BRCA2 promotes DNA‐RNA hybrid resolution by DDX5 helicase at DNA breaks to facilitate their repair‡
2021; Springer Nature; Volume: 40; Issue: 7 Linguagem: Inglês
10.15252/embj.2020106018
ISSN1460-2075
AutoresGaetana Sessa, Belén Gómez‐González, Sónia Silva, Carmen Pérez‐Calero, Romane Beaurepere, Sónia Barroso, Sylvain Martineau, Charlotte Martin, Åsa Ehlén, Juan S. Martinez, Bérangère Lombard, Damarys Loew, Stéphan Vagner, Andrés Aguilera, Aura Carreira,
Tópico(s)Chromosomal and Genetic Variations
ResumoArticle26 February 2021free access Source DataTransparent process BRCA2 promotes DNA-RNA hybrid resolution by DDX5 helicase at DNA breaks to facilitate their repair‡ Gaetana Sessa Gaetana Sessa Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, FranceThese authors contributed equally to this work Search for more papers by this author Belén Gómez-González Belén Gómez-González orcid.org/0000-0003-1655-8407 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, SpainThese authors contributed equally to this work Search for more papers by this author Sonia Silva Sonia Silva Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Carmen Pérez-Calero Carmen Pérez-Calero Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Romane Beaurepere Romane Beaurepere Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Sonia Barroso Sonia Barroso orcid.org/0000-0002-0062-2016 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Sylvain Martineau Sylvain Martineau Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Charlotte Martin Charlotte Martin Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Åsa Ehlén Åsa Ehlén Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Juan S Martínez Juan S Martínez Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Bérangère Lombard Bérangère Lombard Laboratoire de Spectrométrie de Masse Protéomique, Institut Curie, PSL Research University, Paris, France Search for more papers by this author Damarys Loew Damarys Loew orcid.org/0000-0002-9111-8842 Laboratoire de Spectrométrie de Masse Protéomique, Institut Curie, PSL Research University, Paris, France Search for more papers by this author Stephan Vagner Stephan Vagner Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Andrés Aguilera Corresponding Author Andrés Aguilera [email protected] orcid.org/0000-0003-4782-1714 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Aura Carreira Corresponding Author Aura Carreira [email protected] orcid.org/0000-0001-5489-4343 Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Gaetana Sessa Gaetana Sessa Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, FranceThese authors contributed equally to this work Search for more papers by this author Belén Gómez-González Belén Gómez-González orcid.org/0000-0003-1655-8407 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, SpainThese authors contributed equally to this work Search for more papers by this author Sonia Silva Sonia Silva Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Carmen Pérez-Calero Carmen Pérez-Calero Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Romane Beaurepere Romane Beaurepere Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Sonia Barroso Sonia Barroso orcid.org/0000-0002-0062-2016 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Sylvain Martineau Sylvain Martineau Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Charlotte Martin Charlotte Martin Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Åsa Ehlén Åsa Ehlén Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Juan S Martínez Juan S Martínez Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Bérangère Lombard Bérangère Lombard Laboratoire de Spectrométrie de Masse Protéomique, Institut Curie, PSL Research University, Paris, France Search for more papers by this author Damarys Loew Damarys Loew orcid.org/0000-0002-9111-8842 Laboratoire de Spectrométrie de Masse Protéomique, Institut Curie, PSL Research University, Paris, France Search for more papers by this author Stephan Vagner Stephan Vagner Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Andrés Aguilera Corresponding Author Andrés Aguilera [email protected] orcid.org/0000-0003-4782-1714 Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain Search for more papers by this author Aura Carreira Corresponding Author Aura Carreira [email protected] orcid.org/0000-0001-5489-4343 Institut Curie, Université PSL, CNRS UMR3348, Orsay, France Université Paris-Saclay, CNRS UMR3348, Orsay, France