BRCA2 Prevents R-loop Accumulation and Associates with TREX-2 mRNA Export Factor PCID2

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1038/nature13374”. ! 1! BRCA2 prevents R-loop accumulation and associates with the TREX-2 mRNA export factor PCID2 Vaibhav Bhatia, Sonia I. Barroso, María L. García-Rubio, Emanuela Tumini, Emilia Herrera-Moyano and Andrés Aguilera * Centro Andaluz de Biología Molecular y Medicina Regenerativa CABIMER, Universidad de Sevilla, Av. Américo Vespucio s/n, 41092 Seville, Spain * Correspondence to: [email protected] ! 2! Genome instability is central to aging, cancer and other diseases. It is not only proteins involved in DNA replication or the DNA damage response (DDR) that are important for maintaining genome integrity. From yeast to higher eukaryotes, mutations in genes involved in splicing, mRNP biogenesis and export also induce DNA damage and genome instability. Such an instability is frequently mediated by R-loops formed by DNA:RNA hybrids and a displaced single-strand DNA (ssDNA) 1. Here we show that the human TREX-2 complex, involved in mRNP biogenesis and export, prevents genome instability as determined by γH2AX and 53BP1 foci accumulation and single-cell electrophoresis in cells depleted of the TREX-2-subunits PCID2, GANP and DSS1. With an EGFP-tagged Hybrid-Binding Domain of RNaseH1 and the S9.6 antibody we did not detect R-loops in TREX-2-depleted cells. Notably, the BRCA2 repair factor, which binds to DSS1, also associates with PCID2 in the cell, and BRCA2-depleted cells accumulate high levels of R-loops. The results indicate that R-loops are frequently formed in cells and that BRCA2 is required for their processing. This novel link between BRCA2 and RNA-mediated genome instability suggests that R-loops may be a major source of replication stress and cancer-associated instability. R-loops can negatively affect transcription elongation 2,3, and they have also been involved in promoter proximal pausing and transcription termination 4-6. However, R- loops also mediate Replication Fork (RF) stalling as a source of genome instability 7-9, and evidence suggests that they form at common fragile sites 10. Factors like Senataxin (SETX) or RNase H1 (RNH1) are implicated in R-loop dissolution 1 but the mechanisms used by the cellular machinery to resolve R-loops are hardly understood. Importantly, a number of mRNP biogenesis factors prevent R-loop formation and transcriptionassociated genome instability, a paradigmatic example being the conserved eukaryotic THO complex, involved in transcription elongation and mRNA processing and export 2,11-13. THSC/TREX-2 is another well-conserved protein complex working in RNA export, preferentially located at the nuclear pore complex (NPC) 14-16 (Extended Data Fig. S1). In yeast it has a similar effect on genome instability as that of THO 17 and in humans is composed of PCID2, GANP, CENP and DSS1 14,18. However, no information exists about the role of human TREX-2 in genome integrity. ! 3! To answer this question we analyzed genome stability in HeLa cells siRNA- depleted of PCID2, GANP and DSS1 (Extended Data Fig. S2) by determining γH2AX and 53BP1 foci (Fig. 1a). These foci increased upon depletion of both PCID2 and DSS1. Consistent with DNA break accumulation, alkaline single-cell electrophoresis revealed an increase in tail moment (Fig. 1b). Depletion of the SAC3 homolog in human-TREX-2 GANP also induced genomic instability but to a lesser extent (Extended Data Fig. S3a, b). Since one cause of genome instability is transcription-replication collisions, which are augmented by R-loops 7,9, we assayed the effect of TREX-2 depletion on RF progression by DNA combing. RF velocity of PCID2-and GANP- depleted