Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Dec;19(12):1400-1411.
doi: 10.1038/ncb3643. Epub 2017 Nov 27.

Damage-induced lncRNAs Control the DNA Damage Response Through Interaction With DDRNAs at Individual Double-Strand Breaks

Affiliations
Free PMC article

Damage-induced lncRNAs Control the DNA Damage Response Through Interaction With DDRNAs at Individual Double-Strand Breaks

Flavia Michelini et al. Nat Cell Biol. .
Free PMC article

Abstract

The DNA damage response (DDR) preserves genomic integrity. Small non-coding RNAs termed DDRNAs are generated at DNA double-strand breaks (DSBs) and are critical for DDR activation. Here we show that active DDRNAs specifically localize to their damaged homologous genomic sites in a transcription-dependent manner. Following DNA damage, RNA polymerase II (RNAPII) binds to the MRE11-RAD50-NBS1 complex, is recruited to DSBs and synthesizes damage-induced long non-coding RNAs (dilncRNAs) from and towards DNA ends. DilncRNAs act both as DDRNA precursors and by recruiting DDRNAs through RNA-RNA pairing. Together, dilncRNAs and DDRNAs fuel DDR focus formation and associate with 53BP1. Accordingly, inhibition of RNAPII prevents DDRNA recruitment, DDR activation and DNA repair. Antisense oligonucleotides matching dilncRNAs and DDRNAs impair site-specific DDR focus formation and DNA repair. We propose that DDR signalling sites, in addition to sharing a common pool of proteins, individually host a unique set of site-specific RNAs necessary for DDR activation.

