Coordinated nuclease activities counteract Ku at single-ended DNA double-strand breaks

Abstract

Repair of single-ended DNA double-strand breaks (seDSBs) by homologous recombination (HR) requires the generation of a 3' single-strand DNA overhang by exonuclease activities in a process called DNA resection. However, it is anticipated that the highly abundant DNA end-binding protein Ku sequesters seDSBs and shields them from exonuclease activities. Despite pioneering works in yeast, it is unclear how mammalian cells counteract Ku at seDSBs to allow HR to proceed. Here we show that in human cells, ATM-dependent phosphorylation of CtIP and the epistatic and coordinated actions of MRE11 and CtIP nuclease activities are required to limit the stable loading of Ku on seDSBs. We also provide evidence for a hitherto unsuspected additional mechanism that contributes to prevent Ku accumulation at seDSBs, acting downstream of MRE11 endonuclease activity and in parallel with MRE11 exonuclease activity. Finally, we show that Ku persistence at seDSBs compromises Rad51 focus assembly but not DNA resection.

Figure 1. RPA32 S4/S8 phosphorylation is a mark of transient Ku association with seDSBs.

(a) Immunoblotting of extracts from U2OS cells pre-treated with the replication inhibitor aphidicolin (APH) or with DNA-PK inhibitor (DNA-PKi), respectively, before being CPT treated. (b) Representative micrographs of P-RPAS4/S8 and RPA70 foci detected by immunofluorescence in U2OS cells pre-treated with dimethylsulfoxide (DMSO) or DNA-PKi before being treated with CPT. After treatment, cells were pre-extracted with CSK+R (see Methods section) before fixation and immunodetection. (c) Quantification of RPA70 and P-RPAS4/S8 foci number per cell. Cells were treated and processed as in b, and foci quantified as described in Methods. (d) Graph representing the distribution of individual P-RPAS4/S8 focus intensity. Cells were treated and processed as in b. 6,684 and 8,462 foci were analysed for the CPT and CPT+DNA-PKi conditions, respectively. (e) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant wild-type (WT) or Mut6E GFP-FLAG-Ku70, and transfected with the indicated siRNA. (f) Representative micrographs of P-RPAS4/S8 and RPA70 foci detected by immunofluorescence in U2OS T-REx transfected as in e and treated with CPT. After treatment, cells were pre-extracted with CSK+R before fixation and immunodetection. (g) Quantification of RPA70 and P-RPAS4/S8 foci number per cell. Cells were treated and processed as in f and foci quantified as described in Methods. (h) Graph representing the distribution of individual P-RPAS4/S8 foci intensity in U2OS T-REx cells treated and processed as in f. 9,970 and 19,995 foci were analysed for the Ku70-WT and Ku70-Mut6E conditions, respectively. (i) Model depicting the proposed structure transiently forming at seDSB induced by CPT. (j) U2OS cells were treated with ionizing radiation (IR) or CPT before being pre-extracted with CSK+R and processed for immunodetection of Ku80 and the replication marker proliferating cell nuclear antigen (PCNA). White scale bars represent 10 μm; insets represent × 3 magnification. Error bars are s.d. Significant differences between specified pairs of conditions, as judged by t-test, are highlighted by stars (*P<0.05). NS, non-significant difference.

Figure 2. CtIP and ATM-dependent CtIP phosphorylations are required to antagonize Ku at seDSBs.

(a) Immunoblotting of extracts from U2OS cells transfected with the indicated siRNA. (b) Representative micrographs of Ku foci visualized by immunofluorescence in U2OS cells transfected as in a and treated with CPT (white scale bar represents 10 μm; insets represent × 3 magnification). (c) Quantification of Ku foci in PCNA-positive cells. U2OS cells were transfected with the indicated siRNA and pre-treated with dimethylsulfoxide (DMSO) or ATM inhibitor (ATMi) for 1 h, before being CPT treated and processed for Ku foci detection by immunofluorescence. (d) Graph representing cell viability as determined by clonogenical assays of U2OS cells transfected with the indicated siRNA and treated for 18 h with CPT in presence of DMSO or DNA-PK inhibitor. T-tests were used to determine significant differences to siCtrl condition. (e) Schematic of CtIP domains indicating the position of the mutated ATM-dependent phosphorylation sites. (f) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant wild-type (WT) or S→A phosphomutant HA-CtIP expressing plasmids and transfected with the indicated siRNA. (g) Quantification of Ku foci in replicating U2OS T-REx cells complemented with HA-CtIP (as in f) and CPT treated. Error bars on figures correspond to s.d.’s. Significant differences between specified pairs of conditions, as judged by t-test, are highlighted by stars (*P<0.05; **P<0.01; ***P<0.0005). NS, non-significant difference.

Figure 3. MRE11 nuclease activities control Ku accumulation at seDSB.