cells was slightly faster and track lengths longer than in control cells (Figs. 1c, Extended Data Fig. S3c), a phenomenon previously observed in cells depleted of other RNA-processing factors 11. Instead, DSS1-depletion led to slower moving forks. Interorigin distance was longer, implying lower density of active replicons, especially in PCID2-depleted cells (Fig. 1c). The results suggest that the causes of genome instability involved different replication-profile alterations. In addition to PCID2, DSS1 binds and stabilizes BRCA2, a double-strand break (DSB) repair factor 19. Therefore, we assayed whether PCID2 and BRCA2 interact in vivo using Proximity Ligation Assay (PLA), which detects cellular proteinprotein interactions in situ 20. We found a close association of the two proteins in the cells (Figs. 1d, Extended Data Fig. S4). Next, we assessed whether or not DNA- damage accumulation in PCID2-depleted cells was due to a defect in RAD51 loading. As expected, BRCA2-depletion, used as control, and DSS1-depletion, but not PCID2-depletion, reduced RAD51 foci (Fig. 1e), indicating that genomic instability in PCID2-depleted cells is not related to a functional defect of RAD51. Next, we asked if RNA-DNA hybrids accumulated in the absence of PCID2, DSS1 and BRCA2, as a source of instability. For this, a fusion protein of the 52 amino acid DNA:RNA Hybrid Binding (HB) domain of RNH1 and EGFP was constructed, thereafter referred to as HB-GFP (Fig. 2a). We confirmed that HB-GFP was able to detect RNA-DNA hybrids in vivo by different means. Using HeLa cells transfected with vectors encoding HB-GFP or EGFP, we show by confocal microscopy that unlike EGFP, HB-GFP preferentially localize in the nucleus (Fig. 2b). When cells were permeabilized with detergent, EGFP was totally washed out whereas HB-GFP was retained in the nucleus (Fig. 2c). Furthermore, size-exclusion chromatography of HB-GFP expressed in HEK293 cells showed a fraction of the cellular protein pool as part of a multimeric complex, and reversible cross-link immunoprecipitation revealed HB-GFP interaction with histone H3 and ssDNA- ! 4! binding Replication Protein A (RPA), consistent with HB-GFP associating with chromatin (Extended Data Fig. S5a, b). Immunostaining with anti-GFP in HB-GFP- transfected HeLa cells and the anti-DNA:RNA S9.6 antibody in HeLa cells, showed nuclear localization and an intense staining in the nucleolar region (Fig. 2d), consistent with the predicted accumulation of R-loops in the rDNA region 21, which we confirmed by ChIP and which shows increased HB-GFP recruitment upon TOP1- depletion (Extended Data Fig. S5c) as expected. Interestingly, a pool of cells immunostained with anti-GFP, and S9.6 also showed accumulation of signal in the proximity of the nuclear periphery (Fig. 2e). Further, HB-GFP formaldehydecrosslinked to pelleted chromatin could be relieved in the supernatant of cell lysates by treating with RNH1 (Extended Data Fig. S5d). THOC1-depleted cells, known to accumulate DNA:RNA hybrids 11 were used as a positive control with the expected positive results. Altogether the results confirm that HB-GFP efficiently detects R- loops in vivo. Then, we used flow cytometry to quantitatively show that after permeabilization HB-GFP retention leads to a significant number of fluorescent positive cells, unlike EGFP, which is completely washed out. Importantly, overexpression of active RNH1, which competes with HB-GFP, reduced the number of HB-GFP positive cells (Fig. 2f, Extended Data Fig. S5e). Therefore, we used this assay to quantify DNA:RNA hybrids as the number of HB-GFP positive cells without RNH1-overexpression relative to the number with RNH1-overexpression (Fig. 2g). DNA:RNA hybrids