REVIEW ARTICLE Histone H2A phosphorylation in DNA double-strand break repair Elinor R. Foster and Jessica A. Downs Department of Biochemistry, Cambridge University, UK DNA repair must, by definition, occur within the con- text of chromatin. The most basic unit of chromatin is the nucleosome, consisting of two copies of each of four core histones around which DNA is wrapped in two left-handed superhelical turns [1]. This level of DNA compaction, often referred to as ‘beads on a string’, can be folded into numerous higher-order lev- els of chromatin condensation [2]. The shift towards more condensed structures is facilitated by the pres- ence of the linker histone [2]. This packaging of DNA, while essential for compressing a very long, highly neg- atively charged molecule into a relatively small space, is also inhibitory to processes that require the mani- pulation of DNA, such as replication, transcription, and repair. It is therefore not surprising that histones and proteins that modulate chromatin structure are integral players in these processes. The four core histones that make up the nucleosome structure are H2A, H2B, H3, and H4. Each of these proteins contains a histone fold domain, which is cen- tral to the nucleosome core structure. In addition, the histones all have a flexible N-terminal domain that protrudes from the nucleosome core particle. Histones H2A and H2B are unique in having significant sequence on the C-terminal side of the histone fold. The C-terminal domain of H2B forms an a-helix, and lies along the side of the nucleosome. Interestingly, the H2A C-terminal domain, like the histone N-terminal domains, is partly flexible and protrudes from the nucleosome core (Fig. 1). It is located at the point of the nucleosome where the linker DNA enters and exits the structure, and this is the area to which the linker histones bind. There are a number of mechanisms by which chro- matin structure and composition can be manipulated to facilitate events such as DNA replication and repair. These include covalent modifications of histone proteins, ATP-dependent chromatin remodelling, and the exchange of histone variants into and out of chro- matin. Notably, these processes are sometimes inter- related, for example, the replacement of H2A with the H2AZ variant (described in more detail below) is Keywords chromatin; DNA repair; H2AX; histone H2A Correspondence J. A. Downs, Department of Biochemistry, Cambridge University, 80 Tennis Court Road, Cambridge CB2 1GA, UK Fax: +44 1223 766 002 Tel: +44 1223 333 663 E-mail: jad32@mole.bio.cam.ac.uk (Received 11 February 2005, revised 4 April 2005, accepted 28 April 2005) doi:10.1111/j.1742-4658.2005.04741.x DNA repair must take place within the context of chromatin, and it is therefore not surprising that many aspects of both chromatin components and proteins that modify chromatin have been implicated in this process. One of the best-characterized chromatin modification events in DNA-dam- age responses is the phosphorylation of the SQ motif found in histone H2A or the H2AX histone variant in higher eukaryotes. This modification is an early response to the induction of DNA damage, and occurs in a wide range of eukaryotic organisms, suggesting an important conserved func- tion. One function that histone modifications can have is to provide a unique binding site for interacting factors. Here, we review the proteins and protein complexes that have been identified as H2AS129ph (budding yeast) or H2AXS139ph (human) binding partners and discuss the implica- tions of these interactions. Abbreviations ph, phosphorylation; PIKK, phosphatidylinositol 3-kinase-like kinase. FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 3231 catalyzed by an ATP-dependent chromatin remodelling complex. Histones are heavily modified by post-trans- lational modifications in vivo, and the known modifica- tions include acetylation of lysine residues, methylation of lysine and arginine residues, ubiquitination and sumoylation of lysines, and phosphorylation of serine and threonine residues. These modifications can alter the charge of the histone proteins and affect the ability to effectively condense DNA, or they can create or remove binding interfaces for chromatin-associated proteins (reviewed in [3]). Proteins and protein complexes that use the energy of ATP hydrolysis to alter chromatin structure are recognizable by their sequence similarity of the cata- lytic subunit to the founding member: Swi2⁄ Snf2. The exact consequence of this ATP-dependent alteration is variable, and can include the exchange of all histones from one DNA molecule to another, the sliding of the histone octamer to a new position, alteration of his- tone–DNA interactions, removal of H2A–H2B dimers, and exchange of histones in the nucleosome, including swapping core histones for histone variants [4,5]. The majority of human histone genes are clustered in the genome, and their transcription is tightly linked to replication. Histone variants are found outside of these gene clusters and transcriptionally regulated in a replication-independent manner. The incorporation of histone variants into chromatin can have profound structural and physiological consequences [6]. Interest- ingly, the histone H2A family has the largest number of described variants, including H2AX [6]. Here, we focus on the role of H2A in DNA-damage responses. As discussed in more detail below, this phe- nomenon touches on all three mechanisms of chromatin modulation. First, the crucial event in H2A DNA-dam- age responses is the phosphorylation of an SQ motif in the C-terminal tail. This appears to function, at least in part, to recruit chromatin remodelling complexes