Mammalian Ino80 Mediates Double-Strand Break Repair through Its Role in DNA End Strand Resection

Abstract

Chromatin modifications/remodeling are important mechanisms by which cells regulate various functions through providing accessibility to chromatin DNA. Recent studies implicated INO80, a conserved chromatin-remodeling complex, in the process of DNA repair. However, the precise underlying mechanism by which this complex mediates repair in mammalian cells remains enigmatic. Here, we studied the effect of silencing of the Ino80 subunit of the complex on double-strand break repair in mammalian cells. Comet assay and homologous recombination repair reporter system analyses indicated that Ino80 is required for efficient double-strand break repair. Ino80 association with chromatin surrounding double-strand breaks suggested the direct involvement of INO80 in the repair process. Ino80 depletion impaired focal recruitment of 53BP1 but did not impede Rad51 focus formation, suggesting that Ino80 is required for the early steps of repair. Further analysis by using bromodeoxyuridine (BrdU)-labeled single-stranded DNA and replication protein A (RPA) immunofluorescent staining showed that INO80 mediates 5′-3′ resection of double-strand break ends.

INTRODUCTION

Efficient repair of DNA double-strand breaks (DSBs) is required to prevent genomic instability and subsequent oncogenic transformation. Repair of DSBs is carried out by nonhomologous end joining (NHEJ) and homologous recombination (HR) (10, 37, 46). These processes are confronted with DNA being packaged into nucleosomes, which presents a natural barrier to the access of the repair machinery. Thus, to carry out repair, chromatin surrounding DNA breaks needs to be modified and remodeled. While histone modifications have been recognized as important to provide accessibility to DNA lesions and binding interfaces for the recruitment and retention of repair factors (15), studies in yeast have suggested that chromatin remodeling may also play an important role in repair. Chromatin-remodeling complexes use the energy of ATP hydrolysis to alter histone-DNA interactions and to reposition nucleosomes along DNA. The yeast INO80 chromatin-remodeling complex has been the most studied in this respect. Yeast strains deleted for subunits of the INO80 complex are hypersensitive to DSB-inducing agents (35). There is evidence suggesting that INO80 directly participates in both HR and NHEJ repair pathways (20, 40, 42, 43) and facilitates HR-mediated recovery of stalled replication forks (24, 36). In addition, a study by Wu et al. (45) showed that depletion of the Polycomb group protein Yin Yang1 (YY1) and/or Ino80 impaired UV-mediated DNA damage response and reduced HR rates in mammalian cells, suggesting that YYI mediates HDR through its interaction with INO80 complex, although the precise molecular events involving YY1-INO80 that govern HR repair have not been revealed. However, there are ambiguous results about its role in HR and chromatin processing of DSB ends. Thus, it has been shown that in Saccharomyces cerevisiae, the INO80 complex participated in nucleosome eviction in the vicinity of a DSB at the MAT locus, facilitating the recruitment of RAD51. However, in these experiments, HR does not take place because of the lack of the homologous donor sequences HMR and HML (40). It was subsequently shown that the activity of the INO80 complex was not necessary for repair by HR when homologous donor sequences were provided and HR was allowed to proceed (42). Further, one study indicated that strains deficient in components of the INO80 complex were less efficient in the initial 5′-3′ resection at DSB ends (43), while another study found normal strand resection and coating of the single-stranded DNA (ssDNA) with replication protein A (RPA) (40). In yeast, the INO80 complex is recruited to DSBs by binding to phosphorylated H2AX (20, 40, 43), and mutants lacking INO80 subunits are defective in Mre11, Ku80, and Mec1 recruitment in regions surrounding the DSBs, suggesting a role for INO80 in cell cycle checkpoint adaptation. The remarkable conservation of the INO80 complex composition between yeast, plants (8) and humans (11) suggests the importance of the INO80-mediated histone dynamics during repair of DSBs. Recent data in a mammalian system have demonstrated that INO80-depleted cells are deficient for Rad54B and XRCC3 expression and display DSB repair defect due to the transcriptional role of the INO80 complex. However, the repair defect of these cells was not completely rescued by reexpression of Rad54B and XRCC3, suggesting other mechanisms by which INO80 mediates DSB repair (25). Participation of the INO80 complex in DSB repair is also implied by the observation that the YY1 protein, a unique component of the mammalian INO80 complex, binds Holliday junction-like structures in vitro, suggesting a role of INO80 in the later stages of HR-mediated repair of DSBs. These studies strongly implicate INO80 in repair of DSBs in mammalian cells; however, the precise underlying mechanisms remain poorly understood.

To gain insight into how the INO80 complex contributes to DSB repair in mammalian cells, we examined the impact of depletion of the mammalian Ino80 protein (an ATPase that functions as the core of the complex) (3) on the efficiency and early steps of DSB repair in mammalian cells. We found that Ino80 is needed for efficient repair of DSBs by homologous recombination. In addition, our study revealed that Ino80 associates with chromatin upon induction of DSBs and that it is involved in 5′-3′ resection of DSB ends as well as the loading of the mediator protein 53BP1.

MATERIALS AND METHODS

Cell lines, culture conditions, and transfection.