slightly increased in the siTOP1 control (Extended Data Fig. S5f), did not increase in PCID2-depleted cells and, surprisingly, strongly increased in BRCA2- and DSS1-depleted cells, similarly to the siTHOC1 control (Fig. 2g). Interestingly, BRCA1-depleted cells also increased HB-GFP retention although to a lesser extent. Next, we analyzed DNA:RNA hybrid accumulation at various genes at the molecular level. Notably, ChIP using HB-GFP showed significant accumulation of hybrids in BRCA2-depleted cells in the three genes assayed whereas RNH1- overexpression reduced the signal (Fig. 3a, Extended Data Fig. S6). The siTHOC1 control behaved consistently. DNA-RNA immuno-precipitation (DRIP) using the S9.6 antibody 22 shows that BRCA2-depletion increased DNA-RNA hybrids at four actively transcribed genes, while the siTHOC1 cells showed a heterogeneous profile (Fig. 3b) consistent with this being dependent on parameters such as transcription levels or gene length in yeast 13. PCID2-depleted cells did not increase R-loop accumulation (Fig 3a, b), though when co-depleted with BRCA2 it slightly increased HB-GFP ! 5! recruitment (Extended Data Fig. S7), consistent with the idea that both TREX2 and BRCA2 could at least in part cooperate to prevent R-loop formation. As BRCA2 functions at DSBs by recruiting Rad51, and at RFs as part of the Fanconi Anemia (FA) pathway, we asked whether R-loops accumulated after Rad51 or BRCA1 depletion to get further insight into BRCA2’s role in R-loops. Interestingly, BRCA1- but not RAD51-depleted cells accumulated R-loops as detected by DRIP (Fig. 3c), suggesting that BRCA2 prevents R-loops independently of the Rad51 homologous recombination function. Indeed, R-loops are accumulated in non-replicating and replicating BRCA2-depleted cells as detected by immunofluorescence, in the latter case at a substantially higher level (Fig. 4a). Importantly, spontaneous recruitment of BRCA2 at different genes was clearly reduced by RNH1 overexpression as detected by ChIP in 4 genes analyzed (Fig. 4b), indicating that BRCA2 is recruited to DNARNA hybrid regions. We then hypothesized that expression of the inactive HB-GFP should bind and stabilize DNA:RNA hybrids, consequently increasing genome instability by strengthening a putative RF block. As anticipated, BRCA2, DSS1 and THOC1, but not PCID2-depleted cells showed elevated levels of γH2AX foci in HB-GFP expressing cells (Fig. 4c). Chromosomal abnormalities are a hallmark of cancer and BRCA2-depleted cells, and we found that HB-GFP expression increased chromosome breaks and reduced spontaneous sister chromatid exchanges (Extended Data Fig. S8a, b), as shown for BRCA2-deficient mouse fibroblasts 23 and FA cells 24, even though BRCA2-deficient tumor cells show increased SCE 24. In addition, the genomic instability in BRCA2-depleted and HB-GFP-expressing RPE cells correlated with increased adhesion-independent proliferation, a hallmark of cell transformation (Extended Data Fig. S8c), which is consistent with a potential role of R-loops as a cancer risk intermediate. This is consistent with an increase in RNADNA hybrids in BRCA2-deficient cells enhancing HB-GFP binding and reducing the efficiency of repair. Finally, we extended our study to the BRCA2-defective pancreatic adenocarcinoma-derived CAPAN-1 cells. These cells show pan-nuclear staining of γH2AX, as a signal of replication stress, which disappeared after RNH1 overexpression (Fig. 4d). Interestingly, any reminiscent γH2AX observed in CAPAN- 1 cells with RNH1 overexpression, was consistently observed in proximity to the nuclear periphery, consistent with data in Fig. 2e. Altogether, our results demonstrate that R-loops accumulate in BRCA2- depleted cells at high levels. Such accumulation is seen in replicating and nonreplicating cells. As TREX-2/PCID2 works on RNA export close to the NPC 14,16, it is ! 