
The repair of DNA must, of course, occur in the context of chromatin, and therefore, it is reasonable to speculate that one of the first steps in the DNA repair process is the reorganization of chromatin structure to allow access and processing by the repair machinery. Interestingly, chromatin remodeling complexes, such as Ino80 and Swr1 (S.cerevisiae) and Tip60 (Homo sapiens) have been implicated in DNA repair (6). Moreover, chromatin assembly factors, namely Asf1 (S.cerevisiae) and CAF-1 (S.cerevisiae and H.sapiens), have also been implicated in DNA repair. Yeast strains with mutations in ASF1 are sensitive to DSBs (7) while CAF-I defective strains are sensitive to UV when compared with their wild-type counterparts (8). Both factors assemble H3 and H4 into nucleosomes during DNA replication and possibly perform the same task on newly synthesized repaired DNA. Consistent with this view, CAF-I is recruited onto DNA after UV irradiation of human cells (8). In addition to chromatin remodeling, post-translational modifications of histones have been implicated in DNA repair activities. It is important to view these modifications in the context of other chromatin activities. For instance, in addition to effects on the repair machinery, histone modifications may also function to facilitate chromatin remodeling or assembly activities.
|
Talking to histones |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
An astonishing number of biochemical and genetic techniques are available to study chromatin behavior. The vast majority of these approaches have been developed by researchers examining the role of chromatin in the regulation of gene expression and are both highly applicable and easily adaptable to the studies on the role of chromatin in DNA repair and recombination. Therefore, before summarizing the existing data on post-translational modifications of histones in DNA repair and recombination, we will briefly review a few of the most widespread techniques for investigating chromatin behavior.
Model system
The functions of chromatin compaction and genome maintenance are vital to our survival. As such, many proteins involved in both of these activities have been very strongly conserved throughout the eukaryotes. The best characterized DNA damage dependent phosphorylation event (phosphorylation of the H2A C-terminal tail, described in more detail below) appears to be conserved from yeast to humans. Moreover, many in vitro studies of chromatin modulating activities use factors isolated from a variety of organisms [e.g. yeast ATP-dependent chromatin remodeling factors can be analyzed using chromatin composed of human histones and sea urchin rDNA sequences (9)], demonstrating extraordinary structural and functional conservation. Nevertheless, there is variability in the data generated from different labs. It remains to be seen, which of these differences are due to evolutionary variation and which of these are simply due to the different approaches used in different labs. Therefore, while it is reasonable to extrapolate many fundamental insights generated in one species to other eukaryotic organisms, there will inevitably be important differences.
Genetic studies
Historically, the first step in determining gene function is to inactivate or impair the gene and examine the phenotype of the resulting strain. Subsequent analyses, such as complementation, suppression and epistasis studies with other mutations can yield enormous insights into the function of a gene. Unfortunately, genes encoding for histones are essential, and therefore, full disruption of the genes is not feasible. Of course, mutagenesis of specific residues or motifs is possible, and can be an excellent way to very precisely define the role of a specific residue in an activity, such as DNA repair. This approach, however, is complicated by the fact that eukaryotes typically have more than one copy of each histone gene—up to as many as 600 in sea urchins and newts. Barring the analysis of dominant mutations, this makes the mutagenesis of histones challenging.
Using genetics to analyze histones in higher eukaryotes is somewhat limited. A notable exception to this is the potential to analyze the roles of non-essential, single-copy histone variants, such as H2AX, discussed in more detail below. In lower eukaryotes, such as the budding yeast, the already powerful case of genetic manipulation is enhanced by the fact that there are only two copies of each core histone.
One common approach to genetic analysis of histones in the budding yeast is to use strains in which both copies of the core histone(s) of interest are disrupted and the strains are kept alive using one copy of the histone(s) on a plasmid under the control of the endogenous promoter(s). These strains are viable, and when using a plasmid with a selectable marker that can be selected both for and against, allow the wild-type histone plasmid to be easily swapped with mutagenized histone copies on a plasmid with a different selectable marker. Interestingly, however, there are slight differences in both the promoters and the coding sequences of the two copies of the core histones, and it is possible that there will be subtle effects of the absence of the second copy, particularly when analyzing responses to something as pleiotropic and stressful as ionizing radiation (IR). To that end, it is possible to introduce specific mutations into both genomic histone genes, which may aid subsequent genetic analyses.
Once created, the strains can be analyzed for their sensitivity to DNA DSBs to determine whether they are important for DNA damage responses. On account of the myriad of cellular responses to DNA damage, any implicated motifs or enzymes may be important for preventing DNA lesions in the first place, regulating the transcription of DNA damage responsive genes or directly facilitating repair at the site of the lesion.
Biochemical studies
A rapid approach to screen for covalent modifications, such as phosphorylation, methylation, acetylation, sumoylation or ubiquitination, occurring in histones is through the use of antibodies designed to specifically recognize modified residues. This method is commonly used toward several histone modifications, but unfortunately, it presents some limitations. Covalent modifications of neighboring, non-target residues can inhibit the recognition by antibodies designed against single modifications, thus giving an inaccurate reflection of the modified state of the residue of interest. In addition, cross-recognition of covalently modified residues in similar sequence contexts can be a problem.
Alternatively, some protein modifications can give rise to mobility shifts during polyacrylamide electrophoresis, which provides good methods to identify the presence of conditions that induce covalent modifications. The use of specialized electrophoresis approaches, such as 2D and AUT gels, can help to resolve differentially modified species when straightforward SDS–PAGE is insufficient.
