Transcriptional repression facilitates RNA:DNA hybrid accumulation at DNA double-strand breaks

Abstract

RNA:DNA hybrids accumulate at DNA double-strand breaks (DSBs) and were shown to regulate homologous recombination repair. The mechanism responsible for the formation of these non-canonical RNA:DNA structures remains unclear although they were proposed to arise consequently to RNA polymerase II or III loading followed by DSB-induced de novo transcription at the break site. Here, we found no evidence of RNA polymerase recruitment at DSBs. Rather, strand-specific R-loop mapping revealed that RNA:DNA hybrids are mainly generated at DSBs occurring in transcribing loci, from the hybridization of pre-existing RNA to the 3' overhang left by DNA end resection. We further identified the H3K4me3 reader spindlin 1 and the transcriptional regulator PAF1 as factors promoting RNA:DNA hybrid accumulation at DSBs, through their role in mediating transcriptional repression in cis to DSBs. Altogether, we provide evidence that RNA:DNA hybrids accumulate at DSBs occurring in transcribing loci as a result of DSB-induced transcriptional shut down.

Fig. 1. RNA polymerases II and III are not recruited at breaks post-DSB induction.

a, Average profiles of RNAPII ChIP-seq signals using two different antibodies (405A and D8L4Y) before (−DSB) or after DSB induction (+DSB) around 80 best-cleaved DSBs and around 80 random sites. b, Genomic tracks (hg19) of BLESS, RNAPII, RNAPIII (POLR3A and POLR3E) ChIP-seq and RNA-seq around five DSBs located either in the promoter, exon, or intron of transcriptionally active genes, within an intergenic region or within a transcriptionally silent locus in the absence (−DSB) or presence of DSB (+DSB) induction. c, Quantification of RNAPII ChIP-seq signal on TC-DSBs (n = 65) and silent DSBs (n = 15) on a ±500-bp window around DSBs. Centre line shows the median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points indicate outliers. P values were calculated using paired two-sided nonparametric Wilcoxon tests. d, Genomic tracks (hg19) of RNAPII ChIP-seq (405A) and RNA-seq on the RBMXL1 gene showing the position of sgRNA1 (promoter) and sgRNA2 (intron) (top) and on an intergenic locus (bottom) with the position of sgRNA3. The position of primers for ChIP–qPCR in e and g is represented with red bars (A–E for sgRNA1 and sgRNA2, F–H for sgRNA3). e, RNAPII (405A) and γH2AX ChIP–qPCR before (no gRNA) or after DSBs induction with either sgRNA1, sgRNA2 or sgRNA3 normalized by Ctrl2. Mean and s.e.m. of three independent biological experiments are shown. P values calculated from paired two-sided t-tests comparing no sgRNA with sgRNA conditions are only indicated if significant (P < 0.05). f, POLR3E ChIP–qPCR before (no gRNA) or after CRISPR-generated DSBs with either sgRNA1, sgRNA2 or sgRNA3 normalized by Ctrl2. Mean and s.e.m. of n = 3 independent biological experiments are shown. P values from paired two-sided t-tests comparing no sgRNA with sgRNA conditions are only indicated if significant (P < 0.05). g, Average profiles of ChIP-seq signals for two different subunits of RNAPIII (POLR3A and POLR3E) before (−DSB) and after DSB induction (+DSB) around 80 best-cleaved DSBs in DIvA cells. Source numerical data are available in Source data. Source data

Fig. 2. No evidence of de novo transcription at DSBs.