The male embryonic stem (ES) cell line containing direct repeat-green fluorescent protein (DR-GFP) construct integrated in the hprt locus (clone 18-1/10) has been described previously (21). ES cells were initially grown on “feeders” and routinely on gelatin in Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% fetal calf serum (FCS), leukemia inhibitory factor, penicillin-streptomycin, and glutamine, pyruvate, and nonessential amino acids in 5% CO₂. 3T3 mouse fibroblast cells and human PC3 and U20S-AsiSI-ER cells (containing AsiSI expression construct) were grown in DMEM supplemented with 10% FCS, penicillin-streptomycin, and pyruvate. Medium for U20S-AsiSI-ER cells was supplemented with 1 μg/ml puromycin to ensure the retention of the pBABE hemagglutinin-AsiSI restrictase-estrogen receptor fusion (HA-AsiSI-ER) construct. Transfection of plasmids was carried out using Lipofectamine 2000, following the manufacturer's recommendations.

esiRNA knockdown.

Endoribonuclease-prepared small interfering RNAs (esiRNAs) targeting the coding regions of mouse Ino80 (3197 to 3564, transcript NM_026574.3), human Ino80 (3440 to 3894, transcript NM_017553.1), human Arp8 (485 to 916, transcript NM_022899.3) luciferase, or GFP (132 to 591) were synthesized as previously described (13, 48). Primers used to amplify the targeted regions were selected using the Riddle database (14) and were as follows: mIno80 (5′-TCACTATAGGGAGAGGTCGCAGATCCCAGTTCTTC; 5′-TCACTATAGGGAGACTCACTGTGTCTGCAGCTGTG), hIno80 (5′-TCACTATAGGGAGAGTGTGGAGCATCAGACCTCAG; 5′-CACTATAGGGAGACCCTGCTTTGTCTGCCCTAAG), and hArp8 (5′-TCACTATAGGGAGAGGGCACGCTCCTACAATAAGC; 5′-TCACTATAGGGAGACGTGCTGCTTAAGCCACTTCC).

Quantities of Lipofectamine and esiRNAs for efficient knockdown were optimized using esiRNA against Eg5 (Kif11). Typically, 60 pmol of esiRNA and 3 μl of Lipofectamine 2000 were used per well in a 24-well plate (500-μl transfection volume).

Primers used for reverse transcription-PCR (RT-PCR) to assess Ino80 transcript level in ES cells were 5′-GGATGCAAGATGCCACACTA and 5′-TGCTGCATACGGGTACAATG. Primers to assess the transcript level of human Arp8 were 5′-TGATGGCCGGCAACGATTCCG and 5′-TTCCATGCAATCAGCCGGGGG.

Western blots and biochemical fractionation.

Protein lysates were resolved on SDS-PAGE gels (6% or 12.5%, as appropriate) and blotted on nitrocellulose membrane (Bio-Rad). The primary antibodies (Abs) used were rabbit anti-Ino80 antibody (1) (1:500; a gift from C. Wu), rabbit anti-RAD51 antibody (1:1,000; Oncogene), and mouse anti-actin (1:10,000; ICN Biomedicals, Inc., Irvine, CA). Proteins were visualized with horseradish peroxidase (HRP)-conjugated anti-rabbit or anti-mouse IgG (1:10,000; Dako, Trappes, France), followed by use of the ECL chemiluminescence system (Amersham) or using Li-Cor Odyssey infrared imaging system with appropriate IRDye-labeled secondary antibodies (Li-Cor Biosciences, Lincoln, NE).

Fractionation was carried out essentially as described in reference 17. Cells were washed twice with phosphate-buffered saline (PBS), scraped, and resuspended in solution A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 0.34 M sucrose, 10% glycerol, 1 mM dithiothreitol, and protease inhibitors) at about 3 × 10⁷ to 4 × 10⁷ cells/ml. Triton X-100 was added to a final concentration of 0.1%, and the cells were incubated for 5 min on ice. The fraction of soluble proteins was separated from the chromatin fraction by centrifugation (5 min, 1,250 × g) and further clarified by centrifugation at 14,000 × g for 5 min. The chromatin fraction was incubated in buffer B (3 mM EDTA, 0.2 mM EGTA, 1 mM dithiothreitol, protease inhibitors) for 30 min, centrifuged for 5 min at 1,500 × g, and washed once again in solution B, and the pellet was resuspended and sonicated.

Comet assay.

DSB repair was assayed by alkaline single-cell agarose gel electrophoresis as described previously (22). Briefly, cells were irradiated with 20 Gy, harvested (∼100,000 cells per pellet), mixed with low-gelling-temperature agarose (Sigma; type VII), and layered onto agarose-coated glass slides. All further steps were carried out in the dark at 4°C. After gelling, the slides were submerged in lysis buffer (2.5 M NaCl, 0.1 M EDTA, 10 mM Trizma base, 1% Triton X-100, 10% dimethyl sulfoxide [DMSO]) for 1.5 h, washed with Tris buffer, and incubated for 45 min in alkaline electrophoresis buffer (300 mM NaOH, 1 mM EDTA at pH 10). After electrophoresis (∼50 min, 25 V), air-dried and neutralized slides were stained with 30 μl of ethidium bromide (20 μg ml⁻¹). Tail moment was scored by using Comet Imager software (MetaSystems, Altlussheim, Germany).

