Human RECQ5 helicase promotes repair of DNA double-strand breaks by synthesis-dependent strand annealing

Abstract

Most mitotic homologous recombination (HR) events proceed via a synthesis-dependent strand annealing mechanism to avoid crossing over, which may give rise to chromosomal rearrangements and loss of heterozygosity. The molecular mechanisms controlling HR sub-pathway choice are poorly understood. Here, we show that human RECQ5, a DNA helicase that can disrupt RAD51 nucleoprotein filaments, promotes formation of non-crossover products during DNA double-strand break-induced HR and counteracts the inhibitory effect of RAD51 on RAD52-mediated DNA annealing in vitro and in vivo. Moreover, we demonstrate that RECQ5 deficiency is associated with an increased occupancy of RAD51 at a double-strand break site, and it also causes an elevation of sister chromatid exchanges on inactivation of the Holliday junction dissolution pathway or on induction of a high load of DNA damage in the cell. Collectively, our findings suggest that RECQ5 acts during the post-synaptic phase of synthesis-dependent strand annealing to prevent formation of aberrant RAD51 filaments on the extended invading strand, thus limiting its channeling into potentially hazardous crossover pathway of HR.

Figure 1.

RECQ5 promotes homologous recombination with non-crossover outcome. (A) Scheme of the DR-GFP reporter cassette. The site-specific DSB in the reporter cassette is generated by I-SceI endonuclease. Only NCO events give rise to a functional GFP allele. (B) Western blot analysis of extracts from HEK293/DR-GFP and U2OS/DR-GFP cells transfected with indicated siRNAs. Blots were probed with indicated antibodies. (C) Efficiency of HR-mediated repair of I-SceI-induced DSB in HEK293/DR-GFP cells treated with indicated siRNAs. Cells were transfected with appropriate siRNA (40 nM) two days before transfection of I-SceI-expressing plasmid. Percentage of GFP-positive cells was measured by flow cytometry 2 days after DSB induction and taken as a measure of DSBR efficiency. Values plotted represent relative repair efficiency calculated as a percentage of repair efficiency measured in cells transfected with control siRNA (siLuc; 100%). All data points represent an average of at least three replicates with error bars indicating standard deviation. (D) Efficiency of HR-mediated repair of I-SceI-induced DSB in U2OS/DR-GFP cells treated with indicated siRNAs as compared with cells transfected with control siRNA. Experiments were performed as in (C) except that the flow cytometry analysis was performed 3 days after I-SceI transfection. HR, homologous recombination; DSB, double-strand break; NCO, non-crossover; and GFP, green fluorescent protein.

Figure 2.

RECQ5 suppresses inhibitory effect of RAD51 on DNA DSBR by SSA. (A) Scheme of the SA-GFP reporter cassette. SSA-mediated repair of I-SceI-generated DSB results in the formation of a functional GFP allele. (B) Western blot analysis of extracts from HEK293/SA-GFP cells transfected with indicated siRNAs. The blots were probed with indicated antibodies. (C) Efficiency of SSA-mediated repair of I-SceI-induced DSB in HEK293/SA-GFP cells transfected with indicated siRNAs. Cells were transfected with appropriate siRNA (40 nM) 2 days before transfection of I-SceI-expressing plasmid. Percentage of GFP-positive cells was determined by flow cytometry 2 days after DSB induction and taken as a measure of repair efficiency. Values plotted represent relative repair efficiency calculated as a percentage of repair efficiency measured in cells transfected with control siRNA (siLuc; 100%). All data points represent an average of at least three replicates with error bars indicating standard deviation. (D) Efficiency of SSA-mediated repair of I-SceI-induced DSB in U2OS/SA-GFP cells transfected with indicated siRNAs. Experiments were performed as in (C) except that the percentage of GFP-positive cells was determined 3 days after DSB induction. (E) Western blot analysis of extracts from U2OS/SA-GFP cells transfected with pcDNA3.1/HisC-based vectors expressing wild-type RECQ5 or its mutants, K58R and F666A, respectively, as fusions with an Omni-tag. The blots were probed with indicated antibodies. (F) Effect of overexpression of the wild-type and mutant forms of RECQ5 on the efficiency of SSA-mediated repair of I-SceI-induced DSB in U2OS/SA-GFP. Cells were transfected with appropriate RECQ5 expression vector in combination with the plasmid expressing I-SceI. Percentage of GFP-positive cells was determined 3 days after plasmid transfection. Values plotted represent relative repair efficiency calculated as a percentage of repair efficiency measured in cells transfected with empty vector. All data points represent an average of at least three replicates with error bars indicating standard deviation. SSA, single-strand annealing; DSB, double-strand break; and GFP, green fluorescent protein.