6! possible that the topological constrains resulting from the physical tethering of transcribed genes to the NPC 25 promote R-loop accumulation and replication stress. TREX-2 depletion could loose the connection of the transcribed site to the NPC 15, minimally contributing to R-loop accumulation in cells. It may also be that, in the absence of TREX-2, other RNA binding factors help prevent R-loop accumulation to a level detectable in the conditions tested here. With these results, we propose that TREX-2 helps recruit BRCA2 to the proximity of transcribed regions where BRCA2 would bind to the branched structure formed by the displaced ssDNA and the DNARNA hybrids of naturally formed R-loops (Extended Data Fig. S9a). Such a structure may mimic the replicative intermediate to which BRCA2 is believed to bind in vivo 26,27. This could help expose the DNA-RNA hybrid for a better action of enzymes that remove the RNA chain such as ribonucleases (RNH1) or DNA-RNA helicases (SETX). This would be supported by the observations that BRCA2 is recruited to RNA-DNA hybrid regions and that HB-GFP expression enhances or has a synergistic effect on genome instability in normal and BRCA2-depleted cells, respectively. In replicating chromatin, as BRCA2 and BRCA1 works in the FA pathway, which removes crosslinks and replication-blocking lesions 26-29, RFs encountering an R-loop could target the action of BRCA2, BRCA1 and possibly other FA proteins to prevent fork collapse and reversal, likely impeding R-loop extension, and promote RF restart and R-loop dissolution (Extended Data Fig. S9b). Thus, TREX-2 could function at cotranscriptional R-loops encountering RFs, by regulating the stability and/or recruitment of BRCA2 to the ssDNA substrates generated at the sites of collisions 30. Our results, therefore, reveal a new and unexpected role for tumor suppressors in preventing R-loops, suggesting that R-loops may be a major cause of replication stress and, hence, cell death and cancer. Methods Summary ON-TARGET SMART pools of standard siRNAs were used from Thermo Scientific. DNA damage foci were analyzed by Immunofluorescence and quantitated using Metamorph (Molecular Probes) image analysis software. Scale bars shown in all images is 20 µm. DNA combing was performed as described 11, with minor modifications. For HB-GFP-retention analysis HeLa cells were permeabilized prior to fixation with 0.05% TritonX100, and were processed for microscopy or FACS. For S9.6 staining fixed cells were pre-treated with 0.5%SDS of ice cold Methanol ! 7! followed by acetone. DNA:RNA immunoprecipitation (DRIP) was performed as described 22. For ChIP with HB–GFP, HeLa cells carrying the tet-regulated HB-GFP fusion were used. After 48 hr of siRNA-transfection, HB-GFP expression was induced by 2µg/ml doxycycline for 24 hr. For RNH1-positive samples, cells were transfected with the RNH1-encoding plasmid and processed for ChIP. The number of biological repeats (n) performed in each cased is indicated in Figure Legends. More details in Methods. Acknowledgements We thank JC Reyes and AG Rondón for comments on the manuscript and D Haun for style supervision. Research was funded by grants from the Spanish Ministry of Economy and Competitiveness (Consolider CSD2007-00015 and BFU2010-16372), the Junta de Andalucía (CVI4567) and the European Union (FEDER). Contributions VB, SB, MGR, ET and EHM performed the experiments; VB and AA designed the experiments and wrote the paper. Author information Correspondence and request for materials should be addressed to AA, at [email protected] Literature cited 1. Aguilera, A. & García-Muse, T. R loops: from transcription byproducts to threats to genome stability. Mol. Cell 46, 115-124 (2012). 