In looking to see whether a modification, residue or motif is required for the recruitment of DNA repair, recombination or remodeling proteins, one very powerful tool is the chromatin immunoprecipitation (ChIP) assay. Following the cross-linking between DNA and proteins associated to it, specific proteins can be immunoprecipitated and the complexes to it bound can be analyzed. ChIP assays can also be used to determine the timing and location of histone modifications at DNA lesions by using antibodies specific to histone modifications.
|
Generating double strand breaks |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
The approaches described above are in widespread use in the gene regulation field, and are easily adapted to studies of DNA repair and recombination. However, unlike the studies of a defined promoter or open reading frame, DNA repair and recombination can occur anywhere in the genome, and so some special considerations must be made. For example, in order to perform ChIP assays, the site of the DNA DSB must be defined. The mechanisms of generating DNA DSBs are outlined below.
DNA damaging agents
A rapid approach for generating DNA damage is the use of mutagenic chemicals or radiation (Table I). These can, in addition to generating random damage throughout the genomic DNA, also damage various cellular components, which can make interpretation of phenotypic assays complicated. Nevertheless, one major advantage of using mutagenic chemicals or radiation is that they can be applied to any organism, without the need for specialized vectors, expression constructs or technology.
|
DNA DSBs can be created by a variety of agents, including IR, and radiomimetic chemicals, such as phleomycin. DSBs may also arise from replication past single-strand breaks (SSBs), processing of other DNA lesions, nuclease activity or torsional strain. Thus, using a combination of chemicals and cell-cycle analysis can provide great insights into the effects of mutations on DNA repair and signaling pathways.
Two commonly used methods of creating DNA damage specifically during S phase are type I topoisomerase inhibitors and ribonucleotide reductase inhibitors. Type I topoisomerases can be very effectively inhibited by camptothecin (CPT) or its analogs irinotecan and topotecan, and this results in defective DNA replication, leading to fork arrest and ultimately S-phase specific DSBs (10). Hydroxyurea interacts with the free radical in ribonucleotide reductase, the enzyme responsible for the formation of dNTPs. This leads to lowered dNTP pools, and therefore, the generation of DNA DSBs specifically in S phase owing to the collapse of the replication fork (11).
Alkylating agents are electrophiles that add methyl (MMS), ethyl (EMS) and more complicated alkyl groups to nucleic acid bases. Interestingly, N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) in vivo becomes a highly reactive methylating agent. Nitrogen and sulfur mustards link bases on opposite DNA strands, creating intra-strand cross-links. N-Methyl-N-nitrosourea is a potent mutagen and carcinogen that reacts directly with DNA producing methylated bases. These adducts produce GC to AT transitions after two rounds of replication. While none of the lesions created as a direct result of exposure to these alkylating chemicals is a DNA DSB, processing of the lesions, particularly MMS and MNNG-induced DNA alkylation, appears to result in the generation of significant levels of DNA DSBs (11).
Ultraviolet (UV) radiation mainly causes dimerization of the rings of adjacent thymines. DSB induction might arise after UV treatment during excision repair of closely opposed lesions or replication through a partially processed lesion (11).
IR, such as X rays or 
rays, has been used extensively to induce DSBs, although DSBs are only a minor component of the radiation-induced damage. IR damages DNA either directly or indirectly through active oxygen species, causing a spectrum of DNA lesions: SSBs, DSBs, DNA–protein crosslinks and base substitutions. However, studies in mammalian cells indicate that the DSBs are the most relevant lesions with respect to biological effects (11), making IR and IR-mimetic chemicals one of the most useful tools in studying DNA DSB responses.
In summary, most DNA damaging agents have pleiotropic effects in the cell and with few exceptions, the exact mechanisms of action, as well as the relative importance of each of the spectrum of changes caused, are not completely clear. Therefore a thorough understanding of the role of a protein in DNA repair might be facilitated by the use of multiple DNA damaging agents, ideally also in combination with other approaches for creating DNA DSBs (described below).
The homothalic endonuclease
The homothalic (HO) endonuclease creates a deliberate single DSB in the yeast genome at the MAT locus, which is the initial step in the process of mating type conversion. This DSB is normally repaired by Rad52-dependent HR with either of the silent mating loci, HML or HMR, as the source of homologous sequence.
Not surprisingly, this has become a very powerful and widely used tool to study DNA DSB repair in S.cerevisiae. Sensitivity of a mutant strain to HO endonuclease expression directly implicates DNA DSB responses, as, unlike chemical exposure, there is no other cellular damage caused. The most straightforward approach to inducing HO cleavage is by inducing HO expression in plate assays. In these assays, the HO endonuclease is under the control of a galactose inducible promoter and strains are grown on galactose-containing (inducing) plates or glucose-containing (inhibiting) plates. The consequence of this is continual expression of the HO endonuclease, and the cells are, therefore, constantly exposed to the generation of DSBs, allowing surival under these conditions to be a reflection of DNA DSB repair abilities.
To physically monitor the recombination repair of a single round of HO-induced DSBs, members of the Haber laboratory developed a methodology named in vivo biochemistry [(12) and references therein]. In this methodology, mid-log phase liquid cultures are exposed to a brief (1 h) pulse of HO endonuclease expression; cells can then be harvested at varying time points post-expression. The process of mating type conversion can be monitored by Southern blotting, and the order of association of HR proteins can be analyzed in conjunction using ChIP analyses.
In addition to using the HO endonuclease to study DSB repair by HR, induction of the endonuclease in a strain lacking the HML or HMR donor regions can be used to assess DNA DSB repair by NHEJ. Moreover, the HO recognition and cleavage sequence can be placed in other regions of the genome that have or do not have regions of homology elsewhere in the genome in order to probe more precisely the influences of variables, such as chromatin structure or transcriptional activity on DNA repair activities.
The I-SceI endonuclease
I-SceI is a budding yeast mitochondrial endonuclease, which recognizes an 18 bp consensus sequences with high specificity. Therefore, it constitutes a powerful tool that has been widely used in cells from many organisms, including mammals, to study DNA DSB repair in vivo. As with the HO endonuclease, it is possible to introduce artificial I-SceI sites at different chromatin contexts, induce the endonuclease expression from a suitable vector and monitor the chain of events following the introduction of a localized DSB (13–15).
