An R-loop-initiated CSB-RAD52-POLD3 pathway suppresses ROS-induced telomeric DNA breaks

Abstract

Reactive oxygen species (ROS) inflict multiple types of lesions in DNA, threatening genomic integrity. How cells respond to ROS-induced DNA damage at telomeres is still largely unknown. Here, we show that ROS-induced DNA damage at telomeres triggers R-loop accumulation in a TERRA- and TRF2-dependent manner. Both ROS-induced single- and double-strand DNA breaks (SSBs and DSBs) contribute to R-loop induction, promoting the localization of CSB and RAD52 to damaged telomeres. RAD52 is recruited to telomeric R-loops through its interactions with both CSB and DNA:RNA hybrids. Both CSB and RAD52 are required for the efficient repair of ROS-induced telomeric DSBs. The function of RAD52 in telomere repair is dependent on its ability to bind and recruit POLD3, a protein critical for break-induced DNA replication (BIR). Thus, ROS-induced telomeric R-loops promote repair of telomeric DSBs through CSB-RAD52-POLD3-mediated BIR, a previously unknown pathway protecting telomeres from ROS. ROS-induced telomeric SSBs may not only give rise to DSBs indirectly, but also promote DSB repair by inducing R-loops, revealing an unexpected interplay between distinct ROS-induced DNA lesions.

Figure 1.

ROS induced SSB trigger R-loop formation at telomere dependently on TERRA and TRF2. (A) Schematic model of KR-TRF1. Co-localization of GFP-XRCC1 and γ-H2AX with KR-TRF1 in U2OS cells after exposing to light for 30 min are shown below the schematic model. The scale bars represent 5 μm. (B) The staining of S9.6 at KR-TRF1 and RFP-TRF1 after exposing to light for 30 min in U2OS cells and SAOS2 cells. (C) The staining of S9.6 at KR-TRF1 in U2OS cells after treating with DRB (20 μM, 24 h) and α-amanitin (100 μg/ml, 2 h), overexpressed with HA-RNaseH WT, D210N mutant, or TRF2 knockdown by siRNA. (D) Schematic model of FOK1-TRF1. The co-localization of GFP-XRCC1 and γ-H2AX with FOK1-TRF1 in U2OS cells after transfection are showed below the schematic model. (E) The staining of S9.6 at KR-TRF1 and FOK1-TRF1 in U2OS cells after exposing to light for 30 min and recovery for a half hour. (F) RNA FISH confirmed TERRA depletion in U2OS cells after 12 h of LNA treatment. Quantification of the number of TERRA signals in each cell. Error bar represents over 50 cells. (G) The staining of S9.6 at KR-TRF1 in U2OS cells after TERRA knockdown by LNA. (H) Schematic of R-loop formation after inducing ROS damage. For A–G, mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.

Figure 2.

CSB is recruited to telomeric R-loops and contributes to cell survival and telomeric integrity. (A) Colony formation assays for U2OS and U2OS CSB KO cells with transiently expressing KR-TRF1 or RFP-TRF1. n = 3, error bars represent SD. (B) Telomere aberrations were found in U2OS and U2OS CSB KO cells with or without transiently expressing KR-TRF1. (C) Recruitment of CSB in KR-TRF1 or RFP-TRF1 expressing U2OS cells with or without light activation. Quantification of co-location frequency of myc-CSB and KR-TRF1/RFP-TRF1. (D) CSB foci co-localization with KR-TRF1 in U2OS cells treated with KU55933 (ATM inhibitor, 10 uM, 24 h), Nu7026 (DNA-PK inhibitor, 20 uM, 24 h), Olaparib (10 uM, 24 h), XAV-933(Tankyrase inhibitor,20 uM, 24 h), ATR inhibitor (10 uM, 24 h), and CDK2 inhibitor (5 uM, 24 h). (E) The staining of CSB at KR-TRF1 in U2OS cells after treating with overexpressed with HA-RNaseH WT, D210N mutant, DRB (20 μM, 24 h) and α-amanitin (100 μg/ml, 2 h), or TERRA knockdown by LNA. (F) CSB foci co-localization with KR-TRF1 in U2OS cells treated with TRF2 knockdown. (G) The staining of S9.6 at KR-TRF1 in U2OS cells after treating with CSB knockdown by siRNA. For (D)–(H), mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test, **P < 0.01, ***P < 0.001.

Figure 3.

