The helicase domain of Polθ counteracts RPA to promote alt-NHEJ

Abstract

Mammalian polymerase theta (Polθ) is a multifunctional enzyme that promotes error-prone DNA repair by alternative nonhomologous end joining (alt-NHEJ). Here we present structure-function analyses that reveal that, in addition to the polymerase domain, Polθ-helicase activity plays a central role during double-strand break (DSB) repair. Our results show that the helicase domain promotes chromosomal translocations by alt-NHEJ in mouse embryonic stem cells and also suppresses CRISPR-Cas9- mediated gene targeting by homologous recombination (HR). In vitro assays demonstrate that Polθ-helicase activity facilitates the removal of RPA from resected DSBs to allow their annealing and subsequent joining by alt-NHEJ. Consistent with an antagonistic role for RPA during alt-NHEJ, inhibition of RPA1 enhances end joining and suppresses recombination. Taken together, our results reveal that the balance between HR and alt-NHEJ is controlled by opposing activities of Polθ and RPA, providing further insight into the regulation of repair-pathway choice in mammalian cells.

Fig. 1. Structure-function analysis reveals the function of Polθ–helicase during alt-NHEJ

(a) Top – schematic representation depicting the different domains of Polθ. Three Rad51 binding domains were identified in human POLQ, however, only one motif is conserved in mouse (highlighted in blue). The polymerase domain contains three conserved unique insertion loops (highlighted in grey) that are involved in DNA synthesis. Bottom – CRISPR/Cas9 gene targeting was carried out in mouse ES cells (CCE) to introduce two base substitutions (Asp2494 and Glu2495) that inactivate the polymerase domain (referred to as Polq^ΔPol thereon). Independent gene editing was carried out to inactivate the ATPase activity (Lys120) (Polq^ΔHel) and delete the conserved Rad51 interaction motif (residues 844–890; generating Polq^ΔRad51). (b) Quantitative analysis of colony formation assay in mES cells carrying the indicated Polq mutant alleles and treated with shBRCA1. In each cell line, colonies were normalized to cells treated with shControl. Bars represent mean ± S.D. from four independent experiments with two technical replicates/experiment. Two clonally-derived cells lines were analyzed for each mutation. * p<0.05; two-tailed Student’s t-test. n.s.; not-significant (c) Schematic of chromosomal translocation assay in which DSBs are induced at the Rosa26 and H3f3b loci. Generation of derivative chromosomes Der(6) is detected by nested PCR. (d) Frequency of chromosomal translocation in Polq^+/+, Polq^ΔPol, Polq^ΔHel, and Polq^ΔRad51 cells. Bars represent mean ± S.D. from six independent experiments in the case of Polq^+/+, four experiments for Polq^ΔPol and Polq^ΔHel, and three experiments for Polq^ΔRad51 cells. Two clonally-derived cells lines were analyzed for each mutation. *p<0.05; two-tailed Student’s t-test. n.s; not-significant (e) Table summarizing the analysis of nucleotide composition at the junction of translocation events in (d). Source data for Figures 1b and 1d–e are available with the paper online.

Fig. 2. Polθ inhibition increase the efficiency of HR-mediated CRISPR-Cas9 gene targeting

(a) Scheme depicting gene targeting assay at the Hsp90 locus. The donor plasmid contains a Zsgreen coding sequence and 600 base pairs of homology arms. (b) (FACS) analysis to determine the percentage of ZsGreen positive cells. Three distinct populations were isolated (highlighted as 1, 2 and 3). (c) Genotyping PCR for Hsp90 on DNA corresponding to the three highlighted populations of cells (indicated in (b)). Uncropped gel image is shown in Supplementary Data Set 1. (d) Gene-targeting efficiency at the Hsp90 locus in Polq^−/− and Polq^+/+ MEFs. Bars represent mean ± S.D. from four independent experiments. (e) Gene-targeting efficiency at the Hsp90 locus in mES cells with the indicated genotype. Bars represent mean ± S.D. from four independent experiments and performed with two clonally-isolated Polq^ΔPol, Polq^ΔHel cell lines. **p<0.01; two-tailed Student’s t-test. Source data for Figures 2d–e are available with the paper online.