The EcoRI endonuclease
The EcoRI restriction endonuclease has been adapted for use in S.cerevisiae (16). Because of the much smaller 6 bp recognition sequence, by using this enzyme, multiple DSBs can be simultaneously generated in the yeast genome, which may be more reflective of exposure to higher doses of DNA damage, and allow the study of cellular responses to numerous DNA lesions. However, unlike HO and I-SceI, EcoRI has a low cutting frequency and consequently, only 5% of the cells will have a cut at a defined chromosome location. This makes assays that measure events at a defined location, such as ChIPs and Southern blotting, more problematic since the majority of cells in the population have no DNA lesion at a given location.
Laser scissors
The laser scissors technique is used to generate DSBs in a defined nuclear domain, which can then be monitored through optic techniques to assess the modification or movement of factors around the sites of damage. It makes use of a finely controlled laser-dissecting microscope, where UVA-mediated excitation of a dye incorporated into the DNA of living cells introduces DNA DSBs along the path of the laser (17–19). This technique is frequently applied to the observation of DSBs in mammalian cells. However, it is less suitable for use in lower eukaryotes such as budding yeast, owing to their minute size.
Most DSB-generating approaches can be used in combination with the monitoring of foci formation, which is the recruitment to or modification of factors at the site of a break. Foci are monitored through immunofluorescence studies, and can provide good clues to the kinetics of factors acting during the repair of a lesion.
Unlike EcoRI, I-SceI and HO-induced DSBs, which can be used for both microscopy studies as well as survival assays and more sensitive biochemical approaches, the laser scissors approach can be used only in conjunction with immunofluorescence. Obviously, in order to be detected in these assays, multiple proteins (or their modifications) need to be amassed in a small region. Only those histone modifications or modifying enzymes that work directly at the site of the DNA DSB will be detected. Moreover, responses to damage by a small number of proteins or protein modifications can be missed in all immunofluorescence-based approaches, including laser scissors.
Plasmid repair assay
The plasmid repair assay consists of enzymatically cleaving a plasmid in vitro and introducing this defined linear DNA template into yeast cells. The number of colonies formed with respect to cells transformed with equivalent amounts of undigested plasmid DNA gives an indication of the cells to recircularize the plasmid, which is performed by the DNA DSB machinery (20). By using plasmids that share no significant homology to the yeast genome in the region of the introduced DNA DSB, this method provides a good indication of DNA repair by NHEJ. Unfortunately, since it is unclear to date whether the linear plasmid introduced into the host is chromatinized at the time of repair, caution should be taken when extrapolating the results obtained in this assay to those expected to take place in a chromatin environment. Nevertheless, at least some strains lacking histone motifs or histone modifying enzymes that have defects in the plasmid repair assay are also sensitive to DNA damaging agents (21–24), suggesting that the assay may be physiologically relevant.
|
H2A |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
The H2A family
In addition to the histone H2A subfamily that makes up the bulk of mammalian H2A (H2A1 and H2A2, encoded by 11 different genes), there are a number of other histone H2A variants in the human genome, including, but not limited to, H2AX and H2AZ, each encoded by a single gene (25
). H2AX represents between 2 and 25% of the total cellular H2A, whereas H2AZ appears to account for 10% (25). Interestingly, the sequence motifs that define H2AX and H2AZ as separate variants from the bulk H2A species are found in many eukaryotic organisms, suggestive of a conserved role. H2AX variants contain a conserved SQ(E/D) motif in the C-terminal tail, and interestingly, this motif appears to be somewhat mobile in evolution. For example, in Drosophila, the histone H2A variant, H2Av, is clearly a member of the H2AZ family, but contains an SQ motif at its C-terminus. Moreover, in S.cerevisiae, the two genes that encode the bulk of cellular H2A, HTA1 and HTA2, also have the SQE motif at the C-terminus of the protein in common with mammalian H2AX molecules.
Both H2AZ and H2AX have been implicated in DNA damage responses (26–28), although H2AZ appears to also be important for aspects of gene regulation (25). However, no known covalent modifications of histone H2AZ have been reported in conjunction with this activity, so we will focus exclusively on the role of H2AX (and H2A proteins containing X tail motifs) in DNA DSB responses.
H2AX phosphorylation
As mentioned above, H2AX variants possess a highly conserved serine residue, located four amino acids from the C-terminus followed by a glutamine, and this motif also exists on the major replication linked H2A species in many lower eukaryotes. The SQ motif of H2AX is rapidly phosphorylated upon DNA damage (a modification commonly referred to as -H2AX) at sites of DNA DSBs, suggestive of a direct role in the detection, signaling or repair of the lesion itself [(29) and references therein]. The presence of phosphorylated H2AX in mammalian cells is now extensively used as a measure of DNA DSBs. However, caution should be used in this interpretation, as several recent reports indicate that H2AX phosphorylation can also occur in response to cellular events that do not result in DNA DSBs (30,31).
It has subsequently been shown that the C-terminal SQ motifs in numerous eukaryotic organisms are phosphorylated in response to DNA damage, including S.cerevisiae, S.pombe, Xenopus laevis and Drosophila melanogaster (21,32–36). In these studies, phosphorylation was investigated after induction of DNA damage, but consistent with the recent mammalian H2AX findings, at least one case of H2A phosphorylation of the SQ motif has been demonstrated in the absence of DNA lesions (37). Although it has not yet been investigated in great depth, phosphorylation of this motif may occur in response to a variety of cellular stresses, including, but not limited to, DNA DSBs.
