PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Mol Cell. Author manuscript; available in PMC 2010 May 15.
Published in final edited form as:
PMCID: PMC2744111
NIHMSID: NIHMS112078

Formation of Dynamic γ-H2AX Domains along Broken DNA Strands is Distinctly Regulated by ATM and MDC1 and Dependent upon H2AX Densities in Chromatin

SUMMARY

A hallmark of the cellular response to DNA double strand breaks (DSBs) is histone H2AX phosphorylation in chromatin to generate γ-H2AX. Here, we demonstrate that γ-H2AX densities increase transiently along DNA strands as they are broken and repaired in G1 phase cells. The region across which γ-H2AX forms does not spread as DSBs persist, rather γ-H2AX densities equilibrate at distinct levels within a fixed distance from DNA ends. Although both ATM and DNA-PKcs generate γ-H2AX, only ATM promotes γ-H2AX to maximal distance and maintains γ-H2AX densities. MDC1 is essential for γ-H2AX formation at high densities near DSBs, but not for generation of γ-H2AX over distal sequences. Reduced H2AX levels in chromatin impair the density, but not the distance of γ-H2AX formed. Our data indicate that H2AX fuels a γ-H2AX self-reinforcing mechanism that retains MDC1 and activated ATM in chromatin near DSBs and promotes continued local phosphorylation of H2AX.

INTRODUCTION

An early component of the evolutionarily conserved DNA damage response (DDR) is the phosphorylation of H2A histones in chromatin around breakage sites (Kinner et al., 2008; Rogakou et al., 1998). In mammalian cells, the H2AX histone variant comprises 2–25% of the H2A pool and is unevenly distributed throughout chromatin (Bewersdorf et al., 2006; Rogakou et al., 1998). H2AX in chromatin is rapidly phosphorylated on serine 139 (to form γ-H2AX) for distances estimated to extend 1–2 megabases around DNA lesions (Rogakou et al., 1999; Rogakou et al., 1998). H2AX phosphorylation can be visually detected by immune-fluorescence with the read-out discrete nuclear regions (“foci”) where the antibody accumulates (Rogakou et al., 1999). Upon radiation induced DSBs, ATM is required for γ-H2AX foci of normal size and intensity (Burma et al., 2001; Paull et al., 2000), though DNA-PKcs can substitute for ATM (Hickson et al., 2004; Stiff et al., 2004). Formation of γ-H2AX creates binding sites for DDR factors (Kinner et al., 2008), some of which promote γ-H2AX post-translational modifications that form binding sites for additional DDR factors (Huen et al., 2007; Ikura et al., 2007; Mailand et al., 2007). The H2AX dependent assembly of DDR factors into complexes around DNA breakage sites may promote accessibility of broken DNA ends, stabilize disrupted DNA strands, and amplify DDR signals (Rogakou et al., 1998; Stucki and Jackson, 2006; Kinner et al., 2008).

H2AX is a dosage-dependent genomic caretaker and tumor suppressor. H2ax−/− cells exhibit increased sensitivity to radiation, elevated levels of genomic instability, defective chromosomal DSB repair, and an impaired G2/M checkpoint (Bassing et al., 2002a; Celeste et al., 2002; Fernandez-Capetillo et al., 2002; Franco et al., 2006; Ramiro et al., 2006; Xie et al., 2004). DDR factors re-localize to DNA breakage sites in H2ax−/− cells, however formation of γ-H2AX is essential for the retention of these factors in chromatin around DSBs (Celeste et al., 2003b). H2ax+/− cells express H2AX in chromatin at 50% the level of H2ax+/+ cells and exhibit similar phenotypes as H2ax−/− cells (Bassing et al., 2003; Celeste et al., 2003a), indicating that expression of both allelic copies of H2AX is required for a fully effective DDR. H2ax+/− mice are not tumor-prone (Bassing et al., 2003; Celeste et al., 2003a), but H2ax−/− mice exhibit an increased predisposition to thymic lymphomas (Bassing et al., 2003). Moreover, H2ax+/−p53−/− and H2ax−/−p53−/− mice develop lymphomas and solid tumors with clonal translocations (Bassing et al., 2003; Celeste et al., 2003a). The human H2AX gene (H2AFX) maps to a cytogenetic region (11q23) often deleted on one allele in cancers (Bassing et al., 2003), suggesting similar dosage-dependent functions of H2AX in man.

Lymphocyte development involves the assembly of antigen receptor genes from germline variable (V), diversity (D), and joining (J) gene segments. The lymphocyte-specific RAG1/RAG2 (RAG) endonuclease initiates V(D)J recombination through the induction of DSBs between participating gene segments and their flanking RAG recognition sequences (Bassing et al., 2002b). RAG cleavage results in generation of a pair of blunt signal ends (SEs) and a pair of hairpin-sealed coding ends (CEs) that are processed and joined by the classical non-homologous end-joining (NHEJ) proteins to form a signal join (SJ) and coding join (CJ), respectively (Bassing et al., 2002b). DNA-PKcs forms an endonuclease with Artemis that is required to open CEs (Ma et al., 2002), ATM helps maintain CEs in proximity while they are joined (Bredemeyer et al., 2006). H2AX is not required for CJ formation to the extent of DNA-PKcs and ATM (Bassing et al., 2002a; Celeste et al., 2002). However, H2ax−/− mice develop lymphocytes with elevated levels of Tcr translocations and succumb to RAG-dependent pro-B lymphomas on a p53-deficient background (Bassing et al., 2007; Bassing et al., 2003; Celeste et al., 2003a; Celeste et al., 2002), consistent with the notion that H2AX coordinates DSB repair, signaling, and surveillance (Chen et al., 2000).

Since H2AX deficiency and haploinsufficiency lead to increased cancer predisposition in mice (Bassing et al., 2003; Celeste et al., 2003a), elucidating mechanisms by which H2AX functions is important to determine how mammalian cells integrate DSB repair with cellular proliferation and survival. Despite intense efforts, the molecular nature and function of the γ-H2AX domain formed around DNA breakage sites and the manner by which DDR factors regulate this γ-H2AX domain remain to be determined. The ATM/MDC1 dependent generation of γ-H2AX across large chromosomal distances may be important to coordinate DSB repair with cellular proliferation and survival by forming a sufficient number of binding sites for DDR proteins in chromatin around DNA breakage sites (Lou et al., 2006; Stucki and Jackson, 2006). However, considering that H2ax+/− cells contain H2AX in chromatin at 50% normal levels and exhibit increased genomic instability (Bassing et al., 2003; Celeste et al., 2003a), the density of H2AX, and consequently γ-H2AX, in chromatin around DNA breakage sites also may be critical for a fully effective DDR. Therefore, we used chromatin immunoprecipitation (ChIP) to measure γ-H2AX and H2AX densities in nucleosomes along chromosomal DNA strands as they are broken during V(D)J recombination in G1 phase lymphocytes.