Homologous recombination assay.

For the HR assay, the I-SceI expression vector pCBASce34 was transfected into ES cells by using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Plasmid pCAGGS (without the I-SceI gene) and GFP-expressing plasmid pEGFP-N3 were used for mock and transfection efficiency controls, respectively. Flow cytometry analysis was performed 48 h after transfection using a FACScalibur apparatus with Cellquest software (Becton Dickinson, San Jose, CA).

Immunofluorescence.

For immunofluorescence, cells were grown on coverslips, washed in PBS, fixed with 2% formaldehyde in PBS for 5 min at room temperature, permeabilized with 0.2% Triton X-100 in PBS for 5 min, washed with PBS, and blocked in 3% bovine serum albumin (BSA) in PBS for 1 h. Staining was done using rabbit 53BP1 antibody or rabbit Rad51 Ab (Oncogene) diluted 1:100 overnight at 4°C. Staining for γ-H2AX was done using a mouse primary antibody (Ab18311; Abcam, Cambridge, United Kingdom). To stain cells for RPA, coverslips were washed with cytoskeletal (CSK) buffer {10 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)]-KOH [pH 6.8], 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 1 mM EGTA, 0.5% Triton X-100} for 5 min on ice, fixed, and stained with anti-RPA primary antibody (Ab2175; Abcam). Slides were washed 3 times for 5 min each in PBS, and secondary antibodies were applied. Secondary fluorescein isothiocyanate (FITC)-conjugated anti-rabbit IgG (Dako) or anti-mouse Texas Red-conjugated IgG (Sigma) was used at a 1:200 dilution for 1 h at room temperature, and after three 5-min washes with PBS, slides were mounted using Vectashield mounting medium (Vector Laboratories, Burlingame, CA).

For ssDNA staining, cells were grown in the presence of 30 μg/ml bromodeoxyoridine (BrdU) for 36 h and seeded on coverslips to achieve about 60% confluence. Cells were washed in PBS, permeabilized in CSK buffer for 5 min on ice, incubated in cytoskeletal stripping buffer (10 mM Tris, pH 7.4; 10 mM NaCl; 3 mM MgCl₂; 1% Tween 20; 0.5% Na deoxycholate) for 5 min on ice, washed 3 times with PBS, and fixed in 2% formaldehyde-2% sucrose in PBS for 10 min. Cells were permeabilized with 0.2% Triton X-100 in PBS for 5 min, washed three times for 5 min each in PBS, and blocked in 3% bovine serum albumin in PBS for 1 h. Staining was done with a mouse anti-BrdU antibody (catalog number 347580; Becton Dickinson) diluted 1:5 overnight at 4°C, and after three 5-min washes in PBS, the secondary FITC-conjugated anti-mouse secondary Ab was applied for 1 h at room temperature. Following 3 or 4 washes with PBS, slides were mounted using VectaShield mounting medium.

ChIP assay.

U20S-AsiSI-ER cells were grown in 15-cm dishes, serum starved for 24 h, and released in normal medium with or without 4-hydroxytamoxifen (4-OHT) for the times indicated to induce translocation of the AsiSI-ER fusion. Cutting efficiency was assessed by real-time PCR using genomic DNA and the following primers spanning the AsiSI recognition site: 5′-GAACTGTTCGGTCCCCTTCT and 5′-CTACCTCCAGGGGCTGCAT. To cross-link proteins to DNA, formaldehyde was added directly to the culture medium to a final concentration of 1% for 10 min at room temperature. Cross-linking was stopped by the addition of glycine to a final concentration of 0.125 M, followed by an additional incubation for 5 min. Fixed cells were washed twice with PBS and harvested in SDS buffer (50 mM Tris at pH 8.1, 0.5% SDS, 100 mM NaCl, 5 mM EDTA, and protease inhibitors). Sonication was carried out in an ∼200-μl volume on ice, aiming to achieve genomic DNA fragments with a bulk size of 400 to 500 bp. Lysates were clarified by centrifugation for 10 min at maximal speed in a refrigerating Eppendorf centrifuge. Cells were pelleted by centrifugation and diluted 10 times in chromatin immunoprecipitation (ChIP) dilution buffer (16.7 mM Tris at pH 8.1, 167 mM NaCl, 0.01% SDS, 1.1% Triton X-100, and 2 mM EDTA). One milliliter of diluted lysate was precleared by the addition of 50 μl of blocked protein A beads (50% slurry protein A-Sepharose, Amersham; 0.5 mg/ml bovine serum albumin, Sigma; and 0.2 μg/ml salmon sperm DNA in TE). Samples were immunoprecipitated overnight at 4°C with polyclonal Ino80 antibody and immune complexes recovered by adding 50 μl of blocked protein A beads and incubated for 2 to 4 h at 4°C. Beads were washed successively in low-salt wash buffer (0.1% SDS, 1% Triton X-100; 2 mM EDTA; 20 mM Tris-HCl, pH 8.1; 150 mM NaCl), high-salt wash buffer (0.1% SDS, 1% Triton X-100; 2 mM EDTA; 20 mM Tris-HCl, pH 8.1; 500 mM NaCl), LiCl wash buffer (0.25 M LiCl; 1% NP-40; 1% Na deoxycholate; 1 mM EDTA; 10 mM Tris-HCl, pH 8.1), and twice in TE (10 mM Tris-HCl, pH 8.1; 1 mM EDTA) and eluted in elution buffer (1% SDS; 0.1 M NaHCO₃). Cross-linking was reversed by overnight incubation at 65°C after adding NaCl to 0.2 M, and DNA was isolated. Primers used for PCR were as follows: at 0.5 kb away from the AsiSI site, 5′-CCCTGGAGGTAGGTCTGGTT and 5′-CGCACACTCACTGGTTCCT; at 10 kb, 5′-CCTGACCTTGTGATCCACCT and 5′-TGAATGACTGTGGCTGTGGT; and in the human β-globin locus, 5′-TAACTTCCAAAGAACAAGTGC and 5′-GCGGCTAAAAGACCAGA. To calculate enrichment of Ino80, the amount of template from the IP reactions (minus no-antibody control) was normalized to the template amount of the input and expressed as a percentage of the nontreated control.

