ATR and H2AX cooperate in maintaining genome stability under replication stress

Abstract

Chromosomal abnormalities are frequently caused by problems encountered during DNA replication. Although the ATR-Chk1 pathway has previously been implicated in preventing the collapse of stalled replication forks into double-strand breaks (DSB), the importance of the response to fork collapse in ATR-deficient cells has not been well characterized. Herein, we demonstrate that, upon stalled replication, ATR deficiency leads to the phosphorylation of H2AX by ATM and DNA-PKcs and to the focal accumulation of Rad51, a marker of homologous recombination and fork restart. Because H2AX has been shown to play a facilitative role in homologous recombination, we hypothesized that H2AX participates in Rad51-mediated suppression of DSBs generated in the absence of ATR. Consistent with this model, increased Rad51 focal accumulation in ATR-deficient cells is largely dependent on H2AX, and dual deficiencies in ATR and H2AX lead to synergistic increases in chromatid breaks and translocations. Importantly, the ATM and DNA-PK phosphorylation site on H2AX (Ser(139)) is required for genome stabilization in the absence of ATR; therefore, phosphorylation of H2AX by ATM and DNA-PKcs plays a pivotal role in suppressing DSBs during DNA synthesis in instances of ATR pathway failure. These results imply that ATR-dependent fork stabilization and H2AX/ATM/DNA-PKcs-dependent restart pathways cooperatively suppress double-strand breaks as a layered response network when replication stalls.

Figures

FIGURE 1.

Partial and complete suppression of ATR causes increased H2AX phosphorylation and increased reliance on H2AX for cellular viability upon replication fork stalling. A, quantification of ATR mRNA levels in ATR+/+ and ATR+/- MEFs following shRNA-mediated knockdown. G0-enriched MEFs were infected with shRNA-expressing lentivirus (“Experimental Procedures”) and stimulated in enter S phase. RNA was then isolated and ATR mRNA was quantified by quantitative PCR, normalized to β-actin control, and shown relative to wild-type ATR levels. Standard errors are represented by bars at the top of each column. B, ATR protein level as detected by immunoblot following shRNA-mediated knockdown, as described in A. mTOR was used as a loading control. C, detection of H2AX Ser139 phosphorylation following replication stress and varying degrees of ATR knockdown. MEFs with ATR knockdown were untreated, or treated for 1 or 2 h with aphidicolin prior to collection. ATRΔ/- cells were generated from ATRflox/- CreERT2+ cells treated with 4-OHT as described (“Experimental Procedures”). Immunoblots were detected for phospho-H2AX and non-phospho-H2AX as a protein level control. As positive and negative controls, H2AX-/- and wild-type MEFs were treated with 10 gray ionizing radiation (IR) and harvested 45 min later. D, cell proliferation in the presence and absence of low doses of aphidicolin. Population doublings of wild-type, ATRhypo, H2AX-/-, and ATRhypoH2AX-/- MEFs in the absence or presence of 0.2 or 0.4 μm aphidicolin after 4 days in culture following serum stimulation. The average values and S.E. (bars) from three independent experiments are indicated.

FIGURE 2.

H2AX phosphorylation in response to ATR deficiency occurs in S phase and is not caused by premature mitotic entry. A and B, cell cycle flow cytometric detection of phospho-H2AX. Wild-type and ATRΔ/- cells untreated or treated for 3 h with aphidicolin were detected for phospho-H2AX and phosphohistone H3, to quantify M phase cells (A), or propidium iodide for DNA content (B). As positive controls, wild-type cells were treated for 6 h with 0.5 μm nocodazole (noco) or exposed to 10 gray ionizing radiation (IR) and harvested 45 min later. Percentage of phospho-H2AX-positive cells is indicated for each condition, and frequency of both single-positive and double-positive cells for phospho-H2AX and phospho-histone H3 staining are indicated in A.

FIGURE 3.