2. Huertas, P. & Aguilera, A. Cotranscriptionally formed DNA:RNA hybrids mediate transcription elongation impairment and transcription-associated recombination. Mol. Cell 12, 711-721 (2003). 3. Tous, C. & Aguilera, A. Impairment of transcription elongation by R-loops in vitro. Biochem. Biophys. Res. Comm. 360, 428-432 (2007). ! 8! 4. Kaneko, S., Chu, C., Shatkin, A. & Manley, J. Human capping enzyme promotes formation of transcriptional R loops in vitro. Proc. Natl. Acad. Sci. USA 104, 17620-17625 (2007). 5. Skourti-Stathaki, K., Proudfoot, N. & Gromak, N. Human senataxin resolves RNA/DNA hybrids formed at transcriptional pause sites to promote Xrn2- dependent termination. Mol. Cell 42, 794-805 (2011). 6. Mischo, H. et al. Yeast Sen1 helicase protects the genome from transcriptionassociated instability. Mol. Cell 41, 21-32 (2011). 7. Tuduri, S. et al. Topoisomerase I suppresses genomic instability by preventing interference between replication and transcription. Nat. Cell Biol. 11, 1315-1324 (2009). 8. Wellinger, R., Prado, F. & Aguilera, A. Replication fork progression is impaired by transcription in hyperrecombinant yeast cells lacking a functional THO complex. Mol. Cell. Biol. 26, 3327-3334 (2006). 9. Gan, W. et al. R-loop-mediated genomic instability is caused by impairment of replication fork progression. Genes Dev. 25, 2041-2056 (2011). 10. Helmrich, A., Ballarino, M. & Tora, L. Collisions between replication and transcription complexes cause common fragile site instability at the longest human genes. Mol. Cell 44, 966-977 (2011). 11. Domínguez-Sánchez, M., Barroso, S., Gómez-González, B., Luna, R. & Aguilera, A. Genome instability and transcription elongation impairment in human cells depleted of THO/TREX. PLoS Genet. 7(12), e1002386 (2011). 12. Castellano-Pozo, M., García-Muse, T. & Aguilera, A. R-loops cause replication impairment and genome instability during meiosis. EMBO Rep. 13, 923-929 (2012). 13. Gómez-González, B. et al. Genome-wide function of THO/TREX in active genes prevents R-loop-dependent replication obstacles. EMBO J. 30, 3106-3119 (2011). 14. Jani, D. et al. Functional and structural characterization of the mammalian TREX-2 complex that links transcription with nuclear messenger RNA export. Nucleic Acids Res. 40, 4562-4573 (2012). 15. Cabal, G. et al. SAGA interacting factors confine sub-diffusion of transcribed genes to the nuclear envelope. Nature 441, 770-773 (2006). 16. Umlauf, D. et al. The human TREX-2 complex is stably associated with the nuclear pore basket. J. Cell Sci. 126, 2656-67 (2013). 17. Gallardo, M., Luna, R., Erdjument-Bromage, H., Tempst, P. & Aguilera, A. Nab2p and the Thp1p-Sac3p complex functionally interact at the interface between transcription and mRNA metabolism. J. Biol. Chem. 278, 24225-24232 (2003). 18. Ellisdon, A., Dimitrova, L., Hurt, E. & Stewart, M. Structural basis for the assembly and nucleic acid binding of the TREX-2 transcription-export complex. Nat. Struct. Mol. Biol. 19, 328-336 (2012). 19. Li, J. et al. DSS1 is required for the stability of BRCA2. Oncogene 25, 1186-1194 (2006). 20. Söderberg, O. et al. Direct observation of individual endogenous protein complexes in situ by proximity ligation. Nat. Methods 3, 995-1000 (2006). 21. El Hage, A., French, S., Beyer, A. & Tollervey, D. Loss of Topoisomerase I leads to R-loop-mediated transcriptional blocks during ribosomal RNA synthesis. Genes Dev. 24, 1546-1558 (2010). 22. Ginno, P., Lott, P., Christensen, H., Korf, I. & Chédin, F. R-loop formation is a distinctive characteristic of unmethylated human CpG island promoters. Mol. Cell 45, 814-825 (2012). 