Human DNA Topoisomerase I Poisoning Causes R Loop-mediated Genome Instability Attenuated by Transcription Factor IIS

Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 1 of 21 MOLECULAR BIOLOGY Human DNA topoisomerase I poisoning causes R loop–mediated genome instability attenuated by transcription factor IIS Renée C. Duardo1†, Jessica Marinello1†, Marco Russo1†, Sara Morelli1, Simona Pepe1, Federico Guerra1, Belén Gómez- González2,3, Andrés Aguilera2,3, Giovanni Capranico1* DNA topoisomerase I can contribute to cancer genome instability. During catalytic activity, topoisomerase I forms a transient intermediate, topoisomerase I–DNA cleavage complex (Top1cc) to allow strand rotation and duplex relaxation, which can lead to elevated levels of DNARNA hybrids and micronuclei. To comprehend the underlying mechanisms, we have integrated genomic data of Top1cc- triggered hybrids and DNA doublestrand breaks (DSBs) shortly after Top1cc induction, revealing that Top1ccs increase hybrid levels with different mechanisms. DSBs are at highly transcribed genes in early replicating initiation zones and overlap with hybrids downstream of accumulated RNA polymerase II (RNAPII) at gene 5′- ends. A transcription factor IIS mutant impairing transcription elongation further increased RNAPII accumulation likely due to backtracking. Moreover, Top1ccs can trigger micronuclei when occurring during late G1 or early/mid S, but not during late S. As micronuclei and transcriptionreplication conflicts are attenuated by transcription factor IIS, our results support a role of RNAPII arrest in Top1cc- induced transcriptionreplication conflicts leading to DSBs and micronuclei. INTRODUCTION DNA topoisomerases are a family of essential nuclear enzymes that regulate DNA topology, chromatin structures, and basic processes such as transcription and replication (1). DNA topoisomerase I (Top1) couples DNA strand cleavageligation reaction with the rotation of the cut strand around the other, therefore reducing the torsional tension of DNA duplexes (2, 3). During the catalytic cycle, a reaction intermediate called Top1- DNA cleavage complex (Top1cc) forms, wherein Top1 is covalently linked to the 3′- end of the cut strand. However, under certain circumstances, Top1ccs can trigger genome instability (4). Top1ccs can be stabilized by several factors, including U.S. Food and Drug Administration–approved antitumor Top1 poisons, such as camptothecin (CPT) analogs. CPT selectively targets Top1 in living cells (2, 3), leading to a rapid (2 to 5min) increase in Top1ccs (5, 6). Within minutes, they affect RNA and DNA synthesis markedly (2, 3), impair DNA transactions and epigenomic features (5–10), and activate ubiquitindependent Top1 degradation (3, 11, 12). Moreover, CPT- induced Top1ccs increase transcriptionreplication conflicts (TRCs) (13–16) and DNA doublestrand breaks (DSBs) (3, 4, 17, 18). The generation of DSBs may be due to replication run off at Top1cc sites (19, 20) or to the activity of endonucleases at stalled forks (21–25). TRCs occur more often in cancer than normal cells due to high oncogeneinduced transcription rates and can thus be at the basis of high genome instability in cancer (13–15). Top1ccs can also occur in unperturbed cells as they can arise when the DNA substrate is damaged (4, 26). Top1ccs were shown to be a general pathogenic feature of neurodegenerative disorders derived from mutations of DNA repair factors (27). A mutation of Tyrosyl-DNA phosphodiesterase 1 (TDP1), a specific repair factor of Top1ccs, can cause a rare type of ataxia spinocerebellar ataxia with axonal neuropathy 1 (SCAN1) (28) characterized by cell hypersensitivity to CPT (29). Moreover, Top1 can introduce short indels in the genome at sites of misincorporated ribonucleotides (30, 31). Thus, impaired Top1 activity can be dangerous to cell life and differentiation when nuclear Top1cc levels become elevated because of exogenous or endogenous factors. However, the mechanism elicited by Top1ccs is not yet fully understood. Top1ccs can effectively induce TRCs with a yet undefined R loop– mediated mechanism. R loops are threestrand nucleic acid structures containing a hybrid DNA:RNA duplex (15, 32). Top1 poisons cause a transient increase in R loops in cancer cells (5, 18, 33, 34). However, genomic R loop maps are not yet available upon Top1cc increase. Recently, we showed that micronuclei induced by Top1ccs are dependent on R loops (18). In addition, mutations of ribonuclease (RNase) H2, an endonuclease specifically targeting RNA strands of hybrid duplexes, can cause the human Aicardi- Goutières syndrome and an increase in micronuclei triggering a persistent inflammatory activation, likely at the basis of the syndrome (35–37). However, the mechanism of hybriddependent TRCs, DNA cleavage, and micronucleus formation remains largely unknown in human cancer cells. In this study, we uncover a main mechanism by which Top1ccs induce R loops and TRCs. By determining the genomic maps of DNARNA hybrid and DSB loci upon Top1cc induction, we provide evidence for the role of transcription elongation factor IIS (TFIIS) in attenuating TRCs and genome instability. Moreover, analysis of cell cycle phase–dependent triggering of micronuclei allowed us to show that Top1ccs affect genome stability specifically during late G1/early S phases. Our findings establish an early replicationspecific and transcriptiondependent mechanism of genome instability elicited by Top1ccs and provide the main genomic sites of transcription/replication conflicts (TCRs) induced by Top1ccs. These findings can aid in the development of strategies for neurodegeneration prevention and cancer treatment. 1Department of Pharmacy and Biotechnology, Alma Mater Studiorum–University of Bologna, via Selmi 3, 40126, Bologna, italy. 2centro Andaluz de Biología Molecular y Medicina Regenerativa—cABiMeR, Universidad de Sevilla–cSic, calle Américo vespucio 24, 41092 Seville, Spain. 3Departamento de Genetica, Facultad de Biología, Universidad de Sevilla, 41012 Seville, Spain. *corresponding author. email: giovanni. capranico@ unibo. it †these authors contributed equally to this work. copyright © 2024 the Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. no claim to original U.S. Government Works. Distributed under a creative commons Attribution license 4.0 (cc BY). Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 2 of 21 RESULTS Top1ccs induce stable and transient hybrid gains at highly transcribed genes To establish the role of R loops in genome instability by Top1ccs, we first determined the genomic maps of DNARNA hybrids. Cell treatment with CPT increases Top1ccs at the highest levels within 2 to 5min (5, 6) and DNARNA hybrid levels after 5 to 10 min (5, 18) in human cancer cells (fig.S1, A and B). Thus, we determined the hybrid maps immediately upon Top1cc increase, after 5- min CPT treatment, compared with a longer treatment (60min). After read normalization (fig.S1, C and D, and tableS1), the results showed that Top1ccs induced either a reduction (loss) or an increase (gain) in hybrid levels at thousands of loci in a timedependent manner (Fig.1, A and B). Hybrid peaks were then split into seven categories based on their kinetics: Stable gains or losses showed hybrid changes found at 5min that persisted at 60min; transient gains and losses showed hybrid changes found at 5min that did not persist at 60min; late gains and losses showed hybrid changes detected only at 60min; and, last, no change hybrids that did not show any significant alteration at either times (Fig.1, A and B). Thus, Top1ccs can dynamically alter nuclear hybrid levels in either direction. Top1cc- altered hybrid peaks show size differences (fig.S1E) and were largely localized along genes, particularly at introns (Fig.1C and fig.S1F). Specifically, stable hybrid gains and losses are enriched at 5′ untranslated region (5′UTR)/upstream and 3′UTR/downstream