Flow cytometry.

For analysis of GFP expression, cells were harvested by trypsinization and analyzed by a FACSCalibur apparatus with Cellquest software (Becton Dickinson). To analyze cell cycle profiles, cells were harvested by trypsinization and fixed in 70% ethanol. Before analysis, cells were resuspended in PBS, treated with RNase A (20 μg/ml), and stained with propidium iodide (20 μg/ml), and analysis was carried out by a FACSCalibur apparatus with Cellquest software (Becton Dickinson).

RESULTS

Ino80 knockdown cells exhibit DNA repair defects.

To study the effect of Ino80 on double-strand break repair, we employed RNA interference by esiRNA to deplete the protein in mouse ES cells, 3T3 fibroblasts, and the human PC3 cell line. RNA interference by esiRNA has been reported to cause effective knockdown of protein expression (6) and exhibits minimal off-target effects (14, 48). Three days after esiRNA transfection in mouse ES cells, the amount of Ino80 transcript was almost undetectable, whereas the Ino80 protein level was greatly reduced in knockdown compared to control cells (Fig. 1A and B). Similar reduction of the protein levels of Ino80 was obtained in the other two cell lines (see Fig. 5C and 6D). Under these conditions of Ino80 silencing, no changes in cell viability were observed (data not shown).

Fig. 1.

Reduced repair of DNA breaks in INO80-depleted cells. (A) Seventy-two hours after ES cells were transfected with esiRNA against luciferase (mock) or Ino80 (siIno80), mRNA levels were assessed by RT-PCR. (B) Western blot of total extracts from ES cells transfected with esiRNAs against luciferase (mock) or Ino80 (siIno80) with an antibody against Ino80 3 days after transfection. (C) ES cells were transfected with esiRNA against Ino80 (siIno80) or luciferase (mock). Three days after transfection cells were irradiated with 20 Gy of ionizing radiation, left to recover for the indicated times, and subjected to alkaline comet assay. Representative images of comets are given for each time point. NI, not irradiated. (D) Tail moment of at least 100 comets was measured in each of two independent experiments; bars represent distribution of mean tail moments and error bars show standard deviation from the mean (SDM). (E) Cell cycle profiles of control and Ino80-depleted cells irradiated with 20 Gy of IR after different times of recovery.

Fig. 5.

ssDNA formation in Ino80-silenced cells. (A) 3T3 cells were silenced for 72 h and labeled with 30 μg/ml BrdU during the last 36 h. Cells were irradiated with 20 Gy of IR, fixed, and stained at the indicated times after irradiation. (B) The percentage of BrdU-positive cells was estimated (black bars, control mock-silenced cells; gray bars, Ino80-silenced cells). Data are from 2 independent experiments, and error bars show SDM. (C) Western blot of total extracts from 3T3 cells transfected with esiRNAs against GFP (mock) or Ino80 (siIno80) with an antibody against Ino80 3 days after transfection. DAPI, 4′,6-diamidino-2-phenylindole.

Fig. 6.

RPA focus formation in Ino80-silenced cells. (A) PC3 cells were silenced for 72 h with esiRNA against Ino80 or GFP as the control. Following irradiation with 8 Gy of IR, cells were fixed at indicated times and stained with an anti-RPA32 antibody. (B) The number of cells containing foci were counted. Results are from 3 experiments ± SDM. (C) Number of foci per cell containing foci. (D) Western blot of total extracts from PC3 cells transfected with esiRNAs against GFP (mock) or Ino80 (siIno80) with an antibody against Ino80 3 days after transfection.