23. Urtishak, K., Smith, K., Chanoux, R., Greenberg, R., Johnson, F., and Brown, E. Timeless Maintains Genomic Stability and Suppresses Sister Chromatid ! 15! 41.5 ml of 20 %SDS and 5 ml of proteinase K (Roche). DNA was extracted gently with phenol-chloroform. Precipitated DNA was spooled on a glass rod, washed with 70% EtOH, resuspended gently in TE and digested overnight with 50U of HindIII, EcoRI, BsrGI, XbaI and SspI, 2mM spermidine and bovine serum albumin (BSA). For the negative control, half of the DNA was treated with 3 µl RNase H (M0297, New England BioLabs) overnight. 5µg of the digested DNA was bound to 10µl of S9.6 antibody (1mg/ml) in 500ul binding buffer (10mM NaPO4, 140 mM NaCl, 0.05% Triton X-100) overnight a 4ºC. DNA-antibody complexes were immunoprecipitated using Dynabeads Protein A (Invitrogen) for 2hrs at 4ºC and washed 3 times with binding buffer. DNA was eluted with 50mM Tris pH 8.0, 10mM EDTA, 0.5% SDS, treated 45 min with 7ml proteinase K at 55ºC and cleaned with phenol-chloroform. Real time-qPCR was performed at the indicated regions. Signal intensity plotted is the relative abundance of DNA:RNA hybrid immunoprecipitated in each region which was normalized to input values and to the signal at the negative control region SNRPN gene (see supplementary Table). RNaseH1 dependent release of HB-GFP from chromatin of HeLa cells. Transfected cells, expressing HB-GFP, were washed with 0.2% TritonX-100, treated with 2% Formaldehyde, lysed in Lysis buffer (1%Triton, 0.2% SDS,50mM Tris pH-8; 50mM NaCl, 5mM MgCl2, 1mM DTT, 20ug/BSA , 5% Glycerol ) for 15 min and sonicated and split into two halves. RNaseH1 (NEB MO297) was added to one half and incubated at 22°C for 30 min and subjected to Western. Metaphase spread and Sister Chromatid Exchange analysis SCE assay was performed as described 32, with minor modifications. Briefly, cells were transfected with indicated plasmid and incubated with 10 µM BrdU for 42 hours followed by 3 hours treatment with 0.1µg/ml of Colcemid. Cells were harvested at 48 hrs after plasmid transfection. Metaphase spreads were prepared and the slides were incubated with 20 µg/ml Hoechst solution for 20 min, then exposed for 1 hr to UVA irradiation in SSC 2X and incubated in SSC 2X for 20 min before Giemsa staining was performed. Metaphases were scored using a 100X objective and at least 25 images were taken randomly from each condition. Adhesion-independent cell proliferation Retinal Pigment Epithelial (RPE) cells were used for adhesion-independent cell growth. 0.5% DNA-grade agarose/ DMEM F-12 medium was added to 96-well plates to prevent cell attachment to plastic base forming of monolayer. The plates were ! 16! stored at 4°C. RPE cells were transfected with siRNA and/or plasmid as mentioned before. After 72 hrs, on day 4, cells were dissociated with Trypsin- EDTA (Gibco) and resuspended in medium. Cells were counted, centrifuged, and resuspended in complete DMEM F-12 medium with 0.3% agar, prepared at 42°C, and dispensed into a 96-well plate pre-warmed in a 37°C incubator. Cells were incubated at 37°C and quantitated on day-9 using the WST-1 reagent (Roche) according to the manufacturer's protocol. Absorbance was read in a Varioskan Flash Multimode Reader. ! ! 