The SQ motif found in these histone variants is, in fact, a very good consensus site for the DNA damage dependent PIKK family of kinases described earlier (38). Not surprisingly, it was found that the DNA-damage dependent phosphorylation was being carried out by the PIKK homologs Mec1 and Tel1 in budding yeast (21,32), and by ATM, ATR and DNA-PK in mammalian cells (18,39,40). There is some evidence from the higher eukaryotes that the different PIKKs may play slightly different roles in phosphorylating H2AX. For example, Stiff et al. (41) have shown that ATM and DNA-PK function redundantly to phosphorylated H2AXS139 after exposure to IR (in human, mouse and chicken cells). However, ATM appears to be the dominant kinase at least in the early periods of post-irradiation. In contrast, ATR was shown to be important for H2AXS139 phosphorylation induced by UV (42).
H2AX phosphorylation and genome stability
The fact that H2A and H2AX phosphorylation of the SQ motif is so highly conserved is strongly suggestive of a central role in DNA damage responses. Consistent with this hypothesis, yeast strains lacking this motif are sensitive to DNA damaging agents such as phleomycin, MMS and CPT (21,32). Interestingly, however, they are not sensitive to other types of DNA damage, such as UV irradiation or EMS (21). These data suggest that the SQ motif may be important only for the repair of DNA DSBs. Genetic analyses of the H2A SQ motif with known DNA DSB repair genes suggest that the H2A SQ motif may impinge on both HR and NHEJ activities (21). Notably, however, the DNA damage sensitivity of a strain lacking the SQ motif is drastically less than that of a strain lacking genes required for either NHEJ or HR. This suggests that appropriate DNA DSB repair responses in yeast are facilitated by, but not dependent on, the H2A SQ motif.
To determine the physiological role of H2AX in mammalian cells, two groups (28,43) produced targeted disruptions of mouse H2AX (H2AX–/–). Consistent with results obtained in yeast, H2AX–/– deficient embryonic stem cells were hypersensitive to IR; but H2AX was not essential for survival. H2AX–/– mice were, however, growth retarded, and embryo fibroblasts from them proliferated poorly in vitro owing to premature senescence. Interestingly, these phenotypes resembled the ones of mice deficient for Ku80, Ku70 or ATM. As with Ku or ATM deficient murine fibroblasts, H2AX–/– cells exhibited elevated levels of both spontaneous- and IR-induced genomic instability. The absence of H2AX impaired DNA repair caused by IR, probably accounting for an increased radiation sensitivity of H2AX–/– mice (28,43). Examining lymphoid development in H2AX–/– mice, no severe impediment in V(D)J recombination, which is NHEJ dependent, was observed suggesting that H2AX is not required for NHEJ in that particular context (28). By contrast, male H2AX–/– mice were infertile, as spermatocytes arrested in the pachytene stage of meiosis I and underwent apoptosis, suggestive of a defect in HR (43).
These two groups further investigated the role of H2AX in mammalian DNA DSB responses by examining the phenotypes of the offsprings of H2AX–/+ and p53–/+ mice (27,44). While H2AX–/– mice exhibited only a modest predisposition to lymphomas, either haploid or diploid mutations in H2AX in combination with the absence of the tumor suppressor p53, severely predisposed mice to various forms of cancer. H2AX–/+/p53–/– and H2AX–/–/p53–/– exhibited dramatic genomic instability, which led to clonal translocations. The finding that haploid insufficiency also results in increased genomic instability and cancer incidence in the absence of p53 is particularly intriguing, and indicates that H2AX expression levels are crucial for its functions. Interestingly, human H2AX maps to a genomic region that exhibits ‘loss of homology’ in a large number of human cancers. This region has been proposed to possess an unidentified tumor suppressor gene, making H2AX an excellent candidate.
Through the analysis of tumors from H2AX–/+ p53–/– and H2AX–/– p53–/– mice, Celeste et al. and Bassing et al. (27,44) showed that H2AX is required for normal processing of G1 phase DSBs in the context of V(D)J recombination, demonstrating a role for H2AX in NHEJ. Taken together, H2AX, like the H2A SQ motif in yeast, may impinge on both major DNA DSB repair pathways.
H2AX phosphorylation and foci formation
Several factors known to be involved in DNA repair and signaling the presence of damage have been shown to accumulate in large nuclear domains (foci) after DNA double-strand breakage. This response is not yet fully understood but has been suggested to be a visual indication of DNA repair centres. Interestingly, several studies have demonstrated that H2AXS139 phosphorylation is important for foci formation under a vast range of conditions where DSBs are formed (45).
Following H2AXS139 phosphorylation, the product of the BRCA1 tumor suppressor gene and later, the Rad50 and Rad51 repair factors colocalize with phospho-H2AX foci (46). DNA DSBs generated by either a laser scissors apparatus or a 137Cs source of IR, show that the initial pattern of phospho-H2AX molecules formed in the nucleus corresponded to Brca1, Rad50 and Rad51 IR-induced foci (IRIF), which appear subsequently in the recovery. Interference with H2AX phosphorylation by the use of the PIKK inhibitor wortmannin or the use of a kinase defective cell line (DNA-PK absent and ATM at extremely low levels), inhibits the initiation of focus formation (46), proving further evidence for the role of H2AX phosphorylation in the process.
H2AX works with mediator of DNA damage checkpoint protein 1 (MDC1) to promote the recruitment of repair proteins to the sites of DNA breaks, besides controlling the damage-induced cell-cycle arrest checkpoints (47,48). H2AX is also important for both MDC1 and 53BP1 damage-induced phosphorylation; and peptides representing the C-tail of H2AX specifically recruit MDC1 and 53BP1 proteins in a phosphorylation-dependent manner. Interestingly, depleting cells of MDC1 protein by siRNA, significantly affected H2AX phosphorylation and phospho-H2AX foci formation, after exposure to both IR and UV (47). Therefore, in response to DNA damage, MDC1 and H2AX appear to form a complex at sites of DNA lesions and are phosphorylated in a mutually dependent fashion.