ATM and DNA-PKcs are responsible for H2AX phosphorylation in ATR-deficient cells. A, the effect of ATM and DNA-PKcs chemical inhibition on H2AX phosphorylation in wild-type and ATRhypo cells. G0-enriched and restimulated MEFs were untreated or treated with aphidicolin for 1 or 2 h at peak S phase (16-19 h of FBS) prior to SDS-PAGE sample collection. Phospho-H2AX was detected by immunoblot; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was detected as a loading control. B, phosphorylation of H2AX in ATRhypo, ATRhypoATM-/-, and DNA-PKcs-inhibited ATRhypoATM-/- cells. C, phosphorylation of H2AX in ATRhypo, ATRhypoDNA-PKcs-/-, and ATM-inhibited ATRhypoDNA-PKcs-/- cells. Samples for B and C were generated and detected for phospho-H2AX and glyceraldehyde-3-phosphate dehydrogenase as described in A.

FIGURE 4.

ATR-deficient cells have increased Rad51 foci, but cells deficient for both ATR and H2AX do not. A, representative images of Rad51 foci in wild-type, ATRΔ/-, H2AX-/-, and ATRΔ/-H2AX-/- MEFs. Cells were stimulated to enter S phase following ATR deletion as described (“Experimental Procedures”), fixed, and detected for Rad51 by immunocytochemistry. Equivalent S phase entry was confirmed by BrdUrd staining. Nuclei were visualized by 4′,6-diamidino-2-phenylindole staining. B, quantification of the average number of Rad51 foci per cell. Data were collected from 4 independent experiments, and the mean ± S.E. were calculated. ATRΔ/- (ATRflox/-Cre-ERT2+) and ATRΔ/-H2AX-/- (ATRflox/-H2AX-/-Cre-ERT2+) cells were treated with 4-OHT for 48 h to recombine the ATRflox allele. As controls, wild-type (ATR+/+Cre-ERT2+) and H2AX-/- MEFs were similarly treated with 4-OHT. For all cells, 4-OHT was removed 16-19 h prior to fixation. Standard errors are represented by bars at the top of each column and p values were calculated by Student's t test.

FIGURE 5.

Combined loss of ATR and H2AX leads to increased DSBs. A, representative images of metaphase spreads from wild-type, ATRhypo, ATRhypoH2AX-/-, and aphidicolin-treated ATRhypoH2AX-/- cells. shRNA-mediated suppression of ATR was conducted as described in the legends to Figs. 1 and 3. Cells were collected 4 h after nocodazole treatment and metaphase spreads were prepared. B, average number of chromatid breaks/metaphase upon suppression of ATR and H2AX. MEFs of the indicated genotypes were G0-enriched, treated to suppress ATR levels, and stimulated to enter the cell cycle as described under “Experimental Procedures” and in the legend to Fig. 4. Mitotic spreads were isolated without prior treatment or following pulse treatment with 5 μm aphidicolin for 2 h, followed by a 2-h recovery period and subsequent collection in M phase (4 h nocodazole treatment). Identical procedures were performed on ATRhypoH2AX-/- MEFs complemented with wild-type H2AX or a serine 139 to alanine H2AX mutant (H2AXS139A), generated as described under “Experimental Procedures.” Average values from three independent experiments are depicted. Standard errors are represented by bars at the top of each column, and p values were calculated by Student's t test.

FIGURE 6.

Loss of ATR and H2AX leads to an increase in chromatid translocation events. A, example of a chromatid translocation observed in an ATRΔ/-H2AX-/- cell as determined by spectral karyotyping. Images of 4′,6-diamidino-2-phenylindole (DAPI)-stained and fluorescently-labeled chromosomes (spectral karyotyping, SKY) are shown. B, quantification of translocation events in wild-type, ATRΔ/-, H2AX-/-, and ATRΔ/-H2AX-/- MEFs. MEFs were G0-enriched, treated to delete ATR, and stimulated to enter the cell cycle as described under “Experimental Procedures” and in the legend to Fig. 4. For each genotype, 160 metaphases were examined, except wild-type MEFs (80 metaphases). p values were calculated by Fisher's exact test.

FIGURE 7.

A layered response network to replication stress. Model depicts the proposed relationships between ATR-dependent fork stabilization and ATM/DNA-PKcs/H2AX-enforced restart pathways. Deficiencies in ATR-mediated fork stability lead to DSB generation and the ATM- and DNA-PKcs-mediated phosphorylation of H2AX on Ser139. H2AX is required for accelerated Rad51 accumulation, which assists replication fork restart.