Article RAD51-dependent recruitment of TERRA IncRN A to telomeres through R-loops https://doi.org/10.1038/s41586-020-2815-6 Received: 13 October 2019 Accepted: 16 July 2020 Published online: 14 October 2020 "*>j Check for updates Marianna Feretzaki1, Michaela Pospisilova2, Rita Valador Fernandes1, Thomas Lunardi1, Lumir Krejci23H & Joachim Lingner,H Telomeres-repeated, noncoding nucleotide motifs and associated proteinsthat are found at the ends of eukaryotic chromosomes-mediate genome stability and determine cellular lifespan1. Telomeric-repeat-containing RNA (TERRA) is a class of long noncoding RNAs (IncRNAs) that are transcribed from chromosome ends2,3; these RNAs in turn regulate telomeric chromatin structure and telomere maintenance through the telomere-extending enzyme telomerase4 6 and homology-directed DNA repair7,8. The mechanisms by which TERRA is recruited to chromosome ends remain poorly defined. Here we develop a reporter system with which to dissect the underlying mechanisms, and show that the UUAGGG repeats of TERRA are both necessary and sufficient to target TERRA to chromosome ends. TERRA preferentially associates with short telomeres through the formation of telomeric DNA- RNA hybrid (R-loop) structures that can form in trans. Telomere association and R-loop formation trigger telomere fragility and are promoted by the recombinase RAD51 and its interacting partner BRCA2, but counteracted by the RNA-surveillance factors RNaseHl and TRF1. RAD51 physically interacts with TERRA and catalyses R-loop formation with TERRA in vitro, suggesting a direct involvement of this DNA recombinase in the recruitment of TERRA by strand invasion. Together, our findings reveal a RAD51-dependent pathway that governs TERRA-mediated R-loop formation after transcription, providing a mechanism for the recruitment of IncRNAs to new loci in trans. TERRA is transcribed from numerous chromosome ends, and comprises both subtelomeric sequences and telomeric repeats. More than 50% of TERRA is associated with chromatin9. To investigate how TERRA is recruited to or retained at telomeres, we generated a plasmid encoding 24 copies of the stem-loop of phage PP7 (ref.10) under the control of the tetracycline-inducible (TET) promoter,followed by 90TTAGGG repeats (Fig. la). To generate full-length TERRA transcripts, we also cloned the human chromosome Xq and 15q subtelomeric regions containing the TERRA start sites between the PP7 stem-loops and the TTAGGG repeats. The constructs were then transiently transfected into HeLa clones that were constitutively expressing the PP7 coat protein fused to GFP (PCP GFP) and a nuclear-localization signal. PCP GFP exhibited a diffuse signal in the nucleus but formed nuclear foci upon expression of the PP7 stem-loops, which are bound by PCP and can gather up to 48 PCP GFP molecules per RNA. These foci did not co-localize with telomeres (Fig. lb). The fusion of the subtelomeric region of 15q or Xq TERRA to the stem-loops did not promote substantial trafficking of the PP7 foci to telomeres. However, when the telomeric TTAGGG repeatswerefused downstream of PP7,co-localization with telomeres occurred, as analysed by conventional and confocal imaging (Fig. lb), indicating that the 5'-UUAGGG-3' repeats of TERRA drive telomere association. The full-length PP7-tagged 15q and Xq chimaeric TERRA also showed marked co-localization with telomeres (Fig. lb and Extended Data Fig. la). Therefore, chimaeric TERRAs that originated from a plasmid were directed to telomeres in trans. To eliminate possible confounding effects due to the high plasmid copy number or increased levels of transgenic TERRA, we used CRISPR Cas9 technology to integrate the chimaeric TERRA constructs into the genome at the adeno-associated-virus integration site 1 (AAVS1) on chromosome 19, which represents a safe harbour for transgene expression11 (Extended Data Fig. lb). Following isolation of clones, we confirmed monoallelic site-specific integration of the full constructs by polymerase chain reaction (PCR) and sequencing. These TERRA expression levels were lower than the levels of expression from plas-mids, giving one to three foci-indicative of displacement from the transcription site. But, similar to the results obtained upon transient transfection, the PP7 loops formed nuclear foci, and only when fused to 5'-UUAGGG-3' repeats did the chimaeric RNAs co-localize with telomeres (Extended Data Fig. lc, d). Shorter telomeres recruit more TERRA In Saccharomyces cerevisiae and Schlzosaccharomycespom.be, short telomeres recruit more TERRA, possibly to facilitate telomere maintenance through recombination or telomerase recruitment5,6,8. To explore the putative roles of telomere length in TERRA recruitment 'Swiss Institute for Experimental Cancer Research (ISREC), School of Life Sciences, Ecole Polytechnique Federate de Lausanne (EPFL), Lausanne, Switzerland, department of Biology and National Centre for Biomolecular Research, Masaryk University, Brno, Czech Republic, international Clinical Research Center, St Anne's University Hospital, Brno, Czech Republic, ^e-mail: lkrejci@chemi.muni.cz; Joachim.lingner@epfl.ch Nature | Vol 587 I 12 November 2020 | 303 Article Long telomeres (10 kb) Short telomeres (3 kb) 0 < 1 01 100' ^ 90' 2 80' 70' 60' 50 40' 30 20' 10' 0 Short telomeres (3 kb) 4f Fig. 11 Transgenic TERRA associates with telomeres in a manner that depends on RAD51. a, Chimaeric TERRA construct (top) and assay for TERRA localization (bottom). Using lipofectamine, plasmids were transfected into HeLa cells constitutively expressing PCP-GFP protein. After 24 h, transcription of TERRA was induced with doxycycline. Chimaeric TERRA, which is recognized and bound by PCP-GFP, was analysed 24 h after induction. CM V, cytomegalovirus. b, Fluorescence in situ hybridization (FISH) analysis of telomeric DNA (red) and immunofluorescence of GFP (green) were used to assess the co-localization of transiently expressed PP7 constructs with telomeres. Confocal images are shown. White arrows indicate co-localization of PP7 foci with telomeric signals. c, To identify proteins involved in the localization of TERRA at telomeres, we used siRNAs to target factors implicated in TERRA and telomere biology. HeLa clones with long (10-kilobase average) and short (3 kb) telomeres were transfected first with siRNApools and then with chimaeric TERRA. The percentage of co-localizing TERRA foci was assessed by telomeric FISH combined with GFP immunofluorescence, n = 2 biologically independent experiments; more than 40 nuclei were analysed per condition; data are means ± s.d. One-way analysis of variance (ANOVA) with Dunnett's multiple comparisons test was used, comparing all conditions with control siRNA (siControl): *P< 0.05; **P< 0.01; *"P< 0.001; **"P< 0.0001. d, The enzymatic activity of RAD51 is required to recruit TERRA. Endogenous RAD51 was depleted with siRN A, and wild-type RAD51 or the RAD51II3A mutant was expressed from plasmid containing complementary DNA. The co-localization of TERRA with telomeres was assessed as in c. n = 3 biologically independent experiments; more than 80 nuclei were analysed per condition; data are means ± s.d. A two-tailed unpaired f-test was used to calculate P-values: *P< 0.05; "P< 0.01; *"P< 0.001; **"P< 0.0001. in human cells, we isolated individual HeLa clones that constitutively expressed the PCP-GFP and measured telomere lengths by telomere restriction fragment length (TRF) analysis (Extended Data Fig. 2). Cells carrying short telomeres recruited TERRA much more efficiently than cells with long telomeres, as seen upon transient or stable expression of TERRA (Extended Data Fig. 2). Therefore, short telomeres must be 304 I Nature | Vol 587 I 12 November 2020 more accessible to recruitment or retention of TERRA; alternatively, long telomeres might contain active systems that expel TERRA. In both experiments theoverall expression levelsofchimaeric TERRA varied in individual clones, but this did not correlate with telomere length and telomere recruitment of TERRA (Extended Data Fig. 2b, d). Recombination factors enable TERRA recruitment To identify proteins involved in the localization of TERRA at telomeres, we performed