Search for more papers by this author Author Information Gaetana Sessa1,2, Belén Gómez-González3,4, Sonia Silva3,4, Carmen Pérez-Calero3,4, Romane Beaurepere1,2, Sonia Barroso3,4, Sylvain Martineau1,2, Charlotte Martin1,2, Åsa Ehlén1,2, Juan S Martínez1,2, Bérangère Lombard5, Damarys Loew5, Stephan Vagner1,2, Andrés Aguilera *,3,4 and Aura Carreira *,1,2 1Institut Curie, Université PSL, CNRS UMR3348, Orsay, France 2Université Paris-Saclay, CNRS UMR3348, Orsay, France 3Andalusian Molecular Biology and Regenerative Medicine Centre-CABIMER, University of Seville-CSIC, Seville, Spain 4Departamento de Genética, Facultad de Biología, University of Seville, Seville, Spain 5Laboratoire de Spectrométrie de Masse Protéomique, Institut Curie, PSL Research University, Paris, France *Corresponding author. Tel: +34 954 468 372; E-mail: [email protected] *Corresponding author. Tel: +33 169863082; E-mail: [email protected] The EMBO Journal (2021)40:e106018https://doi.org/10.15252/embj.2020106018 ‡Correction added on 1 April 2021, after first online publication: The title was changed from ‘BRCA2 promotes R-loop resolution by DDX5 helicase at DNA breaks to facilitate their repair by homologous recombination’. PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract The BRCA2 tumor suppressor is a DNA double-strand break (DSB) repair factor essential for maintaining genome integrity. BRCA2-deficient cells spontaneously accumulate DNA-RNA hybrids, a known source of genome instability. However, the specific role of BRCA2 on these structures remains poorly understood. Here we identified the DEAD-box RNA helicase DDX5 as a BRCA2-interacting protein. DDX5 associates with DNA-RNA hybrids that form in the vicinity of DSBs, and this association is enhanced by BRCA2. Notably, BRCA2 stimulates the DNA-RNA hybrid-unwinding activity of DDX5 helicase. An impaired BRCA2-DDX5 interaction, as observed in cells expressing the breast cancer variant BRCA2-T207A, reduces the association of DDX5 with DNA-RNA hybrids, decreases the number of RPA foci, and alters the kinetics of appearance of RAD51 foci upon irradiation. Our findings are consistent with DNA-RNA hybrids constituting an impediment for the repair of DSBs by homologous recombination and reveal BRCA2 and DDX5 as active players in their removal. SYNOPSIS Cells lacking the BRCA2 tumor suppressor spontaneously accumulate DNA-RNA hybrids, but the specific BRCA2 role in their suppression has remained unclear. Here, RNA helicase DDX5 is found to interact with BRCA2 and participate in R-loop ressolution to allow homologous recombination (HR) repair. BRCA2 and DDX5 cooperate to unwind DNA double-strand break-associated R-loops at transcribed regions. BRCA2 supports DDX5 association with DNA double-strand breaks. Interaction between BRCA2 and DDX5 facilitates timely repair of DNA damage by HR. The breast-cancer-associated T207A mutation alters BRCA2 interaction with DDX5 and slows the kinetics of HR repair. Introduction BRCA2 tumor suppressor protein is involved in genome maintenance mechanisms including DNA repair by homologous recombination (HR) (Moynahan et al, 2001; Jensen et al, 2010), protection of stalled replication forks (RFs) (Schlacher et al, 2011), and faithful segregation of chromosomes (Daniels et al, 2004; Ehlén et al, 2020). Recent reports have revealed that BRCA2-deficient cells accumulate DNA-RNA hybrids or R-loops (Bhatia et al, 2014; Tan et al, 2017). Unscheduled hybrids may form during transcription representing an important source of genome instability by either the subsequent action of nucleases acting on the displaced ssDNA strand or, mainly, by blocking RF progression leading to transcription–replication conflicts (García-Muse & Aguilera, 2019). On the other hand, DNA-RNA hybrid accumulation is enhanced by both single-strand DNA breaks (SSBs) and double-strand DNA breaks (DSBs) (Aguilera & Gómez-González, 2017) and recent reports indicate that DNA-RNA hybrids accumulate in the proximity of DSBs (Li et al, 2016; Ohle et al, 2016; Cohen et al, 2018; Lu et al, 2018; Yasuhara et al, 2018). Given the ability of R-loops to compromise genome integrity, cells have developed different strategies to prevent the detrimental accumulation of these structures. Among these are particularly relevant nucleases such as RNases H1 and H2 and a number of recently characterized RNA helicases (García-Muse & Aguilera, 2019). The latter include, in addition to Senataxin (Skourti-Stathaki