to sites of DNA damage. Finally, this motif is present on a histone variant, H2AX, in higher eukaryotes. Phosphorylation of the SQ motif of histone H2A or H2AX In higher eukaryotes, the H2AX variants are charac- terized by a longer C-terminal tail that has an SQ motif followed by two additional amino-acid residues before the stop codon (Fig. 2). This SQ motif, Fig. 1. The nucleosome core particle from budding yeast [54] seen from two different angles (A and B). The histone proteins are in the centre of the structure, and the DNA is wrapped around the proteins in two left- handed superhelical turns. H2A, yellow and green; H2B, H3 and H4, grey. The two H2A C-termini of the H2A tails are indicated with arrows and are in the region where the DNA enters and exits the structure. The entire C-terminal tails were not solved in the crystal structure; one H2A molecule ends with H2AT126 (yellow) and the other with H2AK121 (green). Consequently, none of the C-terminal serine residues discussed in the text (highlighted in Fig. 2) are present in the solved structure. Fig. 2. Alignment of H2A and H2AX C-terminal sequences downstream of the histone fold from the indicated organisms. Budding yeast H2AS122 and analogous residues from other organisms are highlighted in green. Budding yeast H2AT126 and the H2AX upstream SQ ⁄ TQ motifs, which may or may not be functioning analogously in DNA-damage responses, are highlighted in blue. The major DNA-damage phos- phorylation SQE motif present in the budding yeast core H2A (S129), mammalian H2AX (S139), and Drosophila H2Av (S137) is highlighted in red. H2A phosphorylation and DNA repair E. R. Foster and J. A. Downs 3232 FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS H2AXS139, has been shown to be phosphorylated in response to DNA damage [7]. Here, we will use the recently proposed unifying nomenclature for histone modifications [8], but note that this phosphorylation event is often referred to as c-H2AX in the literature. As alluded to previously, this variant is a single-copy gene that is outside of the histone gene cluster, and its transcription is regulated differently from that of the core histone H2A genes [9,10]. In lower eukaryotes, there is an SQ motif present in H2A proteins in the same position relative to the stop codon (Fig. 2), and this motif is also phosphorylated in response to DNA damage [11–13]. However, the H2A tails are not exten- ded, and the genes encoding these proteins are linked to replication and make up the majority of histone H2A in the cell, like the core H2A genes in higher euk- aryotes, suggesting that these are more closely related to core H2A than H2AX variants. Moreover, an SQ motif exists in the same position of the H2AZ variant in Drosophila (Fig. 2) and is also phosphorylated in response to DNA damage [14]. Together, these results suggest that the phosphorylation of the SQ motif plays a role in DNA-damage responses regardless of which histone variant it is found on. The invariant character- istic appears to be the position of the SQ motif relative to the end of the protein, and it would be interesting to see whether this is crucial for function in vivo. The SQ motif is a good consensus target site for the phosphatidylinositol 3-kinase-like kinases (PIKKs), and members of this family of kinases have been impli- cated in DNA-damage responses throughout the euk- aryotes [15]. Not surprisingly, it has been shown that the PIKK family members Mec1 and Tel1 in budding yeast [11,12,16] and ATM, ATR and DNA-PK in higher eukaryotes [17–20] are responsible for phos- phorylation of this motif in response to DNA damage. In the systems studied, the phosphorylation takes place very rapidly after DNA damage, within minutes of exposure to c-irradiation [7,12]. In mammalian cells, the phosphorylation is detected in the vicinity of the DNA lesions by immunofluorescence in combination with ‘laser scissors’ [17,21], and is present in megabase chromatin domains [21]. By chromatin immuno- precipitation approaches, it was found that phosphory- lation of the budding yeast H2A SQ motif (H2AS129) occurs in cis on the DNA extending from an induced double-strand break in both directions [16,22] covering regions of 50–100 kb of chromatin [23]. Notably, the levels of phosphorylation immediately adjacent to the DNA break are not particularly high [16,22], which rai- ses the possibility that phosphorylation does not occur to a great extent on these nucleosomes. However, it is equally plausible that the antibody is not recognizing the epitope in the nucleosomes immediately adjacent to the break because of occluding proteins or additional H2A tail modifications. Indeed, there is evidence that budding yeast H2A residues S122 and T126 are both phosphorylated and important for DNA-damage responses [25,26], making them potential culprits for loss of H2AS129ph recognition by the antibody. Never- theless, these results are consistent with a direct role in facilitating repair at the site of the break, in contrast with an indirect role in the appropriate transcriptional regulation of genes necessary for survival after DNA damage [11]. As discussed above, the consequences of covalent modifications of histone proteins such as phosphoryla- tion include a direct structural effect on the ability of chromatin to fold as well as the creation or removal of binding interfaces for interacting proteins. Although there is some evidence that phosphorylation of H2A(X) can affect chromatin structure in vivo [11,27], there are no biochemical data to indicate whether this is a direct effect. In contrast, numerous H2AS129ph and H2AXS139ph binding partners have been reported in the literature. Therefore, we will