regions, respectively (Fig.1C and fig.S1F). Gene size was slightly different among hybrid categories (fig.S1G). As determined by global runon sequencing (GRO- seq) data (38), all Top1cc- altered hybrids, except late loss peaks, were in transcribed regions, and stable gain peaks showed a strong association with highest transcription rates (Fig. 1D). Stable hybrid gain peaks were found in open chromatin, as determined by H3K4me3 and H3K27ac histone marks (fig.S1H). Gain but not loss peaks were closer to laminaassociated domains (fig.S1I), whereas all hybrid peaks were similarly distant from heterochromatin sites, as determined by H3K9me3 (fig.S1J) [histone markers are from ENCODE database of HCT116 cells (39)]. Thus, we conclude that hybrid gains, in particular, stable ones, are at highly transcribed genes in an open chromatin context that can favor R loop formation. Consistently, significant levels of RNA polymerase II (RNAPII) and Top1 [datasets from (6)] were found at altered DNARNA hybrid peaks with highest levels found at stable gain peaks (Fig.1D and fig.S2A). Instead, Top1ccs [datasets from (6)] were at the highest levels at stable and transient gain peaks (Fig.1D and fig.S2, A and B). Consistently, considering all hybrid peaks at 5 min, Top1cc levels showed a linear association with the gain rate with respect to untreated cells (Fig.2A), strongly supporting that hybrid transient and stable gains are directly induced by Top1ccs. We notice that, in untreated cells, transient and stable loss peaks showed hybrid levels higher than gain peaks and high Top1 levels but undetectable Top1ccs (Fig.1D and fig.S2A), suggesting a very low Top1 activity. Thus, a high Top1 activity is likely critical to keep low hybrid levels in untreated cells at gain peak regions, which showed high transcription. This is consistent with a major requirement of Top1 activity at highly transcribed genes that, in turn, would prevent hybrid formation. Top1ccs are close to hybrid gain peaks To better define the relation between Top1cc and DNARNA hybrid gains, we measured the amount of Top1ccs by Top1cc K+/SDS precipitation (40, 41) and hybrid levels by DNA/RNA immunoprecipitation (DRIP)–quantitative polymerase chain reaction (qPCR) at peaks showing stable and transient gains. Top1ccs accumulated at all these sites (Fig.2B), but hybrids were barely detected when cell lysis was performed immediately after treatment (Fig. 2B). Conversely, a recovery time of 15 min before cell lysis allowed Top1cc reversion (Fig.2B), as expected (2, 3), and detection of high hybrid levels (Fig.2B). A recovery step of ~15 min was similarly included in DRIP sequencing (DRIP- seq) protocol. Stable hybrid gains could still be detected although to a significantly lower extent, whereas transient gains were fully lost with no recovery (Fig.2B), indicating that stable and transient gains can correspond to structurally different R loops. Together, the results show that increased hybrids can be destabilized by the presence of Top1ccs during the DRIP protocol, in agreement with the observation that DNA nicks can affect hybrid stability invitro (42). Thus, hybrids can be lost during the DNA extraction step of the DRIP protocol if lysis is performed immediately upon drug removal. To confirm this hypothesis, we measured hybrid levels with the R- loop- chromatin immunoprecipitation (R- ChIP) technique, which does not require DNA extraction (43). We checked R- ChIP efficiency with positive and negative loci (fig.S2C) (43). The results demonstrated that hybrids increased at all gain peak sites immediately upon Top1cc induction (Fig.2C), supporting that Top1ccs are close enough to gain peaks to affect hybrid formation in living cells and during the DRIP protocol. Proximity ligation assay (PLA) results further supported that Top1ccs are close to hybrids, as inferred from positive signals obtained with specific antibodies against Top1ccs or V5- tagged RNase H1 (Fig.2D). Thus, hybrid gains are caused locally by Top1ccs at highly transcribed genes where Top1 activity is high to maintain a low level of R loop structures and a high RNAPII elongation rate (6, 44). As Top1cc induction by CPT increases negative supercoils at promoters of transcribed genes (10), high R loop levels can be due to increased unwinding tension, particularly at 5′- ends of highly transcribed genes. However, other models are possible, particularly along the gene body, as Top1ccs may provide a free DNA end to allow DNARNA annealing (32). Stable hybrid gains are downstream of RNAPII peaks, enhanced by a dominant negative TFIIS mutation Metaplot analyses of aligned hybrid regions showed a marked increase in RNAPII levels upon Top1cc induction at sites with stable hybrid gains, but not at other regions (Fig.1D), suggesting a role for RNAPII elongation in stable hybrid increase. Thus, we next compared the relative orientation of RNAPII and Top1cc levels and hybrid regions for transient and stable hybrid gains (Fig.3A and fig.S3A). Splitting hybrids based on genic location, we found that while levels of Top1cc in transient and stable hybrid regions are similar in each peak category, transient hybrid gains peaked upstream of accumulated RNAPII for promoter- and terminatorassociated hybrids (Fig.3A and fig.S3A), as expected for nascent RNA hybridizing to the template behind the RNAPII (32, 45). In the case of gene bodyassociated transient gains, a welldefined RNAPII peak was not detected (Fig.3A). However, for stable hybrid gains at gene bodies, RNAPII peaks were fully upstream of hybrids (Fig.3A, middle) and at gene termination regions, RNAPII peaks were, in part, upstream of the hybrids for stable gains (Fig.3A, right). RNAPII increased in Top1cc induction mainly upstream of hybrids (gene body and termination regions) or at the downstream side of peaks Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 3 of 21 5′ CPT60′ CPTNT Stable gain N = 3596) Transient N = 966) Late gain N = 7049) Stable loss N = 1740) Transient N = 833) Late loss N = 2678) No Change N =11,626) 5′ CPT + RH5′ CPT 60′ CPT + RH60′ CPT NT + RHNT DRIP units DRIP 0 2 4 6 8 −3 c 3 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 −3 3−33−3 3kbp 01 2345678 B gain loss ccc Gain Late Loss Late 0255075 100 Percentage (%) Stable gain Transient gain Late gain Stable loss Transient loss Late loss No change C D Intergenic Downstream 10 kb 3′UTR Intron Exon 5′UTR Upstream 10 kb 0 5 10 15 20 −3 036 Log2 fold change 28,488 total peaks 0 5 10 15 20 −6 −3 036 Log2 fold change −6 A P value (−Log10) P value (−Log10) DRIP peaks 5′ vs. NT DRIP peaks 60′ vs. NT Gain NT 10′ CPT NT RSS 8 50 100 150 5 10 15 20 DRIPGRORNAPII 4′ CPT Top1cc Top1 1.5 2.0 2.5 3.0 3.5 NT 1.0 1.5 2 4 6 RSS 8 RSS 8 RSS 8RSS 8 RSS 8 RSS 8 Stable Transient Stable Transient Loss No Change kbp 60′ CPT NT 5′ CPT (n.r.d.)(n.r.d.)(n.r.d.)(n.r.d.)(n.r.d.)(n.r.d.)(n.r.d.) (n.r.d.)(n.r.d.) (TPM × 10³)(TPM × 10³)(TPM × 10³) Fig. 1. Top1ccs induce dynamic changes of hybrid levels at specific genomic sites in HCT116 cells. (A) volcano plots of differential analyses of DRiP- seq signals between cells treated with cPt for 5min (5′ cPt) versus untreated cells (nt) (top) and cells treated with cPt for 60min (60’ cPt) versus nt (bottom). each dot represents a DRiP- seq peak. Axis values are log2(fold changes) of DRiP- seq levels (x axis) and −log10(P values) of limma test (y axis). Red dots indicate peaks with P value less than 0.01, shown as a dashed line. (B) Metaplots (left) and heatmap (right) of DRiP- seq normalized read density (n.r.d.) in seven R loop categories distinguished by different kinetics. line colors of metaplots correspond to the means of two biological replicates as reported in the legend. heatmap colors represent DRiP levels as in legend. in both metaplots and heatmaps, “c” indicates DRiP peak center in a window of + or –3 kilo–base pairs (kbp). Rh indicates DRiP samples treated with Rnase h1 before immunoprecipitation. (C) Fractions of DRiP peaks over genomic features, as shown in the legend. (D) Metaplots of DRiP- seq (5- and 60- min treatment with cPt), GRO- seq (38), RnAPii chiP- seq (10- min treatment with cPt) (6), top1 chiP- seq (6), and top1cc- seq (4- min treatment with cPt) (6) normalized levels for each R loop kinetic category. RSS is the start site of R loop peaks. the graph is in a window of −1/+8 kbp with respect to RSS. line colors as in legend. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 4 of 21 (promoters) (Fig.3A and fig.S3A). Thus, Top1ccs can induce hybrid gains that are differently located relative to RNAPII peaks. Together, our results raise the question of whether most of stable hybrid gains correspond to R loops downstream of RNAPII, possibly following RNAPII backtracking (fig.S3B, top) (23), or, alternatively, RNAPII peaks upstream of stable hybrids may reflect stalled RNA polymerases behind a hybrid formed by the nascent RNA of another RNAPII downstream (fig. S3B, bottom). To distinguish between these two possibilities, we studied the effects of TFIIS as this factor can bind to backtracked, arrested RNAPIIs to restart transcription elongation by transcript cleavage and has been reported to prevent R loop downstream of it (46). Therefore, we expressed a dominantnegative TFIIS mutant (TFIISm; D290A and E291A) in human embryonic kidney (HEK) 293T cells (46) (Fig. 3B and fig. S3C) and measured its effects on RNAPII levels at sites of Top1cc- induced RNAPII accumulation by ChIP (fig.S3D). The results confirmed that 0.00 0.01 0.02 0.03 0.04 0.05 Loss peaks Gain peaks DRIP level fold change (5′ vs. NT) R = 0.48 P < 2.2 × 10−16 B CASP 71 YWHAG UpDDX18 NONO ELOC C1D RNF43 FAM177A1 SF3B3 CASP 71 YWHAG UpDDX18 NONO ELOC C1D RNF43 FAM177A1 SF3B3 −4 −2 024 Top1cc R - loop ** * * ** ** * * * Recovery No recovery NEAT1 TSS PMS2 TSS CLPSN TSS JUN TSS TMEM107 EED PCGF6 DCBLD2 LRRC8B RBM15 C1D RNF43 FAM177A1 UpDDX18 H1-2 TCERG1 0123 D210N WKKD RH-WT * * * ** * * C * * * * * * Top1cc levels in DRIP peaks (normalized counts) A Stable gain Transient gain Stable gain Transient gain Stable gain Transient gain D 0 5 10 15 20 Mock RH-WT D210N Mock RH-WT D210N NT CPT PLA foci (N. foci/nucleus) * ** No recovery Recovery D210N NT CPT Mock RH-WT NT CPT No recovery Recovery Log2 fold change Fold change Fig. 2. Top1ccs are close to hybrid gain peaks. (A) violin plot of top1cc levels (y axis), computed using top1cc- seq experiment, in R loop peaks divided in 15 groups on the base of 5mincPt versus untreated (nt) log2(fold change) of DRiP- seq levels (x axis). Spearman correlation coefficient and relative P values are reported in the graph. (B) heatmap showing top1cc and hybrid levels at stable and transient gain hybrid peaks in 5- min cPt–treated samples. cells were lysed directly after drug administration (no recovery) or after 15min of recovery in drugfree medium (recovery). DnA enrichment over input after K+/SDS (top1cc) and DRiP (R loop) assays was quantified by realtime PcR and reported as mean fold change (log2) of treated versus untreated samples. DRiP enrichment over input was normalized on pFc53 plasmid spikein [R loop fragment (RF) amplicon]. At least three biological replicates are reported. (C) heatmap showing R- chiP analysis at stable and transient gain hybrid peaks in 5- min cPt–treated samples. transcription start site (tSS) of neAt (neAt1 tSS), PMS2 (PMS2 tSS), clPSn (clPSn tSS), and JUn (JUn tSS) genes were used as R- chiP–positive controls (43). DnA enrichment over input was quantified by realtime PcR and reported as mean fold change (log2) of treated versus untreated samples. At least three biological replicates are reported. (D) Representative images and PlA focus analysis of untreated and 5- min cPt–treated samples. PlA of top1cc (clone1.1A, #MABe1084) with v5- tagged Rnase h1 (#ab15828) was performed. cells were fixed directly after drug administration (no recovery) or after 15min of recovery in drugfree medium (recovery). each bar represents the mean values ± SeM from two biological replicates. the average number of cells analyzed is 200. For all data, statistical significance was performed on fold change values with onetailed ratio paired t test. *P<0.05; **P<0.01. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 5 of 21 NT doxy − 5′ CPT doxy − NT doxy + 5′ CPT doxy + * * * ** ** B A TFIIS FLAG C -1 0 1 −0.5 1.5 −0.4 −0.6 0.3 0.9 0.2 −1.4 -0.1 0.8 0.7 −1.4 0.2 0.9 0.3 −1.4 0.2 0.4 0.9 −1.4 0.0 1.0 0.4 −1.4 −0.4 0.9 0.8 −1.2 −0.5 1.3 0.3 −1.1 0.2 0.7 0.6 −1.5 −0.3 1.3 0.0 −1.1 RNAPII (Z score) CLTA RFT1 UBE2G2 TMEM209 ATP5MC2 DPM1 PTCD1 AREL1 AP1G1 NIN Stable gain Transient gain − CPT − 1 0 1 2 + − + Doxy − − + + 0.8 −0.9 1.0 −0.8 1.1 −0.8 0.6 −0.9 1.1 −0.8 0.6 −0−9 1.3 −0.6 0.2 −1.0 1.4 −0.6 0.1 −0.8 1.4 −0.4 −0.2 − 0.8 0.6 −0.9 1.1 −0−8 0.9 −0.9 0.8 −0−8 0.8 −0.9 0.9 −0.9 0.9 −1.0 0.8 −0.8 PTCD1 AREL1 AP1G1 NIN Stable gain Transient gain CPT Doxy TFIIS (Z score) D 55 kDa 35 kDa 55 kDa 35 kDa 24 48 TFIISm Doxy −++ Hours 0 CLTA RFT1 UBE2G2 TMEM209 ATP5MC2 DPM1 0 2 4 6 8 0 10 20 30 5 10 15 20 5 10 15 20 DRIPRNAPII Promoter-associated Peaks Genebody-associated Peaks Terminator-associated Peaks 5′ CPT 60′ CPT NT 10′ CPT NT Stable GAIN (n = 1303) Transient GAIN (n = 140) Stable GAIN (n = 1458) Transient GAIN (n = 492) Stable GAIN (n = 367) Transient GAIN (n = 73) 0 2 4 6 8 0 2 4 6 8 1.0 1.5 2.0 −3 8−38−3 8−38−3 8− 8RSS RSS RSS RSS RSS RSS kbp 1.00 1.25 1.50 1.75 2.00 1.0 1.5 2.0 2.5 Top1cc 4′ CPT 0.00 0.02 0.04 0.06 0.08 Relative DRIP signal (normalized on spike-in) ELOCNONOUpDDX18YWHAGCASP7 * ** * SF3B3FAM177A1RNF43C1D Stable gain * * * ** * * * * * * Transient gain E − + − +− − + + * **** ** ** * * 4′ CPT (random) (TPM × 10³) (TPM × 10³) (n.r.d.) Fig. 3. Stable hybrid gains are downstream of RNAPII peaks. (A) Metaplots of DRiP- seq, RnAPii chiP- seq, and top1cc- seq normalized levels at stable and transient R loop gain divided in promoterassociated (up to −10 kbp from promoter and 5′UtR), gene body–associated (located in exon and intron) and terminatorassociated (located in 3′UtR and downstream 10 kbp) regions of genes. Dashed lines in top1cc- seq metaplots show background levels of top1cc after R loop gain randomization. RSS represents the R loop peak start site, in a window of −1/+8 kbp. line colors as in legend. (B) expression by Western blot of the dominant negative tFiiSm upon doxycycline (doxy) exposure (+) in heK293 cells. (C and D) heatmaps showing RnAPii and tFiiS levels at selected loci, respectively, determined by chiP (antibodies sc- 47701 and ab185947, respectively). the tested loci overlap with top1cc- induced RnAPii accumulation and the indicated hybrid gain regions, which are between 300 and 500 bp downstream of the tSS of the indicated genes. heK293 cells were treated with cPt for 10min. tFiiSm was induced for 48 hours with doxycycline. DnA enrichment over input was quantified by realtime PcR and reported as Z score (shown in each cell). Data are from at least three biological replicates. Statistical significance was calculated by comparing treated over nontreated samples using multiple unpaired t test. (E) R loop levels determined by DRiP- qPcR at indicated loci in tFiiSm induced and uninduced heK293 cells. DRiP enrichment over input was normalized on pFc53 plasmid spikein (RF amplicon). each bar represents mean values ± SeM of at least three biological replicates. Statistical tests were performed with onetailed ratio paired t test. in all panels, *P<0.05; **P<0.01. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 6 of 21 Top1ccs induced an accumulation of RNAPII at the tested sites located at 300 to 500base pairs (bp) downstream of the transcription start site (TSS) of genes with stable or transient hybrid gains (Fig.3C and fig.S3D), but not at silent genes or untranscribed sites (fig.S3D). Top1ccs induced a higher accumulation of RNAPII at almost all stable, but not transient, hybrid gains upon TFIISm expression (Fig.3C), showing that RNAPII accumulation can be attenuated by wildtype (WT) TFIIS. Using an antibody against the WT factor, TFIIS accumulated at all the tested sites upon TFIISm expression in CPT- untreated cells (Fig.3D and fig.S3E), in agreement with published data showing higher recruitment of TFIIS within 1000 bp from TSS in the presence of expressed TFIISm [see figure 1F in (47)]. Top1ccs induced higher TFIIS levels in cells not expressing TFIISm, whereas they decreased TFIIS levels upon TFIISm expression (Fig. 3D and fig. S3E). In addition, TFIISm expression increased hybrid levels markedly at stable gains, but very slightly at transient gains, either in untreated or CPT- treated cells (Fig.3E and fig. S3F), suggesting that WT TFIIS activity can reduce stable Top1cc- induced R loops. Thus, accumulation of RNAPII or WT TFIIS at sites of Top1ccs following TFIISm expression or CPT treatment, respectively, shows that transcript cleavage by WT TFIIS can restart transcription elongation and reduce Top1- induced RNAPII accumulation and R loop levels at genes with stable hybrid gains, suggesting a role for RNAPII backtracking in R loop increase. Top1ccs cause DSBs downstream of accumulated RNAPII at highly transcribed genes Next, to get insights into the mechanism of TRCs induced by Top1ccs, we determined genomic sites of DSBs after Top1cc induction using END- seq technique (48). To determine DSBs immediately induced by Top1ccs, we mapped DSBs after 10 and 20min of CPT treatment since DSBs can already be detected at these short times [see figure2 in (18) and fig.S4, A and B]. We performed two biological replicates in HCT116 cells comparing treated to untreated cells (tableS1). After read normalization over library size and spikein and peak calling for each replicate, we considered strandspecific peaks that were closer than 150 bp to each other as a signal of a DSB (fig. S4C). Then, overlapping DSB sites present in both replicates were merged defining a DSB cluster (fig.S4, D and E; see also Materials and Methods for full description). Since Top1ccs likely revert during the initial cellhandling steps of the END- seq procedure (see Fig.2B) (2, 3), DSBs are not primarily caused by a Top1cc in one strand and an endonucleasedependent nick on the other. Read profiles of metaplots of DSB clusters were consistent with END- seq procedures and indicated that a DSB on average was present at the center of the cluster (fig.S4F). However, the clusters were heterogenous including one or a few paired strandspecific peaks (fig.S4G). We identified three kinetic groups of DSB clusters: 1243 DSB clusters were persistent, whereas 398 and 358 were transient and late, respectively (Fig.4A and fig.S4, E and H). In addition, we also detected isolated peaks in one strand only at 10 and 20min, which are consistent with singleended DSBs (seDSBs) (fig.S5, A and B). Genomic distribution of persistent DSB clusters correlated well with chromosome gene density (Fig.4B), consistent with a genic localization of almost all of them (Fig.4C). However, epigenomic annotations were different among persistent, late, and transient DSB clusters (Fig. 4D). In particular, transient DSB clusters were prevalently found in silent or repressed chromatin, as determined by specific histone marks and deoxyribonuclease (DNase) I hypersensitivity (Fig.4D). Top1cc levels [mapped previously in HCT116 cells (6)] were only slightly higher for persistent DSBs supporting a causative role of them in DSB formation for all the three kinetic groups (Fig.4E). Persistent DSB clusters were instead associated with higher transcription levels and RNAPII density [mapped previously in HCT116 cells (6); Fig. 4E]. Persistent and late DSB clusters were found downstream of RNAPII peaks, whereas they coincided precisely with GRO- seq [mapped previously in HCT116 cells (38)] and Top1cc signal peaks (Fig.4E). As Top1cc induction by CPT causes an increase in RNAPII density at promoterproximal pause sites (6), these findings show that persistent DSBs likely form at specific sites of highly transcribed genes downstream of accumulated RNAPII. Top1cc- induced DSBs are enriched at stable, but not transient, hybrid gains downstream of RNAPII peaks Next, to assess whether persistent DSBs were associated with R loops, we computed overlap enrichments and found that DSB clusters were strongly and significantly associated with stable gains only (Fig.5A, n=477). In addition, seDSB peaks were also enriched with stable gains (fig.S5C). Stable gains overlapping (n=444) with at least one DSB cluster (DSB+ stable gains) had only slightly higher hybrid levels than those not overlapping (n=3108) any DSB (DSB− stable gains) (fig.S5D). However, they showed a much higher transcription rate, RNAPII accumulation, and Top1cc signals (fig.S5, E to G). Thus, DSB+ stable gains were marked by an elevated transcription suggesting a mechanism linking high transcription to DNA cleavage induced by Top1ccs (13). To determine the relative position of hybrids, DSBs, and RNAPII, we made metaplot analyses aligning the corresponding genes at the TSS and transcription end site (TES) comparing stable to transient hybrid gains. RNAPII peaks were localized close to TSS and overlapping with hybrids for both stable and transient gains (Fig.5B). Stable gains were downstream of RNAPII, whereas transient hybrid levels were mainly overlapping with and upstream of RNAPII (Fig.5, B and C, and fig.S6A). RNAPII increase was higher downstream to TSS, suggesting an accumulation of RNAPII induced by Top1ccs (Fig.5B). Last, DSB tag profiles showed elevated levels in genes with stable gains, showing that DSB clusters, on average, localized remarkably close and downstream of accumulated RNAPII and overlapped with stable hybrid gains (Fig.5D). Therefore, although our DRIP protocol used enzyme digestion to fragment genomic DNA resulting in an average peak size of 1700 bp (fig.S1E) (49), metaplot analyses are consistent with stable hybrid gains being downstream of accumulated RNAPII and often overlapping with DSB clusters. Persistent DSB clusters are enriched at early replicating initiation zones and topologically associating domain boundaries As Top1cc- induced DNA cleavage is known to occur mainly in S phase (2, 3), we next investigated the relationships of persistent DSB clusters with replication origins and timing using short nascent DNA strand (SNS)–seq (50) and repliseq (51) annotations, respectively. As high SNS signals [from (22)] correspond to replication origin sites (22, 50), persistent DSB clusters did not overlap with replication origins as the SNS signal was depleted (fig.S6B). This is consistent with persistent DSB clusters localized in gene bodies (Fig.4C) where replication origins are suppressed by transcription (50, 52–55). Repliseq annotations show the timing of replication of large replicating initiation zones (IZs) commonly grouped into early, early/mid, and Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 7 of 21 R=0.8 chr1 chr12 chr17 chr3 chr8 100 150 chr10 chr11 chr13 chr14 chr15 chr16 chr18 chr19 chr2 chr20 chr21 chr22 chr4 chr5 chr6 chr7 chr9 chrX 0 50 0 2000 4000 Number of genes per chromosome Number of persistent DSB peaks per chromosome 68101214161820 4 6 8 10 12 14 10 12 14 Transient DSB clusters (n = 398) −3 c 4 6 8 4 6 8 10 12 14 3 Late DSB clusters (n = 358) Persistent DSB clusters (n = 1243) −3 c3−3 c3−3 c3 kbp Normalized DSB units 0 100 200 300 400 500 10′ CPT NT RNAPII −5 c 5 0 0.5 1 1.5 2 2.5 3 −5 c 5 −5 c 5 kbp 4′ CPT Transient DSB clusters (n = 323) Late DSB clusters (n = 332) Persistent DSB clusters (n = 1199) NT 0 10 20 30 40 0 20 40 60 80 H3K4me3 0 10 20 30 40 50 60 70 H3K4me2 0.5 1 1.5 2 2.5 3 H3K4me1 0 0 10 20 30 40 H3K27ac −10 c 10 0 0.05 0.10 0.15 0.20 0.25 0.30 −10 c 10 −10 c 10 kbp H3K9me2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Transient DSB clusters (n = 323) Late DSB clusters (n = 332) Persistent DSB clusters (n = 1199) Transient Late Persistent 0255075 100 Percentage (%) Upstream 10 kb 5′UTR Exon Intron 3′UTR Downstream 10 kb Intergenic GROTOP1cc DNase-seq NT 10′ CPT 20′ CPT NT 10′ CPT 20′ CPT A B C E D (n.r.d.) (n.r.d.)