To test whether DNA repair efficiency is affected in cells lacking Ino80, Ino80-silenced ES cells and mock-silenced cells (transfected with esiRNA against luciferase) were irradiated with 20 Gy, and the repair kinetics was measured using the alkaline comet assay. The comet assay is a single-cell gel electrophoresis technique that is widely used as a quick and reliable method of analyzing and quantifying DNA repair capacity (16). As shown in Fig. 1C and D, following exposure to irradiation, the initial induction of DNA breaks in Ino80-depleted cells was similar to that in Ino80-expressing cells, as revealed by similar tail moments 15 min postirradiation. However, the Ino80-depleted cells exhibited a lower recovery rate, as more cells had higher tail moment at later time points. These results suggest that the cells lacking Ino80 exhibit a repair defect. To assess whether the compromised repair was due to defects in damage signaling, we analyzed cell cycle distribution in mock-silenced and Ino80-silenced cells after irradiation. As shown in Fig. 1E, in both Ino80-proficient and Ino80 knockdown cells, G₂/M cell cycle block was induced 8 h after irradiation, which was overcome by the 24th hour, but the cell cycle profiles were comparable in the two populations at all time points analyzed (Fig. 1E). The results argue that in mammalian cells, Ino80 depletion does not affect DNA damage-induced checkpoint activation.

Requirement for Ino80 in HR repair.

In mammalian cells, NHEJ is the predominant form of DSB repair operating in all phases of the cell cycle, whereas HR functions only in late S and G₂ phases (38). Data in yeast (40, 43) show discrepancies regarding the participation of Ino80 in homology-directed repair. To determine whether Ino80 participates in HR repair, we took advantage of the DR-GFP construct integrated in the hprt locus of mouse ES cells (21, 27). Expression of I-SceI endonuclease induces a single DSB in an out-of-frame GFP reporter, which, when repaired by error-free HR using a homologous GFP sequence in the reporter, restores the expression of functional GFP that can be detected as fluorescence and quantified by flow cytometry. Ino80-depleted cells and Ino80-containing cells were electroporated with either an empty vector or the I-SceI-expressing vector 72 h following transfection with esiRNAs, and the fractions of cells that had undergone HR repair events were quantified 48 h thereafter by flow cytometry. After electroporation with the empty vector, only a few cells were GFP positive in either Ino80-depleted or mock-silenced cells (Fig. 2A, top). After introduction of I-SceI, GFP-positive cells were readily detected in each cell population (Fig. 2A, bottom), indicative of DSB-induced gene conversion. However, we observed that the recombinant fraction in Ino80 knockdown cells was significantly reduced (∼2-fold) in comparison with the Ino80-containing cells (Fig. 2A, bottom, and B). The reduction of the GFP-positive population could not be attributed to differences in transfection efficiency or significant effects of Ino80 knockdown on transcription from the cytomegalovirus (CMV) promoter, since transfection with pEGFP-N3 plasmid, where GFP expression is driven by the same promoter as that of I-SceI, resulted in very similar GFP-positive population sizes (Fig. 2C). These data further strengthen the conclusion that in mammalian cells Ino80 is necessary for efficient homology-directed repair.

Fig. 2.

Reduced HR repair rate in INO80-depleted cells. (A) Mock- and Ino80-silenced ES cells (72 h after transfection with esiRNA) were transfected with I-SceI-expressing plasmid (+I-SceI) or empty vector (−I-SceI) and analyzed by fluorescence-activated cell sorting (FACS) 48 h later. The GFP-positive population is surrounded by ellipses; data are means of 3 independent experiments, and error bars show SDM. (B) Repair efficiency of Ino80-depleted cells, relative to the control. (C) ES cells were mock silenced or Ino80 silenced and 72 h after esiRNA transfection were transfected with pEGFP-N3 plasmid. Forty-eight hours after plasmid transfection, the percentage of GFP-positive cells was assessed by flow cytometry and is given as the expression efficiency of the CMV promoter.

Ino80 localizes to sites of double-strand breaks following induction of DNA damage.

A prerequisite for direct involvement is the association of the protein with sites of repair. To examine whether induction of DSBs results in an increased association of Ino80 with chromatin, we analyzed the distribution of the protein among the soluble and chromatin-bound fractions following biochemical fractionation of mouse 3T3 cells exposed to ionizing radiation (IR). Western blot analysis using antibodies against Ino80 revealed that after irradiation, the percentage of soluble Ino80 was decreased, while the chromatin-bound protein was markedly increased (over 3-fold change compared to nonirradiated cells) (Fig. 3A and B). These results show that DNA damage induces a significant redistribution of Ino80 between the soluble and chromatin-bound fractions.

Fig. 3.

Association of human Ino80 with DSB. (A) Mouse 3T3 cells were irradiated with 20 Gy of IR and fractionated into Triton-soluble (soluble) and chromatin-bound (chr-bound) fractions. Fractions from control and irradiated cells were analyzed by Western blotting. (B) After densitometry, the relative amount of Ino80 was estimated. Data are means of two experiments, and error bars show SDM. (C) U20S-AsiSI-ER cells were treated with different concentrations of 4-OHT. Genomic DNA was extracted after 3 h of treatment, and the efficiency of AsiSI cutting was assessed by real-time PCR with primers spanning the recognition site on chromosome 1. NT, no treatment. (D) U20S-AsiSI-ER cells were treated with 1 μM 4-OHT for the times indicated, lysed, and subjected to ChIP with an antibody against Ino80. PCR was carried out with primers amplifying a region 500 bp away from the AsiSI recognition site. (E) ChIP was carried out as in D, and the amounts of precipitated DNA from the site 0.5 kb from the AsiSI site (white bars), from 10 kb away from that site (light gray bars), and from a region in the β-globin locus (dark bars) were analyzed by real-time PCR. The data from 2 experiments are shown as fold enrichment of Ino80 at the respective sites.