a, Genomic tracks (hg19) of ChIP-seq against RNAPII CTD phosphorylation on serine 5 (S5P) (3E8 or D9NI antibodies), serine 7 (S7P), tyrosine 1 (Y1P) and serine 2 (S2P) (3E10 or E1Z3G antibodies) before and after DSB induction on two TC-DSBs (DSB 379 and DSB 526). b, Average profile of ChIP-seq against RNAPII phosphorylated forms at silent and TC-DSBs. c, Quantification of ChIP-seq signal of RNAPII CTD phosphorylations at silent (n = 15) and TC-DSBs (n = 65) on a ±500 bp window around DSBs. Centre line shows the median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points indicate outliers. P values were calculated using paired two-sided nonparametric Wilcoxon tests. d, Genomic tracks (hg19) of BLESS and TT_chem-seq signal at silent and TC-DSBs before (−DSB) and after DSB induction. e, Average profile of TT_chem-seq signal at silent and TC-DSBs. The signal was oriented according to gene directionality. f, Quantification of TT_chem-seq on a ±500-bp window around silent (n = 15) and TC-DSBs (n = 65). Centre line shows median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points indicate outliers. P values were calculated using paired two-sided nonparametric Wilcoxon tests. g, Average profile of TT_chem-seq signal on damaged and control genes in −DSB and +DSB conditions. h, Quantification of TT_chem-seq signal at damaged (n = 76) and control (n = 73) genes. Centre line shows median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points indicate outliers. P values were calculated using paired two-sided nonparametric Wilcoxon tests. i, Genomic tracks (hg19) of ChIP-seq against SPT5 in −DSB and +DSB conditions at two TC-DSBs as in a. j, Average profile of SPT5 ChIP-seq signal on damaged and control genes in −DSB and +DSB conditions. Source data

Fig. 3. Strand-specific accumulation of RNA:DNA hybrids on single-stranded DNA resulting from DNA end resection.

a, Average qDRIP-seq profiles on ±5 kb around the 80 best-cleaved DSBs. b, Genomic tracks (hg19) of BLESS and strand-specific qDRIP-seq at a TC-DSB (DSB 473). c, Top: schematic representation of DNA end resection at DSBs: resection triggers the elimination of the 5′-terminated strand (brown) leaving the 3′-terminated strand (blue) intact. Heatmaps of qDRIP-seq signal on 5′-terminated strands (brown, left, both sides of the DSB combined) and on 3′-terminated strand (blue, right, both sides of the DSB combined) for the 80 best-cleaved DSBs (bottom). d, Quantification of qDRIP-seq signal at DSBs (n = 80), represented in c for the 5′ (brown) or 3′-terminated strands (blue). Centre line shows median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points indicate outliers. P values, paired two-sided nonparametric Wilcoxon tests. e, RNA:DNA hybrid levels measured by DRIP–qPCR upon CtIP depletion or MRE11 inhibition (endonuclease activity, PFM01 or exonuclease activity, Mirin), before and after DSB induction at a negative control (Ctrl Neg) region and at a TC-DSB (DSB 526). Mean and s.e.m. of independent biological experiments (n ≥ 4) are shown (normalized to −DSB condition). P values, paired two-sided t-test. f, Average profiles of qDRIP-seq, RPA ChIP-seq and END-seq after 4 h or 24 h of DSB induction. The dotted grey boxes highlight the extend of resection at 4 h and 24 h post-DSB induction. g, Genomic tracks (hg19) of the strand-specific qDRIP-seq, RPA ChIP-seq and END-seq at a TC-DSB (DSB 526). Arrows show the local increase of qDRIP-seq and RPA ChIP-seq signals adjacent to the DSB. h, Zoom of the Crick strand of g for the qDRIP-seq and RPA ChIP-seq tracks (hg19). i, Average profile of qDRIP-seq and RPA ChIP-seq oriented by RNA:DNA hybrid directionality on a ±2-kb window around the 20 best hybrid-accumulating DSBs. The arrow points at the local increase in RPA and RNA:DNA hybrids. j, Average END-seq profile in siRNA Ctrl- (siCtrl) or siRNA SETX- (siSETX) transfected cells on ±10 kb around the 80 best-cleaved DSBs. Arrows indicate where DNA end resection reaches its furthest point. Source numerical data are available in Source data. Source data

Fig. 4. RNA:DNA hybrids result from the hybridization of pre-existing RNA on the template strand.

a, Genomic tracks (hg19) showing strand-specific qDRIP-seq and TT_chem-seq, on the Watson and Crick strands and RNA-seq in −DSB and +DSB conditions at TC-DSBs located in a promoter, exon or intron and in a silent DSB located in a transcriptionally inactive gene. b, Schematic representation and average profile of qDRIP-seq signal on the template (red) or nontemplate (orange) strands of TC-DSBs located within genes. The signal was oriented according to gene directionality. c, Quantification of the qDRIP-seq signal on a ±250-bp window around TC-DSBs located within genes (n = 45). Centre line shows the median; box limits show first and third quartiles; whiskers show maximum and minimum without outliers; points show outliers. P values, paired two-sided nonparametric Wilcoxon tests. d, DRIP–qPCR detection of RNA:DNA hybrids around CRISPR-induced DSBs induced at different positions of the RBMXL1 gene (sgRNA1, promoter, sgRNA2, intron, sgRNA4, exon) or in an intergenic region (sgRNA3) normalized by the FOS positive control. Mean and s.e.m. of n = 3 biological replicate are shown. P values, paired two-sided t-test. Source numerical data are available in Source data. Source data