(n.r.d.) (n.r.d.) (TPM × 10³)(TPM × 10³) (signal P value)(signal P value)(signal P value)(signal P value)(signal P value)(n.r.d.) END P = 5.2 × 10−6 Fig. 4. Top1ccs induce DSB clusters after short times in HCT116 cells. (A) Metaplot (left) and heat map (right) of enD- seq levels normalized read density at transient, late, and persistent DSB regions, showing kinetics of DSBs induced by cPtmediated top1ccs at control conditions (nt) and 10min (10′ cPt) and 20min (20′ cPt) of treatment. line colors as in legend. “c” indicates the center of the region in a window of ±3 kbp. All signals are reported as means of two biological replicates normalized levels. (B) Scatter plot of correlation between number of genes (y axis) and number of persistent DSBs (x axis) per chromosome. Spearman correlation coefficient and P values are reported. (C) Proportion of transient, late, and persistent DSBs over genomic features, as percentage. colors as in legend. (D) Metaplot showing chiP- seq normalized signals of euchromatin histone markers (h3K4me3, h3K4me2, h3K4me1, and h3K27ac), heterochromatin histone marker h3K9me2, and Dnaseseq signals (39) at transient, late, and persistent DSB clusters, excluding those localized at intergenic loci. histone marker chiP- seq signal is reported as −log10(P value). Dnaseseq signals are reported as normalized readdepth levels. “c” indicates the center of the region in a window of ±10 kbp. (E) Metaplot showing GRO- seq (nontreated), RnAPii chiP- seq (nontreated and 10- min cPttreated), and top1cc- seq (4- min cPt) normalized levels at transient, late, and persistent DSB regions, excluding regions localized at intergenic loci. line colors as in legend. “c” indicates the center of the region in a window of ±5 kbp. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 8 of 21 late IZs (50, 51). The analyses showed that persistent DSB clusters and seDSBs were strongly and highly enriched at early IZs (fig.S6, C and D, respectively), indicating that significant levels of DNA damage triggered by Top1ccs occur during early S phase. We also observed a significant enrichment of all hybrid kinetic categories, but not transient gains, at early IZs (fig.S6E), suggesting that early IZs are enriched with DNARNA hybrids. Next, we wondered whether stable hybrid gains with persistent DSB clusters had a preferential genomic location. The results show that DSB+ stable gains (263) were strongly enriched at early or mid IZs (51), whereas stable gains without a DSB cluster were depleted (Fig.6A). Since topologically associating domain (TAD) boundaries can overlap early replication regions (51, 56), we also measured the enrichments of DSBs and hybrid gains at these chromatin elements. Class 1 TAD boundaries, which have been shown to be enriched at early replicating IZs (51), showed a significant enrichment of R loops, which are not enriched at other TAD boundary classes (fig.S6F). Moreover, Top1cc induction overall decreased hybrid levels at class 1 TADs (fig.S7, A and B). Enrichment analyses showed that DSB+ stable gains (146), but not DSB- stable gains, were clearly associated with early or early/mid IZs of class 1 TAD boundaries (51) (Fig. 6B). In contrast, transient gains with or without DSBs were not enriched in any IZ or TAD class (Fig.6, A and B). Metagene analyses of DSB+/− stable gains overlapping early versus early/ mid IZs found in class 1 TAD boundaries showed that hybrid levels were increased by Top1ccs, and their levels were higher in early versus early/mid IZs (fig.S7C). RNAPII density and DSB levels also increased in Top1cc induction and both peaked close to gene 477 25 120 45 24 8 180 0 50 100 150 Stable Gain Late No change A Enriched Depleted D TSS TES 2 3 4 5 0 2 4 6 8 10 12 Stable gain genes (n = 3300) Transient gain genes (n = 347) 25 P value(−log10) 10′ CPT20′ CPTNT TSS TES −1.51.5 −1.5 1.5 kb END 10′ CPT 20′ CPT NT 1 2 3 4 5 5′ CPT60′ CPTNT 0 25 50 75 100 TSS TES TSS TES B Stable gain genes (n = 3300) Transient gain genes (n = 347) Stable gain genes (n = 3300) Transient gain genes (n = 347) C 10′ CPT NT −1.5 1.5 −1.51.5 kb 10′ CPT NT 5′ CPT 60′ CPT NT −1.5 1.5 −1.51.5 kb TSS TES TSS TES 0 20 40 60 80 100 RNAPII DRIP Loss Transient Stable Late Transient 2 4 6 8 10 0 (TPM × 10³) (n.r.d.) (n.r.d.) Fig. 5. Top1cc- induced DSBs are enriched at stable hybrid gains and are localized immediately downstream of accumulated RNAPII peaks. (A) enrichment analysis of observed persistent DSBs at DRiP categories versus expected by genome randomization (n=100). Bar plots report binomial test P values (y axis) of enrichment. − log10(P values) are referred as “enriched” for a positive overlap enrichment over expected or “depleted” for a depletion over expected. the number of observed overlaps at each category is reported. (B to D) Metaplots of RnAPii chiP- seq (B), DRiP- seq (c), and enD- seq (D) in genes with annotated stable and transient gain DRiP peaks in a window of ±1.5 kbp. teS, transcription end site. the dashed lines downstream of tSS indicate a 1000- bp unscaled region. the shaded yellow area downstream of the unscaled region indicates a stable gain hybrid region anterior to RnAPii accumulation. All signals are reported as normalized levels. line colors as in the legend. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 9 of 21 AB 153 110 11 7 4 1 0 50 100 150 200 310 110 20 456 216 30 0 50 100 150 200 25 25 96 50 514 21 1 1 1 5 4 0 40 80 120 32 6 2516 2 1 3 1 82 11 3933 3 1 41 1 0 40 80 120 20 20 Enriched Depleted Enriched Depleted Gain with DSB 181 272 Early IZ Early/mid IZ Late IZ Stable Transient Gain w/o DSB Stable Transient 123456 123456 Stable Transient Gain with DS BG ain w/o DSB 123456 123456 Stable Transient Early IZ Early/mid IZ Late IZ D F Late G1 phase NT CPT CPT-RH PDS − − ** 0 2 4 6 8 10 Fold change on NT − *** + G NT CPT CPT-RH PDS Early/mid S phase Fold change on NT ++ − +− + − − 1 ** **** ******* ** *** 0 2 4 6 8 0 MNi EdU MNi BrdU ++− +−+ − − 0 10 20 30 40 * *** **** ** ** Drug EdU Hours MNi detection 00.5 1.5 2.5 24 Drug EdU Hours MNi detection 00.5 1.5 2.5 245 BrdU H Late S phase − −− + MNi EdU MNi BrdU 0 10 20 30 40 50 60 ** Fold change on NT NT CPT CPT-RH PDS 0 2 4 6 8 10 ** ** + −− − * E DAPI BrdU EdU Early/mid S phase Non S phase Late S phase Late G 1 phase 0 5 10 15 % PLA-positive cells ≥ 3 foci Doxy +− NT 10′ CPT 60′ CPT Ctrl RNAPII Ctrl PCNA * * ** 1.6 3.2 3.5 3.8 5.3 1.9 2.3 * * C EdU+ EdU− * 0 2 4 6 γH2AX mean fluorescence (fold chnage on NT) I γH2AX mean fluorescence (fold change on NT) Early/mid S phase Late S phase Late G 1 phase ** 0 2 4 6 * NT CPT J NT CPT HEK293 HeLa HeLa HeLa HeLa HeLa − 0 10 20 30 *** + MNi BrdU MNi EdU −− * Micronuclei/100 cells Micronuclei/100 cells Micronuclei/100 cells P value (−log10) P value (−log10) Fig. 6. Top1cc- induced DSBs are enriched at early replicating IZs during early/mid S phase causing micronucleus formation. (A and B) enrichment analysis of observed R loop gain with and without (w/o) DSBs at early, early/mid, and late iZs (A) and at early, early/mid, and late iZs and in tAD classes (B) assessed as in Fig.5A. (C) PlA assay of PcnA with RnAPii in cells expressing (+ doxy) or not (− doxy) tFiiSm. ctRl, control samples using the indicated antibody only. Fold changes of cPttreated versus untreated and induced versus uninduced samples are reported over each bar. Mean of analyzed cells is 800. Statistical significance was calculated with onetailed paired t test. (D) experimental protocol of 1- hour cPt treatment with single (top) or dual (bottom) labeling of hela cells. (E) Representative images of the possible cases of dual labeled cells. (F to H) levels of micronuclei triggered in late G1, late S, or early/mid S phase [as in (e)], respectively. Micronucleus levels are reported as micronuclei (Mni) per 100 cells (left) or as fold change in Rnase h1–expressing (Rh) versus control cells (right). Means of analyzed cells are 250 (F), 150 (G), and 400 (h). Statistical significance was calculated comparing micronucleus distribution of treated over untreated samples with twotailed Mann- Whitney test. (I and J) γh2AX detected by immunofluorescence microscopy at the end of edU administration (i) or at the end of BrdU administration (J) as in (D). levels of nuclear fluorescence are reported as fold change over untreated cells (number of analyzed cells is 450). Statistical significance was calculated considering treated over untreated samples with onetailed ratio paired t test. in (c) and (F) to (J), bars show means ± SeM of three biological replicates. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 16 of 21 For neutral comet, lysis buffer was drained, and slides were immersed in neutral electrophoresis buffer [100 mM tris (pH 9) and 300 mM sodium acetate] for 30min at RT; the electrophoresis was performed at 4°C for 45min (21 V). Slides were immersed in DNA precipitation solution (1 M ammonium acetate in 95% ethanol solution) for 30min at RT in 70% ethanol for additional 30min and lastly dried at 37°C for 15 min. For both procedures, dried agarose circles were stained using SYBR Gold Nucleic Acid Gel Stain (Thermo Fisher Scientific, #S11494) for 30min at RT and lastly rinsed in water. Slides were visualized by Nikon Eclipse 90i Microscope. Tail moment and DNA in tail quantification analysis were performed using ImageJ OpenComet software. Western blot Cells, collected and resuspended in Laemmli buffer (4% SDS, 20% glycerol, 0.125 M tris, and 1× protease inhibitor), were sonicated for 20 min with Bioruptor Sonicator (Diagenode) and quantified by Lowry method. After incubation at 100°C for 10min, samples were loaded in Bolt 4 to 12% bistris plus (Thermo Fisher Scientific, #NW04120BOX) and transferred onto nitrocellulose membrane (Thermo Fisher Scientific, #88018). Proteins transfer was confirmed by Ponceau staining (RNase H1 Western blot) or by No- Stain Protein Labeling Reagent (Invitrogen, #A44449) according to the manufacturer’s instructions (TFIIS western blots). Membrane was blocked with trisbuffered saline (20 mM tris and 150 mM NaCl)– 0.1% Tween and 5% BSA for 1 hour at RT and incubated with the following antibodies O/N at 4°C: anti- DYKDDDK antibody (1:1000; Cell Signaling Technology, #2368), anti- Top1 (1:1000; Santa Cruz Biotechnology, catalog no. sc- 5342), and anti- TFIIS (1:1000; Abcam, ab185947). After three washes in trisbuffered saline–0.1% Tween, membrane was incubated for 1 hour at RT with secondary antibodies: antirabbit (Abcam, #ab205718), antimouse (Invitrogen, #sc2005), and antigoat (Invitrogen, #sc2922). Membrane was captured by using enhanced chemiluminescence (Thermo Fisher Scientific, #32132) at ChemiDoc imaging system (Bio- Rad). Immunoblot of BRCA2 was carried out as described in (69). Micronucleus detection For micronucleus detection assay, HeLa, U2OS or U2OS- RH, HEK293m, and HCT116 cells were seeded at a density of 2 × 105 cells per well in 35- mm dishes containing a 24- mm by 24- mm cover glass. After 24hours, cells were treated with 10 μM and/or 25 μM drugs in fresh medium for 60min. To distinguish cells in the S phase, cells were incubated with 10 μM EdU 30min before, during the treatment, and 1 hour later for a total of 2.5 hours. To determine cell cycle subpopulations other than S phase, immediately after drug administration and EdU incubation, cells were washed with 1× PBS and incubated with 10 μM BrdU (Thermo Fisher Scientific, #B23151) solution for further 2.5 hours. Upon drug and EdU/BrdU administration, cells were leaved in drugfree medium for further 24 hours for micronucleus detection at the next mitosis (although CPT induces S phase blockage, 24 hours of recovery ensures that ~90% of cells have passed through the cell division phase). At this point, cells were fixed with formaldehyde 4% in 1× PBS for 15min at room RT, washed twice with 3% BSA in 1× PBS (3% BSA/PBS), and permeabilized by 20min incubation in 0.5% Triton X- 100 in 1× PBS at RT. Cells were then washed once with 3% BSA/PBS, and EdU detection was performed by ClickiT EdU Assay (Thermo Fisher Scientific, #C10339). After 30min of incubation at RT, the reaction cocktail was removed, and slides were washed with 1ml of 3% BSA/ PBS. Last, cells were incubated with a 4′,6- diamidino- 2- phenylindole (DAPI) solution (3.3 ng/μl in water) for 30min and then washed with water before being mounted with Mowiol upside down on microscope slides. For BrdU staining, cells were denatured immediately before EdU staining. Denaturation protocol consists in 8min of incubation with 4 N of HCl at 25°C, two washing steps (5min each) with 1× PBS, 5min of incubation in phosphate/citric acid (pH 7.4) (0.2 M Na2H- PO4 and 0.1 M citric acid) and four more washing steps (two with 1× PBS and two with 3% BSA/PBS). In this case, after denaturation and EdU staining, cells were blocked with 3% BSA/PBS for 1hour at RT and subsequently incubated O/N at 4°C with an anti- BrdU antibody (Thermo Fisher Scientific, #B35130) diluted 1:200 in blocking buffer plus 0.1% Triton X- 100. Cells were then incubated with antimouse Alexa Fluor 488 secondary antibody (Thermo Fisher Scientific, #A11001) diluted 1:500 in 3% BSA/PBS for 1hour at RT. After incubation, slides were washed with 3% BSA/PBS and incubated with a DAPI solution (3.3 ng/μl in water) for 30min. A final washing step in water was performed before mounting slides with Mowiol upside down on microscope slides. When micronucleus analysis was performed in cells overexpressing RNase H1, the overexpression of the protein was assessed by immunofluorescence. For HeLa cells transfected with pcDNA3– RNase H1 plasmid, cells were immunolabeled for hemagglutinintag before BrdU staining. Cells were incubated for 1hour with an anti–hemagglutinintag antibody (Cell Signaling Technology, #3724) diluted 1:1000 in 3% BSA/PBS at RT. After washing in 3% BSA/PBS, cells were incubated for 1 hour with antirabbit Alexa Fluor 594 secondary antibody (Thermo Fisher Scientific, #A11037) diluted 1:1000 in 3% BSA/PBS. Cells were washed and then stained for BrdU detection. For U2OS- RH cell line, CPT and EdU were administrated to cells previously induced with doxycycline for 48hours. Fixation, permeabilization, and EdU detection were performed as described above. To detect RNase H1, a primary antibody against a FLAG- tag fused to the enzyme was used. Cells were incubated O/N at 4°C with anti- DYKDDDK antibody (Cell Signaling Technology, #2368) diluted 1:800 in blocking buffer. After washing with 1× PBS, cells were incubated with antirabbit Alexa Fluor 488 secondary antibody (Thermo Fisher Scientific, #A11008) in blocking buffer for 1 hour at RT. Next, DAPI staining and mounting were performed as described above. For micronucleus detection in TFIISmexpressing HEK293 cell line, HEK293 cells were seeded at a density of 2 × 105 cells per well in 35- mm dishes containing a 24- mm by 24- mm polyl- lysine–coated cover glass. After treatment with doxycycline for 48 hours to induce TFIISm expression, 10 or 25 μM CPT and EdU/BrdU were administrated to cells, and immunofluorescence assay was performed as described above. Anaphase bridges and lagging chromosome detection For the detection of anaphase bridges and lagging chromosomes, HeLa, U2OS, or U2OS- RH cells were seeded at a density of 1 × 106 cells in T25 flasks. After 24 hours, cells were treated with 10μM CPT in fresh medium for 60min and labeled with EdU as previously described. After EdU administration, cells were incubated for 20 hours with nocodazole (50 ng/ml; Merck, #M1404). At the end, G2- M cells were collected by mitotic shakeoff and released in fresh medium for further 50 min in 35- mm dishes