We next examined whether redistribution of Ino80 after DNA damage is associated with the recruitment of the protein to DSB ends. For this, we used the human cell line U20S, which contains stably integrated construct expressing the AsiSI restrictase-estrogen receptor fusion (AsiSI-ER), to perform chromatin immunoprecipitation (ChIP) assays at a defined AsiSI cleavage site. Upon addition of 4-hydroxytamoxifen (4-OHT), the fusion protein is translocated to the nucleus and cuts AsiSI recognition sites. As shown in Fig. 3C, 3 h after the addition of 4-OHT, the cutting efficiency was more than 80%, as judged by quantitative PCR. The ChIP assay revealed that induction of a DNA break resulted in an increase of Ino80 associated with the chromatin surrounding the AsiSI recognition site (Fig. 3D and E). Real-time PCR showed up to a 3-fold increase of Ino80 associated at 0.5 kb from the DSB and a less pronounced accumulation at 10 kb from the break site. No accumulation was observed using primers to amplify a genomic fragment from the β-globin locus situated at about 1.2 Mb from the nearest AsiSI site (Fig. 3E). These findings demonstrate that Ino80 is recruited to the sites surrounding DNA double-strand breaks and suggests direct involvement of the protein in the repair process.

Ino80 depletion impairs recruitment of early repair proteins to the sites of DSB.

The repair proteins form nuclear complexes at sites of DNA damage that are microscopically detectable as foci. These foci appear in response to genotoxic agents including IR and are thought to be required for the repair of DNA DSBs (31). To assess whether Ino80 is required early or later in the repair process, we monitored the accumulation of 53BP1 and Rad51 at DSBs under conditions of Ino80 depletion. 53BP1 is considered as a mediator/adaptor of the DNA damage response and is recruited to nuclear foci within minutes following genotoxic insult (7, 30, 34). In contrast to 53BP1, Rad51 forms nuclear foci at least 1 h after damage induction. It functions as a helical nucleoprotein filament loaded on the single-stranded DNA 3′ overhangs and constitutes the core of the HR repair reaction (44). To monitor recruitment and accumulation of these proteins at sites of DNA damage, control and Ino80 knockdown cells were irradiated, and at different times thereafter, the appearance of 53BP1 and Rad51 repair foci was examined. 53BP1 foci were readily induced in Ino80-containing cells, as revealed by a rapid transformation of diffuse nuclear staining into distinct foci already at 30 min after the treatment (Fig. 4A and B). In Ino80-deficient cells, a striking reduction in 53BP1 focus formation was observed 30 min after IR when the decrease of the cells with foci was 4-fold. The differences in both the percentage of cells with foci and the number of foci per cell were diminished at later time points (Fig. 4A to C). The defect in 53BP1 focus formation was not due to changes in the expression of 53BP1 caused by Ino80 depletion (Fig. 4F). These results indicate that depletion of Ino80 affects an early event in the repair of DSBs. We next examined whether Ino80 is required for IR-induced focus formation of Rad51. Immunofluorescence staining for Rad51 failed to reveal differences in focus formation in Ino80-deficient cells, with the exception of a slight (about 10%) increase of the cells with Rad51 foci at the 6th hour after irradiation (Fig. 4D and E). Rad51 protein level was not affected by knockdown of Ino80 (Fig. 4G). These results show that depletion of Ino80 does not impede accumulation of Rad51 to sites of DNA breaks, suggesting that Ino80 may be dispensable for this later stage of the repair process.

Fig. 4.

Repair focus formation in Ino80-silenced cells. Mock- or Ino80-silenced ES cells were irradiated with 8 Gy and 0.5, 1, or 2 h later were fixed and stained with an antibody against 53BP1. Representative images (A) and quantification of cells with foci (B) are shown (black bars, mock-silenced cells; gray bars, Ino80-silenced cells) of two independent experiments are shown ± SDM. (C) Number of foci per cell containing foci; at least 100 cells were counted for each point, and error bars represent SDM. (D) Control and Ino80-silenced cells were irradiated with 8 Gy, left to recover for the times indicated, and stained with an antibody against Rad51. (E) Cells with more than 5 foci were counted. The data are from 3 independent experiments. (F and G) Protein levels of 53BP1 and Rad51 in mock- and Ino80-silenced cells.

Effect of Ino80 depletion on ssDNA formation during 5′-3′ resection.