Surprisingly, H2AX is actually dispensable for the initial recruitment of DNA repair factors to sites of DSB (17). H2AX is important, however, for the retention and the subsequent increase in the concentration of repair factors at sites of DNA damage, which can be visualized as IRIF. Although the physiological function of IRIF is not clear, the concentration of DNA DSB repair and signaling factors in the vicinity of a DSB is thought to facilitate DNA repair and amplify the damage induced checkpoint signal.
H2A and checkpoints
The PIKK family of kinases is vital to instigating a signal transduction cascade resulting in the arrest of progression through the cell cycle as well as transcriptional upregulation of a subset of genes. Therefore, the DNA damage sensitivity detected in strains lacking either the SQ motif or the entire H2AX gene may be the result of impaired signal transduction responses. The transcription of well characterized DNA damage responsive genes was examined in yeast lacking the H2A SQ motif and was found to be normal (21), suggesting that this function is not affected. Moreover, experiments where budding yeast lacking a phosphorylatable H2AS129, were treated with the topoisomerase inhibitor CPT, suggested that H2AS129 phosphorylation is not involved in the activation of the intra-S checkpoint but rather in the efficiency of DNA repair (32). H2AS129 phosphorylation has a central and important function in the S-phase repair of DNA lesions that do not activate the intra S-phase checkpoint (32). Similarly, no detectable G2/M checkpoint defect of S.cerevisiae H2AS129stop mutants exposed to MMS was detected (21). Taken together, these data suggest that the PIKK-dependent signal transduction cascade is unimpaired in the absence of the H2A SQ motif. However, it was recently reported that H2A phosphorylation in fission yeast is important for prolonged checkpoint arrest in response to IR and bleomycin (34).
Consistent with the results obtained in budding yeast, the IR-induced G1/S and G2/M checkpoint functions in mammalian cells lacking H2AX were found to be intact after high doses. Interestingly, however, at lower doses of IR, H2AX was required for G2 arrest (49). It is possible that the different results obtained are due to evolutionary differences between the organisms tested, different mutations in the SQ motif or different assay conditions. In any event, it will be of great interest to further explore the potential role of the H2A SQ motif in checkpoint responses.
Chromatin modulation
There are no data available to indicate whether H2AX phosphorylation will directly impinge on chromatin structure. Nevertheless, there is evidence to suggest that chromatin is altered in vivo in a manner dependent on H2AX. For example, H2AX is required for chromatin condensation and transcriptional silencing of the sex chromosomes during spermatogenesis (50). In addition, changing the phosphorylated serine to a glutamine residue in budding yeast results not only in wild-type levels of survival in the presence of DNA damage but also in less condensed chromatin (21).
Interestingly, phosphorylated yeast H2AS129 but not unphosphorylated H2A interacts specifically with Arp4, a protein present in the NuA4, SwrC and Ino80 chromatin modifying complexes (51), raising the possibility that H2A phosphorylation may indirectly affect chromatin structure by recruiting proteins which modify the local environment. Indeed, evidence shows that all three of these complexes are recruited to the sites of DNA damage in an Arp4-dependent manner (51). Consistent with these results, two recent reports also found that the Ino80 complex is recruited to the sites of DNA damage in a manner dependent on the phosphorylation status of H2A (52,53), although different conclusions as to the mechanism of recruitment were obtained (54).
If different chromatin modifying complexes are recruited to different locations depending on the stage of cell cycle, differentiation or tissue type, then phosphorylation of H2AX could mediate both condensation and decondensation of chromatin. This may partially explain the opposing conclusions reached by studies performed in yeast and mice described above.
Other histone H2A modifications
The phosphorylation of the SQ motif is the most well studied DNA damage-dependent histone modification. However, there is evidence that other covalent modifications of histone H2A proteins will be important for DNA damage responses. First, in addition to H2AX S139 DSB induced phosphorylation, mammalian H2AX is also phosphorylated at S136 upon DSB induction, although to a lesser extent (25). The significance of this, as well as of the responsible kinase is unclear, although a reasonable hypothesis would involve the PIKK family of kinases, as S136 is also followed by a glutamine residue. It is important to note that the studies of H2AX knockout mice have provided us with enormous insights into the role that the gene is playing in DNA repair responses, but that they may not be reflective of the exclusive role of H2AX S139 in these events.
Recently, Drosophila H2A has been shown to be phosphorylated on T119 (55). While the authors did not demonstrate a role for this residue in DNA DSB responses, we have recently found that mutation of the analogous residue, S122 in S.cerevisiae H2A, results in hypersensitivity to phleomycin and defective sporulation (56). Furthermore, by performing 2D gel electrophoresis of steady-state yeast cells labeled with 32P, Wyatt et al. (24) found that H2A S122 is indeed phosphorylated, along with T126 and S129. Moreover, Wyatt et al. found that strains with a threonine to alanine mutation at position 126 of H2A are hypersensitive to bleomycin. Together, these data suggest that covalent modifications of other residues in the C-terminal tail of H2A are important for DNA damage responses, and there might be a complex interplay between the phosphorylation events in their ability to facilitate survival. Yeast H2A T126 may be the analogous residue to the phosphorylated mammalian H2AX S126, although in yeast it is not followed by a glutamine residue, raising the possibility that it is not a target for the PIKKs. Notably, however, yeast H2A S122 is not a part of the X-type tail and is conserved as a phosphorylatable residue, not only in Drosophila but also throughout eukaryotes, in the major replication-linked H2A1 and H2A2 families as well as being present in a number of other H2A variants.
|
H2B |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
The N-terminal tail of histone H2B is flexible and protrudes out of the nucleosomes like the other histone N-terminal tails (57
,58). In contrast, histone H2B has a C-terminal extension of similar size to histone H2A, but in structural studies, this ‘tail’ forms an 
-helix. Nevertheless, the C-terminal -helix protrudes from the nucleosome core and is thus accessible for interactions with DNA, adjacent nucleosomes or regulatory factors.