focus on these potential binding partners and their activities in the DNA-damage response. Proteins that interact with H2AS129ph and H2AXS139ph NuA4 The NuA4 histone acetyltransferase complex was iden- tified in an effort to identify proteins that bound spe- cifically to the budding yeast H2A tail when the SQ motif was phosphorylated [22]. The studies made use of a synthetic peptide to demonstrate a direct inter- action with NuA4 in vitro. Furthermore, NuA4 was found to be directly associated with chromatin in the vicinity of an induced DNA break by chromatin immunoprecipitation assays [22,28]. Ino80 With the use of recombinant subunits of NuA4, the actin-related protein Arp4 was found to directly inter- act with the H2AS129ph peptide. This protein is also a component of the Ino80 and SwrC (see below) ATP- dependent chromatin remodelling complexes. Like NuA4, subunits present in the Ino80 complex were also demonstrated to be present at the site of an induced DNA break [22,29,30], but appeared to localize to the area subsequent to the appearance of NuA4 [22]. Moreover, this appearance at the DNA E. R. Foster and J. A. Downs H2A phosphorylation and DNA repair FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 3233 break was impaired in strains with mutations in either H2A S129 or the Mec1 and Tel1 kinases [29,30]. SwrC SwrC, the third Arp4-containing complex in budding yeast, was also found to interact specifically with the H2AS129ph peptide in vitro, and subunits present in SwrC were found to localize to the site of an induced DNA break in vivo [22]. Together, these data suggest that the binding interface created by phosphorylation of the budding yeast SQ motif is important for the interaction of all three Arp4-containing complexes at sites of DNA damage. However, as mentioned above, subunits present in both Ino80 and SwrC were detected at the site of the DNA break with different kinetics to NuA4. If the interface between Arp4 and H2AS129ph was the sole defining recruitment mechanism, the beha- viour of the three complexes should be comparable. As it is not, these data suggest additional mechanisms for appropriate recruitment and retention to DNA breaks. At least for Ino80, this may involve the Nhp10 protein, identified as important for Ino80-binding chromatin near DNA breaks [30]. As the Ino80 complex from an nhp10 mutant strain contains Arp4 but is no longer able to bind to soluble H2AS129ph [30], this raises the possibility that Nhp10 (or Ies3, which is also missing from the mutant complex) facilitates the interaction between Arp4 and H2AS129ph. Tip60 In higher eukaryotes, homologues of subunits found in NuA4, Ino80 and SwrC exist, but do not form three distinct homologous complexes. Instead, homologues of a subset of these complexes are found together in the Tip60 complex, which contains both HAT activity and ATP-dependent chromatin remodelling activity (reviewed in [31]). In a study from Workman and col- leagues [32], Tip60 was shown to preferentially acety- late chromatin in which the SQ-motif-containing H2Av variant contained a phosphomimetic glutamic acid residue in place of the serine residue (H2AvS137E). Once the H2AvS137E-containing nucleo- somes were acetylated, the Tip60 complex was found to remodel the chromatin and remove the ‘phosphoryl- ated’ H2Av–H2B dimer and replace it with a fresh H2Av–H2B dimer [32]. These results are consistent with the order of appearance at a DNA break in bud- ding yeast of NuA4 first [22], which then acetylates chromatin in the vicinity [28], and subsequently Ino80 and ⁄ or SwrC are detectably present [22]. Cohesin Although there is no evidence for a direct interaction between phosphorylated H2A and cohesin, a recent report demonstrated that cohesin is localized to regions surrounding an induced DNA break in a man- ner that is dependent on the presence of a phosphory- latable H2A SQ motif in budding yeast [23]. Like the complexes described above, there is a detectable delay between the appearance of H2AS129ph and the appearance of cohesin. Unlike the complexes described above, however, cohesin appears to localize to a much broader region of chromatin, more closely matching the pattern seen with H2AS129ph. The wild-type pat- tern of appearance of cohesin over these regions depends not only on H2A phosphorylation, but also on Mre11 and Rad53. As H2A phosphorylation is nor- mal in the absence of both Mre11 and Rad53 [11,16], this suggests that, even if there is a direct interaction between cohesin and H2AS129ph, there are additional requirements for recruitment and ⁄ or retention at the site of DNA damage. The M ⁄ R ⁄ N complex The mammalian Mre11 ⁄ Rad50 ⁄ Nbs1 (M ⁄ R ⁄ N) com- plex (Mre11 ⁄ Rad50 ⁄ Xrs2 in budding yeast) accumu- lates into foci after exposure to ionizing radiation, and this was shown to be dependent on the phosphoryla- tion of H2AX S139 [17,33]. In support of a role for phosphorylation of H2AX in recruitment of this com- plex to sites of DNA damage, a direct interaction between Nbs1 and phosphorylated, but not unphos- phorylated, H2AX was demonstrated in vitro [34]. However, a subsequent study demonstrated that the initial appearance of Nbs1 at sites of DNA damage was unaffected by the loss of H2AX [35], suggesting that H2AX phosphorylation is instead specifically important for the subsequent accumulation of this complex. Furthermore, the budding yeast Mre11 pro- tein (part of the Mre11 ⁄ Rad50 ⁄ Xrs2 complex) localizes to sites of DNA damage normally in the absence of H2A phosphorylation [16]. There are a number of possible interpretations of these data. One possibility is that the physical interaction between the proteins detected in vitro is not physiologically