et al, 2011), AQR (Sollier et al, 2014), members of the DEAD-box family of RNA helicases such as DDX1 (Li et al, 2008), DDX5 (Mersaoui et al, 2019), DDX21 (Song et al, 2017), DDX19 (Hodroj et al, 2017), UAP56/DDX39B (Pérez-Calero et al, 2020) or DHX9 (Chakraborty & Grosse, 2011). Arguably, their mechanism of action is not completely elucidated and their functional specificity might be determined by the nucleic acid structural context and the co-factors they interact with. Several DNA repair proteins have been proposed to act in concert with helicases and nucleases to direct DNA-RNA hybrid resolution. For example, BRCA2 and other related proteins such as BRCA1 or the Fanconi anemia (FA) canonical factors FANCD2, FANCJ, and FANCM reduce DNA-RNA hybrids at transcription–replication conflicts (García-Rubio et al, 2015; Schwab et al, 2015; Madireddy et al, 2016). Both BRCA1 and BRCA2 have also been reported to regulate RNA pol II transcription elongation (Shivji et al, 2018) or termination (Hatchi et al, 2015), which when defective result in R-loop-mediated DNA breaks. Interestingly, a connection between FA factors and splicing has been recently revealed (Moriel-Carretero et al, 2017). In this study, we find that BRCA2 interacts with DDX5, a known DEAD-box RNA helicase (Hirling et al, 1989; Xing et al, 2017), and their association is particularly enriched in DNA damage conditions. BRCA2 stimulates the DNA-RNA hybrid-unwinding activity of DDX5 in vitro and promotes its association with DNA-RNA hybrids located in the vicinity of DSBs. Both DDX5-depleted cells and cells bearing a breast cancer missense variant (T207A), which reduces BRCA2 interaction with DDX5, exhibit increased DNA damage-associated DNA-RNA hybrids and delays kinetics of HR-mediated DSB repair. Our results indicate that DNA-RNA hybrids are an impediment for the repair of DSBs and reveal that BRCA2 and DDX5 are active players in their removal. Results BRCA2 physically interacts with DDX5 The N-terminal region of BRCA2 is highly disordered (Julien et al, 2020). To get insight on its function, we used a mass spectrometry screen to identify the nuclear interacting partners of this region using HEK293T cells overexpressing a fusion protein comprising the first 1,000 aa of BRCA2 fused to a N-terminal 2xMBP tag followed by two nuclear localization signals (NLS) (hereafter BRCA2NT) or the 2xMBP-NLS alone (Fig EV1A). Among the potential protein partners, we found several RNA helicases including the DEAD-box RNA helicase DDX5 (Xing et al, 2017), recently reported to suppress R-loops (Mersaoui et al, 2019; Fig EV1B, Table EV1). In order to validate the interaction between BRCA2 and DDX5, we performed a pull-down assay and Western blots from HEK293T whole cell extracts that showed an interaction between overexpressed BRCA2NT and endogenous DDX5 (Fig 1A). Exposure of the cells to DNA damage induced by γ-irradiation (6 Gy) enhanced the interaction although the increase was moderate (Fig 1A). We then confirmed the interaction with the endogenous proteins BRCA2 and DDX5 by co-immunoprecipitation (co-IP) in both unchallenged or 4 h post-irradiation (γ-irradiation, 6 Gy) (Fig 1B). The association of the endogenous BRCA2 and DDX5 was not mediated by DNA or RNA as was not affected by benzonase (Fig 1B). While we could validate DDX5 interaction, we failed to confirm the interaction with other RNA-binding proteins that were enriched by emPAI quantification (Ishihama et al, 2005; Fig EV1B) such as RBMX and DDX21 (Fig EV1C); thus, we focused on BRCA2-DDX5 interaction. Consistently, using in situ proximity ligation assay (PLA) and specific antibodies and extraction conditions to reveal co-localization specific to chromatin, we found that BRCA2 and DDX5 colocalized in U2OS cells and that their proximity was enhanced in cells exposed to γ-irradiation (Fig 1C). Click here to expand this figure. Figure EV1. Related to Fig 1. DEAD-box proteins identified in the proteomics mass spectrometry screen Amylose pull-down from HEK293T nuclear cell extracts expressing 2xMBP-BRCA2NT (BRCA2NT) and 2XMBP, detected by immunoblot, showing the samples for mass spectrometry experiment. The loading for input and unbound fractions is 1%, for the elution fraction is 8%, and for boiled bead fraction is 35%. DEAD-box helicases enriched in the BRCA2NT interactome. Label-free protein quantification. (Left) BRCA2 (in italic) and DDX Protein ID present in the proteomics mass spectrometry