Figures

Figure 1
Figure 1. Sequence-specific localization of DDRNAs at DNA damage sites is transcription-dependent.
(A) Images of NIH2/4 cells expressing GFP-LacR, microinjected with double-stranded DDRNA-Cy5, artificial CXCR4-Cy5 miRNA (Ctrl RNA 1) or let-7a-Cy5 miRNA (Ctrl RNA 2), together with BSA (-) or I-SceI restriction enzyme (+) and imaged 4 h post injection. Scale bar 5 µm. Inset is a magnified view of the boxed region. Images from one out of 3 experiments with similar results. (B) Quantification of (A) showing the number of fluorophore-labeled RNA molecules at the locus as measured by single-molecule analysis based on stepwise photobleaching. Dots represent individual cells. The black line represents the mean ± SEM (data are shown as pool of n=3 independent experiments). (C) DDRNAs localize at the damage site to restore DDR focus formation. NIH2/4 cells knocked-down for Dicer and Drosha were mildly permeabilized and incubated with DDRNA-Cy5 or CXCR4-Cy5 (Ctrl RNA 1). The bar plot shows the percentage of cells positive for co-localization of 53BP1 with TetR, of RNA-Cy5 with TetR and the triple co-localization of 53BP1, RNA-Cy5 and TetR. Error bars indicate SEM (for siLuc and siDic n=4, for siDro n=3 independent experiments, ≥70 cells analysed in total per condition). (D) NIH2/4 cells expressing YFP-TetR and inducible I-SceI were treated with AM, DRB or ACTD at low and high doses or vehicle alone for 2 h before cut induction, then mildly permeabilized and incubated with DDRNA-Cy5. The bar plots show the percentage of cells in which DDRNA signal co-localizes with the TetR spot. Error bars indicate SEM (n=3 independent experiments, ≥80 cells analysed in total per condition). (E) NIH2/4 cells expressing GFP-LacR were microinjected with double-stranded DDRNA-Cy5, together with I-SceI protein and AM and imaged 4 h post injection. The plot shows the number of DDRNA molecules at the locus as measured by single-molecule counting based on stepwise photobleaching. Dots represent individual cells. The black line represents the mean ± SEM (data are shown as pool of n=3 independent experiments). (B,E) P values were calculated using two-tailed t-test. (C,D) P values were calculated using chi-squared test. ***P<0.001, ****P<0.0001, ns not significant.
Figure 2
Figure 2. DSBs induce dilncRNAs that interact with DDRNAs.
(A) Schematic of the four potential dilncRNAs induced upon DSB in NIH2/4 and of smFISH probes used for detection. (B) Induction of dilncRNAs in NIH2/4 measured by smFISH. Relative intensity of the indicated probes at uncut or cut locus. Dots represent individual cells. Black bar represents mean ± SEM (data are shown as pool of n=3 independent experiments). (C-F) Induction of dilncRNAs in NIH2/4, U2OS19ptight, HeLa111 and I-PpoI HeLa cells measured by strand-specific RT-qPCR. Bar plots show the mean relative enrichment of indicated RNA sets upon cut. Uncut sample of each RNA set was used as reference. Error bars indicate SEM (C: for Uni Lac-from n=4, for the other dilncRNAs n=5; D: n=3; E: n=4; F: n=4 independent experiments). (G) NIH2/4 knocked-down for Drosha (siDro), Dicer (siDic), Translin (siTsn), Dicer and Translin (siDic+Tsn) or Luciferase (siLuc) were transfected with I-SceI-expressing vector (+) or empty vector (-). Bar plots show the mean relative enrichment of Lac-from dilncRNA relative to uncut siLuc by strand-specific RT-qPCR. Error bars indicate SEM (for siLuc, siDic and siDro n=6, for the other conditions n=3 independent experiments). (H) NIH2/4 knocked-down for Drosha (siDro), Dicer (siDic) or Luciferase (siLuc) were transfected with I-SceI-expressing vector (+) or empty vector (-). RNA fractions of 15-40 nt in lenght were recovered by gel-extraction. Bar plot shows the mean relative enrichment of let7a miRNA and DDRNAs matching Lac sequences, relative to uncut siLuc. Error bars indicate SEM (n=3 independent experiments). (I) Biotinylated DDRNA (btn-L1 and btn-U1) or biotinylated miRNA (btn-let-7a) were transfected into NIH2/4 cells expressing I-SceI and RNA pull-down was performed. Bar plot shows dilncRNA Lac-from and Rplp0 mRNA levels, assessed by strand-specific RT-qPCR, as relative to input. Values are expressed as mean of 2 independent experiments. (J) Time course of γH2AX and 53BP1 focal accumulation by immunofluorescence, dilncRNA generation by smFISH and fluorescent DDRNA localization at the DSB in NIH2/4. Plot shows the percentage of cells bearing signals co-localizing with LacR. The dotted line represents best fit of data to a single exponential function. Error bars indicate SD (n=3 independent experiments, ≥12 cells analysed in total per time point). P values were calculated using two-tailed t-test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, ns not significant. Statistical source data are provided in Supplementary Table 4.