Fig. 5. PAF1-dependent transcriptional repression results in RNA:DNA hybrid accumulation at TC-DSBs.

a, DRIP–qPCR before and after DSB induction upon ATM inhibition (ATMi) at a Ctrl Neg region and at a TC-DSB in the RBMXL1 gene (DSB 526). Mean and s.e.m. for n = 4 biological replicates are shown. P values, paired two-sided t-test. b, same as a for cells transfected with Ctrl or NELF-E siRNA with DRIP–qPCR normalized by the RPL13A positive control. c, Measurement of elongation rates by DRB/TT_chem-seq across genes >60 kb for all genes (left, n = 5,644) or damaged genes (right, n = 49) before (−DSB) and after DSB induction (+DSB). Centre line shows the median; box limits represent first and third quartiles; whiskers show maximum and minimum without outliers; points show outliers. P values nonparametric paired two-samples Wilcoxon tests. d, RT–qPCR quantifying cDNA levels normalized by TBP before (−DSB) and after DSB induction (+DSB) for three damaged genes (RBMXL1, MIS12 and KLF7) carrying TC-DSBs in cells transfected with control (Ctrl), PAF1 or SKI8 siRNA. Mean and s.e.m. for n = 5 biological replicates are shown. P values, paired two-sided t-tests. e, Genomic tracks (hg19) of PAF1 ChIP-seq before and after DSB induction as well as the log₂ fold change ratio (+DSB/−DSB) at a TC-DSB (DSB 608). f, Average profiles of PAF1 ChIP-seq enrichment following DSB induction as log₂ fold change ratio (+DSB/−DSB) for damaged and control genes. g, DRIP–qPCR normalized by the LYRM1 positive control before and after DSB induction at a Ctrl Neg region and at DSB 526 in cells transfected with Ctrl, PAF1 or SKI8 siRNA. Mean and s.e.m. for n = 3 biological replicates are shown. P values, paired two-sided t-tests. Source numerical data are available in Source data. Source data

Fig. 6. SPIN1-dependent transcriptional repression results in RNA:DNA hybrid accumulation at TC-DSBs.

a, Average profiles of SPIN1 ChIP-seq enrichment following DSB induction as log₂ fold change ratio (+DSB/−DSB) for damaged and control genes. b, Genomic tracks (hg19) of SPIN1 ChIP-seq before and after DSB induction as well as the log₂ fold change ratio (+DSB/−DSB) at two TC-DSBs (DSB 379 and DSB 765). c, RT–qPCR quantifying cDNA levels normalized by RPLP0 before (−DSB) and after DSB induction (+DSB) for three genes (MIS12, KLF7 and TRIM37) carrying TC-DSBs in cells transfected with control (Ctrl) or SPIN1 siRNA. Mean and s.e.m. for n = 3 biological replicates are shown. P values, paired two-sided t-tests. d, Same as c for cells treated with MS31. n = 4 biological replicates. e, Genomic tracks (hg19) showing qDRIP-seq signal in siCtrl and siSPIN1-transfected cells on Watson and Crick strands in −DSB and +DSB conditions at TC-DSBs located in a promoter, exon or intron and in a silent DSB. f, Average profile of the qDRIP-seq signal in siCtrl and siSPIN1 cells on the Watson and Crick strands on ± 5 kb around the 80 best-cleaved DSBs. Source numerical data are available in Source data. Source data

Fig. 7. RNA:DNA hybrids induced by transcriptional repression at DSBs promote resection and toxicity.