screen. (Center) Heat-map showing fold enrichment of each protein in BRCA2NT/2xMBP. Infinite-fold indicates proteins that are only present in BRCA2NT sample and not in pull-down performed with the 2xMBP. (Bottom) Heat-map log2 color scale. (Right) Bar graph showing protein abundance in molar fraction percentage (mol %) in each pull-down (yellow in 2xMBP, blue in BRCA2NT) based on label-free emPAI quantification (see Materials and Methods section). Immunoprecipitation (IP) of endogenous BRCA2 from benzonase-treated HEK293T whole cell lysates treated or not with IR (6 Gy), as indicated. Normal mouse IgG was used as negative control. Immunoblot of DDX5, DDX21 and RBMX and BRCA2. Stain-Free images of the gels before transfer were used as loading control (cropped images are shown). Download figure Download PowerPoint Figure 1. BRCA2 physically interacts with DDX5 Amylose pull-down from benzonase-treated HEK293T cell lysates expressing 2xMBP-BRCA2NT in untreated or irradiated cells (6Gy; +IR). DDX5 and BRCA2NT (MBP) detected by immunoblot. Stain-Free images of the gels before transfer were used as loading control (cropped image is shown). Immunoprecipitation (IP) of endogenous BRCA2 from benzonase-treated HEK293T cell lysates left untreated or treated with IR (6 Gy) and harvested 4 h post-IR, as indicated. Mouse IgG was used as negative control. Immunoblot of DDX5 and BRCA2. Stain-Free image of the gels before transfer was used as loading control (cropped image is shown). Asterisk (*) indicates a non-specific band detected by anti-DDX5 antibody. Left: Representative images of in situ proximity ligation assay (PLA) between BRCA2 and DDX5 antibodies in U2OS cells either left untreated (−) or irradiated (4 h post-IR; 6 Gy). Nuclei as defined by auto threshold plugin on the DAPI image (ImageJ) are outlined in yellow. When indicated, cells were transfected with a plasmid expressing RNase H1 (RH) 24 h before or treated with cordycepin (Cordy) for 2 h at 37°C before fixation. Single antibody controls from untreated siC cells are shown. Scale bar indicates 10 µm. Right: Quantification of the number of PLA spots per nucleus. For statistical comparison of the differences between the samples, we applied a Kruskal–Wallis test followed by Dunn’s multiple comparison test and the P-values show significant differences. The red line in the plot indicates the median, and each symbol represents a single PLA spot. Diagram showing the BRCA2 N-terminal truncations used in this study and amylose pull-down from HEK293T whole cells extracts overexpressing the indicated BRCA2 N-terminal truncations (BRCA2T1, BRCA2LT2, BRCA2LT3) or the 2xMBP tag. DDX5 and BRCA2 truncations were detected using specific antibodies against DDX5 and MBP, respectively. Stain-Free images of the gels before transfer were used as loading control (cropped image is shown). Left: GST pull-down assay using purified BRCA2T1 and DDX5; MBP antibody was used for the detection of both proteins. UB: unbound; E: eluate. Right: SDS–PAGE showing 300 ng of purified MBP-DDX5-GST and of 2xMBP-tagged BRCA2T1 used in the pull-down assay. Source data are available online for this figure. Source Data for Figure 1 [embj2020106018-sup-0004-SDataFig1.pdf] Download figure Download PowerPoint Given that both BRCA2- and DDX5-deficient cells accumulate DNA-RNA hybrids (Bhatia et al, 2014; Mersaoui et al, 2019), we assessed whether the interaction could be promoted by DNA-RNA hybrids. As shown in Fig 1C, the proximity of BRCA2 and DDX5 in both untreated and irradiated cells was reduced after overexpression of RNase H1, a nuclease that specifically degrades the RNA moiety of DNA-RNA hybrids, the effect being stronger under irradiated conditions. In addition, inhibition of transcription with cordycepin led to a substantial reduction in the proximity of BRCA2 and DDX5 in both untreated and irradiated conditions suggesting that their co-localization is transcription-dependent (Fig 1C). Next, to define a smaller region of BRCA2 sufficient to bind DDX5 we used a series of truncated fragments contained in the BRCA2NT used in the proteomic mass spectrometry screen. We overexpressed three 2xMBP-NLS-tagged fragments comprising either BRCA2 aa 1–250, 1–500, or 1–750 or the 2XMBP-NLS alone as control and performed an amylose pull-down for the detection of DDX5 in complex with these fragments of BRCA2 (Fig 1D). Three BRCA2 fragments but not the control 2xMBP-NLS were able to form a benzonase-resistant