screens using short interfering RNAs (siRNAs) to target selected factors implicated in TERRA and telomere biology. Using cell lines with longor short telomeres, we transfected first siRNA poolsand then the 15q TERRA construct. After inducing TERRA expression with doxycycline, we analysed the cells (Fig. lc) and evaluated the level of depletion of each factor by using reverse transcription with quantitative PCR(RT-qPCR) or western blotting (Extended Data Fig. 3a, b). We found that the number of transgenic TERRA foci was not affected by individual depletions, and we observed no striking effects on levels of selected endogenous TERRA molecules (Extended Data Fig. 3c, d). Among tested factors, depletion of telomeric repeat factor 1 (TRF1) significantly increased TERRA co-localization at short and long telomeres, while removal of TRF2 led to a milder increase in recruitment (Fig. lc). Depletion of the nonsense-mediated decay (NMD) factors also stimulated co-localization of TERRA at long telomeres, supporting their crucial role in displacing TERRA from chromosome ends2. Similarly, removal of RNaseHl resulted in a substantial accumulation of TERRA at chromosome ends in cells with long telomeres (Fig. lc). This result indicated that TERRA recruitment to, or retention at, long telomeres involves the formation of DNA-RNA hybrids. In cells with short telomeres, depletion of RNaseHl only marginally increased the co-localization of TERRA with telomeres. The roles of NMD factors in cells with short telomeres could not be analysed, as their depletion caused cell death. Notably, depletion of RAD51-which facilitates strand invasion of DNA molecules during homology-directed repair (HDR)-led to a substantial decrease in TERRA recruitment to both long and short telomeres (Fig. lc, d). The involvement of RAD51 in TERRA recruitment prompted us to interrogate the role of the BRCA2 protein in TERRA trafficking. BRCA2 loads RAD51 and promotes the displacement of replication protein A (RPA), allowing the formation of stable filaments of RAD51 on single-stranded DNA (ssDNA) that are capable of homology search to facilitate HDR during double-strand-breakrepair;BRCA2 also protects stalled replication forks1213. We found that depletion of BRCA2 led to a marginal decrease in TERRA recruitment at long, but more prominent reduction at short, telomeres (Extended Data Fig. 4a). However, removal of BRCA2 also diminished total RAD51 protein levels in both cell lines (Extended Data Fig. 4a). Finally, we tested whether RAD51 enzymatic activity is required for TERRA recruitment, taking advantage of a mutation (RAD51II3A) that allows the protein to retain its DNA-binding but not its strand-invasion activity14. In siRAD51-treated cells, the expression of wild-type RAD51 from complementary DNA largely rescued TERRA co-localization with telomeres, but expression of RAD51II3A did not (Fig. Id and Extended Data Fig. 4b). Therefore, the enzymatic activity of RAD51 is required for TERRA to associate with telomeres. Overall, these data suggest that the HDR machinery promotes the recruitment of TERRA to telomeres. TERRA forms R-loops causing telomere fragility As the association of TERRA with long telomeres was increased upon depletion of RNaseHl, we hypothesized that the transgenic TERRA may form R-loops with telomeres in trans. To explore this possibility, we applied the DNA-RNA immunoprecipitation protocol (DRIP), in which the specificity of the S9.6 monoclonal antibody for eight to nine base pairs of DNA-RNA hybrids is exploited15. Precipitated nucleic acids were probed for telomeric repeats, and, as a control for specificity, isolated nucleicacids were treated in vitro with RNaseHl before immunoprecipitation. Abolishment of the signal upon pretreatment with RNaseHl confirmed the specificity of the assay for telomeric R-loops (Fig. 2a). As expected, we detected R-loops at telomeres in wild-type cells (Fig. 2a). Depletion of RNaseHl led to an increase in the number of R-loops, while its overexpression to a decrease in R-loops at both long and short telomeres (Fig. 2a and Extended Data Fig. 5a). Overexpression of chimaeric TERRA further increased the frequency of R-loops in cells with both long and short telomeres (Fig. 2a); this frequency again increased upon depletion of RNaseHl and decreased upon its overexpression (Fig. 2a and Extended Data Fig. 5a), indicating that the transgenic TERRA formed DNA-RNA hybrids. The DRIP assay (Fig. 2a) could not distinguish to what extent the DRIP signalsfor transgenic TERRA were derived from R-loopsforming within the transgenic plasmid during transcription, or R-loops forming after transcription in trans with telomeres. To measure R-loops specifically at telomeres, we used the DRIP samples derived from HeLa cells with long telomeres (Fig. 2a), and determined the presence of four specific chromosome ends by qPCR using subtelomere-specific primers residing in immediate proximity to the terminal 5'-TTAGGG-3'repeats (Fig. 2b). Telomeric R-loops became detectable at the ends of all four chromosomes upon depletion of RNaseHl (Fig. 2b). Specifically, lq, lOq and 13q subtelomeric DNA increased strongly in abundance upon expression of transgenic PP7-15qTERRA, indicating that R-loops had formed at these telomeres with PP7-15qTERRA. The 15q subtelomeric signal was enhanced even more extensively, presumably because of R-loops forming with plasmid DNA containing the 15q sequence. As with wild-type TERRA, overexpression of RNaseHl almost completely abolished the signals, whereas RNaseHl depletion increased R-loop abundance. We sequenced the qPCR products (Extended Data Fig. 5b), verifying the identity of products for each chromosomeend. Together, these data confirmed that transgenic PP7-15qTERRA associated with telomeres in trans through the formation of R-loop structures. TERRA R-loops have been implicated in interfering with telomere replication71"8; this is manifested in telomere fragility19, characterized by the accumulation of telomeric signals in metaphase chromosomes with a smeary or discontinuous appearance (Extended Data Fig. 6a). We found that transgenic TERRA increased telomere fragility (Fig. 2c and Extended Data Fig. 6b), which was suppressed by depletion of RAD51 or overexpression of RNaseHl but increased by depletion of RNaseHl (Fig. 2d and Extended Data Fig. 6c, d). These results confirm that, after transcription from plasmids, TERRA forms R-loops at telomeres in trans. RAD51 promotes telomeric R-loop formation We next used the DRIP assay to test the role of RAD51 in the formation of hybrids containing endogenous TERRA at telomeres in wild-type cell lines (Fig. 2eand Extended Data Fig. 7a). While depletion of RNaseHl led to the expected mild increase in R-loops, depletion of RAD51 caused a substantial decrease in hybrid accumulation. Even stronger reduction of R-loops was observed in RAD51-depleted cells with short telomeres, which are characterized by higher levels of DNA-RNA hybrids. Therefore, RAD51 promotes the association of endogenous TERRA with telomeres through R-loop formation. We hypothesized either that RAD51 binds TERRA to catalyse its strand invasion into the telomeric repeats (Extended Data Fig. 7b, lower panel), or that TERRA might hybridize to exposed single-stranded telomeric DNA during RAD51-mediated HDR between telomeric DNA molecules, even in the absence of a physical interaction between TERRA with RAD51 (Extended Data Fig. 7b, upper panel). To explore these hypotheses, we performed native RNA immunoprecipitations using anti-RAD51 antibodies in HeLa cells. As a control, we also included antibodies specific for hnRNPAl, as this protein binds TERRA4. Immunoprecipitation of Nature | Vol 587 I 12 November 2020 | 305 Article s<^ s<«" s<^ v<«- \°j> «g* ^ ^ ff Telomere fragility Long telomeres Endogenous TERRA 3 K V*' •* PP7 + 15qTERRA Long telomeres Short telomeres Fig. 21 TERRA forms R-loops in trans, inducing telomere fragility in a RADSl-dependent way. a, Left panels, detection of telomeric hybrids on dot-blots. Telomeric hybrids were isolated from cells with long or short telomeres by DRIP using the S9.6 antibody. Cells expressed either the PP7 stem-loops or the PP7-15qTERRA, and