a, Schematic representation of the resection assay (top) and quantification of ssDNA after DSB in siCtrl, siCtIP, siPAF1, siSPIN1 or upon MS31 treatment at 200 bp away from a TC-DSB (DSB 657). Mean and s.e.m. for n ≥ 3 biological replicates are shown. P values, paired two-sided t-test. b, RAD51 ChIP–qPCR efficiency (normalized to undamaged Ctrl1) before (−DSB) or after DSB (+DSB) in control or SPIN1 siRNA-treated cells at 100 bp and 800 bp from TC-DSBs (DSB 600, 379 and 657) and silent DSBs (DSB 72 and 871). Mean and s.e.m. for n = 3 biological replicates are shown. P values, paired two-sided t-tests. c, Representative example showing the number of RAD51 foci in siCtrl- or siPAF1-transfected cells in −DSB and +DSB conditions. Red shows the median. P values, paired two-sided nonparametric Wilcoxon tests. d, Clonogenic assay in control (Ctrl) or SPIN1 siRNA-treated DIvA-AID cells. A representative experiment is shown on the top panel. Mean and s.e.m. for n = 7 biological replicates are shown (bottom). P values, paired two-sided t-tests. e, Illegitimate rejoining frequency between two TC-DSBs (DSB 343 and DSB 379) in DIvA-AID cells (+DSB +repair) and normalized to siCtrl upon siPAF1, siSETX or siPAF1+ siSETX (left) or upon siCtrl, siSPIN1, siSETX or siSPIN1+siSETX (right). Mean and s.e.m. for n = 4 biological replicates are shown. P values, paired two-sided t-tests. f, Model for RNA:DNA hybrid formation at TC-DSBs. Upon DSB induction within a transcribing gene, an ATM-dependent pathway, which entails the recruitment of PAF1 and of the H3K4me3 reader SPIN1, as well as the eviction of SPT5, triggers transcriptional repression via a decrease in PPP release. As a consequence, R-loops accumulate on the template strand of the damaged gene. The displaced ssDNA further becomes a substrate for endonucleolytic cleavage (flap removal), thus simultaneously initiating resection and converting the R-loop into a more stable RNA:DNA hybrid. This R-loop-mediated, alternative resection initiation pathway would commit the DSB to HR repair. RNA:DNA hybrids shall further be removed by SETX and potentially other RNA:DNA helicases, to allow HR repair and to avoid RNA:DNA hybrid-dependent toxicity. Source numerical data are available in Source data. Source data

Extended Data Fig. 1. RNAPII levels at DSBs and control loci.

(a) Distribution of 80 best AsiSI-induced DSBs. There are 15 ‘silent’ DSBs (which comprise intergenic DSBs and DSBs falling within transcriptionally silent genes) and 65 TC-DSBs (top panel). Most TC-DSBs fall within 1 kb of the TSS (88%, bottom left panel) and are distributed across promoters, exons and introns (respectively 41, 25 and 34%, bottom right panel). (b) Average profile and heatmap of BLESS signal after break induction on −/+ 5 kb around silent and TC-DSBs. (c) Average profile and heatmap of RNAPII ChIP-seq signal (using 405 A and D8L4Y antibodies) on all genes −/+ 2 kb (scaled read count). (d) Browser tracks (hg19) of RNAPII ChIP-seq (using 405 A and D8L4Y antibodies) before (-DSB) and after DSB induction (+DSB) and Log2 fold change ratio (+DSB/-DSB) at the CDKN1A (encoding p21) DDR gene. (e) Average profile and heatmap of RNAPII ChIP-seq signals (using 405 A and D8L4Y antibodies) on −/+ 5 kb around silent and TC-DSBs (scaled read count). (f) Average profile and heatmap of RNAPII ChIP-seq signals (using 405 A and D8L4Y antibodies) on −/+ 5 kb around the 80 best DSBs classified according to RNAPII level before breakage (high, medium, low). (g) Quantification of RNAPII ChIP-seq signals on a −/+ 1 kb window around DSBs in exons (n = 19) or introns (n = 26). Centre line: median; Box limits: 1st and 3rd quartiles; Whiskers: Maximum and minimum without outliers; Points: outliers. P values, paired two-sided nonparametric Wilcoxon tests. (h) Genomic tracks (hg19) of RNAPII ChIP-seq and RNA-seq Log2 fold change ratio +DSB/-DSB at a TC-DSB in the RBMXL1 gene (DSB 526). Source numerical data are available in source data. Source data

Extended Data Fig. 2. RNAPII and RNAPIII levels at DSBs and control loci.