complex with DDX5 indicating that the first 250 aa of BRCA2 (hereafter BRCA2T1) are sufficient to bind DDX5 (Fig 1D). To find out if the interaction was direct, we purified 2XMBP-BRCA2T1 from HEK293T cells as we previously reported (von Nicolai et al, 2016) and MBP-DDX5-GST from bacteria as previously described (Xing et al, 2017) and performed a GST pull-down assay. Importantly, BRCA2T1 was readily eluted from the glutathione resin only in the reaction containing GST-DDX5-MBP indicating that the interaction between BRCA2 and DDX5 is direct (Fig 1E). Altogether, these results indicate that BRCA2 and DDX5 interact directly through the first 250 aa of BRCA2 and suggest that the interaction is enhanced particularly at DNA-RNA hybrids and in cells exposed to γ-irradiation. DDX5 depletion leads to an increase of DNA-RNA hybrids It has previously been shown that depletion of BRCA2 (Bhatia et al, 2014) or DDX5 (Mersaoui et al, 2019) leads to DNA-RNA hybrids accumulation; accordingly, we observed DNA-RNA accumulation in the nucleus of U2OS depleted of BRCA2 or DDX5 visualized by immunofluorescence (IF) using the DNA-RNA hybrid marker S9.6 (Boguslawski et al, 1986) after nuclei pre-extraction and excluding signal from nucleoli (Fig EV2A). This signal was specific as it was sensitive to RNaseH1 treatment. Consistently, DDX5 overexpression rescued the DNA-RNA hybrid accumulation observed in DDX5-depleted cells but also of BRCA2-depleted cells (Figs 2A and EV2B) confirming its role suppressing these hybrids. Click here to expand this figure. Figure EV2. Related to Figs 1 and 2. DNA-RNA hybrids levels in BRCA2- and DDX5-depleted cells and localization of DDX5 at DNA-RNA hybrids Left: Quantification of the relative intensity of S9.6 staining. The data represent at least 500 cells per condition from three independent experiments. The red line in the scatter plot represents the median. For statistical comparison of the differences between the samples, we applied Kruskal–Wallis test followed by Dunn’s multiple comparison test and the P-values show significant differences. Right: Representative immunofluorescence images of U2OS cells depleted of DDX5 (siDDX5), BRCA2 (siBRCA2), or control cells (siC) and stained with S9.6 antibody (DNA-RNA hybrids) and counterstained with DAPI. When indicated, cells were transfected/treated with RNase H1 (RH) 24 h before fixation. Scale bar indicates 10 µm. Left: Representative images of S9.6 immunofluorescence of U2OS cells depleted of DDX5 (siDDX5) or control cells (siC) expressing RNaseH1-GFP and/or DDX5-GFP. Scale bar indicates 10 µm. Right: Quantification of S9.6 average nuclear intensity of U2OS cells depleted of DDX5 (siDDX5) or control cells (siC) expressing RNaseH1-GFP and/or DDX5-GFP. The red line in the plot indicates the median, and each symbol represents the value of a single cell. The statistical significance of the difference was calculated with Mann–Whitney U-test; the P-values show the significant difference. The data represent at least 170 cells per condition from one single experiment. Left: Representative images of in situ PLA experiment performed between DDX5 and S9.6 antibodies in U2OS cells. When indicated, cells were treated with cordycepin (Cordy) for 2 h at 37°C before fixation. Single antibody controls from untreated cells are shown. Scale bar indicates 10 µm. Nuclei as defined by auto threshold plugin on the DAPI image (ImageJ) are outlined in yellow. Right: Quantification of the number of PLA spots per nucleus in different conditions, as indicated. The data represent at least 200 cells per condition from three independent experiments. For statistical comparison of the differences between the samples, we applied a Kruskal–Wallis test followed by Dunn’s multiple comparison test and the P-values show significant differences. The red line in the plot indicates the median, and each symbol represents a single PLA spot. Download figure Download PowerPoint Figure 2. DDX5 depletion leads to a genome-wide accumulation of DNA-RNA hybrids particularly enriched at DSBs Left: Representative images of S9.6 immunofluorescence of U2OS cells depleted of BRCA2 (siBRCA2), DDX5 (siDDX5), or control cells (siC) after transfection with either an empty plasmid or a plasmid expressing DDX5. The merged images show the signal of S9.6, nucleolin (nucleoli) antibodies and DAPI staining. Scale bar indicates 25 µm. Right: Quantification of S9.6 average nuclear intensity of U2OS