RAD52-mediated repair of DNA double-stranded breaks at inactive centromeres leads to subsequent apoptotic cell death

Abstract

Centromeres, where the kinetochore complex binds, are susceptible to damages including DNA double-stranded breaks (DSBs). Here, we report the functional significance and the temporally and spatially distinct regulation of centromeric DSB repair via the three pathways of non-homologous end joining (NHEJ), homologous recombination (HR) and single-strand annealing (SSA). The SSA factor RAD52 is most frequently recruited to centromeric DSB sites compared with the HR factor RAD51 and the NHEJ factor DNA ligase IV (LIG4), indicating that SSA plays predominant roles in centromeric DSB repair. Upon centromeric DSB induction, LIG4 is recruited to both active centromeres, where kinetochore complex binds, and inactive centromeres. In contrast, RAD51 and RAD52 are recruited only to inactive centromeres. These results indicate that DSBs at active centromeres are repaired through NHEJ, whereas the three pathways of NHEJ, HR and SSA are involved in DSB repair at inactive centromeres. Furthermore, siRNA-mediated depletion of either LIG4 or RAD51 promotes cell death after centromeric DSB induction, whereas RAD52 depletion inhibits it, suggesting that HR and NHEJ are required for appropriate centromeric DSB repair, whereas SSA-mediated centromeric DSB repair leads to subsequent cell death. Thus, SSA-mediated DSB repair at inactive centromeres may cause centromere dysfunction through error-prone repair.

Graphical Abstract

Figure 1.

Establishment of an experimental model to elucidate the molecular mechanisms of centromeric DSB repair in cultured human cells. (A, B) Immunofluorescence images of γH2AX and CENP-B in HT1080 (A) and U2OS (B) cells stably expressing Dox–Cas9 nuclease and sgRNA for control (sgCont) or centromere (sgCent). Enlarged images of the region enclosed by a white dotted line are shown. Data are representative of three independent experiments. Scale bar, 5 μm. (C) Percentage of cells containing centromeric γH2AX foci. (D) Experimental scheme for the time course of centromeric DSB repair after Dox withdrawal. (E, F) Percentage of cells expressing Cas9 nuclease (E) and γH2AX-positive cells (F), determined by immunofluorescence analysis of cells treated as shown in (D). (C, E, F) Data represent the mean ± standard deviation (SD) of three independent experiments.

Figure 2.

Cas9-induced centromeric DSBs recruit various repair factors involved in NHEJ, HR and SSA. (A, B) Immunofluorescence images of indicated repair factors and CENP-A/B in HT-sgCent cells treated (Dox +) or untreated (Dox -) with Dox. Enlarged images of the region enclosed by a white dotted line are shown. Data are representative of three independent experiments. Scale bar, 5 μm.

Figure 3.

HR and SSA are employed more frequently for centromeric DSB repair than NHEJ. (A) Percentage of HT-sgCent cells containing centromeric foci of indicated repair factors. (B) Percentage of HT-sgCent cells containing foci of LIG4 and RPA2 concomitantly or individually, treated or untreated with Dox. (A,B) Data represent the mean ± SD of three independent experiments.

Figure 4.

Centromeric DSBs are mainly repaired through SSA. (A,B) Immunofluorescence images of RPA2 and RAD51 (A), and RPA2 and RAD52 (B) in HT-sgCent cells treated with Dox. Enlarged images of the region enclosed by a white dotted line are shown. Data are representative of three independent experiments. Scale bar, 5 μm. (C,D,E) Percentage of HT-sgCent cells containing foci of RPA2 and RAD51 (C), RPA2 and RAD52 (D), and RAD51 and RAD52 (E) concomitantly or individually, treated or untreated with Dox. Data represent the mean ± SD of three independent experiments.

Figure 5.

NHEJ acts on centromeric DSB repair throughout the cell cycle, whereas HR/SSA function in centromeric DSB repair in the S and G2 phases. (A) Experimental scheme to examine the effects of CDKI on the centromeric foci formation of repair factors and the cell cycle upon induction of centromeric DSBs. (B) DNA content analysis by flow cytometry in HT-sgCent cells treated as shown in (A), stained with propidium iodide. (C) Percentage of HT-sgCent cells containing centromeric foci of indicated repair factors, treated as shown in (A). (D, E) Percentage of HT-sgCent cells containing RPA2 (D) and LIG4 (E) foci with or without cyclin A2 (CycA2) expression. (F, G) Immunofluorescence images of CycA2, and RPA2 (F) or LIG4 (G) in HT-sgCent cells. Enlarged images of the region enclosed by a white dotted line are shown. Scale bar, 50 μm. (H) Percentage of each cell cycle population in HT-sgCent cells treated or untreated with Dox. (C, D, E, H) Data represent the mean ± SD of three independent experiments. *P < 0.05. (B, F,G) Data are representative of three independent experiments.

Figure 6.

Recruitments of HR, SSA and NHEJ factors to centromeric DSB sites are spatially separately regulated. (A, B) Immunofluorescence images of RPA2, and LIG4 (A) or 53BP1 (B) in HT-sgCent cells treated with Dox. (C) Immunofluorescence images of LIG4, and myc-RAD51 or myc-RAD52 in HT-sgCent cells transfected with indicated expression vectors and treated with Dox. (D) Immunofluorescence images of RAD51 and RAD52 in HT-sgCent cells treated with Dox. (A–D) Enlarged images of the region enclosed by a white dotted line are shown. Data are representative of three independent experiments. Scale bar, 5 μm.

Figure 7.

NHEJ and HR play crucial roles in appropriate centromeric DSB repair, whereas SSA-mediated repair of centromeric DSBs leads to subsequent apoptotic cell death. (A) Experimental scheme to examine the effects of siRNA-mediated knockdown of repair factors on cell proliferation and apoptotic cell death after Dox withdrawal. (B) Immunoblotting analysis of HT-sgCent cells transfected with indicated siRNAs. Cells were harvested 48 h after transfection and then analyzed. Actin was used as a loading control. (C–E) Cell proliferation curve of HT-sgCent cells treated as shown in (A), determined by WST-8 assay. (F, G) Percentages of apoptotic cells treated as shown in (A). Apoptotic cell death was assessed on day 3 after Dox withdrawal by annexin V staining followed by flow cytometry analysis. (C–G) Data represent the mean ± SD of three independent experiments. *P < 0.05, against siControl.