17! Extended Data Figure Legends Figure S1. Subcellular localization of PCID2- Immunofluorescence of endogenous PCID2 in HeLa cells. a, without permeabilization b, with prepermeabilization (see Full Methods). Figure S2. Validation of siRNAs. a, Relative mRNA quantitation, Mean and SEM is plotted. b, Western blots analysis, of siRNA treated HeLa cells. Figure S3. Effect of GANP-depletion in genomic instability. a, γH2AX and 53BP1 foci. b, Single-cell electrophoresis. c,!DNA-combing analysis in GANP-depleted cells. Details as in Fig. 1. ! Figure S4. Proximity Ligation Assay in PCID2-depleted cells- a, Staining pattern observed with both anti-BRCA2 and anti-PCID2 antibodies in conditions used for PLA in Fig. 1d. b, anti-PCID2 immunofluorescence analysis. c, Proximity Ligation Assay; in Control and PCID2 siRNA-treated cells (see Fig. 1d). Figure S5. HB-GFP interacts with chromatin and chromatin-associated proteins via DNA:RNA hybrids. a, HB-GFP expressing HEK293 lysate fractionated on Superose 6 size-exclusion columns (Gelifesciences-17-5172-01) and analyzed by western. b, HB-GFP co-immunoprecipitated proteins using the ReCLIP method 33 . c, HB-GFP ChIP in rDNA region in TOP1-depleted HeLa cells (n=3). Mean and SEM are plotted for three independent experiments. d) RNH1-dependent release of HBGFP from chromatin of HeLa cells. e, Scheme and representative plot of FACS assays used to quantitate DNA:RNA hybrids. PI is Propidium Iodide (see Fig. 2f). f, FACS assay to quantify DNA:RNA hybrids in TOP1-depleted cells. Mean and SEM are plotted (n=3). Figure S6. HB-GFP ChIP of the UTRN and ACTB gene regions. a, b, Normalized siRNA control-relative values for the immunoprecipitated DNA:RNA hybrids plotted relative to the siRNA control in the UTRN and ACTB genes. Mean and SEM of a minimum of three experiments is shown. Amplicon positions used are indicated in gene diagram above each graph. Mean and SEM are plotted (n=3). ! 18! Figure S7. DNA:RNA hybrids accumulation in cells co-depleted of PCID2 and BRCA2. HeLa cells were treated with siBRCA2 in combination with siRNA control or siPCID2, and processed. Mean and SEM are plotted (n=3). Details as in Fig. 3b. Figure S8. Chromosomal aberrations in cells expressing HB-GFP. a, Metaphase spreads of HeLa cells expressing GFP (Control) or HB-GFP. Fragmentation and SCE events are indicated by arrowheads and arrows, respectively. b, Quantitation of chromosome breaks in RPE cells expressing HB-GFP. c, Adhesion-independent proliferation assay. Cell proliferation relative to control siRNA-treated RPE cells is shown. Mean and SEM are plotted (n=3). * P ≤ 0.05 using two tailed Student’s t-test Figure S9. Model to explain the role of BRCA2 preventing R-loops as a source of genome instability. a, RNA-DNA hybrids may form in the interior as well as in the periphery of the nucleus. mRNP biogenesis factors such as the TREX-2 complex may help recruit or stabilize BRCA2 at the proximity of transcribed regions, whether or not in the proximity of the nuclear pore complex (NPC). BRCA2 and other related proteins could bind to the branched structure generated by the ssDNA displaced in the R-loop facilitating the action of enzymes that remove R-loops, such as specific ribonucleases or DNA-RNA helicases. This could occur in non-replicating chromatin. b, In replicating chromatin, BRCA2 and, presumably, other Fanconi Anemia proteins may act directly at putatively stalled RFs in front of an R-loop to impede RF collapse or reversal, likely impeding R-loop extension. Subsequently, R-loop removal could be promoted by the passage of the RF. Supplementary References 31. ten Asbroek, A., van Groenigen, M., Nooij, M. & Baas, F. The involvement of human ribonucleases H1 and H2 in the variation of response of cells to antisense phosphorothioate oligonucleotides. Eur. J. Biochem. 269, 583-592 (2002). 32. Bayani, J. & Squire, J. A. Sister chromatid exchange. Curr. Protoc. Cell Biol. Chapter 22, Unit 22.7 (2005). 33. Smith, A., Friedman, D., Yu, H., Carnahan, R., and Reynolds, A. ReCLIP (reversible cross-link immuno-precipitation): an efficient method for interrogation of labile protein complexes. PLoS One 6(1):e16206 (2011). !