A prerequisite for recombinational repair of a DSB is the 5′-to-3′ resection of the two ends to yield 3′ single-stranded overhangs, carried out by a partnership between the MRN (Mre11-Rad50-Nbs1) complex and CtIP, followed by a processive step involving helicases and exonucleases (19). Therefore, we next tested whether the recruitment of Ino80 to DSB ends is involved in processing of chromatin at DSB ends by checking 5′-3′ resection efficiency in Ino80-depleted cells. Ino80-containing or -deficient 3T3 cells were incubated with BrdU for 36 h to uniformly label DNA, after which BrdU focus formation was monitored after exposure to IR. As BrdU is detected by the anti-BrdU antibody only in the context of single-stranded DNA, microscopically visualized foci with the BrdU antibody can be used as a measure of single-strand resection (29). Exposure to irradiation induced BrdU foci in both control and Ino80 knockdown cells, and the number of cells containing foci rapidly increased, reaching a maximum at 90 min postirradiation (Fig. 5A and B). However, in Ino80-silenced cells, the number of cells containing BrdU foci was reduced. The reduction was most pronounced (∼2-fold) at 30 min after irradiation, after which the difference diminished (Fig. 5B). These results indicate that Ino80 is important for ssDNA formation.

To further extend the finding that Ino80 participates in 5′-3′ strand resection by an independent approach, we investigated the ssDNA formation using nuclear focus formation of the mammalian single-strand binding protein RPA, a component of the minimal protein complex capable of DNA end resection (2). Human PC3 cells were mock transfected or transfected with Ino80 esiRNA and after 72 h were exposed to IR. As shown in Fig. 6A, RPA focus formation was impaired in Ino80-depleted cells after exposure to irradiation, as both the percentage of cells with foci and the number of foci per cell were reduced about 2-fold in Ino80-depleted cells compared to Ino80-containing cells (Fig. 6B and C). Together, these results demonstrate that in mammalian cells Ino80 is required for HR repair of DNA breaks through its function in resection of DSB ends.

Ino80 depletion and H2AX phosphorylation.

To test whether DNA damage induction, detection, or signaling was affected following depletion of Ino80, we stained mock- and Ino80-silenced PC3 cells with an antibody against γ-H2AX at different times after 8 Gy of IR. We failed to detect differences in the intensity or the pattern of γ-H2AX foci between Ino80-depleted and control cells (Fig. 7). We conclude that the effects of Ino80 depletion on DSB end processing may not be explained by changes in the induction of DSBs.

Fig. 7.

H2AX phosphorylation in Ino80-silenced cells. PC3 cells were silenced for 72 h with esiRNA against Ino80 or GFP as the control. Following irradiation with 8 Gy of IR, cells were fixed at indicated times and stained with a γ-H2AX antibody.

Effect of Arp8 depletion on 5′-3′ resection.

To understand if the participation of Ino80 in 5′-3′ strand resection was specific to the protein or a characteristic function of the whole INO80 complex, we knocked down Arp8, another unique subunit of the complex that mediates its recruitment to DNA damage sites (12). We followed its effect on RPA focal recruitment in irradiated cells. PC3 cells were transfected with esiRNAs against Arp8 or GFP (Fig. 8D) and 3 days thereafter exposed to IR. After staining with an RPA32 antibody at different times postirradiation, we observed an effect that paralleled that in Ino80-depleted cells. There were fewer cells containing RPA foci, and the number of foci per cell were diminished in Arp8-deficient cells (Fig. 8A to C). The fact that depletions of both unique subunits of the INO80 chromatin-remodeling complex have similar outcomes indicates that the remodeler is required for efficient DSB end processing.

Fig. 8.

RPA focus formation in Arp8-silenced cells. (A) PC3 cells were silenced for 72 h with esiRNA against Arp8 or GFP as the control. Following irradiation with 8 Gy of IR, cells were fixed at indicated times and stained with an anti-RPA32 antibody. (B) The number of cells containing foci were counted. Results are from 2 experiments. (C) Number of foci per cell containing foci. (D) RT-PCR to assess Arp8 mRNA levels in PC3 cells transfected with esiRNAs against GFP (mock) or Arp8 (siArp8) 3 days after transfection.

DISCUSSION

The mechanistic aspects of the involvement of the chromatin-remodeling complex INO80 in the repair of DSBs remain poorly understood due to discrepancies in the results obtained for yeast systems and insufficient data for the role of INO80 in DSB repair in mammalian cells. Some of the inconsistencies on the role of Ino80 in DNA repair reported in the previous studies may be due to the different experimental systems and cell types used, as well as to the nature of the genetic modification introduced into the cells. Here, we have attempted to elucidate the role of the INO80 complex in DSB repair by studying the impact of Ino80 depletion in mammalian cells. Our data indicate that repair efficiency of DNA breaks is compromised in cells depleted for Ino80. Ino80 knockdown cells repaired DNA damage at a lower rate, as judged by their higher tail moment when subjected to comet assay. By using cells containing the DR-GFP construct, we observed lower portion of fluorescent cells, meaning that homology-directed repair in Ino80 knockdown cells was compromised. This is in line with previous observations showing that yeast strains lacking INO80 complex subunits are hypersensitive to hydroxyurea and methyl methanesulfonate (MMS) (43) and that Ino80 depletion results in reduced HR rates in mammalian cells (45) and plants (8). We observed recruitment of the Ino80 protein to chromatin in the vicinity of DSBs, which implies conservation of Ino80 behavior in response to damage from yeast to mammals (40, 43). The physical association of Ino80 protein with the sites of DSBs strongly suggests the direct participation of Ino80 in the repair of these lesions, as opposed to transcription-mediated effects. These data are also in agreement with a previous study that showed recruitment of mammalian INO80 complex to DNA damage sites in an Arp8-dependent manner (12).