R464 of CSB contributes to DNA: RNA hybrid binding. (A) Schematic diagram of CSB fragments. The left panel is the quantification of co-location frequency of CSB fragments to KR-TRF1in U2OS cells. (B) Cell survival of CSB WT and KO cells expressing indicated fragments of CSB with 1 h light activation of KR-TRF1 in U2OS cells. n = 3, unpaired t-test, error bars represent SD, *P < 0.05 (C) GFP-AD foci co-localization with KR-TRF1 in U2OS cells treated with α-amanitin (100 μg/ml, 2 h), overexpressed with HA-RNaseH WT, D210N mutant or TERRA depleted with LNA. (D) Schematic diagram of CSB-ADN and ADC. (E) AD WT, ADN and ADC co-localization with KR-TRF1 after light activation. (F) The affinity of GFP-AD, GFP-ADN, and GFP-ADC to DNA: RNA hybrid by biotin-labeled hybrid pulldown using cell lysate from U2OS CSB KO cells. (G) Schematic diagram of CSB-AD 22A, R464A, RRAA, 3RA, K470A and R472A. (H) The recruitment of CSB-AD mutants to KR-TRF1 after light activation. (I) Colony formation assays for U2OS CSB KO cells expressing GFP-AD WT, AD R464A, AD3RA, and AD RRAA after inducing ROS damage by KR-TRF1 with 1.5 light activation. n = 3, unpaired t-test, error bars represent SD, **P < 0.01 (J) The affinity of GFP-ADC and GFP-ADC R464A to DNA: RNA hybrid by biotin-labeled hybrid pulldown using cell lysate from U2OS CSB KO cells. (K) DNA:RNA hybrids EMSA assay of purified CSB-ADC WT and R464A protein. For (C), (E) and (H), Mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.

Figure 4.

CSB-RAD52 cascade contributes to the repair of ROS induced telomeric DSB. (A) Colony formation assays for U2OS CSB KO cells with RAD52 knockdown by siRNA transiently expressing KR-TRF1. (B) The intensity of GFP-XRCC1 foci after CSB or RAD52 knockdown by siRNA at 4h and 24h after light activation. (C) The clearance of γ-H2AX after RAD52 knockdown. (D) The recruitment of RAD52 to RFP/KR/FOK1-TRF1after light activation or transfection. (E) The recruitment of RAD52 to KR-TRF1 after treating with DRB (20 μM, 24 h) and α-amanitin (100 μg/ml, 2 h), TERRA knockdown by LNA, or overexpressed with HA-RNaseH WT, D210N mutant. (F) Telomere aberrations were found in U2OS and U2OS RAD52 KO cells with or without transiently expressing KR-TRF1. For (B), (C), (D) and (E), Mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.

Figure 5.

RAD52 is recruited to telomeric R-loop via CSB and K144-mediated hybrid binding. (A) The recruitment of RAD52 to damaged telomere in U2OS CSB KO cells after light exposure for 30 min and 60 min. (B) Schematic model of RAD52 structure and mutants. Right panel is Monomer structure of RAD52 and mutant sites. Y65A (orange), K141A (yellow), K144A (blue), R153A (purple), R156A (pink). (C) The recruitment of RAD52 mutants to KR-TRF1 after ROS damage. Quantification of the percentage of RAD52 co-localization with KR-TRF1. (D) DNA: RNA hybrids EMSA assay of purified RAD52 WT and RAD52 K144A protein. For (A) and (C), mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.

Figure 6.

RAD52-POLD3 mediated BIR contributes to repair of ROS induced DSBs. (A) EdU incorporation after KR-TRF1-induced damage in U2OS. (B) Schematic of the RAD52 dependent repair pathway. (C) The recruitment of RAD52 and RAD51 to KR-TRF1 after light activation with 30 min recovery. (D) The recruitment of GFP-POLD3 to KR-TRF1 after light activation. (E, F) The recruitment of GFP-POLD3 (E) and POLD3 (F) to KR-TRF1 after RAD52 KO or knockdown by siRNA. For (A)–(F), mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.

Figure 7.

Tyrosine 65 (Y65) of RAD52 is required for the recruitment and the interaction with POLD3. (A) The recruitment of POLD3 to KR-TRF1 in U2OS RAD52 KO cells after transient expressing RAD52 mutants. (B) Co-immunoprecipitation of GFP-RAD52 WT and mutants with POLD3 after KR-TRF1-induced damage for 1 h (left). Co-immunoprecipitation of GFP-RAD52 WT with POLD3 after treating cell lysates with ethidium bromide (0.1 mg/ml) or RNase H (20 U/ml) for 30 min (right). (C) The clearance of γ-H2AX after POLD3 knockdown by siRNA. (D) Colony formation assays for U2OS and U2OS POLD3 knockdown by siRNA and transiently expressing KR-TRF1 or RFP-TRF1 with indicated time to light exposure. n = 3, error bars represent SD. (E) Schematic model of R-loop–CSB–RAD52 mediates BIR contribute to the repair of telomeric DSB induced by ROS. The current model is composed of the following sequential steps: ROS-induced R-loop formation is dependent on TERRA and TRF2; CSB–ADC–R464 recognizes R-loop; K141/144 mediates R-loop binding of RAD52; RAD52 R65 mediated POLD3 interaction promotes BIR. For (a) and (c), Mean values with SD from 50 cells in three independent experiments are given. P-value is calculated by unpaired t-test. ***P < 0.001.