containing a 24- mm Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 17 of 21 by 24- mm polyl- lysine–coated cover glass. At this point, anaphase cells were fixed for 15min with PTEMF buffer [20 mM Pipes (pH 6.8), 1 mM MgCl2, 10 mM EGTA, 0.4% Triton X- 100, and 8% paraformaldehyde) at RT, washed with 1× PBS, and then blocked with 3% BSA in 1× PBST (1× PBS and 0.5% Triton X- 100) at 4°C O/N. Primary antibody incubation was performed O/N at 4°C using an anti- PICH (Millipore, #04- 1540) or an anti- RPA70 antibody (Abcam, #ab79398) respectively diluted 1:100 and 1:750 in blocking buffer. Antimouse Alexa Fluor 488 secondary antibody (Thermo Fisher Scientific, #A11001) was used for secondary antibody incubation (1 hour) and diluted 1:1000 in blocking buffer. After primary or secondary antibodies incubation, slides were washed with 3% BSA/PBS. DAPI staining was then performed as previously described. γH2AX detection For γH2AX detection, HeLa cells were seeded at a density of 2 × 105 cells per well in 35- mm dishes containing a 24- mm by 24- mm cover glass. After 24hours, cells were treated with 10 μM CPT in fresh medium for 60min. Cells were fixed immediately after EdU (and BrdU, where indicated) administration. EdU staining protocol was performed as described above. Cells were then blocked for 30min with 8% BSA/PBS at RT (gently rocking) and incubated for 2 hours at RT with antiγH2AX antibody (Millipore, #05- 636) diluted 1:1000 in 1% BSA/PBS. After primary antibody incubation, slides were washed with 1× PBS and incubated for 1 hour with antimouse Alexa Fluor 488 secondary antibody (Thermo Fisher Scientific, #A11001) diluted 1:1000 in the same buffer used for primary antibody. In case of EdU/BrdU dual labeling, Alexa Fluor 647 azide (Thermo Fisher Scientific, #A10277) was used instead of Alexa Flu- or 594 azide included in the ClickiT EdU Assay kit. BrdU staining was performed after γH2AX staining by using only primary antibody conjugated with Alexa Fluor 488. DAPI staining was then performed as previously described. pATM and pATR detection For S1981- phospjorylated ATM (pATM) and S428- phosphorylated ATR (pATR) detection, 2.5 × 105 of TFIISmexpressing HEK293 cells were seeded in 35- mm dishes containing a 24- mm by 24- mm cover glass. After TFIISm induction, cells were treated with 10 μM CPT in fresh medium for 10 and 60min. At this point, cells were fixed with 4% formaldehyde in 1× PBS for 15min at RT, washed twice with 1× PBS, and permeabilized by a 20- min incubation in 0.5% Triton X- 100 in 1× PBS at RT. Cells were then blocked for 1 hour with 1% BSA/PBS and 5% BSA/PBS, respectively, for pATM and pATR staining at RT (gently rocking). After permeabilization, cells were incubated for 2 hours at RT with antipATM (S1981) (Santa Cruz Biotechnology, sc- 47739) or antipATR (S428) (Cell Signaling Technology, #2853) diluted 1:250 and 1:100, respectively, in blocking buffer. Slides were then washed with 0.1% PBST (0.1% Tween 20 in 1× PBS) and incubated for 1 hour with antimouse Alexa Fluor 488 secondary antibody (Thermo Fisher Scientific, #A11001) or antirabbit Alexa Fluor 594 secondary antibody (Thermo Fisher Scientific, #A11037) diluted 1:1000 in the same buffer used for primary antibody. After three washes with 0.1% PBST, DAPI staining was performed as previously described. Cells were observed with Nikon A1R Confocal Microscope (Eclipse Ti2), and pATM/pATR mean fluorescence was quantified using ImageJ software. Image analysis and representation Fluorescent images, if not differently specified, were acquired by using fluorescence microscope Eclipse 90i (Nikon) and then analyzed with the ImageJ software. Cells were classified in S phase or non–S phase according to their positivity to EdU staining. In case of EdU/BrdU dual labeling, cells were divided into three main classes: G1 or late G1 phase when only BrdU was incorporated; late G1 or S phase as cells entering S phase incorporated both thymidine analogs; and late S or early G2 phase when only EdU was incorporated. Cells that were not replicating during either pulse were conegative for both EdU and BrdU staining. To perform micronucleus or bridge analysis, micronuclei and bridges were counted for each captured image. As already done for cells, micronuclei and bridges were also classified in EdU+/EdU− and BrdU+/BrdU− according to their positivity to EdU and BrdU staining. Micronucleus/bridge data were reported as number of micronuclei/bridges per 100 cells for each cell class and normalized or not with respect to the untreated control. For γH2AX, pATM, and pATR signal, fluorescence mean value was measured for each cell and backgroundsubtracted. In addition, γH2AX signal increase was evaluated considering the different classes of cells previously described. Data were plotted using GraphPad Prism version 9. Statistical analysis was carried out using GraphPad Prism 9. Proximity ligation assay PLA was carried out by the Duolink In Situ PLA Kit (Merck, #DUO92101- 1KT) according to datasheet instructions. Briefly, 2 × 105 of HCT116 cells were seeded in 35- mm dishes containing an 18- mm by 18- mm glass slide. After 24 hours, cells were treated with 10 μM CPT in fresh medium for 5min and fixed with 4% formaldehyde in 1× PBS for 15min at RT. After washing with 1× PBS, cells were permeabilized by incubating them with 0.1% Triton X- 100 in 1× PBS three times for 5min at RT (gently rocking). Anti- Top1cc (Merck, #MABE1084) and anti–tag- V5 (Abcam, #ab15828) antibodies were diluted 1:2000 and 1:3000, respectively, and incubation was performed O/N at 4°C. DAPI staining was performed as described in the “Micronucleus detection” section. For PCNA and RNAPII PLA assay, 2 × 105 of TFIISmexpressing HEK293 cells were seeded in 35- mm dishes containing an 18- mm by 18- mm poly- llysine–coated glass slide. After TFIISm induction, cells were treated with 10 μM CPT in fresh medium for 10 and 60min. For transcription or replication inhibition, cells were treated with 1 μM triptolide, cordycepin (12.5 μg/ml), or 1 μM aphidicolin for 2 hours prior the 10- min CPT treatment. Fixation and permeabilization were performed as described above. Anti- PCNA (PC11; sc- 53407) and anti- RNAPII (H- 224; sc- 9001) antibodies were diluted 1:100, and incubation was performed O/N at 4°C. All the incubations were done by putting each slide upside down on 40μl of each solution previously placed on a parafilm layer. Cells were observed with Nikon A1R Confocal Microscope (Eclipse Ti2), and the number of PLA foci was quantified using ImageJ software (analysis protocol available at https://microscopy.duke.edu/guides/countnuclear- foci- ImageJ). Bioinformatic analysis DRIP- seq analysis DRIP- seq libraries were quality checked using Fastqc, and reads were trimmed using Cutadapt (73). Reads were aligned on both human genome (hg19) and spikein yeast genome (sacCer3) using Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 18 of 21 Burrow- Wheeler Aligner (74). Alignment sorting, filtering for ENCODE blacklist (75) region, and duplicated reads removal were performed using SAMtools (76) and Picard (tableS1). Peak calling and genomic signal computation for each library were performed using macs2 using paired IP- Input libraries as - t and - c, respectively (77). Only peaks that were present in both replicates for at least one biological condition were considered for further analysis. Read counts computation for each peak was performed using bedtools (78) multicov command. Reads counts were normalized on the base of library dimension and percentage of spikein reads in each library. Differential analysis was performed with limma R (version 2.30.0) library (79). First, we performed a comparison between DRIP- seq libraries of immunoprecipitated R loop– and RNase H1– treated libraries to select only truepositive R loop regions. Then, we performed differential analysis between each treatment condition and control. DRIP- seq signal tracks