We failed to observe defects in checkpoint activation following DNA damage, as manifested by the lack of differences in cell cycle profiles and induction of γ-H2AX foci in Ino80-depleted and control cells, unlike data previously reported for yeast (23, 42). However, we note that checkpoint activation defects were not observed in a yeast strain harboring mutation in Arp8, as judged by Rad53 phosphorylation (43). Also, no differences in checkpoint activation and γ-H2AX phosphorylation were observed in a yeast Ino80 mutant strain after hydroxyurea treatment (5).

Earlier work has shown that the mammalian Trrap-TIP60 histone acetylase is required for loading of a subset of repair factors, including 53BP1 and Rad51, to sites of double-strand breaks (9, 21, 47). We studied recruitment of 53BP1 as an early repair protein and Rad51 as a late one to figure out the step of repair for which the INO80 chromatin-remodeling complex is required. Our observations suggest that when the amount of Ino80 protein is reduced, the recruitment of 53BP1 to repair foci is impaired. This may reflect a requirement for Ino80 in the remodeling of chromatin structure at break sites, needed to expose methylated H4K20, that has been reported as a prerequisite for 53BP1 binding (26, 32, 33, 39). Thus, our data indicate for the first time in mammalian cells that in addition to histone posttranslational modifications (7), chromatin remodeling is also required for efficient loading of early repair proteins. It also suggests that Ino80 is involved early in the DSB repair process. A previous study showed that yeast strains deficient in components of the INO80 complex are less effective in the initial 5′-3′ resection at DSB ends prior to strand invasion in the homologous duplex template (43). However, an independent study failed to observe an impaired strand resection in INO80-defective yeast strains (40). Our data using both immunofluorescence for BrdU-labeled ssDNA and RPA demonstrate that Ino80 and Arp8 are required for normal resection of DSB ends. These observations confirm results from yeast showing that INO80 remodeler participates in 5′-3′ resection of DSBs and demonstrate for the first time in mammalian cells that INO80 is required for the processing of DSB ends during the initial stages of repair.

Chromatin immunoprecipitation studies in yeast have shown that INO80 controls the recruitment of repair proteins and that an Arp8Δ mutant yeast strain does not efficiently recruit Rad51 to a HO nuclease-induced break site (40). However, the same authors have shown that when a homologous donor sequence is present, which mimics the situation in diploid mammalian cells, Rad51 is recruited normally (41). We did not observe reduced Rad51 focus formation in Ino80-silenced cells up to 6 h after irradiation, even though DSB end resection was affected. Our interpretation of these data is that Ino80 influences early steps of repair, being strongest at 30 min after induction of breaks, as evidenced by BrdU, 53BP1, and RPA focus formation, but the effect is significantly diminished by the time when Rad51 foci appear. This notion is exemplified by the observation that at 180 min postirradiation when cells with Rad51 foci are at their maximum (Fig. 4E), the difference between Ino80 knockdown and control cells containing BrdU foci has already disappeared (Fig. 5B). Yet, this temporally limited impairment of an early step in the repair clearly influenced the overall repair efficiency (Fig. 2). A similar phenomenon of transient initial interference, which was quickly overcome yet capable of influencing the outcome of DSB repair, has been described for another epigenetic mechanism—that of H4K16 deacetylation by HDAC1 and HDAC2 (18).

Previously, we reported that RUVBL proteins that are components of both Trrap-TIP60 and INO80 modulate Rad51 loading in response to DNA damage (9). Our results suggested that this effect was due to impaired Trrap-TIP60 but not INO80 function in RUVBL knockdown cells, which is in agreement with the data presented here and imply that INO80 may not be required for Rad51 recruitment. On the other hand, the slight increase of cells containing Rad51 foci that we detected at the 6th hour may be interpreted as a delayed disappearance of Rad51 foci in Ino80-silenced cells, presumably caused by a requirement of INO80 in the resolution of Holliday junctions, a notion suggested by in vitro studies (45).

On the basis of our results, we propose a model for the role of the INO80 chromatin-remodeling complex in DNA repair in mammalian cells, in which, upon induction of DSB, Ino80 participates in early steps of homology-directed repair to produce the RPA-coated 3′ overhangs. In this model, the subsequent step involving the displacement of RPA by RAD51 is independent of INO80. These data also demonstrate the evolutionary conservation of INO80-mediated histone dynamics during repair of DSBs (4, 28). Because proficient DNA repair is essential for genomic stability, deregulation of the chromatin-based mechanism that facilitates HR repair of DSBs might result in the induction of mutations, leading to an aberrant inactivation of tumor suppressor genes and the promotion of tumorigenesis.