RAD52 Adjusts Repair of Single-Strand Breaks via Reducing DNA-Damage-Promoted XRCC1/LIG3α Co-localization

Abstract

Radiation sensitive 52 (RAD52) is an important factor for double-strand break repair (DSBR). However, deficiency in vertebrate/mammalian Rad52 has no apparent phenotype. The underlying mechanism remains elusive. Here, we report that RAD52 deficiency increased cell survival after camptothecin (CPT) treatment. CPT generates single-strand breaks (SSBs) that further convert to double-strand breaks (DSBs) if they are not repaired. RAD52 inhibits SSB repair (SSBR) through strong single-strand DNA (ssDNA) and/or poly(ADP-ribose) (PAR) binding affinity to reduce DNA-damage-promoted X-Ray Repair Cross Complementing 1 (XRCC1)/ligase IIIα (LIG3α) co-localization. The inhibitory effects of RAD52 on SSBR neutralize the role of RAD52 in DSBR, suggesting that RAD52 may maintain a balance between cell survival and genomic integrity. Furthermore, we demonstrate that blocking RAD52 oligomerization that disrupts RAD52's DSBR, while retaining its ssDNA binding capacity that is required for RAD52's inhibitory effects on SSBR, sensitizes cells to different DNA-damaging agents. This discovery provides guidance for developing efficient RAD52 inhibitors in cancer therapy.

Keywords: DNA damage; DNA repair; RAD52; XRCC1; single-strand breaks.

Figure 1.. RAD52 Promotes CPT-Induced Vertebrate/Mammalian Cell Killing

(A) WT, Rad52^−/−, Rad54^−/−, or Atm^−/− DT40 cell sensitivities to CPT were measured using the Cell-Titer-Glo Luminescent Cell Viability Assay Kit following our modified protocol and confirmed by clonogenic assay. RAD52 inhibitor (RAD52i) CD1, 2.5 μM, was added to cell culture 1.5 h before adding CPT at 20 nM. Data are mean ± SEM from three independent experiments. **p < 0.01. (B) RAD52 levels were detected by western blotting in WT (mock) and Rad52 knockout MEFs generated using CRISPR-Cas9 (2 from targeting exon 3 and 2 from targeting exon 5). (C) WT, Rad52-deficient (Rad52 d1 targeting exon 3–1; Rad52 d2 targeting exon 3–2), or Ku70-deficient (Ku70 d) MEF sensitivities to CPT using a clonogenic assay. Data are mean ± SEM from three independent experiments. **p < 0.01. (D) RAD52 levels were detected by western blotting in U2OS cells treated with control RNA (Ct RNA) or Rad52 siRNA (for 48 h). (E) U2OS cell sensitivity to CPT after treatment with Ct RNA, Rad52 siRNA, or RAD52i. Data are mean ± SEM from three independent experiments. **p < 0.01.

Figure 2.. RAD52 Suppresses PARP-Mediated Repair of CPT-Induced SSBs

(A) Top: images of RAD52 foci in Rad52-expressing MEFs following 50 nM CPT treatment using a RAD52 antibody (Ochs et al., 2016; scale bar represents 4 μM). Bottom: quantification of RAD52 foci per cell at indicated times after CPT treatment is shown. Data are mean ± SEM (n = 50 cells). ***p < 0.001. (B) Left: images of CPT-induced γ-H2AX foci in WT or Rad52-deficient (Rad52^−/−) MEFs treated with CPT for 30 min. The cells were collected for immunostaining with an anti-γ-H2AX antibody DAPI (scale bar represents 8 μM). Right: percentage of cells with γ-H2AX foci was analyzed. Data are mean ± SEM (n = 50 cells). **p < 0.01. (C) Left: chromatin-bound “C” and whole-cell “W” fractions of different DNA repair proteins from vector or Rad52-expressing Rad52-deficient MEFs were evaluated using western blotting after 50 nM CPT treatment. Right: chromatin-bound protein quantification was based on western blotting image data from 3 independent experiments of 2 clones of Rad52-deficient cells. Data are mean ± SEM from biological triplicates. **p < 0.01. (D) WT, Parp1^−/−, Rad52^−/−, DT40 cell sensitivities to CPT with or without PARPi (olaparib, 1 μM for 1.5 h before CPT treatment). ***p < 0.001. (E) WT, Rad52 d1 (R52–1), or Rad52 d2 (R52–2) MEF sensitivities to CPT. Data are mean ± SEM from three independent experiments. **p < 0.01; ***p < 0.001. (F) U2OS cell sensitivities to CPT after treatment with Ct RNA or Rad52 siRNA. Data are mean ± SEM from three independent experiments. **p < 0.01; ***p < 0.001. (G) Left: WT or Parp1-deficient DT40 cells were treated with CPT for 30 min and then DSBs in these cells were assessed by neutral comet assays (scale bar represents 8 μM). Right: percentage of tail DNA in total DNA (from 100 cells) was analyzed using the CometScore software. Data are mean ± SEM from three independent experiments. ****p < 0.0001.