RESULTS

γ-H2AX Densities Increase along Tcrα Locus DNA Strands During V(D)J Recombination in Primary Thymocytes

RAG-dependent γ-H2AX foci have only been reported to co-localize with endogenous Tcrα loci (Chen et al., 2000). Thus, we first used ChIP to measure relative γ-H2AX densities in chromatin along Tcrα loci during V(D)J recombination in primary thymocytes. The murine Tcrα locus consists of 100 segments that span 1.3 megabases (Mb) and 59 segments that span 60 kilobases (kb) and reside telomeric of the cluster (Figure 1A). Tcrα rearrangement occurs in CD4+/CD8+ double-positive (DP) thymocytes through the coupled cleavage of a and a segment, excision of intervening sequences, and joining of and CEs. RAG mediated Tcrα locus DSBs are induced and repaired in wild-type DP cells, which comprise 80–90% of thymocytes from wild-type mice; in contrast, RAG mediated DSBs are not generated in Tcrβ:Rag2−/− DP cells, which comprise a similar percentage of thymocytes from Tcrβ:Rag2−/− mice. By ChIP, we detected significantly higher γ-H2AX densities at the most telomeric segment (Jα1), genomic locations extending 400–600 kb telomeric of Jα1, and sequences within the cluster in wild-type cells, as compared to in Tcrβ:Rag2−/− cells (Figure 1A). Peak γ-H2AX densities (6–8 fold above background) were formed in chromatin along DNA strands for at least 200 kb telomeric from DSBs induced within the cluster. Increased γ-H2AX densities were not detected over sequences between the and segments, which would be located on extra-chromosomal circles containing / SJs.

Figure 1
γ-H2AX forms along Tcrα locus DNA strands broken during V(D)J recombination in primary cells

A symmetric pattern of H2A phosphorylation forms along DNA strands on both sides of DSBs induced in yeast, but only in mutant backgrounds that cannot repair these genomic lesions (Shroff et al., 2004; Unal et al., 2004). In lymphocytes lacking DNA-PKcs kinase activity or Artemis, RAG generated CEs persist as stable hairpins and activate the DDR (Bredemeyer et al., 2008; Guidos et al., 1996). Thus, we also conducted γ-H2AX ChIP on thymocytes isolated from LckCre:Artemis+/+:Tcrβ:p53Flox/Flox (LβP) or LckCre:Artemis−/−:Tcrβ:p53Flox/Flox (LAβP) mice to determine whether the pattern, distance, and/or density of γ-H2AX formation changes when DSBs persist un-repaired. Conditional p53 inactivation was necessary to allow the survival and differentiation of Artemis−/− thymocytes and prevent development of pro-B lymphomas. RAG mediated Tcrα locus DSBs are induced and repaired in LβP DP cells, which comprise 80–90% of thymocytes from LβP mice; in contrast, RAG mediated DSBs are generated but not repaired in LAβP DP cells, which comprise a similar percentage of thymocytes from LAβP mice. By ChIP, we detected significantly higher γ-H2AX densities at Jα1, genomic locations extending 400–600 kb telomeric of Jα1, and sequences within the cluster in LAβP thymocytes, as compared to in LβP thymocytes (Figure 1B). The pattern and distance of γ-H2AX generated along Tcrα DNA strands were similar among LAβP, LβP, and wild-type thymocytes, however γ-H2AX densities were substantially greater in LAβP cells: 20–40 fold versus 6–8 fold. These data suggest that the maximal distance over which γ-H2AX forms along broken DNA strands may be achieved more rapidly than peak γ-H2AX densities near DNA breakage sites.

Since H2AX appears unevenly incorporated throughout chromatin (Bewersdorf et al., 2006), we sought to determine whether potential differences in H2AX distribution influence the γ-H2AX formed along broken Tcrα loci. Because anti-H2AX and anti-γ-H2AX antibodies bind to the non-phosphorylated and phosphorylated carboxy terminal amino acids of H2AX, respectively, an accurate quantification of H2AX densities along Tcrα can only be conducted in Tcrβ:Rag2−/− thymocytes. By ChIP, we found that H2AX densities were similar at all genomic locations assayed within and adjacent to Tcrα in Tcrβ:Rag2−/− thymocytes (Figure 1C). Despite the caveat mentioned above, we also conducted H2AX ChIP on LβP and LAβP thymocytes and found that relative H2AX densities were significantly lower at some locations where γ-H2AX formed in these cells, i.e. at Jα1 and sequences extending 200 kb telomeric (Figure 1C). These data demonstrate that the pattern, distance, and densities of γ-H2AX formed along Tcrα DNA strands broken during V(D)J recombination do not simply reflect corresponding differential distribution of H2AX in chromatin at these genomic locations prior to DSB induction.

γ-H2AX Densities Increase along Immunoglobulin Locus DNA Strands During V(D)J Recombination in Pre-B Cell Lines

The complexity of the Tcrα locus and the asynchronous induction of RAG DSBs in thymocytes provide obstacles for mechanistic studies of γ-H2AX formation and maintenance. We recently developed a cell line based approach to control and monitor the induction and repair of RAG DSBs (Bredemeyer et al., 2006). Treatment of immortalized murine (abl) pre-B cell lines with the STI571 abl kinase inhibitor promotes G1 arrest, Rag1/Rag2 expression, and immunoglobulin kappa (Igκ) locus transcription and rearrangement (Bredemeyer et al., 2006; Muljo and Schlissel, 2003). The murine Igκ locus consists of 140 s that span 3 Mb and four functional s that span 2 kb and reside telomeric of the cluster (Figure 2A). Igκ rearrangement occurs through coupled and cleavage, excision or inversion of intervening sequences, and joining of and CEs and chromosomal SJs for inversional rearrangements (Bassing et al., 2002b). Due to simplicity of the cluster, Southern blot analysis can be used to monitor the induction and repair of RAG induced Igκ DSBs (Figure 2B).