had been transfected with control siRNA (siC), siRNAs against RNaseHl (RNH1), or an RNHl-overexpressing plasmid. In vitro treatment with RNH1 served as a negative control. Right panels, quantification of signals, b, Left, schematic representation showing the expression of PP7-15qTERRA from a plasmid, forming R-loops in trans at telomeres lq, lOq and 13q. Right, telomeric hybrids arising at the indicated chromosome ends were quantified by qPCR using primers that amplify specific subtelomeric DNA found next to the telomeric tracts in cells with long telomeres, c, Telomere fragility induced upon expression of TERRA from a plasmid. The fraction of fragile telomeres was quantified on metaphase chromosomes stained with a telomeric FISH probe (Extended Data Fig. 6). d, Effects of RNH1 and RAD51 on telomere fragility. In cells expressing PP7-15qTERRA from a plasmid, telomere fragility was quantified upon depletion or overexpression of RNH1, or depletion of RAD51. e, Detection of endogenous telomeric R-loops in cells transfected with siC, siRNHI or siRAD51. For all panels, n = 3 biologically independent experiments; data are means ± s.d. Two-tailed unpaired f-tests were used to calculate P-values: *P< 0.05; "P< 0.01; "*P< 0.001; ""P< 0.0001. For c, d, the numbers of metaphases scored are indicated for each condition in Extended Data Fig. 6b, c. endogenous RAD51 specifically retrieved TERRA and not the nuclear Ul small nuclear RNA (snRNA; Fig. 3a and Extended Data Fig. 7c), and similar results were observed in the U20S ALTcell line (Extended Data Fig. 7d). The TERRA signal was sensitive to treatment with RNaseA, showing a specific recovery of RNA, but the RNA signal was insensitive to treatment with DNasel. Together these results indicate that TERRA associates with RAD51 in vivo. We wished to determine whether purified RAD51 can bind TERRA directly by carryingoutelectrophoretic mobility shift assays (EMSAs). We incubated recombinant RAD51 with fluorescently labelled TERRA 306 I Nature | Vol 587 I 12 November 2020 a 4.0-, Long telomeres • • • • 3.5- + RNaseA 3.0- Short telomeres • 2.5- + RNaseA ZJ 2.0- Probe: TERRA £Z 1.5- Long telomeres 1.0- + RNaseA 0.5- Short telomeres 0- + RNaseA Probe: U1 snRNA Long telomeres Short telomeres 2.0 1.6 1.2 0.8 0.4 0 0 1,300 2,600 3,900 5,200 Concentration of RAD51 (nM) 100 80 ■o "o Ü 60 0 Ü Zi 40 In 20 Q < 0 • TERRA • TelDNA 0 1,500 3,000 4,500 Concentration of RAD51 (nM) Fig. 31RAD51 associates with TERRA and catalyses R-loop formation, a, Native RNA-immunoprecipitation assays using anti-RAD51 and anti-hnRNPAl antibodies were performed in extracts from HeLa cell clones with long or short telomeres. Western blotting was used to eval uate the efficiency of immunoprecipitation of RAD51 and hnRNPAl (and the co-immunoprecipitation of BRCA2; Extended Data Fig. 7c, d). Immunoprecipitation-recovered RNA was analysed for TERRAand Ul snRNA. n = 3 biologically independent experiments; data are means ± s.d. b, The affinity of RAD51 for TERRAand telomericDNA (TelDNA) oligonucleotides was analysed by electrophoretic mobility shift assay (EMSA). Fluorescently labelled or telomeric DNA oligonucleotides, and resolved the reactions by agarose-gel electrophoresis (Fig. 3b). RAD51 bound the TERRA oligonucleotide with threefold higher affinity than the corresponding telomeric DNA sequence (TelDNA), as with TERRA oligonucleotide a lower protein concentration was required to obtain shifted oligonucleo-tide-RAD51 complexes (Fig. 3b). The binding of RAD51 to the TERRA oligonucleotide was tighter than to an unrelated RNA (Extended Data Fig. 7e) and also more stable when compared with telomeric DNA, as shown by a stability assay in which preformed TERRA-RAD51 or Tel-DNA-RAD51 complexes were challenged with an excess of unlabelled 49-mer competitor ssDNA (cDNA) (Extended Data Fig. 7f). To test whether RAD51 promotes R-loop formation in vitro, we carried out a strand-invasion assay using recombinant RAD51 and telomeric RNA and DNA (Fig. 3c-e). Wild-type RAD51, unlike the catalytically dead RAD51II3A, catalysed both R-loop and D-loop formation in a concentration-dependent manner (Extended Data Fig. 8a). As expected, canonical R-loops were sensitive to treatment with RNaseHl and were supershifted with the R-loop recognizing S9.6 antibody when combined with anti-mouse IgG (Extended Data Fig. 8b, c). Together, these data support a direct role of the RAD51 protein in localizing TERRA to initiate strand invasion and promote the formation ofDNA-RNA hybrids at telomeres. Discussion Our data reveal the mechanism by which TERRA is recruited to chromosome ends through RNA strand invasion. The recruitment of 0 1,300 2,600 3,900 5,200 Concentration of RAD51 (nM) (pink star) RNA or DNAsubstrates (20 nM) were incubated with increasing concentrations of RAD51 protein (0 nM, 550 nM, 1,100 nM, 2,200 nM and 4,400 nM). Quantification is shown at the bottom, n = 3 independent experiments; data are means ±s.d. c, Assay for the formation of R-loops and D-loops. d, The formation of R-loops depends on the concentration of RAD51: detection on native gel (left) and quantification (right), n = 3 independent experiments; data are means ±s.d. e, Detection of D-loops on native gel (left) and quantification (right), n = 3 independent experiments; data are means ± s.d. TERRA and formation of R-loops depend on the recombinase RAD51 and the 5'-UUAGGG-3' repeats of TERRA. Although the subtelomeric TERRA sequences were not required for telomere recruitment in our assay, we do not exclude the possibility that they may contribute by targeting endogenous TERRA molecules to their native chromosome ends. The observed base pairing between TERRA and telomeric DNA provides a mechanism for how this long noncoding RNA can encounter its major site of action at chromosome ends. Therefore, the recruitment of TERRA seems to occur through a process that strongly resembles the strand-invasion and homology-search mechanism exploited in all living organisms during DNA repair by HDR, and which is also characteristic of telomere stabilization by the 'alternative lengthening of telomeres' (ALT) mechanism in cancer cells. A direct involvement of RAD51 in the strand-invasion reaction is supported by our finding that RAD51 strongly binds TERRA and catalyses strand invasion into telomeric sequences in vitro. However, it is likely that there are important differences between DNA- and RNA-mediated strand invasion with the regard to the mechanism and the requirement for accessory factors, which must be explored in the future. The increased formation of R-loops and association of RAD51 with short telomeres has previously been shown in yeast8. Our findings in human cells are consistent with these observations. Furthermore, we reveal that TERRA can form R-loops post-transcriptionally at telomeres in trans. R-loops have generally been assumed to form only during transcription, through the unsuccessful removal of native RNA from its DNA template20. However, previous studies in yeast suggested that Nature | Vol 587 I 12 November 2020 | 307 Article R-loops may also form at loci that are distinct from the siteof synthesis of the RNA21. Our work here suggests that R-loop formation in trans plays a major part in telomere homeostasis. Through its mechanism of recruitment, TERRA may represent a scaffold that guides and regulates telomerase, HDR proteins and chromatin modifiers specifically at the chromosome ends that may need their attention. It will be interesting to see which other IncRNAs are recruited to their sites of action through similar mechanisms. Online content Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability areavailableat https://doi.org/10.1038/s41586-020-2815-6. 1. Maciejowski, J. & de Lange, T. Telomeres in cancer: tumour suppression and genome instability. Nat. Rev. Mol. Cell Biol. 18,175-186 (2017). 2. Azzalin, C. M., Reichenbach, P., Khoriauli, L, Giulotto, E. & Lingner, J. Telomeric repeat containing RNA and RNA surveillance factors at mammalian chromosome ends. Science 318,798-801 (2007). 3. Schoeftner, S. & Blasco, M. A. Developmentally regulated transcription of mammalian telomeres by DNA-dependent RNA polymerase II. Nat. Cell Biol. 10, 228-236 (2008). 