Figure 3
Figure 3. Active RNAPII is recruited to DSBs in mammalian cells and in cell extracts.
(A) Detection of γH2AX and RNAPII pSer5 at the DNA damage locus on chromatin spreads by super-resolution microscopy. Representative super-resolution images of chromatin fibers (green), LacR (red), γH2AX (blue) and RNAPII pSer5 (purple) in uncut and cut NIH2/4 cells. The yellow-boxed images correspond to the yellow-boxed inset region. Scale bar 500 nm. (B) The bar plot is the quantification of (A) and represents the percentage of cells showing γH2AX and active RNAPII pSer5 co-localizing with LacR by super-resolution microscopy on chromatin spreads in uncut and cut conditions. Data are shown as mean. Error bars indicate SEM (n=3 independent experiments, ≥20 images per sample per experiment). (C) Accumulation of RNAPII at the damaged locus by ChIP in uncut and cut NIH2/4 cells. Bar plot shows the percentage of enrichment relative to the input of RNAPII, RNAPII pSer5 and RNAPII pSer2 associated with genomic DNA, as detected with primers matching Lac sequences flanking the I-SceI-induced DSB. Data are shown as one representative of 3 independent experiments. (D) Bar plot shows the percentage of enrichment relative to the input of γH2AX and total RNAPII at the endogenous DAB1 locus by ChIP in HeLa cells cut by I-PpoI. Data are shown as one representative of 3 independent experiments. (E) Biotinylated DNA immobilized on streptavidin beads was either cut or not cut by recombinant I-SceI enzyme in vitro and incubated with nuclear cell extract. Input and pull-down samples were probed for total RNAPII. Below, an agarose gel shows equal amounts of uncut or cut DNA employed. This experiment was repeated 3 times independently with similar results. P value was calculated using two-tailed t-test. *P<0.05. Statistical source data are provided in Supplementary Table 4. Unprocessed original blots are shown in Supplementary Figure 9.
Figure 4
Figure 4. DSBs induce bidirectional transcription in cell-free extracts.
(A) Transcriptionally competent human cell-free extracts (CFE, “+” and “-” indicate active or heat-inactivated CFE, respectively) were incubated with a circular or I-SceI-linearized form of pLac-Tet plasmid in the presence of [α-32P]UTP. Where indicated, products were treated with increasing amounts of DNaseI or RNaseA. Products were resolved on a denaturing PAGE, along with a radiolabeled DNA ladder (M, nt indicates nucleotides). This experiment was repeated 3 times independently with similar results. (B) In the same settings described in (A), AM reduces DSB-induced transcription. This experiment was repeated 3 times independently with similar results. (C) Ion Proton sequencer was used to perform deep sequencing of RNA products generated in (A). From the outer to the inner circles, plots display the entire plasmid (brown, 2834 bp with ticks showing 100 bp intervals), the coverage of forward reads (blue, maximum value of the distribution is set to 1300) and reverse reads (red, maximum value of the distribution is set to 8000) for each nucleotide position of both circular and linearized plasmids. (D) Individual 5’ RACE clones (in blue forward, in red reverse) are aligned to the sequence of the DNA substrate used. The site of DSB generation is indicated by the dashed line. Unprocessed original scans of radioactive blots are shown in Supplementary Figure 9.
Figure 5
Figure 5. The MRN complex binds to RNAPII upon DNA damage and it is necessary for RNAPII transcription at DSBs in mammalian cells.