Figure 2
γ-H2AX forms along Igκ locus DNA strands broken during V(D)J recombination in cell lines

To study γ-H2AX formation along DNA strands as they are broken and repaired in G1 phase cells, we generated multiple independent wild-type and Artemis−/−Rag2−/− abl pre-B cell lines and assayed each for Igκ DSBs and γ-H2AX formation over time. By Southern blotting, we detected rearranged Igκ alleles at 24 hours STI571 treatment in wild-type cells, but not in Artemis−/−Rag2−/− cells (Figure 2C). We observed increasing amounts of rearranged Igκ alleles and concomitant loss of germline Igκ alleles at each time-point assayed between 48–120 hours STI571 treatment (Figure 2C). By ChIP at 24 hours STI571 treatment, we detected significantly elevated γ-H2AX densities at all locations assayed within Igκ and 200–400 kb on both sides of Igκ in wild-type cells, but not in Artemis−/−Rag2−/− cells (Figure 2A,D). Peak γ-H2AX densities (6–8 fold above background) were formed along DNA strands for at least 200 kb from DSBs induced within the cluster. Between 48 to 120 hours STI571 treatment, the induction and repair of Igκ DSBs in wild-type cells continued, but at reduced levels (Figure 2C); the RAG-dependent γ-H2AX densities formed at all locations were lower and at each time-point assayed and eventually returned to background levels observed in Artemis−/−Rag2−/− cells (Figure 2D). These data demonstrate that γ-H2AX densities increase transiently across kilobase distances along DNA strands as they are broken and repaired in G1 phase cells.

To investigate whether H2AX densities change upon G1 arrest and concomitant transcription, we conducted H2AX ChIP on un-treated and STI571-treated Artemis−/−Rag2−/− pre-B cells. H2AX densities were similar at all locations assayed in un-treated Artemis−/−Rag2−/− cells (Figure 2E). However, two to three fold increases of H2AX densities at Jκ5 and over sequences 50 kb telomeric and 20 kb centromeric were detected in Artemis−/−Rag2−/− cells treated with STI571 for 72 hours (Figure 2E). These data demonstrate that the pattern, distance, and densities of γ-H2AX formed along Igκ DNA strands are not due to corresponding changes of H2AX densities in chromatin prior to DSB induction. Yet, they also suggest that H2AX may be deposited in chromatin over segments upon activation of transcription.

A Dynamic γ-H2AX Domain is Generated and Maintained along Broken DNA Strands in G1 Phase Cells

To study γ-H2AX distribution along DNA strands as they are broken and remain unrepaired in G1 phase cells, we generated multiple independent Artemis−/− pre-B cell lines and assayed each for Igκ locus DSBs and γ-H2AX formation over time. By Southern blotting, we detected CEs at 24 hours STI571 treatment in Artemis−/− cells (Figure 2C). We observed increasing amounts of CEs and loss of germline Igk alleles at 48 and 72 hours STI571 treatment, but no further increase of either at 96 and 120 hours (Figure 2C). By ChIP at 24 hours STI571 treatment, we detected significantly elevated γ-H2AX densities at all locations assayed within and 200–400 kb on both sides of Igκ in Artemis−/− cells (Figure 3A). The pattern and distance of γ-H2AX generated were similar to those observed at 24 hours STI571 treatment in wild-type cells. However the peak densities at which γ-H2AX formed were substantially greater in Artemis−/− cells: 10–20 fold versus 6–8 fold over sequences 0–70 kb from DSBs. Since the induction of DSBs within a 2 kb wide region can effectively be considered a “site-specific DSB”, we monitored γ-H2AX formation only at and telomeric of Jκ5 sequences at later time points and in all subsequent experiments. The RAG-dependent γ-H2AX densities at these locations were higher at 48 hours STI571 treatment, but then maintained at these levels throughout the remaining time course (Figure 3B,C). Notably, neither the pattern nor distance of elevated γ-H2AX densities changed as more DNA ends were generated and existing DNA ends persisted over time, rather γ-H2AX densities equilibrated at distinct levels over sequences within a fixed distance (Figure 3B,C). These data indicate that factors limit the chromosomal distance across which γ-H2AX spreads from DSBs along DNA strands and also regulate γ-H2AX densities over these sequences to establish and maintain a chromatin domain that marks breakage sites in G1 phase cells.

Figure 3
A dynamic γ-H2AX domain is maintained along broken DNA strands

Since H2AX phosphorylation and γ-H2AX foci are negatively regulated by phosphatases (Chowdhury et al., 2005; Chowdhury et al., 2008; Kimura et al., 2006), we sought to evaluate whether phosphatase activity influences the pattern, distance, and/or density of γ-H2AX formation along broken DNA strands. For this purpose, we treated three pools of Artemis−/− pre-B cells with STI571 for 48 hours and then added the PP2A inhibitor endothall to one pool. At this time, we conducted γ-H2AX and H2AX ChIP on cells from one pool and on un-treated cells. After 24 additional hours, we conducted γ-H2AX and H2AX ChIP on cells of the STI571 and STI571/endothall treated pools. Relative γ-H2AX densities were increased at 5 and sequences extending further than 400 kb telomeric in cells treated only with STI571, but significantly lower at these locations in the cells also treated with endothall (Figure 3D). H2AX densities were not affected by endothall treatment (Figure 3E). These data demonstrate that the maintenance of γ-H2AX at peak levels along broken DNA strands in G1 phase cells requires PP2A phosphatase activity. This finding indicates that the γ-H2AX domain that marks breakage sites in G1 phase cells is dynamic and likely maintained through repeated cycles of H2AX phosphorylation and γ-H2AX de-phosphorylation.