relevant, and that the accumulation at sites of DNA damage in vivo is indirectly affected by H2AX phosphorylation- dependent events. Alternatively, redundant recruitment mechanisms may exist, and ⁄ or the interaction is only important after recruitment for the retention of the proteins. H2A phosphorylation and DNA repair E. R. Foster and J. A. Downs 3234 FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 53BP1(Hs) ⁄ Rad9(Sc) ⁄ Crb2(Sp) The mammalian checkpoint protein 53BP1 was found to bind to H2AXS139ph, but not H2AX [36,37]. Like the M ⁄ R ⁄ N complex, 53BP1 foci formation is also abrogated in the absence of H2AX [33,36,38]. The fis- sion yeast homologue of 53BP1, Crb2, has also been shown to accumulate at sites of DNA damage, and this appears to be dependent on the phosphorylation of H2A [13]. The authors also showed a direct interaction between Crb2 and H2AS129ph in vitro [13]. Taken together, these reports suggest that phosphorylation of H2A ⁄ H2AX recruits 53BP1 (or its homologue) to sites of damage. However, in the study by Celeste et al. [35] described above, the authors also examined 53BP1 behaviour and found that, like Nbs1, the initial recruit- ment of 53BP1 to sites of DNA damage is normal in H2AX – ⁄ – cells, but that subsequent accumulation is impaired. Again, it is possible that there is no physiolo- gical role for the direct interaction detected in vitro, that the interaction is a mechanism for retention, not recruitment, and ⁄ or there are redundant recruitment mechanisms. In support of the existence of redundant mechanisms, two recent reports suggest that methyla- tion of H3 K79 in mammals [39] and of H4 K20 in fission yeast [40] are important for recruitment of 53BP1 and Crb2 recruitment, respectively. Mdc1 The mammalian Mdc1 protein, which has no obvious homologues in lower eukaryotes, is also able to interact with H2AXS139ph peptides in vitro [37]. Like the M ⁄ R ⁄ N complex and 53BP1, Mdc1 accumulates in irra- diation-induced foci, and this foci formation is abro- gated in the absence of H2AX [37]. Interestingly, Mdc1 itself is required for the accumulation of the M ⁄ R ⁄ N complex into foci [37,41]. Recently, it was shown that the H2AXS139ph peptide is able to pull-down the M ⁄ R ⁄ N complex from a whole cell extract only when Mdc1 is present, suggesting that it acts as a bridging factor and that, at least in these assays, there is no direct interaction between Nbs1 and H2AXS139ph [42]. Consequences of H2AS129ph ⁄ H2AXS139ph-mediated recruitment of these protein(s) Increased malleability of DNA One obvious consequence of recruiting both HAT (NuA4, TIP60) and ATP-dependent chromatin remod- elling activities (Ino80, SwrC, TIP60) is the reorganiza- tion of chromatin structure at the site of the break. To repair a DNA lesion, a number of enzymatic activities are required, including exonuclease, polymerase and ligase activities to name just a few. It is conceivable that most, if not all, of the activities required for DNA repair will be inhibited by condensed chromatin struc- tures. Indeed, Gasser and colleagues [29] found that the processing of the DNA break is deficient in either H2A S129A or arp8 (an Ino80 subunit) mutant strains. Notably, the formation of single-stranded DNA as a DNA repair intermediate is important not only for the repair event, but for the appropriate checkpoint response [43]. This may provide an explanation for the checkpoint defects seen under certain circumstances in the absence of H2AS129ph ⁄ H2AXS139ph [13,44]. Exposure of other chromatin modifications In addition to providing a template that is more amen- able to DNA repair protein manipulation, the reorgan- ization of chromatin in the vicinity of a DNA break may expose other chromatin modifications that play a role in DNA-damage responses. Specifically, studies in mammalian cells and fission yeast demonstrated a role for methylation events of H3 K79 and H4 K20, respectively [39,40]. Intriguingly, both studies found that the methylation events are constitutive and do not appear to change in either quantity or location upon DNA damage. Yet, the methylation of these sites appears to be critical for the recruitment of the check- point proteins 53BP1 (mammalian) and Crb2 (fission yeast) to sites of damage in vivo, and the authors of both studies propose that the methylated motifs are ‘uncovered’ in the regions of DNA breaks. One possibility is that recruitment of chromatin-modi- fying activities by H2AS129ph or H2AXS139ph results in an increased number of exposed methylated motifs after DNA damage. If so, the number of Crb2 ⁄ 53BP1- binding sites would be severely reduced (but probably not abrogated) in the absence of H2A(X) phosphoryla- tion. This model would be consistent with the inability, in the absence of H2AX or the H2A SQ motif, of 53BP1 ⁄ Crb2 to accumulate into foci [13,36] and to either properly maintain a checkpoint [13] or initiate a check- point in response to very low doses of ionizing radiation where amplification of the signal might be crucial [44]. Facilitation of the removal of phosphorylated H2A(X) Once the DNA lesion is repaired, it may be deleterious for the cell to maintain phosphorylated H2A or H2AX in undamaged chromatin. For instance, the recruitment E. R. Foster and J. A. Downs H2A phosphorylation and DNA repair FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 3235 of repair and ⁄ or remodelling factors to undamaged DNA may result in misregulation of gene expression or the sequestration of repair factors that are needed else- where. Therefore, it seems likely that there is a phospha- tase that removes the phosphate residue from H2A or H2AX. Another, not mutually exclusive possibility, however, is