4. Redon, S., Zemp, I. & Lingner, J. A three-state model for the regulation of telomerase by TERRA and hnRNPAI. Nucleic Acids Res. 41, 9117-9128 (2013). 5. Cusanelli, E., Romero, C. A. P. & Chartrand, P. Telomeric noncoding RNA TERRA is induced by telomere shortening to nucleate telomerase molecules at short telomeres. Mol. Cell 51, 780-791 (2013). 6. Moravec, M. et al. TERRA promotes telomerase-mediated telomere elongation in Schizosaccharomyces pombe. EMBO Rep. 17,999-1012 (2016). 7. Arora, R. et al. RNaseHI regulates TERRA-telomeric DNA hybrids and telomere maintenance in ALT tumour cells. Nat. Commun. 5,5220 (2014). 8. Graf, M. et al. Telomere length determines TERRA and R-loop regulation through the cell cycle. Cell 170, 72-85 (2017). 9. Porro, A., Feuerhahn, S., Reichenbach, P. & Lingner, J. Molecular dissection of telomeric repeat-containing RNA biogenesis unveils the presence of distinct and multiple regulatory pathways. Mol. Cell. Biol. 30, 4808-4817(2010). 10. Larson, D. R., Zenklusen, D., Wu, B., Chao, J. A. & Singer, R. H. Real-time observation of transcription initiation and elongation on an endogenous yeast gene. Science 332, 475-478(2011). 11. Hockemeyer, D. et al. Efficient targeting of expressed and silent genes in human ESCs and iPSCs using zinc-finger nucleases. Nat. Biotechnol. 27, 851-857 (2009). 12. Jensen, R. B., Carreira, A. & Kowalczykowski, S. C. Purified human BRCA2 stimulates RAD51-mediated recombination. Nature 467, 678-683 (2010). 13. Thorslund, T. et al. The breast cancer tumor suppressor BRCA2 promotes the specific targeting of RAD51 to single-stranded DNA. Nat. Struct. Mol. Biol. 17,1263-1265 (2010). 14. Mason, J. M., Chan, Y.-L., Weichselbaum, R. W. & Bishop, D. K. Non-enzymatic roles of human RAD51 at stalled replication forks. Nat. Commun. 10, 4410 (2019). 15. Boguslawski, S. J. et al. Characterization of monoclonal antibody to DNA.RNA and its application to immunodetection of hybrids. J. Immunol. Methods 89,123-130 (1986). 16. Sagie, S. et al. Telomeres in ICF syndrome cells are vulnerable to DNA damage due to elevated DNA:RNA hybrids. Nat. Commun. 8,14015 (2017). 17. Pfeiffer, V., Crittin, J., Grolimund, L. & Lingner, J. The THO complex component Thp2 counteracts telomeric R-loops and telomere shortening. EMBO J. 32,2861-2871 (2013). 18. Balk, B. et al. Telomeric RNA-DNA hybrids affect telomere-length dynamics and senescence. Nat. Struct. Mol. Biol. 20,1199-1205 (2013). 19. Sfeir, A. etal. Mammalian telomeres resemble fragile sites and require TRF1 for efficient replication. Cell 138, 90-103 (2009). 20. Crossley, M. P., Bocek, M. & Cimprich, K. A. R-loops as cellular regulators and genomic threats. Mol. Cell 73, 398-411 (2019). 21. Wahba, L., Gore, S. K. & KoshLand, D. The homologous recombination machinery modulates the formation of RNA-DNA hybrids and associated chromosome instability. eLlfe2,e00505(2013). Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. © The Author(s), under exclusive licence to Springer Nature Limited 2020 308 | Nature | Vol 587 I 12 November 2020 Methods No statistical methods were used to predetermine sample size. The experiments were not randomized and investigators were not blinded to allocation during experiments and outcome assessment. Cell culture and transfections The telomerase-positive cell line HeLa (cervical cancer) was from the American Type Culture Collection (ATCC), and the ALT cell line U20S (osteosarcoma) from the European Collection of Authenticated Cell Cultures (ECACC). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 U ml 1 of penicillin/streptomycin (Thermo Fisher Scientific). The cell lines were maintained in a 5% C02 incubator at 37 °C and were routinely checked to ensure that they remained mycoplasma-free. HeLa cells were transfected with lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). To induce TERRA transcription, doxycycline (1 \ig mL1) was added to the medium 24 h after transfection, and the cells were harvested 48 h after transfection. To isolate clones containing integrated TERRA transgenes, puromycin (1 \ig mL1) was added to the medium 24 h following transfection using the pSpCas9(BB)-2A-Puro plasmid. Selection was maintained for 5 days. SiRNAs were purchased as pools from Dharmacon (siGENOMESMART-pool; Extended Data Table 1). Cells were transfected with 20 pmol siR-NAs using calcium phosphate transfection in 6-well plates in DMEM supplemented with 10% FBS. Cells were harvested 72 h after transfection. Wild-type RAD51 or the RAD51II3A mutant was expressed from plasmid containing complementary DNA14. Lentivirus production and cell transduction Plasmids pMD2.G (1 ng) and pCMVR8.74 (3 ug) (gifts from D. Trono, EPFL) were mixed with PP7 eGFP (3 \ig) (a gift from D. Larson, National Institutes of Health (NIH)) or pLenti CMV rtTA3 Hygro (Addgene catalogue number 26730) for transfection of HEK293T cells in Optimem medium (Thermo Fisher Scientific) using lipofectamine 2000. The transfection mix was incubated overnight and the medium replaced the next day. Supernatants were collected on the next two consecutive days upon centrifugation, and cleared through a 0.22-um filter unit (Stericup, Millipore), before the viruses were aliquoted and frozen at -80 °C. HeLa cells were transduced with 1 ml of viral particles of pLenti CMV rtTA3 Hygro. Following infection, the cells were selected with hygromycin (200 \ig mf1) for 5 days, before they were infected again with the PP7-GFP viruses. GFP-positive clones with similar GFP intensity were isolated with a FACSAria Fusion sorter by EPFL's flow cytometry core facility. Cloning of TERRA constructs TERRA constructs were cloned in the pTRE2puro vector of Clontech's TETON system without the corresponding polyA region. The PP7 stem-loops were amplified from Addgene plasmid #61762 (a gift from D. Larson) and introduced into pTRE2puro through infusion cloning. To amplify the TTAGGG sequence, we performed PCR using the Phu-sion Green Hot Start II high-fidelity DNA polymerase (F537S) with no template in a reaction containing: 1* GC Phusion buffer, 0.2 mM dNTPs, 0.4 uM primer 5'-TTAGGGTTAGGGTTAGGGTTAGGGTTAGGG-3', 0.4 uM primer 5'- CCCTAACCCTAACCCTAACCCTAACCCTAA-3', and 2 U of polymerase. The PCR consisted of 30 s at 98 °C followed by 10 cycles at 98 °C for 5 s, 60 °C for 10 s, and 72 °C for 15 s. DNA products of variable size were fractionated on a 1.2% agarose gel, and the desired size was excised, extracted with a QIAquick gel extraction kit, and cloned into the pTRE2puro vector containingthe24copiesofPP7 stem-loops. Sub-telomeric sequences were amplified using a high-fidelity polymerase from phenol-chloroform-extracted genomic DNAfrom HeLa cells,and introduced into thecorresponding vectors through infusion cloning. All constructs were verified by restriction digestion and sequencing. The plasmids were amplified in a homologous-recombination-deficient E. coll strain (Stbl3) at 30 °C. All of the primers are listed in Extended Data Table 1. Generation of integrated TERRA cell lines The integration of TERRA constructs into the AAVS1 locus was performed as described22. The guide RNA for AAVS1 (5'-ACCCCACAG TGGGGCCACTA-3') and its complementary strand were annealed, and cloned into the pSpCas9(BB)-2A-Puro plasmid acquired from Addgene (catalogue number 48139). The donor templateencompassed roughly 800 base pairsof homology armsforAAVSl, which wereamplified from HeLa genomic DNA, and cloned into the plasmidscontainingthecontrol PP7 stem-loop construct and the different TERRA transgenes. HeLa cells were transfected with both the gRNA/Cas9 and donor template vectors. Individual clones were screened by PCR for the presence of the transgene in the AAVS1 locus. Immunofluorescence and FISH Cells were grown on glass coverslips by followingthe culture conditions above. The coverslips were washed twice with lx phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde for 10 min at room temperature, and washed twice with lx PBS. The cells were then permea-bilized in lx detergent solution for 5 min (0.1% Triton X-100,0.02% SDS in lx PBS), followed by pre-blocking with 2% bovine serum albumin (BSA) in lx PBS for 10 min. Next the cells were blocked with 10% normal