ATM Generates and Maintains the Dynamic γ-H2AX Domain along Broken DNA Strands

Both ATM and DNA-PKcs can generate γ-H2AX foci in G1 phase cells (Burma et al., 2001; Paull et al., 2000; Stiff et al., 2004), however the relative contribution of either kinase in γ-H2AX formation along broken DNA strands has not been reported. Since DNA-PKcs-mediated phosphorylation of Artemis is required to open CEs (Ma et al., 2002), lymphocytes that lack DNA-PKcs kinase activity are equivalent to Artemis−/− cells with respect to accumulation of unrepaired chromosomal DNA breaks (Rooney et al., 2002; Roth et al., 1992). Thus, to evaluate the contributions of ATM and DNA-PKcs in γ-H2AX formation along broken DNA strands, we generated and analyzed multiple independent pre-B cell lines lacking a functional DNA-PKcs kinase (Scid cells) or both Artemis and Atm (Artemis−/−Atm−/−). We found increased γ-H2AX densities at Jκ5 and sequences further telomeric in STI571-treated Scid and Artemis−/−Atm−/− cells, but not in Scid cells treated with STI571 and the KU55933 ATM kinase inhibitor (Figure 4A), indicating that only ATM and DNA-PKcs can generate γ-H2AX along Igκ broken in G1 phase cells during V(D)J recombination. We next assayed for Igκ DSBs and γ-H2AX formation over time in Scid and Artemis−/−Atm−/− cells. In Scid cells, γ-H2AX densities increased with the accumulation and persistence of un-repaired CEs and then equilibrated (Figure 4B); in Artemis−/−Atm−/− cells, γ-H2AX densities dropped slightly with the accumulation of un-repaired CEs (Figure 4C). The patterns, distances, and densities of γ-H2AX formed along broken DNA strands were similar between Scid and Artemis−/− cells, but the distance and densities of γ-H2AX were significantly lower in Artemis−/−Atm−/− cells. These data demonstrate that ATM is the major kinase that promotes and maintains γ-H2AX formation to peak equilibrium densities and across maximal distance along DNA strands broken in G1 phase cells.

Figure 4
ATM generates and maintains a dynamic γ-H2AX domain along broken DNA strands

To further evaluate the role of ATM in maintenance of γ-H2AX around DSBs, we investigated whether the pharmacological inhibition of ATM after DSB induction would alter the pattern, distance, and/or density of γ-H2AX along broken DNA strands. For this purpose, we employed caffeine to inhibit ATM kinase activity and assayed γ-H2AX formation using the same experimental approach by which we inhibited PP2A. Relative γ-H2AX densities were elevated at 5 and sequences extending further than 400 kb telomeric in Artemis−/− cells treated with STI571, but γ-H2AX densities were significantly lower at these locations in Artemis−/− cells also treated with caffeine (Figure 4D). H2AX densities were not affected by caffeine treatment (data not shown). Our data demonstrates that sustained ATM kinase activity is required to maintain γ-H2AX at peak equilibrium densities along broken DNA strands in G1 phase cells. This finding reinforces the notion that the γ-H2AX domain that marks these DNA breakage sites is dynamic and likely maintained through repeated cycles of H2AX phosphorylation and γ-H2AX de-phosphorylation.

ATM Generates γ-H2AX at Distances from DSBs in an MDC1 Independent Manner

Based upon similar reduction in the size and intensity of γ-H2AX foci upon inactivation of ATM or MDC1, it has been hypothesized that MDC1 mediates a feed-forward mechanism that promotes the ATM dependent spread of H2AX phosphorylation into chromatin further and further away from the initial break (Lou et al., 2006; Stucki and Jackson, 2006). To directly test the contribution of MDC1 on ATM dependent γ-H2AX formation along broken DNA strands, we generated multiple independent Artemis−/−Mdc1−/− pre-B cell lines and conducted γ-H2AX ChIP on each, and in parallel on Artemis−/−Atm−/− and control cells. At 72 hours STI571 treatment, the densities of γ-H2AX at Jκ5 and sequences extending 100 kb telomeric were increased to similar levels in Artemis−/−Mdc1−/− and Artemis−/−Atm−/− cells (Figure 5A,B). In contrast, the densities of γ-H2AX over sequences 100–400 kb telomeric of Jκ5 in Artemis−/−Mdc1−/− cells were significantly higher than the γ-H2AX densities at these locations in Artemis−/−Atm−/− cells (Figure 5A,B). We did not observe significant differences in H2AX densities at Jκ5 or sequences telomeric in untreated cells of any genotype (Figure 5C). However, at 72 hours STI571 treatment, slight two to five fold increases of H2AX densities at Jκ5 and sequences 20 kb telomeric were detectable in Artemis−/− and Artemis−/−Mdc1−/− cells, but not in Artemis−/−Atm−/− cells (Figure 5D). These increases may reflect H2AX deposition or exchange along intact transcribing loci and/or ATM mediated chromatin changes along broken Igκ loci. Regardless, our data reveal that MDC1 is essential for ATM mediated formation of γ-H2AX at high densities along Igκ within 0–200 kb of DSBs, but not for ATM dependent generation of γ-H2AX in chromatin over sequences 200–500 kb from DNA breakage sites. These findings demonstrate that the ATM dependent spread of γ-H2AX formation along DNA strands further and further away from DSBs in G1 phase cells does not require MDC1 function.

Figure 5
ATM generates γ-H2AX at distances from DSBs in an MDC1 independent manner

H2AX Levels in Chromatin Influence γ-H2AX Densities Along Broken DNA Strands

Considering that H2ax+/− cells contain H2AX in chromatin at 50% normal levels and exhibit increased genomic instability (Bassing et al., 2003; Celeste et al., 2003a), the density of H2AX in chromatin could influence the pattern, distance, and/or density of γ-H2AX formed along broken DNA strands. To determine the effect of H2AX density on γ-H2AX formation, we generated multiple independent Artemis−/− pre-B cells lines that contain H2ax flanked by loxP sites on both alleles. We transduced these cells with Tat-Cre and isolated derivative cell lines in which H2ax was either intact (Artemis−/−H2ax+/+) or deleted on a single allele (Artemis−/−H2ax+/−). Southern blot analysis demonstrated that equal levels of Igκ DSBs were induced in Artemis−/−H2ax+/+ and Artemis−/−H2ax+/− cells (data not shown). At 72 hours STI571 treatment of Artemis−/−H2ax+/+ cells, γ-H2AX densities were increased 20-fold or greater at 5 and over sequences extending 200 kb telomeric and still above background levels 400 kb telomeric (Figure 6A,B). In marked contrast, at 72 hours STI571 treatment of Artemis−/−H2ax+/− cells, γ-H2AX densities were increased only three to five fold at 5 and sequences extending 200 kb telomeric and even lower 400 kb telomeric (Figure 6A,B). H2AX densities at all locations assayed in Artemis−/−H2ax+/− cells were approximately 50% of the H2AX densities at the corresponding sequences in Artemis−/−H2ax+/+ cells (Figure 6C). The two to four fold increased H2AX densities at Jκ5 and sequences 20 kb telomeric may reflect H2AX deposition or exchange in chromatin along broken Igκ loci and/or intact transcribing loci. Regardless, these data demonstrate that a two-fold reduction of H2AX density in chromatin can lead to a profound defect in the level to which γ-H2AX densities are increased along broken DNA strands, with no significant decrease in the distance across which γ-H2AX forms from breakage sites.

