Histone H2AX is Phosphorylated at Sites of Retroviral DNA Integration, But is Dispensable for Post-integration Repair*

The histone variant H2AX is rapidly phosphorylated (denoted g H2AX) in large chromatin domains (foci) flanking dsDNA breaks that are produced by ionizing radiation or genotoxic agents, and during V(D)J recombination. H2AX-deficient cells, and mice, demonstrate increased sensitivity to dsDNA break damage, indicating an active role for g H2AX in DNA repair. However, g H2AX formation is not required for V(D)J recombination. The latter finding has suggested a greater dependency on g H2AX for anchoring free broken ends, versus ends that are held together during programmed breakage-joining reactions. Retroviral DNA integration produces a unique intermediate in which a dsDNA break in host DNA is held together by the intervening viral DNA, and such a reaction provides a useful model to distinguish g H2AX functions. We found that integration promotes transient formation of g H2AX at retroviral integration sites, as detected by both immunocytological and chromatin immunoprecipitation (ChIP) methods. These results provide the first direct evidence for association of newly integrated viral DNA with a protein species that is an established marker for the onset of a DNA damage response. We also show that H2AX is not required for repair of the retroviral integration intermediate, as determined by stable transduction. These observations provide independent support of an anchoring model for the function of g H2AX in chromatin repair.


Introduction
The evolutionarily-conserved histone H2AX comprises about 2-25% of the histone H2A pool in mammalian cells and is incorporated randomly into nucleosomes (1). The extended Cterminal tail of H2AX contains a serine (S139), embedded in an invariant SQE motif that is a target for phosphorylation by the PI-3K-related protein kinases, DNA-PK, ATM (ataxia telangiectasia mutated), and ATR (ATM and Rad3-related) (2)(3)(4). This H2AX serine residue is massively and rapidly phosphorylated at sites of double strand breaks (DSBs) and stalled replication forks (3,5,6) forming microscopically visible foci upon staining with a specific antibody. This phosphorylation seems to play an important role in processing or repair of DSBs (7,8). H2AX phosphorylation has also been observed at sites of V(D)J recombination (9), meiotic strand breaks, and other physiological, programmed reactions in which DSBs are formed (10)(11)(12).
Early events in retroviral replication include entry of the viral capsid, with accompanying enzymes reverse transcriptase and integrase (IN), followed by synthesis of a DNA copy of the viral RNA genome to form a pre-integration complex. This complex then enters the nucleus and integration is first detected at about 3-4 h post-infection (13). Retroviral integration is catalyzed by integrase, acting upon specific sequences at the ends of the viral DNA and via a concerted cleavage-ligation reaction that is mechanistically similar to that catalyzed by RAG proteins during V(D)J recombination (14)(15)(16) (Fig. 1A). As a consequence of integrasemediated joining, the host cell DNA suffers a double strand break (DSB), but the ends are held together by single strand links to viral DNA (Fig. 1B). Post-integration repair of this intermediate (Fig. 1B) is essential for the maintenance of host DNA integrity as well as the stable association of retroviral DNA with host chromosomes. Numerous lines of evidence indicate that retroviral DNA elicits a DNA damage response and that the integration intermediate is repaired primarily via components of the non-homologous end-joining (NHEJ) pathway (17)(18)(19)(20). In this study, we asked if H2AX is phosphorylated at sites of retroviral DNA integration and whether this response is essential for repair of this complex lesion as determined by survival of stably transduced cells.

EXPERIMENTAL PROCEDURES
Cells and Viruses -MO59K cells (purchased from ATCC) were carried as described (18,19). Mouse embryo fibroblasts were carried as described (17). The ASV-based vectors, IN + and IN -, and the HIV-1 based vector were described previously (17,19