Fig. 3. Polθ–helicase antagonizes RPA to promote DNA annealing and alt-NHEJ in vitro

(a) Diagram of the annealing assay used to investigate whether Polθ–helicase (Polθ-hel) promotes pairing of complementary ssDNA bound by RPA. ssDNA (38 nucleotides in length) is incubated with RPA, prior to addition of Polθ–helicase. Radio-labeled (asterisk) complementary ssDNA is then added, the reaction terminated, and DNA analyzed in non-denaturing gels following protein degradation. (b) Representative non-denaturing gels displaying ssDNA annealing in the presence and absence of ATP/AMP-PNP and indicated amounts of Polθ–helicase. % dsDNA indicated. (c) Schematic of the assay used to demonstrate that Polθ-hel dissociates RPA-ssDNA complexes independently of ssDNA annealing. Increasing amounts of Polθ-hel are mixed with pre-formed RPA-ssDNA complexes in the presence of ATP or AMP-PNP. ssDNA is radio-labeled (asterisk). (d) Non-denaturing gel showing the effects of Polθ-hel on pre-assembled RPA-ssDNA complexes in the presence of ATP and AMPPNP. Polθ-hel takes the place of RPA on ssDNA exclusively in the presence of ATP. The high molecular weight of the Polθ-hel-ssDNA complex (sixth lane) is likely due to Polθ-helicase tetramer formation. (e) Schematic of Polθ–polymerase (Polθ-pol) mediated alt-NHEJ assay. Partially resected DNA model substrates (12 nucleotides in length) containing 3’ terminal microhomology (6 bases) were used as previously described. (f) Non-denaturing gels showing Polθ–polymerase mediated alt-NHEJ in the presence of the indicated proteins, dNTP, and ATP. % end-joining indicated. Uncropped gel images are shown in Supplementary Data Set 1.

Fig. 4. Mammalian RPA inhibits alt-NHEJ

(a) Quantification of telomere sister-chromatid exchanges (T-SCEs), reflective of telomere recombination events in cells with the indicated treatments. Top – Representative examples of chromosome ends undergoing T-SCE (white arrows) in metaphase spreads from TRF1^F/FTRF2^F/FLig4^−/−Cre-ER^T2 cells treated with 4-OHT. Telomeres in red and green (PNA probes) and chromosomes in blue (DAPI). Bars represent mean ± S.D. from three independent experiments. *P < 0.05, **P < 0.01; two-tailed Student’s t-test. (b) Quantification of telomere fusions by alt-NHEJ in cells with the indicated treatment. Top – Examples of chromosomes fusion events denoted by white arrows. Bars represent mean ± S.D. from three independent experiments. *P < 0.05 and **P < 0.01; two-tailed Student’s t-test. (c) Immunofluorescence for Myc-RPA1 and TRF1-FokI-mCherry in U2OS cells expressing the indicated alleles. TRF1-Fok1-mCherry expression was induced upon treatment with doxycycline, shield-1, and 4-OHT. Scale bars, 10 μm. (d) Graph representing the quantification of Myc-RPA1 accumulation in cells expressing TRF1-FokI (as in c). Bars represent mean ± S.D. from three independent experiments. *P < 0.05 and **P < 0.01; two-tailed Student’s t-test. Source data for Figures 4a–d are available with the paper online.

Fig. 5. The interplay between Polθ and RPA determine the outcome of DSB repair

Schematic depicting our model for the antagonistic relationship between Polθ and RPA during DSB repair. Binding of the RPA complex to resected DSB ends block alt-NHEJ and promote HR. The helicase activity of Polθ perturbs the binding of RPA to ssDNA and stimulate the annealing of resected DSBs. Annealed intermediates are then stabilized upon fill-in synthesis by Polθ-polymerase and ultimately joined by Lig3/Lig1.