Figure 3.. RAD52 Suppression of PARP-Mediated SSBR Is through Interference with XRCC1/LIG3a Co-localization

(A) Whole-cell PAR levels were measured in WT or Rad52-deficient MEFs after 50 nM CPT treatment. (B) Western blot analysis of lysates and IP from HEK293T cells transfected with the indicated genes (GFP-XRCC1 or FLAG-LIG3α) for 27 h and then treated with CPT (20 nM) for the indicated times. (C) PLA plots show the percentage of cells with various numbers of XRCC1/LIG3α foci in vector or WT Rad52 expressing in Rad52^−/− MEFs after CPT or MMS treatment. Data from 5 to 6 randomly selected fields (n = 50 cells) in each group were quantified using ImageJ. (D) PAR polymer dot blot analysis. Incremental amounts of each protein (1, 2, 4, and 8 pMol) were spotted onto a nitrocellulose membrane (Bio-Rad), incubated in PBS-T containing 50 nM pADPr polymer (Trevigen), and subsequently subjected to western blotting using a PAR antibody. (E) IP of PAR polymer incubated with RAD52 and glutathione S-transferase (GST)-XRCC1 proteins at different ratios (12:0, 12:6, 12:12, 0:12, 6:12, and 12:12 [pMol:pMol]) using a PAR antibody. (F) WT and Xrcc1-deficient MEFs were treated with DMSO, RAD52i, or PARPi for 1.5 h and then treated with CPT for 24 h. The survival results were obtained using the clonogenic assay. Data are mean ± SEM from three independent experiments. *p < 0.05. **p < 0.01. ***p < 0.001.

Figure 4.. A Strategy to Enhance the Inhibitory Effects of RAD52 on SSBR and Sensitize Cells to DNA-Damaging Agents

(A and B) Description of key RAD52 mutants (R55A and Y104D) tested in this study. Y104 is conserved across variant species. Y104F (abolished Abelson tyrosine kinase (c-ABL) phosphorylation) was used as a control for Y104D. (C) Expression of FLAG-tagged WT or mutant mouse Rad52 in Rad52-deficient MEFs was measured using western blotting. (D) Sensitivity of vector, WT, or mutant Rad52-expressing cells to CPT (40 nM). Data are mean ± SEM from three independent experiments. **p < 0.01. (E) Left: images of XRCC1/LIG3α foci in R55A or Y104 mutant Rasd52-expressing Rad52-deficient MEFs on slides treated with or without CPT (20 nM) for 5 min and then fixed for PLA (scale bar represents 5 μM). Right: the plots show the percentage of XRCC1/LIG3α foci in R55A or Y104 mutant Rad52-expressing Rad52-deficient MEFs with different numbers of foci/cell after CPT treatment from 5 to 6 randomly selected fields (n = 50 cells) in each group, quantified using ImageJ. (F) Sensitivity of MEFs expressing vector, WT, or mutant Rad52 (R55A or Y104D) to different DNA damage inducers: IR; MMS; or etoposide (Top II inhibitor). Data are mean ± SEM from three independent experiments. For IR- and MMS-treated cells, as compared to empty-vector-expressing cells, **p < 0.01 and ***p < 0.001; for etoposide-treated cells, as compared to WT Rad52-expressing cells, **p < 0.01. (G) Comparison of CPT-induced PARylation levels between WT and mutant Rad52-expressing (R55A or Y104D) HEK293T cells under the same conditions as described in Figure S7A.