goat serum in 2%BSA/lx PBS for 30 min at room temperature. Coverslips were incubated with primary and then secondary antibodies in blocking solution for 90 min each time at room temperature, and fixed with 4% paraformaldehyde for 5 min at room temperature; the samples were then dehydrated with increasing concentrations of ethanol. For FISH staining, the cells were incubated with hybridization solution containing 10 mM Tris-HCI pH 7.4,70% formamide, 0.5% blocking reagent and 1/1,000 Cy3 probe, denatured at 80 °C for 3 min, and hybridized for 3 h at room temperature. The cells were washed with wash 1 (10 mM Tris-HCI pH 7.4,70% formamide) and wash 2 (0.1 M Tris-HCI pH 7.4,0.15 M NaCI, 0.08% Tween-20) twice, stained with 4',6-diamidino-2-phenylindole (DAPI),and dehydrated with ethanol. Images were acquired on an Upright Zeiss Axioplan or on a Zeiss LSM700 equipped with an AxioCam MRm B/W. Images were processed and analysed with ImageJ and Adobe Photoshop. All statistical analysis was performed using GraphPad Prism. All figures were created in Adobe Illustrator. Telomeric FISH on metaphase spreads Cells were treated with 0.05 \ig mL1 demecolcine for 2 h, collected and incubated in hypotonic solution (0.075 M KCI) at 37 °Cfor 8 min. Swollen cells werecollected and fixed in ice-cold methanokacetic acid (3:1) overnight at 4 °C. The next day, 100 \i\ of cell suspension was dropped onto slides, incubated at 70 °C for 1 min and air-dried overnight at room temperature. FISH staining was performed as above. Telomere length analysis HeLa genomic DNA was isolated with the Wizard genomic DNA purification kit according to the manufacturer's instructions (Promega). Then, 6 \ig of genomic DNA was digested with 30 U of Hinfl and Rsa\ overnight at 37 °C. The digested DNA was mixed with 6x MassRuler DNA-loading dye (Thermo Fisher Scientific), loaded on a 0.8% agarose gel in lx Tris-borate-EDTA (TBE) buffer, and fractionated by gel electrophoresis at 2 V cnT1 for 20 h. The gels were dried for two hours at 50 °C in vacuum, treated with denaturation buffer (0.5 M NaOH, 1.5 M NaCI) and neutralization buffer (0.5 M Tris-HCI pH 7.5,1.5 M NaCI), and then pre-hybridized with Church buffer for 1 h at 50 °C. The gels were hybridized overnight at 50 °C with a randomly labelled TeloC probe as described23. The gels were washed for 1 h at 50 °C with 4x saline sodium Article citrate (SSC),4x SSC 0.1% SDS, and 2x SSC 0.1%SDS, exposed to a phos-phorimager screen and analysed on a Typhoon phosphorimager (GE). RT-qPCR For RT-qPCR of TERRA, 3 x 106 cells were harvested following trans-fection of the chimaeric TERRA constructs. RNA was isolated with a NucleoSpin RNA kit (Macherey-Nagel). RT-qPCR for 15qTERRA was carried out using our previously described protocol24. To assess messenger RNA levelsfollowing siRNA transfections, we used ThermoFisher Scientific's Superscript III reverse transcriptase and Power SYBR Green PCR master mix on an Applied Biosystems 7900HT fast real-time system according to the manufacturer's instructions. Western blotting Antibodies are listed in Extended Data Table 2. Protein samples were mixed with 2x Laemmli buffer, boiled for 5 min at 95 °C, run on 4-15% SDS-PAGE precast gels (Mini-PROTEAN TGX Gels, BioRad), transferred to nitrocellulose blotting membranes (Amersham), blocked with blocking solution (3% BSA in lx PBS, 0.1% Tween-20), and incubated overnight at 4 °C with the corresponding primary antibody. Membranes were washed three times for 5 min each with lx PBS plus Tween (PBST) buffer, and then incubated for 1 h at room temperature with horseradish-peroxidase-conjugated secondary antibodies (Promega) in blocking solution. Membranes were washed again three times for 5 min with lx PBST before revealing them with a chemiluminescence detection kit (Western bright electrochemiluminescence, Advansta) and analysing them on a Vilber Fusion FX imaging system. DNA-RNA immunoprecipitation Cells at roughly 40% confluence in 6-well plates were transfected first with siRNAs and then, the next day, with plasmids. The cells were harvested on ice 48 h later, counted on a CASY cell counter and washed with lx PBS; samples were taken for DRIP and western blot analysis. For DRIP, 107 cells were dissolved in 175 pi of ice-cold RLN buffer (50 mM Tris-HCI pH 8.0,140 mM NaCI, 1.5 mM MgCI2,0.5% NP-40,1 mM dithi-othreitol (DTT), and 100 U ml1 RNasIN PLUS), incubated on ice for 5 min, and centrifuged (300g, 2 min, 4 °C). The nuclei were lysed in 500 pi RA1 buffer (NucleoSpin RNA, Macherey-Nagel) containing5 pi of P-mercaptoethanol, and homogenized by passing them through a 20G x 1V2 syringe (0.9 mm x 40 mm). The nucleic-acid-containing extracts were mixed with 250 pi H20 and 750 pi phenol:chloroform:isoamyla Icohol (25:24:1) in a Phase Lock Gel heavy (5PRIME) and centrifuged (13,000g, 5 min, room temperature). The upper aqueous phase was mixed with 750 pi of ice-cold isopropanol, with addition of NaCI to 50 mM, then incubated on ice for 30 min; precipitated nucleic acids were collected by centrifugation at 10,000gfor 30 min at 4 °C. The pellets were washed twice with 70% ice-cold ethanol, air-dried, dissolved in 130 pi of H20, and sonicated on a Covaris system (10% duty factor, 200 cycles per burst, for 180 s, with an AFA intensif ier) to obtain fragments of100-500 bp. Next, 90 pgof sonicated nucleic acids were mixed with RNaseHl or H20 in RNaseHl buffer (20 mM HEPES-KOH pH 7.5,50 mM NaCI, 10 mM MgCI2,1 mM DTT) and incubated at 37 °C for 90 min. The samples were diluted ten times in DIP-1 buffer (10 mM HEPES-KOH pH 7.5,275 mM NaCI, 0.1% Na-deoxycholate, 0.1% SDS, 1% Triton X-100) and pre-cleared with 80 pi of sepharose protein G beads for 1 h, on a rotating wheel, at 4 °C. One per cent of the nucleic acids were kept as input. Half of the samples (roughly 45 pgof nucleic acids) were incubated with 3 pg of S9.6 antibody or mouse IgG and 40 pi of sepharose protein G beads on a rotating wheel at 4 °C overnight. The next day the samples were washed for 5 min on a rotating wheel at 4 °C with DIP-2 (50 mM HEPES-KOH pH 7.5,140 mM NaCI, 1 mM EDTA pH 8.0,1% Triton X-100,0.1% Na-deoxycholate), DIP-3 (50 mM HEPES-KOH pH 7.5,500 mM NaCI, 1 mM EDTA pH 8.0,1% Triton-XlOO, 0.1% Na-deoxycholate), DIP 4 (10 mM Tris-HCI pH 8.0,1 mM EDTA pH 8.0,250 mM LiCI, 1% NP-40,1% Na-deoxycholate), and TE buffer (10 mM Tris-HCI pH 8.0,1 mM EDTA pH 8.0). The samples were digested at 65 °C overnight with 10 pg mL1 RNase (DNase-free (Roche)) in a buffer containing20 mM Tris-HCI pH 8.0,1% SDS, 0.1 M NaHC03,0.5 mM EDTA pH 8.0. The DNA was isolated using the Qiagen PCR clean-up kit. The DNA was pipetted onto a positively charged nylon membrane (Amersham Hybond N+) and telomeric DNA was detected with a telomeric probe as described25. RNA immunoprecipitation Cells were grown to roughly 70%confluence in 15-cm dishes, harvested on ice, counted on a CASY cell counter, washed with lx PBS, and lysed in RNA-immunoprecipitation RLN buffer (50 mM Tris-HCI pH 8.0,140 mM NaCI, 1.5 mM MgCI2,0.5% NP-40,1 mM DTT, 400 U mr1 RNasIN PLUS, and protease inhibitors (Complete, Roche)). The lysates were pre-cleared with Dynabeads plus protein G for 1 h on a rotating wheel at 4 °C, and incubated overnight with 6 pg of the corresponding antibody on a rotating wheel at 4 °C. Next 35 pi of Dynabeads protein G (ThermoFisher Scientific) that had been preblocked with transfer RNA were added to each mixture and incubated for 2 h on a rotating wheel at 4 °C; this was followed by five washes with RLN buffer supplemented with 6mM EDTA pH 8.0. The RNA was eluted in 1% SDS, 5 mM EDTA pH 8.0, and 5 mM P-mercaptoethanol at 42 °C for 30 min, followed by 30 min at 65 °C. The RNA was purified using the RNA clean-up protocol of the NucleoSpin RNA kit (Macherey-Nagel). The RNA was denatured at 65 °C for 3 min and pipetted on a positively charged nylon membrane (Amersham Hybond N+). TERRA and Ul snRNA were detected with corresponding probes as described26. EMSA RAD51 protein was purified27 and diluted in dilution buffer (25 mM Tris-HCI (pH 7.5), 10% (v/v) glycerol, 0.5 mM EDTA, 50 mM KCI, 1 mM DTT and 0.01% NP40). Increasing concentrations of RAD51 were incubated with 20 nM Cy3-labelled 41mer TERRA (5'-UUAGGGUUAGGGUUAGGGUUAGGGUUAGGGUUAGGGUUAGG-3'), 41mer TelDNA(5'-TTAGGGTTAGGGTTAGGGTTAGGGTTAGG GTTAGGGTTAGG-3'),41mernon-TelRNA(5'-AGUAUAUAUGAGUAAACUU GGUCUGACAGUUACCAAUGCUU-3') or 40mer non-TelDNA (5'AAATTAACAAGTATAATAAGAA ATAGAAACAAGAAATAGA-3') substrate at 37 °C for 10 min in 10 pi of buffer D (50 mM Tris-HCI (pH 7.5), 50 mM KCI, 1 mM MgCI2,1 mM ATP). Reaction mixtures were then crosslinked with 0.01% glutaraldehyde for 10 min. Products were resolved using 0.8% TBE agarose gels supplemented with 10 mM KCI at 4 °C for 50 min (6.5 V cm"1). Gels were imaged on a FLA-9000 scanner (Fujifilm) and quantified with Multi Gauge version 3.2 (Fujifilm). Stability EMSA Fluorescently labelled 41mer TERRA or 41mer TelDNA substrates (20 nM) were incubated with indicated concentrations of RAD51 at 37 °C for 10 min in 50 mM Tris-HCI (pH 7.5), 50 mM KCI, 1 mM MgCI2 and 1 mM ATP. To challenge assembled RAD51 ssDNA complexes, increasing concentrations of unlabelled 49mer competitor ssDNA (5'-CCTGTTCAAACGCACATATTAAGCATTTC CTGTCATTG GCGGCTAATTC-3') were added and incubated for another 10 min at 37 °C. Products were crosslinked with 0.005% glutaraldehyde for a further 10 min and resolved using a 0.8% TBE agarose gel supplemented with 10 mM KCI at 4 °C for 50 min (6.5 V cm4). Gels were imaged on a FLA-9000 scanner (Fujifilm) and quantified with Multi Gauge version 3.2 (Fujifilm). Assays for R-loop/D-loop formation Fluorescently labelled 41mer TERRA or 41mer TelDNA (50 nM) were pre-incubated for 10 min at 37 °C with increasing concentrations of RAD51 in 50 mM Tris-HCI (pH 7.5), 1 mM MgCI2 supplemented with 1 mM CaCI2 and 1 mM adenylyl-imidodiphosphate (AMP-PNP). The reaction was started by adding 600 ng pCR4-TOPO vector containing the 15q subtelomeric sequence followed by 15 copies of TTAGGG repeats. The mixture was incubated for another 10 min at 37 °C and then stopped by adding SDS (0.1%) and proteinase K (0.1 mg mr1); this was followed by 3-min incubations at 37 °C. Reaction mixtures were separated on 0.8% agarose gel and analysed as described above for EMSA. For digestion of R-loops by RNaseH, the mixtures were incubated for 1 min with phenylmethylsulfonyl fluoride (PMSF;2 mM) and with EGTA(1.6 mM) at 37 °C to inhibit proteinase Kand chelate calcium ions, respectively. The products were digested with RNaseHl (6.8 mil ul_1, Thermo Scientific) or RNaseH2 (6.8 mU New England Biolabs) for 1 min at 37 °C and resolved as above. For antibody-specific supershift of R-loops, the products were formed as described above for RNaseH digestion (but without EGTA) and then incubated with S9.6 antibodies (0.02 pg uT1) for 2 min at 37 °C and/or anti-mouse horseradish-peroxidase-conjugated antibody. RNaseHl digestion of telomeric DNA-RNA hybrids Fluorescently labelled TERRAor non-telomeric RNAs were mixed with their correspondingcomplementary ssDNA in annealing buffer (10 mM Tris pH 8.5,50 mM NaCI, 1 mM EDTA). To form DNA-RNA hybrids, the mixture was heated for 5 min at 95 °C, and then gradually cooled to room temperature. The DNA-RNA hybrid (40 nM) was cleaved by RNaseHl (6.8 mU |xl_1) in the presence or absence of 1 mM CaCI2 for 10 min at 37 °C in 50 mM Tris (pH 7.5), 1 mM MgCI2 and 1 mM AMP-PNP. The reaction was stopped by addingSDS/proteinase K, and then incubated for 10 min at 37 °C. Reaction mixtures were loaded on 10% PAGE gel, separated by electrophoresis (90 V for 60 min at 4 °C), scanned using an Image Reader FLA-9000 scanner and quantified by MultiGauge version 3.2 software. Supershift of DNA-RNA hybrid with S9.6 antibody Fluorescently labelled TERRA and non-telomeric DNA-RNA hybrids (40 nM) were incubated with 0.015 pg ul"1 of S9.6 antibody for 10 min at 37 °C, then crosslinked with 0.01% glutaraldehyde for 10 min at 37 °C and resolved using 0.8% TBE agarose gel at 4 °C for 50 min (6.5 Vcrrf'). Gels were imaged on a FLA-9000 scanner (Fujifilm) and quantified with Multi Gauge version 3.2 (Fujifilm). Reporting summary Further information on research design is available in the Nature Research Reporting Summary linked to this paper. Data availability The data that support the findings of this study are available from the correspondingauthors upon reasonable request. Source data are provided with this paper. 22. Ran, F. A. et al. Genome engineering using the CRISPR-Cas9 system. Nat. Protocois 8, 2281-2308(2013). 23. Grolimund, L. et al. A quantitative telomeric chromatin isolation protocol identifies different telomeric states. Nat. Commun. 4,2848 (2013). 24. Feretzaki, M. & Lingner, J. A practical qPCR approach to detect TERRA, the elusive telomeric repeat-containing RNA. Methods 114, 39-45 (2017). 25. Porro, A. eta I. Functional characterization of the TERRA transcriptome at damaged telomeres. Nat. Commun. 5, 5379 (2014). 26. Porro, A., Feuerhahn, S. & Lingner, J. TERRA-reinforced association of LSD1 with MRE11 promotes processing of uncapped telomeres. Ceii Rep. 6, 765-776 (2014). 27. Spirek, M. et al. Human RAD51 rapidly forms intrinsically dynamic nucleoprotein filaments modulated by nucleotide binding state. Nucieic Acids Res. 46, 3967-3980 (2018). Acknowledgements We thank D. Larson (NIH), D. Bishop (Univ. Chicago), V. Simanis (EPFL), P. Gonczy (EPFL) and D. Trono (EPFL) for providing material. We also thank the members of the BIOP core facility at EPFL, members of the Gonczy laboratory and M. Spirek for technical support and advice and J. Cibulka for recombinant mutant RAD51 protein. M.F. was supported by the European Union's Horizon 2020 research and innovation programme under Marie Sklodowska-Curie grant agreement 702824. J.L.'s laboratory was supported by the Swiss National Science Foundation (SNFS grant 310030J84718), the SNFS-funded National Centres of Competence in Research (NCCR) RNA and disease network (grant 182880), and EPFL. L.K.'s laboratory was supported by the European Structural and Investment Funds, Operational Programme Research, Development and Education 'Preclinical Progression of New Organic Compounds with Targeted Biological Activity' (Preclinprogress) (CZ.02.1.01/0.0/0.0/16_025/00 07381); a Wellcome Trust Collaborative Grant206292/E/17/Z; the Czech Science Foundation (GACR17-17720S); and the National Program of Sustainability II (MEYS CR, project number LO1605). Both J.L. and L.K. were also supported by the European Union's Horizon 2020 research and innovation programme under grant agreement 812829. Author contributions M.F. and J.L. conceived the study. M.F. and R.V.F. executed all cell and molecular biology experiments. M.P. performed all biochemistry experiments and T.L. some EMSA experiments. L.K. conceived the RAD51-based biochemistry experiments and advised on the text. J.L. and M.F. wrote the paper. Competing interests The authors declare no competing interests. Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41586-020-2815-6. Correspondence and requests for materials should be addressed to L.K. or J.L. Peer review information Nature thanks the anonymous reviewers for their contribution to the peer review of this work. Reprints and permissions information is available at http://www.nature.com/reprints. Article Extended Data Fig. 11 Co -localization of transgenic TERRA with telomeres. a, Quantification of co-localization of transgenic RNA foci expressed from plasmids with telomeres, b, Strategy for targeting the AAVS1 locus through CRISPR-C AS9 to integrate chimaeric TERRA constructs into the genome. LH A and RHA, left and right homology arms, c, Quantification of co-localization of Integrated TERRA transgenic RNA foci with telomeres when TERR A was expressed from the AAVS1 locus, d, Confocal images obtained when TERR A was expressed from the AAVS1 locus. For a, c, n = 3 biologically independent experiments; data are meants.d.; two-tailed unpairedf-test: *"P< 0.001;""P< 0.0001. HeLa Clones 1 2 3 4 5 6 7 8 9101112: CRISPR Clones 3kb 1.5 kb I 1 kb 3 4 5 6 10 kb 7kb 5kb 3kb 1 I II J r 1.5 kb I 1 kb III" '• Transient 15q TERRA Expression Integrated 15q TERRA Expression 4.5 t 3-5 7 3 0 1 2 5 1.5 S i £ 0.5 0 S 8 Telomere length (kb) 3 7 2.5 "c 2 o Si - r- t . S 5 0.5 5 6 Telomere length (kb] Integrated TERRA 4 6 8 10 Telomere length (kb) 4 5 6 Telomere length (kb) Long Telomeres Extended Data Fig. 