(a) ChIP–qPCR efficiency (% input normalized to an undamaged control locus) of Mock (no antibody) or RNAPII before and after DSB induction at an intergenic undamaged locus (Ctrl), three TC-DSBs (DSB 526, 657 and 379), a DSB in a silent locus (DSB 72) and the promoter of the DDR gene CDKN1A gene (encoding p21) as a control. Mean and s.e.m. for n = 4 biological replicates are shown. P values, paired two-sided t-tests. (b) Genomic tracks (hg19) of END-seq before (-Eto) and after Etoposide (+Eto, two replicates Rep1 and Rep2) and RNAPII ChIP-seq (405 A) in the absence of Etoposide (-Eto). Three examples of etoposide-induced genic DSBs (top panels) and etoposide-induced intergenic DSBs (bottom panel) are shown. (c) ChIP–qPCR efficiency (% input normalized to the undamaged control locus Ctrl2) of Mock (no antibody), RNAPII and RNAPIII ChIP–qPCR before (-Eto) or after Etoposide treatment (+Eto) at the promoter of CDKN1A (p21) or at six etoposide-induced DSBs (3 genic and 3 intergenic are shown). Mean and s.e.m. for n = 5 (RNAPII) or n = 4 (RNAPIII) biological replicates are shown. P values, paired two-sided t-tests. Only significant P values between – and + DSB are indicated. (d) Genomic tracks (hg19) of POLR3A and POLR3E ChIP-seq before and after DSB induction at a tRNA cluster on chromosome 1. (e) Average profiles of POLR3A and POLR3E ChIP-seq centred on tRNA genes before (-DSB) and after DSB induction (+DSB). (f) Genomic tracks (hg19) of BLESS, POLR3A and POLR3E ChIP-seq in -DSB and +DSB conditions at sites shown in ref. (NB: these three AsiSI sites did not display cleavage in DIvA cells, see BLESS track). Source numerical data are available in source data. Source data

Extended Data Fig. 3. Levels of RNAPII CTD phosphorylation, nascent transcription and SPT5 at DSBs and control loci.

(a) Metagene profile and heatmaps of Y1P, S7P, S5P and S2P RNAPII CTD phosphorylation before DSB induction on all genes (−/+ 2 kb) using scaled read counts. (b) Genomic tracks (hg19) of Y1P, S7P, S5P and S2P RNAPII ChIP-seq before and after DSB induction on the DDR gene CDKN1A (encoding p21) using scaled read counts. (c) Heatmap of RNAPII CTD phosphorylation before DSB induction at silent and TC-DSBs. (d) Metagene profile of the TT_chem-seq signal without 4SU incorporation and with 4SU incorporation in - and + DSB conditions. (e) Left panel: Genomic tracks (hg19) of stranded TT_chem-seq in the absence of damage showing nascent RNA detection at the lowly expressed DICER-AS gene. Right panel: total TT_chem-seq on CDKN1A in -DSB and +DSB conditions. (f) Heatmaps of TT_chem-seq at silent and TC-DSBs orientated according to gene directionality. (g) Quantification of the TT_chem-seq signal on a −/+ 500 bp window around TC-DSBs falling in a promoter (n = 31), exon (n = 19) or intron (n = 26) (Centre line: median; Box limits: 1st and 3rd quartiles; Whiskers: Maximum and minimum without outliers; Points: outliers. P values, paired two-sided nonparametric Wilcoxon tests). (h) Same as in (e) for SPT5 ChIP-seq. (i) Same as in (a) but for SPT5 ChIP-seq. Source numerical data are available in source data. Source data

Extended Data Fig. 4. DSB-induced RNA:DNA hybrid mapping by qDRIP-seq.