Figure 6
H2AX levels in chromatin influence γ-H2AX densities along broken DNA strands

DISCUSSION

Here, we report the first molecular analysis of γ-H2AX formation in chromatin along chromosomal DNA strands broken at defined genomic locations in mammalian cells. Our data indicates that an evolutionarily conserved component of the DDR is the ATM dependent phosphorylation of H2A histones in chromatin along DNA strands on both sides of DSBs induced in G1 phase cells. ATM and H2AX do not exist in Saccharomyces cerevisiae, rather Tel1, the ATM homologue, phosphorylates the core H2A histone to form γ-H2A in chromatin upon DSB induction (Downs et al., 2000). In G1 arrested cells, Tel1 dependent γ-H2A forms rapidly and equilibrates along DNA strands for approximately 50 kb on both sides of DSBs (Shroff et al., 2004; Unal et al., 2004). Peak γ-H2A levels form 3–5 kb from DSBs, but little γ-H2A is detectable within 1–2 kb of DNA breakage sites (Shroff et al., 2004; Unal et al., 2004). We demonstrate that ATM dependent γ-H2AX densities increase along DNA strands for distances greater than 400 kb from breakage sites, with peak γ-H2AX densities equilibrating over sequences 20–200 kb from DSBs. Due to Tcrα and Igκ locus complexity, we were not able to assay γ-H2AX densities in nucleosomes at broken DNA ends. Yet, we observed lower γ-H2AX densities at Jκ5, which is located 1.5–1.8 kbs from Jκ1 and Jκ2 DSBs. Two molecules of H2A are incorporated into each nuclesosome in yeast; H2AX distribution in mammalian chromatin remains to be determined. Based upon the observation that H2AX comprises 2–25% of the H2A pool in mammalian cells (Rogakou et al., 1998), on average two molecules of H2AX would be incorporated within every six nucleosomes. Thus, the different distances across which γ-H2A versus γ-H2AX forms along broken DNA strands in G1 phase cells may simply reflect that an equal number of H2A histone molecules is phosphorylated in yeast and mammalian cells.

Our data reveal that maintenance of the ATM dependent γ-H2AX chromatin domain along broken DNA strands in G1 phase cells requires sustained activity of ATM and PP2A. In Saccharomyces cerevisiae, Drosophila melanogaster, and mammalian cells, γ-H2AX can be removed from chromatin, de-phosphorylated after removal, and then exchanged for γ-H2AX in chromatin along broken DNA strands (Heo et al., 2008; Ikura et al., 2007; Kimura et al., 2006; Kusch et al., 2004). These mechanisms have been shown or proposed to be important for attenuation of the DDR following repair of DSBs (Heo et al., 2008; Ikura et al., 2007; Kimura et al., 2006; Kusch et al., 2004). Our current observations suggest that repeated cycles of ATM mediated H2AX phosphorylation and γ-H2AX de-phosphorylation, either in chromatin or after γ-H2AX/H2AX exchange, also may be required for effective activation and maintenance of the DDR in G1 phase cells.

Upon induction of DSBs, ATM is activated and distributed between soluble and chromatin bound pools (Andegeko et al., 2001). Normal retention of activated ATM in chromatin around DNA breakage sites requires physical interaction with MDC1, which binds γ-H2AX (Lou et al., 2006; Stewart et al., 2003; Stucki and Jackson, 2006). Based upon similar diminished size and intensity of γ-H2AX foci in cells lacking ATM or MDC1, it has been hypothesized that MDC1 drives a feed-forward mechanism that promotes the ATM dependent spread of H2AX phosphorylation into chromatin further and further away from initial DSBs (Lou et al., 2006; Stucki and Jackson, 2006). However, our data reveal that the predominant manner by which ATM and MDC1 cooperate to generate γ-H2AX along broken DNA strands in G1 phase cells is through the continued phosphorylation of H2AX in chromatin proximal to DNA breakage sites, forming γ-H2AX at high densities near DSBs rather than spreading H2AX phosphorylation further and further away from DNA breaks (Figure 7). The sequences 100–200 kb from Jκ5 where the transition between MDC1 dependent and independent ATM mediated γ-H2AX formation occurs contains an actively transcribed gene, suggesting that features of local chromatin environment may influence the ATM/MDC1 dependent generation of γ-H2AX along broken DNA strands. Our observation that the ATM dependent formation of γ-H2AX along DNA strands 200–600 kb from breakage sites does not require MDC1 indicates that the soluble pool of activated ATM phosphorylates H2AX at distances further away from DSBs along disrupted DNA fibers (Figure 7). Thus, we conclude that the spread of γ-H2AX along broken DNA strands in G1 phase cells is limited by the concentration of activated ATM that emanates from initial DSBs and diffuses throughout a surrounding nuclear volume. Additional studies are required to investigate a potential MDC1 function in fueling the ATM mediated spread of γ-H2AX to accessible H2AX molecules in nucleosomes on any chromosomal DNA strand located within nuclear territories around DNA breakage sites.