H2B ubiquitination
Histone H2B can be mono-ubiquitinated on the C-terminal -helix at K123; a process dependent on the ubiquitin conjugating enzymes Rad6 and Bre1 (59–61). This covalent modification regulates H3 methylation and gene silencing in yeast, whereas H3 does not affect H2B ubiquitination (62), demonstrating the intricate interplay between different histone modifications. This is the first demonstration of a unidirectional ‘trans-tail’ histone modification where a covalent modification of one histone tail is dependent on a different histone tail (63).
Recent findings indicate that H2B ubiquitination affects H3 methylation through the recruitment of the proteosomal ATPases Rpt4 and Rpt6 to chromatin (64). Mutations in Rpt6 and Rpt4 affect therefore, global levels of dimethylated and trimethylated H3K4 and H3K79, leading to loss of telomeric gene silencing, probably by titrating Sir proteins away from heterochromatic regions.
Rad6 is important for DNA damage responses and strains with rad6 mutations are sensitive to UV irradiation. One obvious possibility, therefore, is that H2B ubiquitination is important for mediating DNA repair. In support of this, H2BK123R mutants are unable to sporulate properly (59). However, unlike rad6 mutant strains, H2BK123R mutant strains are not sensitive to UV (59).
Intriguingly, the defect in sporulation is not due to an inability to repair DNA DSBs, but due to the inability to generate them (65). Strains defective in rad6, bre1 or H2BK123 have lower DSB formation during meiosis, yet DSBs ectopically induced by the use of Gal4-Spo11 are normal in these mutant strains. Therefore, the authors speculate that histone H2B ubiquitination might be important for the recruitment and/or stabilization of the DSB forming machinery during meiosis.
Because the DNA damage is never generated in the first place, it is not known whether H2B ubiquitination would also be important in mediating DNA repair events after the induction of DSBs during meiosis. It is plausible to speculate that a modification required for a programmed generation of DSBs might also be involved in facilitating its repair either through the rapid recruitments of repair factors to the break site or by anchoring the DNA ends together. Clearly, it will be interesting to elucidate the potential role of H2B ubiquitination and its downstream effects on H3 methylation, in DNA damage responses, discussed in more detail below.
H2B phosphorylation
H2BS14 phosphorylation is associated with chromatin condensation both in vivo and in vitro. Indeed, Cheung et al. (66) found that H2B is phosphorylated at S14 by the 34 kDa apoptosis-induced H2B kinase (Mst1). Furthermore, a peptide from H2B N-tail has the property of self-aggregating when phosphorylated at S14 (66) and therefore, this modification could play a direct role in regulating chromatin condensation, which is a hallmark event during apoptosis. H2BS14 is conserved in chicken, frog, rat, mouse and human (vertebrates) but absent in fly, worm and yeast. H2BS14 phosphorylation is detectable 2 h following the induction of the apoptotic process (67), in contrast to H2AXS139 phosphorylation, which is virtually instant (18).
Interestingly, mammalian H2BS14 is phosphorylated at the sites of DNA DSBs, indicative of a role for H2B as well as H2AX in mediating DNA repair responses at the site of the lesion (67). However, as with H2AXS139 phosphorylation, H2B S14 phosphorylation forms IRIF with a significant delay and at fewer chromatin locations (67). This may indicate that the involvement of H2BS14 phosphorylation in DNA damage responses is a late event, or that this phosphorylation marks a subset of DSBs; e.g. irrepairable DSBs or abnormal chromosome rearrangements induced by excessive DNA damage. Alternatively, low but physiologically relevant levels of H2BS14 phosphorylation may be present earlier during the process. Consistent with this possibility, H2BS14 phosphorylation was detected as early as 1 min using the laser scissors technique instead of IRIF formation. This overlapped well with H2AXS139 phosphorylation kinetics. Therefore, although the detection of phosphorylated H2BS14 in IRIF is not visible at early time points, phosphorylation of Ser-14 in H2B occurs rapidly at sites of DSBs.
Interestingly, Fernandez-Capetillo et al. (67) found that H2BS14-P shows a similar staining pattern to that of H2AXS139-P in mouse spermatocytes, being particularly enriched in the highly compacted XY chromosome. Therefore it is possible that phosphorylation of H2BS14 can act in concert with H2AX phosphorylation to promote appropriate chromatin structure of the chromosomes in the sex body as well as DNA damage responses elsewhere in the genome.
|
H3 |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
In mammalian cells, there is evidence that phosphorylation of two residues, S10 and S28 of the histone H3 N-terminal tail may be important for mediating UV irradiation responses. Utilizing western blotting approaches in combination with antibodies against specific phospho-residues, work from Dong et al. (68
) generated a series of articles demonstrating the importance of mammalian MAPK cascades for the UVB-induced phosphorylation of H3S10 and H3S28. They found that UVB strongly induced H3 phosphorylation in vivo. Moreover, UVB-induced histone H3S10 phosphorylation was mediated by ERKs and p38, whereas UVB-induced histone H3S28 phosphorylation was mediated by MSK1, Erks, p38 and JNKs pathways (69,70). Interestingly, there appears to be a coupling between mammalian H3S28 phosphorylation and mitotic chromosome condensation (70). In contrast, phosphorylation of H3S10 has been shown to be associated with regions of relatively decondensed chromatin in other studies (71). Whether changes in condensation occur in response to UV-irradiation, and whether these changes, if any, are mediated by H3 phosphorylation is not known. Moreover, it is not known whether these modifications play any functional role in the repair of UV-induced DNA lesions.