(a) Schematic representation of strand-specific qDRIP-seq. For (+) orientated genes, the resulting R-loops are detected on the Crick strand and for (−) orientated genes, R-loops are detected on the Watson strand. (b) Genomic track (hg19) of qDRIP-seq signal on the Watson and Crick strands for the (+) orientated NEAT1 and MALAT1 genes in a previously published qDRIP-seq obtained in HeLa cells (ref. ) and in DIvA cells in the absence of DSB. (c) Genomic track (hg19) of strand-specific qDRIP-seq signals (Watson and Crick strands) at GADD45A. The differential enrichment of the qDRIP-seq signal on the Crick strand following DSB induction (+DSB (−) -DSB) is also shown. (d) Box plot of the qDRIP-seq signal on −/+ 1 kb around DSBs (n = 80) before (-DSB) and after DSB induction (+DSB) (Centre line: median; Box limits: 1st and 3rd quartiles; Whiskers: Maximum and minimum without outliers; Points: outliers; P values paired two-sided nonparametric Wilcoxon tests). (e) Average profiles and heatmaps of qDRIP-seq classified according to the presence (hybrid-positive) or absence of hybrids (hybrid-negative) at DSBs. BLESS signal is also shown for these two categories. (f) Genomic track (hg19) of qDRIP-seq at two hybrid-positive TC-DSBs and at two genes (RPL13A and RPS24) known to accumulate R-loops and whose exact levels have been previously measured by SMRF-seq (ref. ). (g) DRIP–qPCR at the R-loop positive RPL13A gene and at the TC-DSB induced in RBMXL1 (DSB 526). The median value of n = 16 biologically independent experiments is shown (red). (h) Percentage of cleavage (normalized to an undamaged control locus) of DSB 526 as measured using ddPCR. (Mean and s.e.m. of n = 4 biological replicates are shown. P-value, paired two-sided t-test). (i) Left panel: RT–qPCR quantifying CtIP cDNA levels in control (Ctrl) or CtIP siRNA-transfected cells. Mean and s.e.m. for n = 4 biological replicates are shown. P-value, paired two-sided t-test. Right panel: Western blot against CtIP and Tubulin in siCtrl- and siCtIP-transfected cells. A representative experiment is shown (out of n = 2). Source numerical data and unprocessed blots are available in source data. Source data

Extended Data Fig. 5. RNA:DNA hybrids and resection.

(a, b) Heatmaps showing strand-specific signals for END-seq (a) and RPA ChIP-seq (b) on the 80 best-cleaved DSBs in the absence of DSB and after 4 h or 24 h of DSB induction. Heatmaps are sorted by decreasing END-seq or RPA ChIP-seq. (c) Average profiles of END-seq, RPA ChIP-seq and qDRIP-seq (both strands combined) oriented by RNA:DNA hybrid directionality on a −/+ 10 kb window around the 80 best-cleaved DSBs. The arrows show the extent of resection as measured by END-seq and RPA ChIP-seq. (d) Level of SETX and Tubulin measured by Western blot in siCtrl and siSETX transfected cells. A representative experiment is shown (out of n = 3). (e) Average profiles and heatmaps of the differential enrichment of the qDRIP-seq signal after depletion of SETX (siSETX +DSB (-) siCtrl +DSB) on hybrid-positive and -negative DSBs. Unprocessed blots are available in source data. Source data

Extended Data Fig. 6. DSB-induced RNA:DNA hybrids accumulate on the template strand upon transcriptional repression.

(a) Genomic tracks (hg19) of qDRIP-seq on the Watson and Crick strands -DSB and +DSB at a TC-DSB located in an exon and in a silent DSB located in an intergenic region. The strand-specific TT_chem-seq signal is also shown as well as the RNA-seq signal in -DSB and +DSB conditions. (b) Quantification of the qDRIP-seq signal on −/+ 250 bp around TC-DSBs located in promoter (n = 31), exon (n = 19) or intron (n = 26) (Centre line: median; Box limits: 1st and 3rd quartiles; Whiskers: Maximum and minimum without outliers; Points: outliers. P values, paired two-sided nonparametric Wilcoxon tests). (c) Genomic tracks (hg19) of RNAPII ChIP-seq (405 A) and RNA-seq on RBMXL1 showing the position of sgRNA1, 2 and 4 and of the primers (I) used in (d). (d) ChIP–qPCR (% input normalized to the undamaged control locus Ctrl2) of γH2AX before (no gRNA) or after CRISPR-generated DSBs with sgRNA4 (mean and s.e.m. for n = 3 biological replicates are shown. P-value, paired two-sided t-test). (e) RT–qPCR quantifying RBMXL1 cDNA levels normalized to TBP upon no gRNA or CRISPR-generated DSBs with sgRNA1, 2 or 4. Mean and s.e.m. for n = 4 biological replicates are shown. P values, paired two-sided t-tests. (f) DRIP–qPCR detection of RNA:DNA hybrids around DSB induced with sgRNA1 in HAP1 cells normalized to the FOS positive control (mean and s.e.m. for n = 4 biological replicates are shown. P-value, paired two-sided t-test). Source numerical data are available in source data. Source data

Extended Data Fig. 7. ATM, NELF-E, the PAF complex participate in DSB-induced transcriptional repression in cis.