that phosphorylated H2A mediates its own removal from chromatin by recruitment of the SwrC complex [22]. This complex removes H2A–H2B dimers from nucleosomes and replaces them with Htz1–H2B di- mers [45]. As Htz1 has also been genetically linked to H2A S129-dependent DNA-damage responses [22], it is tempting to speculate that SwrC is recruited to sites of DNA damage and swaps the phospho-H2A for Htz1. Consistent with this is the study in Drosophila, described above, in which the homologous complex, Tip60, prefer- entially acts on phospho-mimic-containing nucleosomes [32]. Finally, it is conceivable that the phosphorylated H2A(X) is targeted for degradation. It has recently been demonstrated in budding yeast that the proteasome is present at sites of DNA damage, and that this is neces- sary for appropriate DNA-damage responses [46]. Mediation of sister chromatid cohesion by loading cohesin As mentioned above, in budding yeast, cohesin was found to be associated with DNA after the induction of a double-strand break in a manner dependent on the phosphorylation of H2A S129 [23]. This recruitment to damaged DNA appears to be required for appropriate sister chromatid cohesion and DNA repair [23,24]. Intriguingly, whereas the absence of cohesin results in a severe inability to survive in the presence of DNA dam- age, the absence of S129 results in a strain that is only mildly sensitive to DNA damage relative to the survival patterns seen with other DNA signalling and repair protein mutant strains. This raises the possibility that, although cohesin is not detectably present at DNA breaks by chromatin immunoprecipitation in the S129A mutant strain, there is in fact enough cohesin loaded on to the area to facilitate repair. This possibil- ity again highlights the likely existence of redundant, albeit impaired, recruitment and ⁄ or retention mecha- nisms. Alternatively, cohesin may also facilitate the repair of DNA breaks in a manner that is not depend- ent on its localization to the sites of damage. Mediation of foci formation ⁄ checkpoint responses Regardless of whether the M ⁄ R ⁄ N complex, 53BP1 ⁄ Rad9 ⁄ Crb2 or Mdc1 bind directly to the phosphoryl- ated H2A(X) tail in vivo, it is clear that the accumula- tion of these proteins in DNA-damage-induced foci is dependent on H2A(X) phosphorylation. Notably, both the checkpoint defects and the DNA damage sensitiv- ity seen in yeast H2AS129A or mammalian H2AX – ⁄ – mutant cells are not as dramatic as the loss of check- point proteins such as Brca1 or ATM [11,12,44]. This suggests that the checkpoint proteins may be impaired in their behaviour, but that they can still function to facilitate DNA-damage responses to a reasonable degree in the absence of H2A(X) phosphorylation. Decreased malleability of chromatin ⁄ anchor ends together After the creation of a double-strand break, it could be beneficial to ‘lock down’ the two ends by creating a heterochromatic structure over the broken region. This would prevent transcription or replication machinery from traversing the site. The heterochromatin may also actively facilitate the maintenance of interactions between the two broken ends. This would help to pre- vent inappropriate or inefficient repair of the DNA ends. There is some evidence that H2A(X) phosphoryla- tion may help to form such a heterochromatic struc- ture. First, condensation of the X and Y chromosomes during male mouse meiosis is defective in H2AX – ⁄ – mice [27]. In addition, H2B has been shown to be phosphorylated in response to DNA damage (H2BS14ph), and is located in foci with H2AX phos- phorylation [47]. However, in the absence of H2AX, H2BS14 phosphorylation does not change by Western blotting, but is no longer visible in foci. One interpret- ation of these data proposed by the authors is that, in the absence of H2AX, the broken ends are unable to condense, leaving the H2BS14 phosphorylation signal too diffuse to visualize by immunofluorescence. The H2AZ variants have been shown to have altered stability compared with H2A [48], and FRET-based approaches have shown that H2AZ–H2B dimers are more stably attached to the nucleosome than H2A– H2B dimers [49]. If phosphorylation of H2A directs its own replacement with Htz1 in nucleosomes, as pro- posed above, it is possible that this results in the cre- ation of a less malleable structure in the region of DNA breaks. Even without the formation of heterochromatic structures, H2AX phosphorylation may help to ‘anchor’ broken ends together. By assisting the accu- mulation of 53BP1 ⁄ Rad9 ⁄ Crb2, and, at least in mam- malian cells, the M ⁄ R ⁄ N complex and Mdc1, H2A(X) phosphorylation may assist in the formation of a pro- H2A phosphorylation and DNA repair E. R. Foster and J. A. Downs 3236 FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS tein bridge that prevents the DNA ends from dissoci- ating [50]. In support of this, H2AX-deficient cells show increased levels of translocation [33,38]. Discussion A wide range of proteins have been identified as H2AS129ph or H2AXS139ph binding partners. In many cases, an interaction was demonstrated with the protein in vitro in addition to evidence pointing to colocalization in vivo. This raises the obvious question of whether all of these actually bind to H2AS129ph or H2AXS139ph in vivo. In response to DNA breaks, a great deal of H2A and H2AX is phosphorylated, and this increases in a time-dependent and dose-dependent manner [12,16,22]. Numerous potential binding platforms are being cre- ated, and thus it is reasonable to speculate that a wide variety of binding partners can be physically accom- modated. Notably, the consequences of H2A or H2AX phosphorylation listed above are