The role of the budding yeast acetyltransferase Hat1 and histone H3 in DNA repair was recently investigated, and it was found that both Hat1 and specific residues in the H3 N-tail play a role in DNA DSB repair though recombination repair (22). By substituting H3 lysine residues 9, 14, 18, 23 and 27 for arginine in different combinations and observing cell growth in the presence of MMS, the authors determined that the presence of a single lysine residue either at position 14 or at position 23 was necessary for DNA damage repair at levels close to wild type. When lysines 14 and 23 were replaced with glutamine (H3K14,23Q), which partially mimics acetylation, they found that sensitivity to MMS was decreased, supporting the idea that acetylation of H3K14 or H3K23 is important for the repair of DNA damage. Combining different H3 lysine mutations with a hat1 null mutant, they found that triple mutation in lysines 9,18 and 27 (H3K9,18,27R) was sensitive to MMS when HAT1 was absent. Accordingly, the viability of these mutant strains was significantly reduced when DSBs were induced using EcoRI. Using a plasmid repair assay the authors found that hat1
or hat1 H3K9,18,27R had normal NHEJ, whereas H3K14,23R mutants had the plasmid repair capacity reduced to 60% of wild-type levels. Inducing the HO endonuclease in plate assays, the authors found a high degree of sensitivity of H3K14,23R mutants to HO induction, which is indicative of a role of histone H3 lysines 14 and 23 in DNA DSB repair by HR. Furthermore, in this assay the authors found an increased sensitivity of the H3K14,23R allele to HO induction when combined with a hat1 mutant. As expected, the H3K9,18,27R allele in combination with a hat1 mutant was also sensitive to HO induction. Taking advantage of ‘in vivo biochemistry’ methodologies similar results were obtained, indicating that mutations affecting both Hat1 and the acetylable lysine residues in the histone H3 N-tail compromise the repair of DSBs by HR.
Although not on the histone tail, methylation of histone H3 at lysine 79 has recently been implicated in DNA damage responses (72). Methylation of this residue provides a binding site for interaction with the evolutionarily conserved checkpoint proteins 53BP1 (human) and Rad9 (S.cerevisiae), and this is necessary for the recruitment of these proteins to the sites of DNA damage (72). Perplexingly, histone H3 is constitutively phosphorylated at K79, which is not consistent with a mechanism for regulated recruitment in response to DNA damage. The authors, therefore, propose that the methylated residue is uncovered at the sites of DNA damage, thus presenting the docking site for repair proteins. One intriguing possibility is that this reorganization is dependent on the chromatin modifying activities brought to DNA DSBs by H2A phosphorylation (51–53).
|
H4 |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
Four lysines, at positions 5, 8, 12 and 16, in the N-terminal tail of histone H4 are reversibly acetylated in vivo in all eukaryotes (23
). Interestingly, histone acetyl transferases (HATs) associate physically with the Ku70 DNA repair protein and acetylation of Ku70 influences apoptotic responses [for review, see (73)], which suggests that histone acetylation may be important in DNA damage responses.
Recently, it was shown that yeast strains carrying mutations in all H4 N-tail lysines (hhf1-10) have a pronounced defect in genome integrity (23). The hhf1-10 mutant is markedly hypersensitive to both CPT and MMS but not even modestly sensitive to UV, suggesting the presence of a defect in DNA DSB repair, but not a global defect in all DNA repair processes. Interestingly the reintroduction of a single lysine, even at ectopic locations, is able to rescue hhf1-10 sensitivity to DSB generating agents. This rescue was shown to be due to the lysine acetylation in vivo, demonstrating that lysine acetylation is required, either directly or indirectly, for the correct repair of DNA DSBs. The HAT responsible for this activity is Esa1, an essential protein whose DSB repair role can be separated from its essential function (23). Esa1 is a subunit of the NuA4 HAT complex (74).
In plasmid repair assays, hhf1-10 mutants were defective in NHEJ, and unlike the sensitivity to CPT, mutants containing single lysines in any position could not rescue this defect (23). One interpretation of these results is that there are at least two distinct repair pathways that are defective in hhf1-10 mutant strains: an NHEJ pathway that requires acetylation of more than one lysine at the H4 N-tail, which has a minor role in cellular resistance to CPT or MMS, and a distinct pathway that requires acetylation of any one of the H4 N-tail lysines and is the principal determinant of DSB repair capacity. As the latter pathway is important in the repair of CPT-induced DSB (at the replication fork) but not DSB caused by IR (random), this pathway was named replication-coupled pathway. To further understand the replication-coupled pathway, the authors carried out a suppressor screen to identify genes whose overexpression restored CPT resistance to hhf1-10 and identified Arp4 (23). Interestingly, Arp4 is also a component of the NuA4 HAT complex, and thus independent lines of investigation implicated the NuA4 complex in facilitating DNA repair.
As mentioned above, Arp4 is also a component of the budding yeast ATP-dependent chromatin remodeling complexes Ino80 and SwrC. Interestingly, in higher eukaryotes, including Drosophila and mammals, clear homologs of proteins present in these complexes exist and are found in the TIP60 complex. This complex contains both HAT activity and ATP-dependent chromatin remodeling activity. The mammalian TIP60 histone acetylase complex (which contains many components homologous to the yeast NuA4 complex) was shown to be involved in DNA repair and apoptosis (75). By inducing DSBs in wild-type or mutated TIP60 carrying cells and performing pulse-field gel electrophoresis at various times after -irradiation, the authors demonstrated that the TIP60 complex is important for the efficient repair of DNA DSBs. Interestingly, it was recently demonstrated that the Drosophila dTip60 complex is crucial for the optimal exchange of phosphorylated H2Av/H2B heterodimers from sites of DNA damage (76). Moreover, this exchange was more efficient on substrates that had a phospho-mimicking glutamic acid residue at the SQ motif of H2Av (76), again linking the phosphorylation status of H2AX with these chromatin remodeling activities.