RESULTS
To determine if retroviral infection induces formation of γH2AX foci, we infected cells with an amphotropic ASV vector (17) and examined them by immunofluorescence with an antibody specific for γH2AX. In preliminary experiments (data not included) we observed formation of γH2AX foci in DNA repair-proficient human (HeLa and MO59K) and mouse (3T3) cells early after infection. As a control we infected cells with an integration-deficient (IN -) vector (17) and no increase in foci over background was detected. The data in Figure 2 and Table 1 summarize results from subsequent experiments, which included computer-assisted, quantitative analyses of γH2AX foci in MO59K cells infected at a multiplicity (m.o.i.) of 10 infectious particles/cell. We again observed an increased number of foci in the infected culture ( Fig. 2A), which appeared to peak at 6 h post-infection. A comparison of the distribution of the number foci/cell in uninfected cells and 6 h post-infection is shown in Figure 2B. The bulk (approximately 75%) of the uninfected cells contained no or few foci per cell, whereas a small proportion (ca. 10%) displayed numerous foci, which we speculate may be due to replication stress; such cells were also observed in the infected culture. At 6 h post infection, the infected culture had many fewer cells with no foci, and the percentage of cells with 5-10 foci was substantially higher than in the uninfected culture. As summarized in Table 1 Table 1.
To verify that H2AX is phosphorylated at sites of retroviral DNA integration, we  Figure 3B and C. Figure 3B shows  (Table 1), these results showed that association with γH2AX peaks at ~6 h, shortly after viral DNA is detected in the nuclear extract. We estimate that 60% of the nuclear viral DNA in this experiment is associated with γH2AX, and therefore integrated at this time point. From these results we conclude that γH2AX is a valid marker for sites of retroviral DNA integration. We note that ChIP detects integrated viral DNA both pre-and post-repair (Fig. 1). To examine the functional relevance of H2AX phosphorylation to repair, we performed the transduction assays described below.
Because transduced genes are expressed efficiently only from stably integrated proviruses, retroviral transduction is a readout for successful post-integration repair. For example, we have shown that transduction efficiency of NHEJ-defective, DNA-PKcs-deficient murine cells is 80-90% reduced compared to wild type cells or deficient cells into which DNA-PKcs-expressing DNA was reintroduced (17,20). As the DNA damage induced by integration cannot be repaired efficiently, most NHEJ-deficient cells are unlikely to survive infection and, therefore, cannot give rise to stable transductants. To ask if H2AX phosphorylation is required for post-integration repair, we performed transduction experiments using embryo fibroblast lines (MEFs) obtained from H2AX knockout mice, and derivatives (24,25). Results in Table 2 show that there is no significant difference in the transduction efficiency of an ASV-GFP vector with H2AX+/+ and H2AX-/-MEFs. Similar results were obtained with a VSV-G protein-pseudo-typed HIV-1 GFP vector (not shown). These data indicate that H2AX deficiency has little or no effect on post-integration repair. One possible explanation for this result is that H2AX function in these cells is redundant. We therefore examined transduction in H2AX-/-MEFs that had been complemented with transgenes that express wild type murine H2AX or proteins carrying either neutral, non-modifiable substitutions (S136/139A) or negatively charged substitutions that mimic constitutive phosphorylation (S136/139E) in the C-terminal PI-3K-related protein kinase target sites of H2AX. No significant difference was observed in the efficiencies of transduction of these lines compared to H2AX-/-cells (Table 1). It appears, therefore, that H2AX phosphorylation is dispensable for post-integration repair.

DISCUSSION
In this study, we show that retroviral infection induces the formation of histone γH2AX foci, and chromatin immunoprecipitation assays confirmed that H2AX phosphorylation occurs at sites of retroviral DNA integration. Therefore these results are consistent with our previous findings (17)(18)(19)(20) that cells respond to retroviral DNA integration in a manner similar to DSBs. We also demonstrate that an H2AX deficiency and an inability to phosphorylate this histone has no detectable effect on the efficiency of retroviral transduction of cultured mouse cell lines.
Because efficient expression of transduced genes requires stable retroviral vector integration, we conclude that H2AX phosphorylation is largely dispensable for post-integration repair of chromatin damage.
Although γH2AX is required for the accumulation of a subset of repair and signaling proteins into irradiation-induced foci (7,8,24,26,27), the exact role of H2AX phosphorylation is not well understood. Some DNA damage-sensing and repair proteins have been shown to interact physically with γH2AX (27)(28)(29). A functional role for γH2AX has been indicated, as H2AX deficient cells are hypersensitive to ionizing radiation, exhibit genomic instability, and also show an aberrant checkpoint response under certain conditions (7,8,30). On the other hand, although γH2AX foci form at sites of V(D)J recombination, H2AX appears to be dispensable for this reaction when tested with extra-chromosomal substrates in cultured cells (7,8,30).
The study of H2AX deficient mice has provided further insight into H2AX function. H2AX mice are viable, but DNA repair seems to proceed less efficiently in such animals, which show modest sensitivity to ionizing radiation and impairment in immunoglobulin class-switch recombination (31). In keeping with results from the cell-based assays cited above, these mice show no detectable abnormalities in V(D)J recombination (7,8). However, the genomic caretaker function of H2AX is more fully exposed when cell cycle checkpoints are compromised due to absence of p53 (25,32). In a p53-/-background, even H2AX +/-heterozygotes show  (30) or; 2. γH2AX interaction with such proteins might help to hold broken ends together thereby minimizing the risk of misrepair (25,(32)(33)(34). According to these models differential dependencies on H2AX are expected, with repair of free DSBs formed by ionizing radiation being more dependent on γH2AX than programmed recombination reactions (e.g. V(D)J) in which ends are held in proximity by recombination proteins (35). H2AX -/-spermatocytes show severe defects in meiotic X-Y chromosome pairing (12) and as such, it is tempting to speculate that γ-H2AX plays a bridging function during meiotic recombination.
In summary, our studies have produced two important findings. First they provide direct confirmation that cultured cells respond to retroviral DNA integration in the same way that they respond to DSBs produced by a variety of genotoxic agents or normal programmed events, namely massive phosphorylation of histone H2AX in the vicinity of the damage site. The second finding is that H2AX appears to be dispensable for post-integration repair. These observations lend independent support to a model in which the anchoring of broken DNA ends to facilitate their repair is a critical function of γH2AX. Because chromosome breaks are likely to be held together by the RAG1/2 complex during V(D)J recombination and are held covalently by viral DNA in retroviral integration (Fig. 4), such an anchoring function should not be essential in either reaction.