(A) Co-immunoprecipitation of RNAPII and its phosphorylated forms with the MRN complex upon IR. HEK293T cells were irradiated (+IR) or not (-IR) and samples were collected 10 minutes post IR, followed by immunoprecipitation of the individual components of the MRN complex. Whole cell extract (WCE) and immunoprecipitated samples were analysed by immunoblotting. Mouse (IgG(M)) or rabbit (IgG(R)) immunoglobulins were used as control. This experiment was repeated twice independently with similar results. (B) DSB-induced transcription is dependent on MRN. The bar plot shows the mean relative enrichment of the indicated RNA sets by strand-specific RT-qPCR in cut NIH2/4 cells knocked-down for the three components of the MRN complex (siMRN) or in siLuciferase (siLuc) transfected cells as control. Uncut sample of each RNA set was used as reference. Error bars indicate SEM (for Lac-from n=4, for Tet-from n=3 independent experiments). (C) DSB-induced transcription is reduced by the MRN inhibitor Mirin. The bar plot shows the mean relative enrichment of the indicated RNA sets by strand specific RT-qPCR in NIH2/4 cell treated with Mirin or DMSO as control 2h before cut induction. Uncut sample of each RNA set was used as reference. Error bars indicate SEM (for Lac-from n=3, for Tet-from n=4 independent experiments). (D) Relative intensity of the indicated smFISH probes at the uncut or cut locus, in DMSO or Mirin-treated samples. Dots represent individual cells. The black bar represents the mean ± SEM (data are shown as pool of n=3 independent experiments). (E) Accumulation of active RNAPII at the damaged locus by ChIP in cut NIH2/4 cells is reduced by Mirin. The bar plot shows the percentage of enrichment relative to the input of total RNAPII, RNAPII pSer5 and RNAPII pSer2 associated with genomic DNA as detected by primers matching Lac sequences flanking the I-SceI-induced DSB. Data are shown as one representative of 2 independent experiments. P values were calculated using two-tailed t-test. ****P<0.0001. Statistical source data are provided in Supplementary Table 4. Unprocessed original blots are shown in Supplementary Figure 9.
Figure 6
Figure 6. RNAPII transcription is necessary for DDR focus formation and DNA repair and 53BP1 interacts with DDRNA and dilncRNA through its Tudor domain.
(A) NIH2/4 cells treated with vehicle or an RNAPII inhibitor (AM in these images) before cut induction. Scale bar 5 μm. (B, C) Quantification of (A) showing the percentage of cells positive for DDR markers co-localizing with LacR in NIH2/4 cells treated with AM, DRB, ACTD or vehicle for 2 h before cut induction. Error bars indicate SEM (n=3 independent experiments, ≥70 cells analysed in total per condition). (D) Representative images of human normal fibroblasts (BJ) treated with vehicle or an RNAPII inhibitor (DRB in these images) before IR. Scale bar 5 μm. Quantification is shown in Supplementary Fig. 6L. (E, F). Percentage of DDR-positive HeLa cells treated with AM, DRB, ACTD or vehicle before IR (cells with >10 foci were considered positive). Error bars indicate SEM (n=3 independent experiments, ≥200 cells analysed in total per condition). (G) Bar plot shows the percentage of γH2AX-positive BJ cells pre-treated with DMSO or DRB for 2 h, irradiated (2 Gy) and fixed at the indicated time points (irradiated cells with >30 foci were considered positive). Error bars indicate SEM (n=3 independent experiments, ≥200 cells analysed in total per condition). (H) Representative images of neutral comet assay at 5 h post irradiation. Scale bar 5 μm. Quantification is shown in I. (I) Scatter plot shows tail moment analysis of neutral comet assay of HeLa cells pre-treated with DMSO or DRB for 2 h, irradiated (5 Gy) and collected at the indicated time points. Dots represent individual cells. Black bars indicate mean. Error bars indicate SEM (data are shown as pool of n=3 independent experiments, ≥100 tails analysed in total per condition). (J, K) Endogenous 53BP1 was immunoprecipitated in cut NIH2/4 cells and RNA bound to it was analysed by small RNA-specific RT-qPCR or strand-specific RT-qPCR. Results are shown as mean fraction of input. Error bars indicate SEM (n=3 independent experiments). (L, M) Constructs expressing GFP, GFP-53BP1 or GFP-53BP1 lacking the Tudor domain (GFP-53BP1ΔTUD) were transfected in NIH2/4 cells expressing I-SceI. Immunoprecipitation with anti-GFP antibody and RNA analysis was performed as in L. Results are shown as mean fraction of input. Error bars indicate SEM (n=3 independent experiments). (B-G) P values were calculated using chi-squared test. (I-M) P values were calculated using two-tailed t-test. *P<0.05, ***P<0.001, ****P<0.0001, ns not significant. Images in A, D and H are representative of 3 independent experiments. Statistical source data are provided in Supplementary Table 4.
Figure 7
Figure 7. ASOs preventing DDRNA:dilncRNA interactions affects 53BP1 focus formation.