21 TERRA associates preferentially with short telomeres, a, TRF analysis of HeLa clones used for transient expression of TERRA, b, d, TERRAexpression levels measured by RT-qPCR relative to the levels of TERR A in the clone with the shortest telomeres. n = 3 biologically independent experiments; data are means ± s.d. c, TRF analysis of HeLa clones expressing transgenic TERR Afrom the AAVS1 locus, e, HeLa clones of different telomere lengths were transiently transfected with the PP7-15qTERR A construct and co-localization was assessed by immunofluorescence with FISH (bottom). White arrows indicate co-localization of PP7foci with the telomeric signal. The percentage of TERR A foci that co-localized with telomeres is plotted as a function of telomere length (top). Random colocalization events (roughly 3%) were not subtracted, f, Co-localization events as in e but for PP7-15qTERRA expressed from the AAVS locus on chromosome 19. n = 3 biologically independent experiments; data are means ± s.d. Significant differences are indicated. Two-tailed unpaired f-tests were used to calculate P-values: *P< 0.05; "P< 0.01; *"P< 0.001; ""P< 0.0001. Article SÍRNA transfection TERRA ^ transfection Add ^ Doxycycline Long Telomeres (10 kb) ^ Harvest 130 KDa 70 KDa I siCONTROL I siRNA 130 KDa 55 KDa 130 KDa 35 KDa 130 KDa 55 KDa Vinculin 35 KDa TRF1 hnRNPAl SMG1 Long Telomeres Short Telomeres Long Telomeres Vinculin TRF2 Long Short Telomeres Telomeres Long Short Telomeres Telomeres Vinculin RPA2 35 KDa 130 KDa hnRNPAl UPF1 Long Telomeres é? — -1 Vinculin 130 KDa -lo^-r, 35 KDa Vinculin RNH1 Long Short Telomeres Telomeres 130 KDa 35 KDa Vinculin RAD51 Long Short Telomeres Telomeres Long Short Telomeres Telomeres C Long Telomeres (10 kb) Short Telomeres (3 kb) Long Telomeres (10 kb) „ 350n § £ 300- 8 I 25»" ID U Q. " 20°- X > LU « 150- 0) c > ^ 100- ■4-" O TO ^ ID o ? 50- liliiliilllii ii □ SiCONTROL ■ SÍTRF1 ■ SÍTRF2 ■ SÍPOT1 ■ sísmg1 ■ SÍUPF1 ■ SÍUPF2 ■ siRNaseHI ■ SÍRPA2 I SÍRAD51 Short Telomeres (3 kb) 15q-TERRA 13q-TERRA 10q-TERRA 350- ion TRO 300- ss NO 250- O a "in 200- X > UJ ci c 150- > .E 100- u ela "D O 50- a. i n i lil in n, 15q-TERRA 13q-TERRA 10q-TERRA □ SiCONTROL ■ SÍTRF1 ■ SÍTRF2 ■ SÍPOT1 ■ SiRNaseHI ■ SÍRPA2 SÍRAD51 Extended Data Fig. 31 Depletion of factors regulatingTERRA trafficking. a, Timeline of transfections and cell harvesting. mRN A levels were determined by RT-qPCR for long-telomere and short-telomere cell lines, b, For the second replicate of each siRN A screen, the level of depletion was also evaluated on western blots. Vinculin and hnRNPAl were used as loading controls, c, Numbers of TERR A foci per cell are plotted for each depleted factor, n = 2 biologically independent experiments; at least 40 nuclei were analysed per condition, d, Quantification of endogenous TERR Astemming from the ends of chromosomes 15q, 13q and lOq upon depletion of the indicated factors relative to the negative control (siCONTROL). n = 2 biologically independent experiments. a ,\T ,\Y ^ .ST ,3? 180 kDa 130 kDa MM — - — — - — BRCA2 Vinculin RAD51 Long Telomeres Short Telomeres is TS* IE SS Long Telomeres (10 kb) 5 íT Vinculin RAD51 Long Telomeres Extended Data Fig. 41 Depletion of RAD51 and BRCA2, which regulate the association of TERRA with telomeres, a, HeLa clones with long and short telomeres were transfected with siRAD51 and siBRCA2, and then with chimaeric TERRA constructs. Representative immunoblots show RAD51 and BRCA2 depletion, with vinculin as a loading control. Upon depletion, co-localization of TERRA with telomeres was assessed with immunofluorescence-FISH. n = 2 biologically independent experiments; at least 54 nuclei were analysed per Short Telomeres condition; data are means ± s.d. One-way ANOVA with Dunnett's multiple comparisons test was used, comparing all conditions to siCONTROL: *P< 0.05; "P< 0.01; "*P< 0.001. b, Detection of endogenous and transgenic RAD51 on a western blot. Endogenous RAD51 was depleted with siRN A, and wild-type (WT) RAD51 or the RAD51II3A mutant was expressed from plasmids containing complementary DNA. Article RNH1 OE siRNHI TERRA OE 130 kDa + + 35 kDa 130 kDa 35 kDa Vinculin RNH1 Vinculin RNH1 b DRIP-qPCR 13q siC siRNHI RNH1 OE siC siRNHI RNH1 OE WT TERRA OE Extended Data Fig. 51 RNaseHl-regulated formation of telomeric R-loops for the end of chromosome 13q. Left, quantification of RN A-DN A hybrids is in trans, a, RNH1 depletion and overexpression (OE) was assessed by western expressed as a fraction of input. Right, the amplified DNA was run on a gel, then blotting in cell lines with long and short telomeres and upon TERRA isolated and sequenced, overexpression. b, Representative example of one DRIP assay followed by qPCR a siC + PP7-15qTERRA siRNM + PP7-15qTERRA siRAD51 + PP7-15qTERRA 10 um RNH1 OE + PP7-15qTERRA 10 Mm V Plasmid Metaphases scored T°meTn%±SEM)y PP7 PP7+UUAGGG PP7+15q+UUAGGG PP7+Xq+UUAGGG n=59 n=58 n=6C n=58 2.7 ± 0.2 5.4 ± 0.4 4.6 ± 0.2 5.1 ±0.3 _ .... ij . ■ . Telomere Fragility Condition Metaphases scored (mean % ± SEM) siC SIRNH1 siRAD51 RNH1 OE [1=64 n=69 n=64 n=64 5.3 ± 0.4 7.1 ± 0.3 3.4 ± 0.3 2.8 ± 0.2 Extended Data Fig. 61 Transgenic TERRA expressed from plasmids induces telomere fragility, a, Representative examples of metaphase chromosomes stained with FISH to visualize telomeres. DNA is stained with DAPI. Fragile telomeres are indicated by white arrowheads, b, Quantification of telomere fragility, c, Quantification of telomere fragility in cells expressing PP7-15qTERRA in which expression of RNaseHl(RNHl) and RAD51 was manipulated as indicated. For b, c, the numbers of metaphases scored over three biologically independent experiments are indicated for each condition as n. Article in vitro +RNaseH in vitro +RNaseH siC • 250 kDa siRNHI 0. # m 1 • Of * * 130 kDa SÍRAD51 35 kDa Long Telomeres Short Telomeres ™ SN (0.2%) IP (2%) BRCA2 Vinculin 130 kDa RAD51 35 kDa SN (0.2%) IP (2%) Vinculin hnRNPAl RAD51 ^ binding RNA RAD51 ^Q^^W^ loading e? -J TERRA RIP - U20S 250 kDa 130 kDa »r <^ >.&i? RAD51 (I) plasmid .«? (II) PK V. (III) PMSF+EGTA^"^ ^ RNaseH TERRA TelDNA RAD51 PMSF RNaseHI - + + + 4- - + + + RAD51 + PMSF RNaseHI - + + + +■ - + + Control RNaseHI PMSF ■ RNaseH2 RNaseH2 IhRRA RAD51 or TelDNA (I) plasmid OOP (II) PK (III) PMSF (l)S9.6ab (II) secondary ab TERRA RAD51 PMSF secondary ab S9.6ab S9.6+secondary ab RAD51 PMSF secondary ab S9.6ab S9.6+secondary ab R-loop D-loop D-loop Extended Data Fig. 81 See next page for caption. Extended Data Fig. 8 RAD51catalyses the formation of canonical R-loops. a, Left, the RAD51II3A mutant (lines 2-5; 325 nM, 650 nM, 1,300 nM or 2,600 nM) or wild-type RAD51 (lines 6 and 7; 325 nM or 650 nM) was incubated with TERRAoligonucleotide substrate (50 nM), and then plasmid containing the homologous region was added. Right, quantification of R-loops in the presence of wild-type R AD51 or RAD51II3A. n = 2 independent experiments; data are means ± s.d. b, Top, the R-loop and D-loop assays. After RAD51-mediated R-loop or D-loop formation, proteins were digested with proteinase K (PK), which was then inactivated with PMSF and EGTA. Bottom left, R-loops or D-loops were detected on native gels as indicated. Treatment with RNaseHl, which degrades the RN Amoiety in RNA-DN A hybrid structures, eliminates R-loops but has no effect on D-loops. RNaseH2, which cleaves ribonucleotides in DNA, has no effect, as expected. Right, quantification, n = 3 independent experiments; data are means ± s.d. c, Top, the R-loop and D-loop assays. Middle and bottom, R AD51-mediated R-loops (middle; n = 3 independent experiments; data are means ± s.d.), unlike D-loops (n=l experiment), are recognized by the S9.6 antibody and supershifted in presence of both S9.6 and anti-mouse IgG. Quantification is shown on the right. Article Extended Data Table 11 Oligonucleotides used herein Primers used for Chimeric TERRA constructs Name Forward Reverse Amplification of PP7 stem loops ccgcggccccgaattctatcgatactcgagat cct a gctgactagaggatccacacgcgttctcgataatgaa Amplification of TTAGGG repeats ttagggttagggttagggttagggttaggg ccctaaccctaaccctaaccctaaccctaa Amplification of 15q subtelomere gaacgcgtgtggatcattctcctcaggtcagacccg ctaaccctaaggatcctaaccgtgaccctgaccccg