(a) RT–qPCR quantifying cDNA levels normalized to RPLP0 at 4 damaged genes (RBMXL1, MIS12, KLF7 and TRIM37) before and after DSB upon ATMi treatment. Mean and s.e.m. for n = 4 biological replicates are shown. P values are calculated with paired two-sided t-tests. (b) Levels of NELF-E and Myosin measured by Western blot after depletion of NELF-E. A representative experiment is shown (out of n = 2) (c) same as (a) for siRNA NELF-E-transfected cells. (d) Waves of elongation on all genes measured with DRB/TTchem-seq in -DSB and +DSB conditions. (e) Genomic track (hg19) example of a (+) orientated long gene (PKN2) and a (-) orientated short gene (GTF2B) showing waves of elongation measured by DRB/TTchem-seq on both Watson and Crick strands in -DSB and +DSB conditions. (f) Level of repression measured by RT–qPCR for RBMXL1 following DSB induction upon transfection of cells with different siRNAs (indicated on the y axis). The red dotted line represents the baseline of repression detected with siCtrl. (g) Levels of PAF1 and Tubulin (left panel) and SKI8 and Myosin (right panel) measured by Western blot after depletion of either PAF1 or SKI8 (n = 1). (h) Average profile and heatmap of PAF1 ChIP-seq signal on all genes −/+ 2 kb (normalized read count). (i) Genomic tracks (hg19) of PAF1 ChIP-seq before and after DSB induction as well as the Log2 fold change ratio (+DSB/-DSB) at a TC-DSB (DSB 765). Source numerical data and unprocessed blots are available in source data. Source data

Extended Data Fig. 8. SPIN1 participates in DSB-induced transcriptional repression in cis.

(a) Genomic tracks (hg19) of SPIN1, H3K4me3 and RNAPII (405 A) ChIP-seq as well as RNA-seq before DSB induction showing the similarities between the signals detected. (b) Average profile and heatmap of SPIN1 ChIP-seq signal on all genes −/+ 2 kb (normalized read count). (c) Metagene profiles of ChIP-seq signals for H3K4me3 (left panel) and SPIN1 (right panel) on genes classified into three classes according to their RNAPII levels (high, mid or low) before DSB induction. (d) Levels of SPIN1 and Myosin detected by Western blot after siRNA-mediated depletion of SPIN1. A representative experiment is shown (out of n = 4). (e) Levels of SPIN1, H3K4me3 and Tubulin detected by Western blot after treatment with MS31 (n = 1). (f) ChIP–qPCR efficiency (%input) of either H3K4me3 (left panel) or SPIN1 (right panel) at a control intergenic locus (Ctrl1) and at the promoter of CDKN1A and GAPDH in the presence or absence of MS31 treatment (mean and s.e.m. of technical replicates (n = 4)). (g) Average profiles and heatmaps of the differential enrichment of qDRIP-seq signal after depletion of SPIN1 (siSPIN1 +DSB (-) siCtrl +DSB) on hybrid-positive and -negative DSBs. Source numerical data and unprocessed blots are available in source data. Source data

Extended Data Fig. 9. Factors participating in DSB-induced transcriptional repression participate in HR and cell survival.

(a) Number of BRCA1 or RAD51 foci in the presence or absence of Etoposide (Eto) and/or MS31. Red line: median. P values, paired two-sided nonparametric Wilcoxon tests. (b) Clonogenic assay in MS31-treated DIvA-AID cells after DSB induction (+DSB). A representative experiment is shown on the left panel. Mean and s.e.m. for n = 3 biological replicates are shown on the right panel. P values, paired two-sided t-test. (c) Percentage of cell survival upon treatment with different doses of Etoposide treatment (indicated on the x axis) and in the presence or absence of MS31. (d) Clonogenic assay in control (Ctrl) or NELF-E siRNA-treated DIvA-AID cells in -DSB and +DSB conditions and after IAA treatment allowing DSB repair (+DSB +repair). Mean and s.e.m. for n = 4 biological replicates are shown. P values, paired two-sided t-tests. (e) same as in (d) for control (Ctrl), SPIN1 or SPIN1 + SETX siRNA-treated DIvA-AID cells, n = 3 biological replicates. P values, paired two-sided t-test. Source numerical data are available in source data. Source data