Figure 3.

RECQ5 helicase counteracts the inhibitory effect of RAD51 on RAD52-mediated ssDNA annealing in vitro. (A) Upper panel: reaction scheme depicting the effect of RAD51 (green circles) on annealing of two complementary oligonucleotides (59-mer and 30-mer represented by red and blue lines, respectively) in presence of RAD52 and RPA. RAD52 is depicted as a heptameric ring structure (red circles). The 30-mer oligonucleotide can accommodate binding of one RPA heterotrimer (light blue ovals). Lower panel: all reactions were carried out at 30°C in buffer R supplemented with ATP-regenerating system. Reactions contained 5′ end radiolabeled 59-mer oligonucleotide (2.5 nM), either free or pre-coated with RAD51^K133R (100 nM), a 30-mer oligonucleotide (2.5 nM) complementary to the 5′-half of the 59-mer, RAD52 (60 nM) and RPA (30 nM). Where indicated, RECQ5 or RECQ5^K58R were present at a concentration of 80 nM. Reaction aliquots at indicated time points were subjected to PAGE followed by phosphorimaging as described in ‘Materials and Methods’. (B) Quantification of data shown in (A). Each data point represents the mean of three independent experiments. Error bars represent standard deviation. (C) Upper panel: reaction scheme depicting the effect of RAD51 on annealing of two complementary oligonucleotides in presence of a homologous duplex, RAD52 and RPA. RAD51 filament formed on the radiolabeled oligonucleotide (red line with asterisk) inhibits RAD52/RPA-mediated annealing and promotes strand exchange with the homologous duplex. Lower panel: reactions were carried out and analyzed as in (B). Homologous 59-mer duplex was present at a concentration of 2.5 nM. Schemes of radiolabeled DNA species are shown on left. Radioactive label at the 5′ end is depicted by asterisks. (D) Quantification of data shown in (C). Each data point represents the mean of three independent experiments. Error bars represent standard deviation.

Figure 4.

Effect of RECQ5 deficiency on RAD51 occupancy at chromatin flanking the I-SceI-induced DSB in U2OS/DR-GFP cells. (A) A schematic diagram of a part of the DR-GFP reporter cassette showing the locations of the regions amplified in ChIP-qPCR assays (P1 and P2). The GFP open reading frame with a single I-SceI recognition site is shown as gray box. The numbers correspond to base pairs. (B) Plot of ChIP-qPCR data. Mock-depleted (siLuc) and RECQ5-depleted (siRECQ5#1) cells were transfected with either I-SceI expression vector (+I-SceI) or empty vector, respectively. Chromatin for ChIP analysis was prepared 2 days after I-SceI transfection. Fold enrichment was calculated as a ratio of RAD51 antibody signal versus control IgG. ChIP, chromatin immunoprecipitation; and qPCR, quantitative real-time PCR.

Figure 5.

RECQ5 and BLM act in different pathways to suppress crossovers in human cells. (A) Frequency of spontaneous SCEs in U2OS cells transfected with indicated siRNAs. (B) Frequency of CPT-induced SCEs in U2OS cells transfected with indicated siRNA. Cells were treated with 40-nM CPT for 20 h where indicated. SCE assay and analysis was conducted as described in ‘Materials and Methods’. Each data point represents number of SCEs per chromosome in a single metaphase spread. The 50 metaphase spreads were analyzed from each condition. (C) Western blot analysis of extracts from U2OS cells transfected with indicated siRNAs. Blots were probed with indicated antibodies. SCE, sister chromatid exchange; and CPT, camptothecin.

Figure 6.

Model for the roles of RECQ5 and BLM in suppression of COs during DSBR by HR. RECQ5 promotes SDSA by disrupting aberrant RAD51-filaments formed after unwinding of the extended D-loop. BLM acts as a part of the BTR (BLM-TOPOIIIα-RMI1/2) complex to mediate dissolution of dHJs. RAD51 filaments formed during the post-synaptic phase of SDSA can promote re-invasion of the homologous duplex followed by formation of a dHJ structutre, increasing the risk of COs. RTEL1 promotes SDSA by catalyzing D-loop disruption. SDSA, synthesis-dependent strand annealing; dHJ, double Holliday junction; and DSBR, double-strand break repair.