Amplification of Xq subtelomere gaacgcgtgtggatcccgagttgcgttctcg ctaaccctaaggatcgcacatgaggaatgtgggtg Amplification of AAVS1 left homology arm tgaccggttgctttctctgaccagcattc gtgttaaccactgtggggtggaggggac Amplification of AAVS1 right homology arm cggatatcactagggacaggattggt gtgatatcctgtaggaaggggcaggaga Primers used for RT-qPCR Name Forward Reverse 15q TERRA cagcgagattctcccaagctaag aaccctaaccacatgagcaacg GAPDH agccacatcgctcagacac gcccaatacgaccaaatcc TRF1 cttgccagttgagaacgatataca catcagggctgattccaagg TRF2 gggttatgcagtgtctgtcgcg cagtggtgtgagctcagcct POT1 ccaagctctggatcagtatcatt catagtggtgtcctctc aaatac SMG1 ocaagcaccgttccaggaactg ctctcttgcaccgctttcccag UPF1 cgcagggctacatctccatgag ctcgtcaccaaggtaactgtcctg UPF2 ggctgagtctgcagacacaatgc gcagcaagttgagaggacatggg RNaseHl ggccaggccatcctttaaatgtagg gcttgttcaatggctttgcaggc RPA2 gcagggaactttggtgggaatagc cccttcaggtcttggacaagcc RAD51 gaggaaaggaagaggggaaaccag cccactccatctgcattaatggcg 1 q subtelomere cagcgtcgcaactcaaatg ccctcaccctccatgagtaata 10q subtelomere gcattcctaatgcacacatgac tacccgaacctgaaccctaa 13q subtelomere gcacttgaaccctgcaatacag cctgcgcaccgagattct 15q subtelomere aaccctaaccacatgagcaacg gctgcattaaagggtccagt siRNAs Gene name Dharmacon catalog number siRNA sequences siControl d-001206-13-20 uagcgacuaaacacaucaa, uaaggcuaugaagagauac, auguauuggccuguauuag, augaacgugaauugcucaa siTRFI m-010542-02 caag auaaaccuag ug g ua, g g ug auccaaauucucaua, ggaaacuggucuaaaauac, gccaguugagaacgauaua siTRF2 m-003546-00 gaaguggacuguagaagaa, ggaagcugcugucauuauu, ggaaucagcuaucaaugug, gaagacaguacaaccaaua siPOTI m-004205-01 agaaagaugucaacagcua, gaggcaaagaaucgaaaua, cag g ag uacuag aagccua, ug caag aucuccacg uuaa siSMGI m-005033-01 gugaagauguucccuauga, gagguuagcugcggaaaga, ggucagacauccaccagaa, uaacuuggcucagcuguau siUPFI m-011763-01 g cuccuaccug g ug cag ua, ucaag g ucccug auaauua, ggaagucgaccuccuuuga, caagauaacaucacuguca siUPF2 m-012993-01 ggaacgagaauucuuaaua, gcauguaccuuguguagaa, gaagauauucgauuaggaa, ggucuagagaguugcgaau siRNaseHI m-012595-00 g cg cag agccg uaug caaa, g agcuaaacaaucg g aag a, gccaggccauccuuuaaau, gacauucag uggaugcaug siRPA2 m-017058-00 gaucaaugcacacauggua, caaaauagaugacaugaca, gag ug aag cag g gaacuuu, g ugg aacag ug g auucg aa siRAD51 m-003530-04 gaagcuauguucgccauua, gcagugauguccuggauaa, ccaacgaugugaagaaauu, aagcuauguucgccauuaa SJBRCA2 m-003462-01 gaaacggacuugcuauuua, guaaagaaaugcagaauuc, gguaucagaugcuucauua, gaagaaugcagg uuuaaua Extended Data Table 21 Antibodies used herein Antibody Company Catalogue No Dilution (Technique) Anti-GFP Home-Made 1:1,000 (IF) Anti-GFP Merck Millipore MAB3580 1:1,000 (IF, WB) Anti-RAD51 Santa-Cruz sc-8349 1:1,000 (WB) Anti-RAD51 ABCAM abl33534 6 lag (IP) Anti-BRCA2 Merck Millipore MAB3580 1:1,000 (WB) Anti-Vinculin ABCAM abl29002 1:10,000 (WB) Anti-hnRNPAl Santa-Cruz sc-32301 1:1,000 (WB), 6 ug (IP) Anti-RNaseHl GeneTex GTX-117624 1:1,000 (WB) Anti-TRFl Santa-Cruz SC-6165R 1:1,000 (WB) Anti-TRF2 Merck Millipore 05-521 1:1,000 (WB) Anti-POTl ABCAM abl24784 1:1,000 (WB) Anti-SMGl Abgent AP8055a 1:500 (WB) Anti-UPFl Bethyl Laboratories A300-037A 1:1,000 (WB) Anti-RPA2 ABCAM ab2175 1:1,000 (WB) Anti-DNA-RNA Hybrid [S9.6] Kerafast ENH001 1 ug / 10 |ag nucleic acids (DRIP) Anti-Rabbit IgG (H+L), HRP Conjugate Promega W4011 1:10,000 (WB) Anti-Mouse IgG (H+L), HRP Conjugate Promega W4021 1:10,000 (WB) Goat anti-Rabbit IgG (H+L) Cross- Adsorbed Secondary Antibody, Thermo Fisher A-21070 1:1,000 (IF) Alexa Fluor 633 IF, immunofluorescence; IP, immunoprecipitation; WB, western bLot. nature research Reporting Summary Corresponding author(s): Joachim Lingner Last updated by author(s): Jun 22, 2020 Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist. Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a □ □ □ □ □ Confirmed ^ The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement ^| A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly prpi The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. "j A description of all covariates tested "j A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) ^ AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable "j For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings "j For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes "j Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection Microscopy images were acquired on an Upright Zeiss Axioplan or on a Zeiss LSM700 equipped with an AxioCam MRm B/W. The screens were developed on a Typhoon phosphorimager (GE). Western Blot membranes were developed on a Vilber Fusion FX imaging system I EMSA Gels were imaged on a FLA-9000 scanner (Fujifilm) and quantified with Multi Gauge V3.2 (Fujifilm) Data analysis Images were processed and analyzed with Image J (2.0.0-rc-69/1.52n). Graphs were prepared in GraphPad Prism 7 (7.0d) and Figures assembled using Adobe Photoshop (21.1.0) and Adobe Illustrator (24.0.3) -or manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. Data Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability All materials and data-sets are available upon request. Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. ^ Life sciences ] Behavioural & social sciences ] Ecological, evolutionary & environmental sciences zor a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summarv-flat.pdf Life sciences study design All studies must disclose on these points even when the disclosure is negative Sample size No statistical method was used to pre-determine sample size. Sample sizes were chosen according to commonly used standards in the field. (Azzalin et al. Science. 2007 Nov 2;318(5851):798-801; Vancevska et al. EMBO J. 2020 Apr l;39(7):el02668) The number of independent experiments are indicated in the legend of each Figure. Proper Negative and whenever possible positive controls were used for each [experiment. Data exclusions No data was excluded from the analysis Replication The screen in Figure 2c was performed in two independent biological replicates. All the other experiments were performed in three [independent biological replicates. All the replicates were successfu Randomization All samples were chosen randomly Blinding For Figures lc, Id, 2c, 2d, and Ext. Figure 2e, 2f, 4a the researchers were blinded to image acquiring and analysis. For the rest either the differences between samples were obvious making blinding impossible or blinding was not applicable due to the nature of the experiment. Reporting for specific materials, systems and methods_ We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems Methods n/a Involved in the study n/a Involved in the study □ ^ Antibodies M ~] ChlP-sec □ |^ Eukaryotic cell lines ] Flow cytometry ] Palaeontology ] MRI-based neuroimagin ] Animals and other organisms ] Human research participants ] Clinical data Antibodies Antibodies used See also Table S2: GFP Home-Made GFP Merck Millipore MAB3580 RAD51 Santa-Cruz sc-8349 RAD51 ABCAM abl33534 BRCA2 Merck Millipore MAB3580 Vinculin ABCAM abl29002 hnRNPAl Santa-Cruz sc-32301 RNaseHl GeneTex GTX-117624 TRF1 Santa-Cruz sc-6165R TRF2 Merck Millipore 05-521 P0T1 ABCAM abl24784 SMG1 Abgent AP8055a UPF1 Bethul Laboratories A300-037A RPA2 ABCAM ab2175 Anti-DNA-RNA Hybrid [S9.6] Kerafast ENH001 Anti-Rabbit IgG (H+L), HRP conjugate, Promega, W4011 Anti-Mouse IgG (H+L), HRP conjugate, Promega, W4021 Goat Anti-Rabbit IgG (H+L) Cross-Absorbed Secondary Antibody, Alexa Fluor 633, ThermoFisher, A-21070 Validation I Validation for each antibody is provided in the manuscript. The specificity of most of them (including the home-made GFP 2 antibody) was tested upon disappearance of the corresponding band upon siRNA knock down in this study. S9.6 specificity for DNA-RNA hybrids is confirmed by in-vitro treatment with RNAseHl. Eukaryotic cell lines Policy information about cell lines Cell line source(s) HeLa and HEK293T cell lines were purchased from ATCC. U20S was purchased from ECACC Authentication HeLa and HEK293Tcell lines have not been authenticated by us but the supplier. U20S cell line was authenticated by measuring the production of c-circles (a known ALT cell line phenotype) Mycoplasma contamination All the cell lines were tested negative for mycoplasma Commonly misidentified lines no commonly misidentified cell lines were usee (See JCLAC register)