Figure 7
Models depicting how γ-H2AX forms along DNA strands broken in G1 phase cells

Here, we also demonstrate H2AX haploinsufficient cells exhibit a 50% reduction of H2AX densities in chromatin that results in a six to ten fold decrease of γ-H2AX densities formed along broken DNA strands, without causing a similar impairment to the distance across which γ-H2AX forms. Our data indicates that the ATM and MDC1 dependent generation of γ-H2AX requires a critical density of H2AX, and by extension γ-H2AX, in chromatin along broken DNA strands to achieve maximal γ-H2AX formation around breakage sites. Thus, we conclude that the initial formation of γ-H2AX at DSBs, by either ATM or DNA-PKcs, fuels a self-reinforcing mechanism by which γ-H2AX retains MDC1 and activated ATM in chromatin near initial DNA breakage sites to promote the continued local phosphorylation of H2AX (Figure 7). We propose that this self-reinforcing “concentration” of γ-H2AX around breakage sites is important to: 1) create a dynamic chromatin structure that facilitates accessibility of broken DNA ends, and 2) generate a proper spatial orientation and density of binding sites for DDR factors to assemble into complexes that stabilize disrupted DNA strands and amplify DNA damage signals. Since H2AX is unevenly distributed throughout chromatin (Bewersdorf et al., 2006), our findings also suggest that the kinetics of DSB repair, positional stability of broken DNA strands, and the magnitude of the elicited DDR might depend upon the genomic location of the DNA breakage sites. In this context, the chromatin structure of antigen receptor loci may have evolved through pressure elicited by the necessity to efficiently repair programmed DSBs, preserve cellular viability, maintain genomic integrity, and prevent malignant transformation during the somatic gene rearrangements required for adaptive immunity.

EXPERIMENTAL PROCEDURES

Mice

Wild-type 129SvEv and TCRβRag2−/− mice were purchased (Taconic). Artemis−/− mice (Rooney et al., 2002) were bred with mice containing an endogenous TCRβ rearrangement (Hochedlinger and Jaenisch, 2002), Lck-Cre (Lee et al., 2001), and p53 “floxed” (Jonkers et al., 2001) mice to create LβP and LAβP mice. All experiments in mice were performed in accordance with national guidelines and approved by the Children’s Hospital of Philadelphia IACUC committee.

Cell lines and culture

We previously described the generation and culture of wild-type, Artemis−/−, Artemis−/−Atm−/−, and Scid pre-B cell lines (Bredemeyer et al., 2006; Bredemeyer et al., 2008). The other pre-B cell lines were generated the same way from single or compound mutant Artemis−/− (Rooney et al., 2002), Rag2−/− (Shinkai et al., 1992), H2ax “floxed” (Bassing et al., 2002a), and Mdc1−/− (Lou et al., 2006) mice. For STI571 treatment, 100 × 106 cells were placed in 50 mL of media and treated with STI571 to final concentration of 3 μM, or concomitantly with KU-55933 (Sigma) to final concentration of 40 μM. Caffeine (Sigma) and endothall (EMDBiosciences) were added to final concentrations of 10 mM and 30 μM.

Southern analysis

Southern analysis was performed as described on BamHI-digested DNA isolated from 1 × 107 pre-B cells (Bredemeyer et al., 2008). Cutting was quantified as the loss of the above background pixel intensity of the germline fragment of each individual Southern blot exposure film. For each blot an internal background standard was used in an attempt to eliminate any exposure variances. This number was then compared to the maximum observed loss of band and plotted as the fraction of maximal cutting.

Chromatin immunoprecipitation

Cells (1 × 108) were harvested, washed with PBS, re-suspended in cold PBS, and incubated in PBS with 1% formaldehyde for 10 minutes at RT. Cross-linking reaction was stopped by addition of glycine to 0.125 M. Cells were washed with PBS, re-suspended in ChIP lysis buffer (Upstate) and sonicated to 300–500 bp average DNA fragment size. Debris was removed by centrifugation (13,200 g for 10 minutes at 4°C) and the supernatant diluted 10-fold in immunoprecipitation buffer (150mM NaCl, 20mM Tris-HCl pH=7.5, 0.1% Triton X-100, protease inhibitors (EDTA-free, Roche), phosphatase inhibitors (Cocktail II, Sigma)) and incubated with antibodies for either H4 (Upstate, 05–858), γ-H2AX (Upstate, 05–636), or H2AX (Bethyl laboratories, 300–083) overnight at 4°C with agitation. The next morning, 60μL protein A agarose beads (H4 and H2AX ChIP) or 60μL protein A agarose beads with 20 μg of rabbit anti-mouse antibodies (γ-H2AX ChIP) were added for an hour. Beads were washed six times (H4 and γ-H2AX ChIP) or four times (H2AX ChIP) for 10 minutes at RT with wash buffer (IP buffer + 0.1% sodium lauryl sulphate) and once for 15 minutes at RT with TE buffer. Cross-linked protein-DNA complexes were eluted for 15 minutes at RT with elution buffer (100 mM NaHCO3, 1% SDS) and incubated in NaCl at 100 mM overnight at 65°C to reverse cross-links. DNA was purified using Qiagen QIAquick PCR kit and used as template for q-PCR quantification on the Applied Biosystems 7500 Fast Real Time PCR System. The sequences of the Tcrα and Igκ locus DNA strand primers are listed in a Supplementary Table. In each experiment, a primer pair that specifically amplifies the intronic region of Plac8 on chromosome 5 (Forward primer – TTCCATCCAAAGCATCAATACAAG, Reverse primer –TTGGACAAACCCACAAAAACAG) was used as an internal DNA control since this gene is expressed only in placenta and chromosome 5 lacks any antigen receptor loci. To calculate γ-H2AX density, we normalized the γ-H2AX qPCR signal to the H4 qPCR signal at each location assayed and compared it to the γH2AX/H4 ratio at Plac8 control.

Supplementary Material

Acknowledgments

We thank Steve Reiner, Thomas Curran, and Celeste Simon for helpful discussions of the manuscript. This work was supported by the Cancer Research Institute Training Grant (V.S. and B.Y.); the Training Program in Immune System Development (A.C.C.); the Department of Pathology and the Center for Childhood Cancer Research of the Children’s Hospital of Philadelphia, the Abramson Family Cancer Research Institute, a Pew Scholar in the Biomedical Sciences Award, a grant from the Pennsylvania Department of Health, and the National Institutes of Health Grant R01 CA125195 (C.H.B.).