in some cases directly opposing activities. Yet, the very nature of using cova- lent modifications to create binding interfaces allows different protein complexes to be relevant under differ- ent conditions. For instance, binding partners import- ant for the formation of the condensed X and Y chromosomes [27] may only exist during meiosis or be otherwise regulated so they do not ‘see’ phosphory- lated H2AX in mitotically dividing cells. In addition, the binding and accumulation of chromatin-modifying complexes to DNA breaks in yeast was examined in asynchronous haploid cell cultures. It is conceivable that they are in fact only recruited during a particular phase of the cell cycle. Even the recruitment of pro- teins that both increase and decrease the malleability of chromatin at a single double-strand break are not necessarily mutually exclusive activities, as long as they can be either physically or temporally separated at the DNA lesion (discussed in more detail below). If we are to postulate that all of the factors listed above are physiologically bound to H2AS129ph or H2AXS139ph, then one might hypothesize that they should all show the same timing and location of bind- ing, which would be dictated directly by the appear- ance of H2A(X) phosphorylation. Clearly, this is not the case, but one variable that may help to explain this discrepancy is the existence of other histone modifica- tions in the vicinity of the break. Specifically, in the yeast H2A tail, two other residues in very close proxi- mity to S129 (S122 and T126) have been shown to be phosphorylated in vivo [25], and both of these residues have been implicated in DNA-damage responses [25,26]. This raises the possibility that there is interplay between the residues, which dictates additional specific- ity in the associated binding partner. For example, H2AS122phS129ph and H2AT126phS129ph may bind to two mutually exclusive sets of proteins. In higher eukaryotes, the analogous residue to budding yeast H2A S122 is phosphorylated on core H2A [51] and the site exists on the H2AX variant as well (Fig. 2). More- over, an additional phosphorylation site on H2AX has been identified, which is in the same position as budding yeast T126 relative to the SQ motif (H2AXS136ph; Fig. 2). One obvious difference is that the mammalian residue is followed by a glutamine, making it a potential substrate for the PIKK family of kinases, but the yeast residue is not (Fig. 2). Neverthe- less, substrates for the PIKK family of kinases that do not conform to the consensus sequence have been identified (for example [52]), so it is possible that bud- ding yeast T126 is a target for Mec1 and ⁄ or Tel1. Nev- ertheless, these additional modifications are likely to contribute to the behaviour of the interacting proteins. Of course, modifications of other core histones, such as H4 acetylation by NuA4, are also likely to be involved in the differential recruitment of factors to DNA lesions. Finally, the regulation of the complexes themselves may dictate their differential behaviour at DNA breaks, despite the common existence of their binding platform. Although it is possible to postulate that all of the proteins identified as H2A or H2AX phospho-specific binding partners are physiologically relevant, it is also reasonable to interpret these results more cautiously. For example, many of the experiments demonstrating a direct physical interaction were performed using either synthetic peptides corresponding to the end of the H2A or H2AX tail that were either unmodified or contained a phosphoserine residue in the SQ motif. Proteins containing known phosphoserine ⁄ threonine binding motifs such as BRCT or FHA domains may bind to some degree to any phosphorylated peptide and be detected when these assays are performed, as the amount of peptide is often in vast excess to the proteins being analyzed. In addition, Arp4, the subunit in NuA4, Ino80 and SwrC implicated in H2AS129ph binding, may have affinity for phosphate moieties as it is related to the ATP-binding actin protein, although it did show a preference for the H2AS129ph tail over other phosphoserine-containing peptides [22]. In some cases, such as cohesin, no direct interaction was dem- onstrated [23], which leaves open the possibility that there is an indirect dependency on H2A phosphoryla- tion for binding to sites of DNA damage. As 53BP1 has been shown to bind both phosphorylated H2AX and methylated H3 in vitro [36,39], it would be inter- E. R. Foster and J. A. Downs H2A phosphorylation and DNA repair FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 3237 esting to see what the relative affinities for these sub- strates are. The use of a chromatin template, instead of syn- thetic peptides, in which H2A or H2AX is phosphoryl- ated would be extremely informative about how these proteins interact. The use of chromatin templates may also shed light on an aspect of H2A or H2AX phos- phorylation that is as yet unclear: the potential direct effect of phosphorylation on chromatin structure. If, as discussed above, there are multiple phosphorylation events in the H2A(X) tail after DNA damage, it is conceivable that this change in charge will affect higher-order chromatin structure. Given the location of the tail in the vicinity of the linker DNA (Fig. 1), one tempting possibility is that phosphorylation of the H2A tail may impinge on the behaviour of the linker histone, which has been shown to be inhibitory to DNA repair in yeast [53]. Inevitably, additional proteins that can bind to the phosphorylated SQ motif of H2A and H2AX will be identified in the future, and it is likely that we will also find more covalent modifications of histones that directly impinge on H2A(X) phospho-SQ-dependent functions. Integrating the roles of the known players and understanding the different conditions and situa- tions in which they are relevant will be a key step to elucidating the interplay between chromatin and chromatin-interacting factors during DNA-damage responses. References 1 Luger K, Mader AW, Richmond RK, Sargent DF & Richmond TJ (1997) Crystal structure of the nucleosome core particle at 2.8 A ˚ resolution. Nature 389, 251–260. 