cells depleted of BRCA2 (siBRCA2), DDX5 (siDDX5), or control cells (siC) after transfection with either an empty plasmid or a plasmid expressing DDX5. The red line in the plot indicates the median, and each symbol represents the value of a single cell. The statistical significance of the difference was calculated with Mann–Whitney U-test, and the P-values show the significant difference. The data represent at least 235 cells per condition from three independent experiments. See also Fig EV2B. Top: Representative images of in situ PLA performed with anti-DDX5 and S9.6 antibodies in EdU-labeled U2OS cells. Where indicated, cells were transfected with a plasmid expressing RNase H1 (RH). Nuclei as defined by auto threshold plugin on the DAPI image (ImageJ) are outlined in yellow. Bottom: Quantification of PLA spots per nucleus in each condition as indicated. At least 300 cells per condition were counted from three independent experiments. For statistical comparison of the differences between the samples, we applied a Kruskal–Wallis test followed by Dunn’s multiple comparison test and the P-values show significant differences. The red line in the plot indicates the median, and each symbol represents a single PLA spot. See also Fig EV2C. Representative screenshot of a specific genomic region showing DRIPc-seq profiles at Watson (W) and Crick (C) strands in K562 cells depleted of DDX5 (siDDX5) or control cells (siC) from two independent experiments. See also Fig EV3. DNA-RNA hybrid distribution along protein-coding genes containing DRIPc-seq peaks in both conditions (siC and siDDX5) and replicates. Gene metaplots represent the mean of antisense or sense DRIPc-seq signal from two independent experiments in K562 cells depleted of DDX5 (siDDX5) or control cells (siC) as indicated. DNA-RNA hybrid metaplot distribution over γH2AX ChIP-seq peaks. Peak metaplot shows the mean DRIPc-seq signal from two independent experiments in K562 cells depleted of DDX5 (siDDX5) or control cells (siC). Venn diagram representing the overlap between γH2AX-positive genes in K562 cells (γH2AX ChIP-seq) and genes that specifically accumulate hybrids in control cells (top) or in DDX5-depleted cells (bottom). Download figure Download PowerPoint To test whether DDX5 associates with DNA-RNA hybrids, we performed in situ PLA experiments and found that DDX5 was indeed in close proximity to them (Fig 2B). As expected for an association with DNA-RNA hybrids, the proximity was reduced in cells transfected with a plasmid expressing RNase H1 (Fig 2B). Given that (i) DDX5 depletion leads to increased sensitivity to replication stress (Mersaoui et al, 2019) and (ii) unscheduled DNA-RNA hybrids represent a barrier for replication (Kotsantis et al, 2016; Stork et al, 2016; Gómez-González & Aguilera, 2019), we asked whether this association was particularly enriched in replicating cells. However, in our conditions, DDX5 association with hybrids was independent of replication, since both EdU- and non-EdU-stained cells displayed similar levels of DDX5-S9.6 PLA signal (Fig 2B), but was dependent on transcription (Fig EV2C). To analyze the genome-wide effect of DDX5 depletion on DNA-RNA hybrids, we performed DNA-RNA hybrid immunoprecipitation (DRIP) followed by cDNA conversion coupled to high-throughput sequencing (DRIPc-seq) that provides high-resolution and strand-specific profiling of hybrids (Sanz et al, 2016) in K562 cells. To verify the specificity of the S9.6 immunoprecipitated signal before sequencing, we confirmed the presence of DNA-RNA hybrids in this cell type by DRIP followed by qPCR (DRIP-qPCR) at different loci (Fig EV3A). These included APOE, previously described to be hybrid-prone in several conditions such as BRCA2-depleted HeLa cells (Bhatia et al, 2014), HIST1H2BG, shown to accumulate DNA-RNA hybrids in U2OS cells upon DDX5 loss (Mersaoui et al, 2019), and WDR90, shown to accumulate DNA-RNA hybrids in HeLa cells depleted of DNA damage response (DDR) factors (Barroso et al, 2019). Importantly, all S9.6 signals were severely reduced after in vitro treatment with RNase H1 indicating that S9.6 immunoprecipitation was specific for DNA-RNA hybrids. Consistent with the reliability of the DRIPc-seq method (Sanz & Chédin, 2019), the data obtained from three biological replicates were reproducible (Fig EV3B). We compared the genome-wide strand-specific composite profile between two replicas (Figs 2C and EV3C) as well as with contro
Referência(s)