Footnotes

The authors have no conflicts of interest to disclose.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Andegeko Y, Moyal L, Mittelman L, Tsarfaty I, Shiloh Y, Rotman G. Nuclear retention of ATM at sites of DNA double strand breaks. J Biol Chem. 2001;276:38224–38230. [PubMed]
  • Bassing CH, Chua KF, Sekiguchi J, Suh H, Whitlow SR, Fleming JC, Monroe BC, Ciccone DN, Yan C, Vlasakova K, et al. Increased ionizing radiation sensitivity and genomic instability in the absence of histone H2AX. Proc Natl Acad Sci U S A. 2002a;99:8173–8178. [PubMed]
  • Bassing CH, Ranganath S, Murphy M, Savic V, Gleason M, Alt FW. Aberrant V(D)J recombination is not required for rapid development of H2ax/p53 deficient thymic lymphomas with clonal translocations. Blood. 2007;111:2163–2169. [PubMed]
  • Bassing CH, Suh H, Ferguson DO, Chua KF, Manis J, Eckersdorff M, Gleason M, Bronson R, Lee C, Alt FW. Histone H2AX: a dosage-dependent suppressor of oncogenic translocations and tumors. Cell. 2003;114:359–370. [PubMed]
  • Bassing CH, Swat W, Alt FW. The mechanism and regulation of chromosomal V(D)J recombination. Cell . 2002b;109(Suppl):S45–55. [PubMed]
  • Bewersdorf J, Bennett BT, Knight KL. H2AX chromatin structures and their response to DNA damage revealed by 4Pi microscopy. Proc Natl Acad Sci U S A. 2006;103:18137–18142. [PubMed]
  • Bredemeyer AL, Helmink BA, Innes CL, Calderon B, McGinnis LM, Mahowald GK, Gapud EJ, Walker LM, Collins JB, Weaver BK, et al. DNA double-strand breaks activate a multi-functional genetic program in developing lymphocytes. Nature. 2008;456:819–823. [PMC free article] [PubMed]
  • Bredemeyer AL, Sharma GG, Huang CY, Helmink BA, Walker LM, Khor KC, Nuskey B, Sullivan KE, Pandita TK, Bassing CH, Sleckman BP. ATM stabilizes DNA double-strand-break complexes during V(D)J recombination. Nature. 2006;442:466–470. [PubMed]
  • Burma S, Chen BP, Murphy M, Kurimasa A, Chen DJ. ATM phosphorylates histone H2AX in response to DNA double-strand breaks. J Biol Chem. 2001;276:42462–42467. [PubMed]
  • Celeste A, Difilippantonio S, Difilippantonio MJ, Fernandez-Capetillo O, Pilch DR, Sedelnikova OA, Eckhaus M, Ried T, Bonner WM, Nussenzweig A. H2AX haploinsufficiency modifies genomic stability and tumor susceptibility. Cell. 2003a;114:371–383. [PMC free article] [PubMed]
  • Celeste A, Fernandez-Capetillo O, Kruhlak MJ, Pilch DR, Staudt DW, Lee A, Bonner RF, Bonner WM, Nussenzweig A. Histone H2AX phosphorylation is dispensable for the initial recognition of DNA breaks. Nat Cell Biol. 2003b;5:675–679. [PubMed]
  • Celeste A, Petersen S, Romanienko PJ, Fernandez-Capetillo O, Chen HT, Sedelnikova OA, Reina-San-Martin B, Coppola V, Meffre E, Difilippantonio MJ, et al. Genomic instability in mice lacking histone H2AX. Science. 2002;296:922–927. [PMC free article] [PubMed]
  • Chen HT, Bhandoola A, Difilippantonio MJ, Zhu J, Brown MJ, Tai X, Rogakou EP, Brotz TM, Bonner WM, Ried T, Nussenzweig A. Response to RAG-mediated VDJ cleavage by NBS1 and gamma-H2AX. Science. 2000;290:1962–1965. [PMC free article] [PubMed]
  • Chowdhury D, Keogh MC, Ishii H, Peterson CL, Buratowski S, Lieberman J. gamma-H2AX dephosphorylation by protein phosphatase 2A facilitates DNA double-strand break repair. Mol Cell. 2005;20:801–809. [PubMed]
  • Chowdhury D, Xu X, Zhong X, Ahmed F, Zhong J, Liao J, Dykxhoorn DM, Weinstock DM, Pfeifer GP, Lieberman J. A PP4-phosphatase complex dephosphorylates gamma-H2AX generated during DNA replication. Mol Cell. 2008;31:33–46. [PMC free article] [PubMed]
  • Downs JA, Lowndes NF, Jackson SP. A role for Saccharomyces cerevisiae histone H2A in DNA repair. Nature. 2000;408:1001–1004. [PubMed]
  • Fernandez-Capetillo O, Chen HT, Celeste A, Ward I, Romanienko PJ, Morales JC, Naka K, Xia Z, Camerini-Otero RD, Motoyama N, et al. DNA damage-induced G2-M checkpoint activation by histone H2AX and 53BP1. Nat Cell Biol. 2002;4:993–997. [PubMed]
  • Franco S, Gostissa M, Zha S, Lombard DB, Murphy MM, Zarrin AA, Yan C, Tepsuporn S, Morales JC, Adams MM, et al. H2AX Prevents DNA Breaks from Progressing to Chromosome Breaks and Translocations. Mol Cell. 2006;21:201–214. [PubMed]
  • Guidos CJ, Williams CJ, Grandal I, Knowles G, Huang MT, Danska JS. V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid lymphocyte precursors. Genes Dev. 1996;10:2038–2054. [PubMed]
  • Heo K, Kim H, Choi SH, Choi J, Kim K, Gu J, Lieber MR, Yang AS, An W. FACT-mediated exchange of histone variant H2AX regulated by phosphorylation of H2AX and ADP-ribosylation of Spt16. Mol Cell. 2008;30:86–97. [PubMed]
  • Hickson I, Zhao Y, Richardson CJ, Green SJ, Martin NM, Orr AI, Reaper PM, Jackson SP, Curtin NJ, Smith GC. Identification and characterization of a novel and specific inhibitor of the ataxia-telangiectasia mutated kinase ATM. Cancer Res. 2004;64:9152–9159. [PubMed]
  • Hochedlinger K, Jaenisch R. Monoclonal mice generated by nuclear transfer from mature B and T donor cells. Nature. 2002;415:1035–1038. [PubMed]
  • Huen MS, Grant R, Manke I, Minn K, Yu X, Yaffe MB, Chen J. RNF8 transduces the DNA-damage signal via histone ubiquitylation and checkpoint protein assembly. Cell. 2007;131:901–914. [PMC free article] [PubMed]