2 Hansen JC (2002) Conformational dynamics of the chromatin fiber in solution: determinants, mechanisms, and functions. Annu Rev Biophys Biomol Struct 31, 361– 392. 3 Peterson CL & Laniel MA (2004) Histones and histone modifications. Curr Biol 14, R546–R551. 4 Flaus A & Owen-Hughes T (2004) Mechanisms for ATP-dependent chromatin remodelling: farewell to the tuna-can octamer? Curr Opin Genet Dev 14, 165–173. 5 Owen-Hughes T & Bruno M (2004) Breaking the silence. Science 303, 324–325. 6 Ausio J & Abbott DW (2002) The many tales of a tail: carboxyl-terminal tail heterogeneity specializes histone H2A variants for defined chromatin function. Biochem- istry 41, 5945–5949. 7 Rogakou EP, Pilch DR, Orr AH, Ivanova VS & Bonner WM (1998) DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139. J Biol Chem 273, 5858–5868. 8 Turner BM (2005) Reading signals on the nucleosome with a new nomenclature for modified histones. Nat Struct Mol Biol 12, 110–112. 9 Mannironi C, Bonner WM & Hatch CL (1989) H2A.X, a histone isoprotein with a conserved C-terminal sequence, is encoded by a novel mRNA with both DNA replication type and polyA processing signals. Nucleic Acids Res 17, 9113–9126. 10 Bonner WM, Mannironi C, Orr A, Pilch DR & Hatch CL (1993) Histone H2A.X gene transcription is regu- lated differently than transcription of other replication- linked histone genes. Mol Cell Biol 13, 984–992. 11 Downs JA, Lowndes NF & Jackson SP (2000) A role for Saccharomyces cerevisiae histone H2A in DNA repair. Nature 408, 1001–1004. 12 Redon C, Pilch DR, Rogakou E, Orr AH, Lowndes NF & Bonner WM (2003) Yeast histone 2A serine 129 is essential for the efficient repair of checkpoint-blind DNA damage. EMBO Rep 4, 678–684. 13 Nakamura TML-L, Redon C & Russell P (2004) His- tone H2A phosphorylation controls Crb2 recruitment at DNA breaks, maintains checkpoint arrest, and influ- ences DNA repair in fission yeast. Mol Cell Biol 24, 6215–6230. 14 Madigan JP, Chotkowski HL & Glaser RL (2002) DNA double-strand break-induced phosphorylation of Drosophila histone variant H2Av helps prevent radia- tion-induced apoptosis. Nucleic Acids Res 30, 3698– 3705. 15 Durocher D & Jackson SP (2001) DNA-PK, ATM and ATR as sensors of DNA damage: variations on a theme? Curr Opin Cell Biol 13, 225–231. 16 Shroff R, Arbel-Eden A, Pilch D, Ira G, Bonner WM, Petrini JH, Haber JE & Lichten M (2004) Distribution and dynamics of chromatin modification induced by a defined DNA double-strand break. Curr Biol 14, 1703– 1711. 17 Paull TT, Rogakou EP, Yamazaki V, Kirchgessner CU, Gellert M & Bonner WM (2000) A critical role for his- tone H2AX in recruitment of repair factors to nuclear foci after DNA damage. Curr Biol 10, 886–895. 18 Ward IM & Chen J (2001) Histone H2AX is phos- phorylated in an ATR-dependent manner in response to replicational stress. J Biol Chem 276, 47759–47762. 19 Burma S, Chen BP, Murphy M, Kurimasa A & Chen DJ (2001) ATM phosphorylates histone H2AX in response to DNA double-strand breaks. J Biol Chem 276, 42462–42467. 20 Stiff T, O’Driscoll M, Rief N, Iwabuchi K, Lobrich M & Jeggo PA (2004) ATM and DNA-PK function redun- dantly to phosphorylate H2AX after exposure to ioniz- ing radiation. Cancer Res 64, 2390–2396. H2A phosphorylation and DNA repair E. R. Foster and J. A. Downs 3238 FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 21 Rogakou EP, Boon C, Redon C & Bonner WM (1999) Megabase chromatin domains involved in DNA double- strand breaks in vivo. J Cell Biol 146, 905–915. 22 Downs JA, Allard S, Jobin-Robitaille O, Javaheri A, Auger A, Bouchard N, Kron SJ, Jackson SP & Cote J (2004) Binding of chromatin-modifying activities to phosphorylated histone H2A at DNA damage sites. Mol Cell 16, 979–990. 23 Unal E, Arbel-Eden A, Sattler U, Shroff R, Lichten M, Haber JE & Koshland D (2004) DNA damage response pathway uses histone modification to assemble a dou- ble-strand break-specific cohesin domain. Mol Cell 16, 991–1002. 24 Strom L, Lindroos HB, Shirahige K & Sjogren C (2004) Postreplicative recruitment of cohesin to double-strand breaks is required for DNA repair. Mol Cell 16, 1003– 1015. 25 Wyatt HR, Liaw H, Green GR & Lustig AJ (2003) Multiple roles for Saccharomyces cerevisiae histone H2A in telomere position effect, Spt phenotypes and double- strand-break repair. Genetics 164, 47–64. 26 Harvey AC, Jackson SP & Downs JA (2005) Saccharo- myces cerevisiae histone H2A Ser122 facilitates DNA repair. Genetics 10.1534/genetics.104.038570. 27 Fernandez-Capetillo O, Mahadevaiah SK, Celeste A, Romanienko PJ, Camerini-Otero RD, Bonner WM, Manova K, Burgoyne P & Nussenzweig A (2003) H2AX is required for chromatin remodeling and inacti- vation of sex chromosomes in male meiosis. Dev Cell 4, 497–508. 28 Bird AW, Yu DY, Pray-Grant MG, Qiu Q, Harmon KE, Megee PC, Grant PA, Smith MM & Christman MF (2002) Acetylation of histone H4 by Esa1 is required for DNA double-strand break repair. Nature 419, 411–415. 29 van Attikum H, Fritsch O, Hohn B & Gasser SM (2004) Recruitment of the INO80 complex by H2A phosphorylation links ATP-dependent chromatin remo- deling with DNA double-strand break repair. Cell 119, 777–788. 