(A) Schematic representation of ASOs (red) preventing the interaction between dilncRNAs (light blue) and DDRNAs (dark blue), originating from Lac or Tet sequences flanking the I-SceI site in NIH2/4 cells. For graphic clarity dilncRNA-from, but not the dilncRNA-to and the corresponding complementary DDRNA, are shown. (B) NIH2/4 expressing I-SceI were transfected with control ASO (CTL) or specific ASOs matching Lac sequences (ASOs C, D) and subsequently with biotinylated DDRNA (btn-L1) or a biotinylated miRNA (btn-let-7a). After cut induction, RNA pulled-down was performed. Bar plot shows the fold change of dilncRNA (Lac-from) levels, assessed by strand-specific RT-qPCR as relative to input, with respect to control levels (CTL ASO + btn-L1). Values are expressed as mean. Error bars indicate SEM (n=3 independent experiments). (C) Representative images of cut NIH2/4 cells transfected with control or specific ASOs and probed for γH2AX and 53BP1. White circles mark LacR spot. Scale bar 5 μm. (D) Bar plot is the quantification of (C) and shows the percentage of cut NIH2/4 cells positive for LacR co-localization with γH2AX and 53BP1 in the presence of different sets of ASOs. ASO with a sequence unrelated to the locus (CTL) or pre-annealed (INACTIVE) ASOs were used as control. Error bars indicate SEM (n=3 independent experiments, ≥100 cells analysed in total per condition). (B) P values were calculated using two-tailed t-test. (D) P values were calculated using chi-squared test. **P<0.01, ***P<0.001, ****P<0.0001, ns not significant. Statistical source data are provided in Supplementary Table 4.
Figure 8
Figure 8. Site-specific inhibition of 53BP1 focus formation and DNA repair by ASOs.
(A) Representative images of cut NIH3T3duo cells transfected with control or Tet-specific ASOs and probed for γH2AX and 53BP1. Red circles mark TetR dots, green circles mark LacR dots. Scale bar 5 μm. (B) Bar plots are the quantification of (A) and show the percentage of TetR or LacR co-localization with γH2AX and 53BP1 in the presence of control (CTL) or Tet-specific ASOs. Error bars indicate SEM (n=3 independent experiments, ≥150 Tet loci and ≥70 Lac loci analysed in total per condition). (C) Schematic representation of the sets of ASOs (red) preventing the interaction between dilncRNAs (light blue) and DDRNAs (dark blue) originating from the DSB in the DAB1 locus in HeLa cells cut by I-PpoI, and primers used for RT–qPCR (black). For graphic clarity dilncRNA-from, but not the dilncRNA-to and the corresponding complementary DDRNA, are shown. (D) HeLa cells expressing inducible I-PpoI were transfected with control (CTL ASO) or specific ASOs targeting RNA molecules originated from the DSB within the endogenous DAB1 locus. Bar plot shows the mean fold change normalized to uncut CTL ASO of enrichment relative to input of 53BP1 at the DAB1 locus at 50, 1000 bp from DSB. Error bars indicate SEM (n=3 independent experiments). (E) NIH2/4 cells expressing inducible I-SceI were transfected with specific ASOs or CTL ASO. I-SceI ON: 3 h after induction, I-SceI OFF: 24 h after removal of inducing agent. Bar plot shows the percentage of cells positive for γH2AX-TetR co-localization. Error bars indicate SEM (n=4 independent experiments, ≥80 cells analysed in total per condition). (F) MRN recruits RNAPII at the DSB triggering the bidirectional synthesis of dilncRNA-from (blue) and, less abundantly, of dilncRNA-to (light blue). DROSHA and DICER process the long double-stranded RNA, likely the outcome of paired or folded dilncRNAs, generating DDRNAs which pair with nascent unprocessed single-stranded dilncRNAs; together they bind to 53BP1 and fuel DDR focus formation. Interfering with dilncRNA:DDRNA interactions through ASOs allows site-specific inhibition of DDR. (B, E) P values were calculated using chi-squared test. (D) P values were calculated using two-tailed t-test. *P<0.05, ***P<0.001, ****P<0.0001. Statistical source data are provided in Supplementary Table 4.

Similar articles

See all similar articles

Cited by 45 articles

See all "Cited by" articles

References

    1. Polo SE, Jackson SP. Dynamics of DNA damage response proteins at DNA breaks: a focus on protein modifications. Genes Dev. 2011;25:409–433. - PMC - PubMed
    1. Sharma V, Misteli T. Non-coding RNAs in DNA damage and repair. FEBS Lett. 2013;587:1832–1839. - PMC - PubMed
    1. Derr LK, Strathern JN. A role for reverse transcripts in gene conversion. Nature. 1993;361:170–173. - PubMed
    1. Storici F, Bebenek K, Kunkel TA, Gordenin DA, Resnick MA. RNA-templated DNA repair. Nature. 2007;447:338–341. - PMC - PubMed
    1. Keskin H, et al. Transcript-RNA-templated DNA recombination and repair. Nature. 2014;515:436–439. - PMC - PubMed

MeSH terms

Feedback