  • Ikura T, Tashiro S, Kakino A, Shima H, Jacob N, Amunugama R, Yoder K, Izumi S, Kuraoka I, Tanaka K, et al. DNA damage-dependent acetylation and ubiquitination of H2AX enhances chromatin dynamics. Mol Cell Biol. 2007;27:7028–7040. [PMC free article] [PubMed]
  • Jonkers J, Meuwissen R, van der Gulden H, Peterse H, van der Valk M, Berns A. Synergistic tumor suppressor activity of BRCA2 and p53 in a conditional mouse model for breast cancer. Nat Genet. 2001;29:418–425. [PubMed]
  • Kimura H, Takizawa N, Allemand E, Hori T, Iborra FJ, Nozaki N, Muraki M, Hagiwara M, Krainer AR, Fukagawa T, Okawa K. A novel histone exchange factor, protein phosphatase 2Cgamma, mediates the exchange and dephosphorylation of H2A-H2B. J Cell Biol. 2006;175:389–400. [PMC free article] [PubMed]
  • Kinner A, Wu W, Staudt C, Iliakis G. Gamma-H2AX in recognition and signaling of DNA double-strand breaks in the context of chromatin. Nucleic Acids Res. 2008;36:5678–5694. [PMC free article] [PubMed]
  • Kusch T, Florens L, Macdonald WH, Swanson SK, Glaser RL, Yates JR, 3rd, Abmayr SM, Washburn MP, Workman JL. Acetylation by Tip60 is required for selective histone variant exchange at DNA lesions. Science. 2004;306:2084–2087. [PubMed]
  • Lee PP, Fitzpatrick DR, Beard C, Jessup HK, Lehar S, Makar KW, Perez-Melgosa M, Sweetser MT, Schlissel MS, Nguyen S, et al. A critical role for Dnmt1 and DNA methylation in T cell development, function, and survival. Immunity. 2001;15:763–774. [PubMed]
  • Lou Z, Minter-Dykhouse K, Franco S, Gostissa M, Rivera MA, Celeste A, Manis JP, van Deursen J, Nussenzweig A, Paull TT, et al. MDC1 Maintains Genomic Stability by Participating in the Amplification of ATM-Dependent DNA Damage Signals. Mol Cell. 2006;21:187–200. [PubMed]
  • Ma Y, Pannicke U, Schwarz K, Lieber MR. Hairpin opening and overhang processing by an Artemis/DNA-dependent protein kinase complex in nonhomologous end joining and V(D)J recombination. Cell. 2002;108:781–794. [PubMed]
  • Mailand N, Bekker-Jensen S, Faustrup H, Melander F, Bartek J, Lukas C, Lukas J. RNF8 ubiquitylates histones at DNA double-strand breaks and promotes assembly of repair proteins. Cell. 2007;131:887–900. [PubMed]
  • Muljo SA, Schlissel MS. A small molecule Abl kinase inhibitor induces differentiation of Abelson virus-transformed pre-B cell lines. Nat Immunol. 2003;4:31–37. [PubMed]
  • Paull TT, Rogakou EP, Yamazaki V, Kirchgessner CU, Gellert M, Bonner WM. A critical role for histone H2AX in recruitment of repair factors to nuclear foci after DNA damage. Curr Biol. 2000;10:886–895. [PubMed]
  • Ramiro AR, Jankovic M, Callen E, Difilippantonio S, Chen HT, McBride KM, Eisenreich TR, Chen J, Dickins RA, Lowe SW, et al. Role of genomic instability and p53 in AID-induced c-myc-Igh translocations. Nature. 2006;440:105–109. [PMC free article] [PubMed]
  • Rogakou EP, Boon C, Redon C, Bonner WM. Megabase chromatin domains involved in DNA double-strand breaks in vivo. J Cell Biol. 1999;146:905–916. [PMC free article] [PubMed]
  • Rogakou EP, Pilch DR, Orr AH, Ivanova VS, Bonner WM. DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139. J Biol Chem. 1998;273:5858–5868. [PubMed]
  • Rooney S, Sekiguchi J, Zhu C, Cheng HL, Manis J, Whitlow S, DeVido J, Foy D, Chaudhuri J, Lombard D, Alt FW. Leaky Scid phenotype associated with defective V(D)J coding end processing in Artemis-deficient mice. Mol Cell. 2002;10:1379–1390. [PubMed]
  • Shinkai Y, Rathbun G, Lam KP, Oltz EM, Stewart V, Mendelsohn M, Charron J, Datta M, Young F, Stall AM, et al. RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell. 1992;68:855–867. [PubMed]
  • Shroff R, Arbel-Eden A, Pilch D, Ira G, Bonner WM, Petrini JH, Haber JE, Lichten M. Distribution and dynamics of chromatin modification induced by a defined DNA double-strand break. Curr Biol. 2004;14:1703–1711. [PMC free article] [PubMed]
  • Stewart GS, Wang B, Bignell CR, Taylor AM, Elledge SJ. MDC1 is a mediator of the mammalian DNA damage checkpoint. Nature. 2003;421:961–966. [PubMed]
  • Stiff T, O’Driscoll M, Rief N, Iwabuchi K, Lobrich M, Jeggo PA. ATM and DNA-PK function redundantly to phosphorylate H2AX after exposure to ionizing radiation. Cancer Res. 2004;64:2390–2396. [PubMed]
  • Stucki M, Clapperton JA, Mohammad D, Yaffe MB, Smerdon SJ, Jackson SP. MDC1 directly binds phosphorylated histone H2AX to regulate cellular responses to DNA double-strand breaks. Cell. 2005;123:1213–1226. [PubMed]
  • Stucki M, Jackson SP. gammaH2AX and MDC1: anchoring the DNA-damage-response machinery to broken chromosomes. DNA Repair (Amst) 2006;5:534–543. [PubMed]
  • Unal E, Arbel-Eden A, Sattler U, Shroff R, Lichten M, Haber JE, Koshland D. DNA damage response pathway uses histone modification to assemble a double-strand break-specific cohesin domain. Mol Cell. 2004;16:991–1002. [PubMed]
  • Xie A, Puget N, Shim I, Odate S, Jarzyna I, Bassing CH, Alt FW, Scully R. Control of sister chromatid recombination by histone H2AX. Mol Cell. 2004;16:1017–1025. [PMC free article] [PubMed]