In addition, recent data from fission yeast have suggested that Set9 methylation of histone H4 at lysine 20 is important for DNA damage responses (77). As with H3 methylation at lysine 79 (72), methylation of H4 K20 creates a binding site for the S.pombe checkpoint protein Crb2 (homolog of budding yeast Rad9 and human 53BP1) and is likewise critical for the recruitment of this protein to sites of DNA damage (77). Also as in H3 K79 methylation, the levels of H4 K20 methylation are not regulated in response to DNA damage, and similar mechanisms of exposing the modified region of the histone after DNA damage are likely to exist.
|
Conclusions |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
There is increasing evidence for a role of chromatin in DNA damage responses, and studies implicating histone modifications and chromatin modulation to genome stability abound.
The number of histone modifications already implicated in DNA DSB repair is clearly enough to provide significant complexity in the histone code, although not all of these will necessarily be located and act at the site of the DNA lesion. Nevertheless, it is clear that there will be more histone modifications with roles in DNA DSB repair uncovered as the field progresses. If one considers that in mammalian cells one single DSB results in the phosphorylation of H2AXS139 as far as 2 Gb from the damaged site, the potential for combinations of modifications and recruitment of factors is enormous. Adding to this is the likelihood that histone modification patterns and kinetics will be influenced by variables such as cell-cycle stage, cellular origin and chromatin context at the site of a break.
Histone motifs and modifications affecting the cellular resistance to DNA damaging agents that do not act directly at the site of the break are of course still of great physiological importance. These can affect the transcription of DNA repair proteins or global chromatin conformation and therefore its susceptibility to damage. It will be extremely informative to examine these roles and their interplay with chromatin activities occurring at the sites of DNA damage.
Ultimately, these studies should have significant impact on the clinical side of DNA damage responses and genomic instability, including the treatment of tumors. Chemical inhibitors of chromatin modulation activities have already been successfully used in the treatment of tumors (78), although it is not known whether these affect DNA DSB repair activities. In addition, H2AX phosphorylation levels in IR treated tumors proved to be a good predictor of tumor sensibility to radiotherapy (79), allowing more precision in the choice of radiotherapy doses. Moreover, the authors found that a peptide mimicking the H2AX C-terminal tail was capable of antagonizing H2AX functions and enhanced cell death in irradiated, radioresistant tumor cells. With the identification of DNA damage-dependent chromatin modifying activities, comes the possibility of new therapeutic or diagnostic approaches.
|
Notes |
|---|
* To whom correspondence should be addressed. Tel: +01223 333 663; Fax: +01223 766 002; Email: jad32{at}mole.bio.cam.ac.uk
|
References |
|---|
|
TopAbstractIntroductionTalking to histonesGenerating double strand breaksH2AH2BH3H4ConclusionsReferences |
|---|
-
1. Flaus,A. and Owen-Hughes,T. (2004) Mechanisms for ATP-dependent chromatin remodelling: farewell to the tuna-can octamer? Curr. Opin. Genet. Dev., 14, 165–173.[CrossRef][ISI][Medline]
2. Jenuwein,T. and Allis,C.D. (2001) Translating the histone code. Science, 293, 1074–1080.
3. West,S.C. (2003) Molecular views of recombination proteins and their control. Nat. Rev. Mol. Cell Biol., 4, 1–11.[Medline]
4. Lieber, M.R., Ma,Y., Pannicke,U. and Schwarz,K. (2003) Mechanism and regulation of human non-homologous DNA end-joining. Nat. Rev. Mol. Cell Biol., 4, 712–720.[CrossRef][ISI][Medline]
5. Jackson,S.P. (2002) Sensing and repairing DNA double-strand breaks. Carcinogenesis, 23, 687–696.
6. Peterson,C.L. and Cote,J. (2004) Cellular machineries for chromosomal DNA repair. Genes Dev., 18, 602–616.
7. Prado,F., Cortes-Ledesma,F. and Aguilera,A. (2004) The absence of the yeast chromatin assembly factor Asf1 increases genomic instability and sister chromatid exchange. EMBO Rep., 5, 497–502.[CrossRef][ISI][Medline]
8. Moggs,J.G., Grandi,P., Quivy,J.P., Jonsson,Z.O., Hubscher,U,. Becker,P.B. and Almouzni,G. (2000) A CAF-1-PCNA-mediated chromatin assembly pathway triggered by sensing DNA damage. Mol. Cell. Biol., 20, 1206–1218.
9. Owen-HughesT.A., Utley,R.T., Cote,J., Peterson,C.L. and Workman,J.L. (1996) Persistent site-specific remodeling of a nucleosome array by transient action of the SWI/SNF complex. Science, 273, 513–516.[Abstract]
10. D'Arpa,P., Beardmore,C. and Liu,L.F. (1990) Involvement of nucleic acid synthesis in cell killing mechanisms of topoisomerase poisons. Cancer Res., 50, 6919–6924.
11. Nickoloff,J.A. and Hoekstra,M.F. (1998) DNA damage and repair. In Nickoloff,J.A. (ed.), Contemporary Cancer Research. Humana Press, Totowa, NJ, Vols I and II, pp. 1–32.
12. Sugawara,N., Wang,X. and Haber,J.E. (2003) In vivo roles of Rad52, Rad54 and Rad55 proteins in Rad51-mediated recombination. Mol. Cell, 12, 209–219.[CrossRef][ISI][Medline]
13. Plessis,A., Perrin,A., Haber,J.E. and Dujon,B. (1992) Site-specific recombination determined by I-SceI, a mitochondrial group I intron-encoded endonuclease expressed in the yeast nucleus. Genetics, 130, 451–460.