(a) Schematic of MRE11 domains with the position of MRE11 H129 and H63 residues. (b) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant wild-type (WT) or H129N HA-MRE11-expressing plasmids and transfected with the indicated siRNA. (c) U2OS or U2OS T-REx cells complemented with MRE11 as in b were transfected with the indicated siRNA, treated with CPT and processed for analysis of DNA resection as monitored by measuring RPA32 association with chromatin using a flow cytometry assay. (d,e) Representative micrographs (d) and quantification (e) of Ku foci in replicating U2OS T-REx cells complemented with MRE11 as in b and treated or not with CPT before being processed for immunofluorescence. Replicating cells were identified using proliferating cell nuclear antigen (PCNA) staining. (f) Quantification of Ku foci in replicating cells. U2OS T-REx cells stably transfected with control or siRNA-resistant WT or H129N HA-MRE11-expressing plasmids were transfected with the indicated siRNA and Ku foci were quantified in PCNA-positive cells. (g) Alignment of scSae2 H59 with human MRE11 H63 revealing that amino acid H59 is conserved in humans and corresponds to H63. (h) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant WT, H63S or H63N HA-MRE11-expressing plasmids, and transfected with the indicated siRNA. (i) 5′ radio-labelled double-stranded DNA substrate used for in vitro nuclease assays. (j) Analysis by denaturating PAGE of the exonuclease activity of WT and mutants MRE11 on the probe depicted in i. (k) Quantification of nuclease activity in each condition relative to the MRE11 WT condition. (l) Immunoblotting of bead-associated complexes used for in vitro nuclease assays. (m) Quantification of Ku foci in replicating U2OS T-REx cells complemented by WT or mutant HA-MRE11 as in f and treated with CPT. Error bars are s.d. Significant differences between specified pairs of conditions, as judged by t-test, are highlighted by stars (**P<0.01; ***P<0.0005; ****P<0.0001). NS, non-significant difference.

Figure 4. CtIP 5′-flap endonuclease activity shows epistasis with MRE11 3′–5′ exonuclease activity in counteracting Ku binding to seDSBs.

(a) Schematic of CtIP domains with the position of N289 and H290 residues. (b) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant wild-type (WT) or NAHA HA-CtIP-expressing plasmids, and transfected with the indicated siRNA. (c) Quantification of Ku focus number in replicating cells transfected as in b and treated with CPT. (d) Immunoblotting of extracts from U2OS T-REx cells stably transfected with control or siRNA-resistant WT HA-CtIP and HA-MRE11 or NAHA HA-CtIP and H63N HA-MRE11, and transfected with the indicated siRNAs. (e) Quantification of Ku foci in U2OS T-REx stably transfected with control or the specified siRNA-resistant constructs and transfected with the indicated siRNAs. Error bars are s.d. Significant differences between specified pairs of conditions, as judged by t-test, are highlighted by stars (*P<0.05; **P<0.01; ***P<0.0005; ****P<0.0001). NS, non-significant difference.

Figure 5. Releasing Ku from seDSBs is critical for RAD51 loading but not for resection.

(a) Analysis by flow cytometry of DNA damage induction and DNA resection. U2OS T-REx cells stably transfected with siRNA-resistant wild-type (WT) or H63N HA-MRE11-expressing plasmids were transfected with siRNA, treated with CPT and processed for analysis of the γH2AX DNA damage marker and RPA32 association to chromatin. (b) Representative micrographs of Ku80 foci showing co-localization with RPA70. U2OS T-REx cells stably transfected with siRNA-resistant H63N HA-MRE11-expressing plasmid were transfected with the indicated siRNA, treated with CPT and processed for immunofluorescence. The green scale bar represents 100 nm. (c,d) Representative micrographs of RAD51 foci (c) and related quantification (d) in U2OS T-REx stably transfected with control or siRNA-resistant WT or H63N HA-MRE11-expressing plasmids, transfected with the indicated siRNA, treated with CPT and post-incubated for 1 h in drug-free medium. The white scale bar represents 10 μm. Error bars are s.d. Significant differences between specified pairs of conditions, as judged by t-test, are highlighted by stars (*P<0.05; **P<0.01). NS, non-significant difference. (e) Proposed model integrating the present findings. Ku and MRN simultaneously recognize seDSBs, with Ku loading at the DNA end and MRN associating to the side of the DSB. Ku recruits DNA-PKcs to form the active DNA-PK complex while MRN recruits ATM that mediates CtIP phosphorylation on multiple residues, including S664, S679 and S745. This activates MRE11 endonuclease activity that creates a nick on the side of the DSB. This nick is extended in the 5′–3′ direction by EXO1 and DNA2 exonuclease activities, which create ssDNA that recruits RPA that is phosphorylated locally on RPA32-S4/S8 by activated DNA-PK. Then, for a subset of seDSBs, MRE11 exonuclease activity processes the DNA flanking Ku and contributes to the generation of a 5′-flap cleaved by CtIP endonuclease activity to release Ku. The generation of a free ssDNA end by this activity is critical for replacement of RPA by RAD51 and HR. Another, as-yet uncharacterized mechanism operates in parallel to release Ku from >50% of seDSBs.