30 Morrison AJ, Highland J, Krogan NJ, Arbel-Eden A, Greenblatt JF, Haber JE & Shen X (2004) INO80 and c-H2AX interaction links ATP-dependent chromatin remodeling to DNA damage repair. Cell 119, 767–775. 31 Doyon Y & Cote J (2004) The highly conserved and multifunctional NuA4 HAT complex. Curr Opin Genet Dev 14, 147–154. 32 Kusch T, Florens L, Macdonald WH, Swanson SK, Glaser RL, Yates JR, Abmayr SM, Washburn MP & Workman JL (2004) Acetylation by Tip60 is required for selective histone variant exchange at DNA lesions. Science 306, 2084–2087. 33 Celeste A, Petersen S, Romanienko PJ, Fernandez- Capetillo O, Chen HT, Sedelnikova OA, Reina-San- Martin B, Coppola V, Meffre E, Difilppantonio MJ, et al. (2002) Genomic instability in mice lacking histone H2AX. Science 296, 922–927. 34 Kobayashi J, Tauchi H, Sakamoto S, Nakamura A, Morishima K, Matsuura S, Kobayashi T, Tamai K, Tanimoto K & Komatsu K (2002) NBS1 localizes to c-H2AX foci through interaction with the FHA ⁄ BRCT domain. Curr Biol 12, 1846–1851. 35 Celeste A, Fernandez-Capetillo O, Kruhlak MJ, Pilch DR, Staudt DW, Lee A, Bonner RF, Bonner WM & Nussenzweig A (2003) Histone H2AX phosphorylation is dispensable for the initial recognition of DNA breaks. Nat Cell Biol 5, 675–679. 36 Ward IM, Minn K, Jorda KG & Chen J (2003) Accu- mulation of checkpoint protein 53BP1 at DNA breaks involves its binding to phosphorylated histone H2AX. J Biol Chem 278, 19579–19582. 37 Stewart GS, Wang B, Bignell CR, Taylor AM & Elledge SJ (2003) MDC1 is a mediator of the mamma- lian DNA damage checkpoint. Nature 421, 961–966. 38 Bassing CH, Chua KF, Sekiguchi J, Suh H, Whitlow SR, Fleming JC, Monroe BC, Ciccone DN, Yan C, Vla- sakova K, et al. (2002) Increased ionizing radiation sen- sitivity and genomic instability in the absence of histone H2AX. Proc Natl Acad Sci USA 99, 8173–8178. 39 Huyen Y, Zgheib O, Ditullio RA Jr, Gorgoulis VG, Zacharatos P, Petty TJ, Sheston EA, Mellert HS, Stav- ridi ES & Halazonetis TD (2004) Methylated lysine 79 of histone H3 targets 53BP1 to DNA double-strand breaks. Nature 432, 406–411. 40 Sanders SL, Portoso M, Mata J, Bahler J, Allshire RC & Kouzarides T (2004) Methylation of histone H4 lysine 20 controls recruitment of Crb2 to sites of DNA damage. Cell 119, 603–614. 41 Goldberg M, Stucki M, Falck J, D’Amours D, Rahman D, Pappin D, Bartek J & Jackson SP (2003) MDC1 is required for the intra-S-phase DNA damage checkpoint. Nature 421, 952–956. 42 Lukas C, Melander F, Stucki M, Falck J, Bekker-Jensen S, Goldberg M, Lerenthal Y, Jackson SP, Bartek J & Lukas J (2004) Mdc1 couples DNA double-strand break recognition by Nbs1 with its H2AX-dependent chroma- tin retention. EMBO J 23, 2674–2683. 43 Wang X & Haber JE (2004) Role of Saccharomyces single-stranded DNA-binding protein RPA in the strand invasion step of double-strand break repair. PLoS Biol 2, 0104–0112. 44 Fernandez-Capetillo O, Chen HT, Celeste A, Ward I, Romanienko PJ, Morales JC, Xia Z, Camerini-Otero RD, Motoyama N, Carpenter PB et al. (2002) DNA damage-induced G2-M checkpoint activation by histone H2AX and 53BP1. Nat Cell Biol 4, 993–997. 45 Mizuguchi G, Shen X, Landry J, Wu W-H, Sen S & Wu C (2004) ATP-driven exchange of histone H2AZ variant catalyzed by SWR1 chromatin remodeling com- plex. Science 303, 343–348. E. R. Foster and J. A. Downs H2A phosphorylation and DNA repair FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS 3239 46 Krogan NJ, Lam MH, Fillingham J, Keogh MC, Gebbia M, Li J, Datta N, Cagney G, Buratowski S, Emili A & Greenblatt JF (2004) Proteasome involve- ment in the repair of DNA double-strand breaks. Mol Cell 16, 1027–1034. 47 Fernandez-Capetillo O, Allis CD & Nussenzweig A (2004) Phosphorylation of histone H2B at DNA dou- ble-strand breaks. J Exp Med 199, 1671–1677. 48 Fan JY, Gordon F, Luger K, Hansen JC & Tremethick DJ (2002) The essential histone variant H2A.Z regulates the equilibrium between different chromatin conforma- tional states. Nat Struct Biol 9, 172–176. 49 Park Y-J, Dyer PN, Tremethick DJ & Luger K (2004) A new FRET approach demonstrates that the histone variant H2AZ stabilizes the histone octamer within the nucleosome. J Biol Chem 279, 24724–24282. 50 Bassing CH & Alt FW (2004) H2AX may function as an anchor to hold broken chromosomal DNA ends in close proximity. Cell Cycle 3, 149–153. 51 Aihara H, Nakagawa T, Yasui K, Ohta T, Hirose S, Dhomae N, Takio K, Kaneko M, Takeshima Y, Muramatsu M et al. (2004) Nucleosomal histone kinase- 1 phosphorylates H2A Thr119 during mitosis in the early Drosophila embryo. Genes Dev 18, 877–888. 52 Yavuzer U, Smith GCM, Bliss T, Werner D & Jackson SP (1998) DNA end-independent activation of DNA– PK mediated via association with the DNA-binding protein C1D. Genes Dev 12, 2188–2199. 53 Downs JA, Kosmidou E, Morgan A & Jackson SP (2003) Suppression of homologous recombination by the Saccharomyces cerevisiae linker histone. Mol Cell 11, 1685–1692. 54 White CL, Suto RK & Luger K (2001) Structure of the yeast nucleosome core particle reveals fundamental changes in internucleosome interactions. EMBO J 20, 5207–5218. H2A phosphorylation and DNA repair E. R. Foster and J. A. Downs 3240 FEBS Journal 272 (2005) 3231–3240 ª 2005 FEBS . to DNA breaks. At least for Ino80, this may involve the Nhp10 protein, identified as important for Ino80-binding chromatin near DNA breaks [30]. As the Ino80 complex from an nhp10 mutant strain. of the H2A or H2AX tail that were either unmodified or contained a phosphoserine residue in the SQ motif. Proteins containing known phosphoserine ⁄ threonine binding motifs such as BRCT or FHA domains. acety- late chromatin in which the SQ-motif-containing H2Av variant contained a phosphomimetic glutamic acid residue in place of the serine residue (H2AvS137E). Once the H2AvS137E-containing nucleo- somes