for each condition were created using deepTools bamCoverage function, using the normalization scaling factor computed using both library dimension and spikein fraction for each library. DRIP- seq peak annotation was performed using DROPA (80) with standard settings. Laminassociated domain coordinates were downloaded from UCSC table browser. R loop peak distance from laminaassociated domains was performed using bedtools closest command. Genomic mapping for Top1- seq, Top1 ChIP- seq, RNAPII ChIP- seq, GRO- seq, and SNS- seq were downloaded as processed and normalized (as in figure legends) bigwig and Wig files. Top1cc levels (Top1- seq), Top1 ChIP- seq, and RNAPII ChIP- seq Wig files were converted to bigWig files using rtracklayer R library (81). To calculate read counts of Top1cc over R loop peaks, Top1- seq libraries were mapped to hg19 human genome using bowtie2 aligner, and read counts computation was performed using bedtools multicov command. Plots of DRIP- seq, Top1- seq, Top1 ChIP- seq, RNAPII ChIP- seq, GRO- seq, and SNS- seq levels over genomic features were performed using deepTools (82) computeMatrix and plotHeatmap functions. TAD boundaries and replication IZs were converted from hg38 to hg19 human genome version using liftOver UCSC tool. Enrichment analysis of R loop regions over TAD boundaries, replication IZs, and END- seq peaks was performed using the command bedtools shuffle and intersect commands (100 times), and then we calculated mean log2(fold change) and −log10(P values) of observed/expected values for each and R loop class, using a binomial test. Barplots of enrichment were made in R (version 4.1.3) using “ggplot2” (version 3.3.6). END- seq data analysis Adapter sequences of paired end reads in fastq format were trimmed, using Trim Galore (version 0.6.5), and then reads were aligned to hg19 reference human genome and to mm10 reference mouse genome, using Bowtie2 with default settings (version 2.5.0). Only properly paired reads with mapping quality (MAPQ) ≥ 30 were kept, including duplicated reads, using SAMtools (version 1.16.1) (83) (tableS1). Prior peak calling–aligned pairedend reads in BAM format were splitted according to firstin pair orientation to discriminate reads aligned on the left side (F2R1, reverseminus strand) and on the right side (F1R2, forwardplus strand) respect with the DSB. Peak calling was performed using MACS2 (version 2.2.7.1) (77) and using control samples as “input.” Peak calling was performed separately for each replicate and according to firstin- pair reads orientation, using the parameters - - nomodel - - nolambda - q 0.05 - - keepdup all - f BAMPE as previously described in (84). After peak calling, to specifically identify forward and reverse peaks resembling the whole END- seq signal, we searched for the closest forward peaks respect with the reverse ones, using bedtools closest (version 2.30.0) (78) and discarded pair of peaks resulting in negative values of distance or positive values of distance above or equal to 150 bp. Then, we retrieved start coordinates of reverse peaks and end coordinates of the closest forward peaks to obtain whole END- seq peaks for each replicate. To normalize target genome reads on spikein genome reads covering “ontarget” locus (TCRβ), as indicated in (48), we firstly created a universe of peaks that were present in both replicates for each condition, excluding ones overlapping hg19- blacklisted regions (85) and obtained reads covering peaks (RiP) using bedtools multicov (version 2.30.0) (78). Ontarget spikein RiP was divided by qualityfiltered reads (RQC) and then by the minimum resulting factor within each replicate. The resulting spikein scaling factors were then multiplied by a scaling factor performed on library size to lastly normalize target genome reads. Differential expression analysis for each condition (treated versus control) was performed using R/Bioconductor package “limma” (version 3.50.3) (79) with default settings, and peaks with log2(fold change) of >1 or log2(fold change) of <−1 and P <0.05 were considered as differentially expressed. Volcano plot of differentially expressed peaks was made with “EnhancedVolcano” R/Bioconductor package (version 1.12.0). Then, as we were interested in regions significantly enriched in treated over control samples, the following steps only regard positively enriched END- seq peaks. First, END- seq peaks that resulted differentially expressed in both 10′ CPT and 20′ CPT treatment samples were included in the “persistent” class (n=3498) to further distinguish peaks that were exclusively enriched at 10′ CPT (“transient,” n=763) from ones enriched at 20′ CPT (“late,” n=948). Second, within each class, overlapping peaks were merged to obtain clusters of DSB. Successively, we also considered as persistent all those transient and late clusters that were overlapping each other (n=273) to further refine kinetic classes of DSB clusters. BigWig files of END- seq signal scores for control and treated conditions were obtained from BAM files using bamCoverage command from deepTools (version 3.5.1) (82) normalizing by the previously calculated scaling factors, and then mean signal of replicates for each condition was obtained using bigwigCompare (version 3.5.1) (82). BigWig files of plus and minus strand signal were obtained from aligned pairedend reads, using samtools view (version 1.16.1) (83), command and - f option to include right and left alignments matching SAM flags for forward or reverse reads according to firstin pair orientation, and then using samtools merge (version 1.16.1) (83), the outputs were merged to obtain a single BAM file for plus and minus strand alignments. Metaplots of END- seq signal scores per genomic regions were performed using computeMatrix and plotHeatmap tools from deepTools (version 3.5.1) (82). Bigwigs of histone markers ChIP- seq and DNaseseq in HCT116 cells for the hg19 assembly were obtained from ENCODE website as processed and normalized tracks. Proportion and enrichment of END- seq peaks over gene features plots were performed using DROPA tool (version 1.0.0) (80) in GRCh37 Ensembl annotation. Randomization and over R loop regions of END- seq peaks were performed using the command bedtools shuffle (version 2.30.0) (78) (100 times), and then we calculated log2(fold change) and −log10(P values) of observed/expected values for each END- seq and R loop class, using Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 19 of 21 a binomial test. Barplots of enrichment were made in R (version 4.1.3) (www.R- project.org) using ggplot2 (version 3.3.6). To identify seDSBs, we analyzed separately peaks according to firstin pair orientation. First, we selected peaks present in both replicates within each condition using bedtools intersect and merged them using bedtools merge. Then, we identified forward and reverse peaks distant more than 150 bp from one another to retrieve “solitary” peaks ascribable to seDSBs. Reads covering peaks were calculated for both universes of peaks and then normalized, as previously described, distinguishing for forward and reverse reads for both spikein and library scaling factors. Then, to remove potential outliers, only counts within the first and the third quartile were retained for the differential expression analysis. Differential expression analysis was performed as previously described, and persistent seDSB was defined as enriched peaks under both condition of treatment versus control condition. Visualization of genomic distribution of DRIP- seq, END- seq, and ChIP- seq data was performed using Integrative Genomics Viewer software (86). Statistical analysis Statistical analyses were performed in Prism v9 (GraphPad Software) using tests shown in each figure legends. All the used analysis are hereafter reported: onetailed ratio paired t test, multiple unpaired t test, onetailed paired t test, twotailed Mann- Whitney test, the nonparametric Mann- Whitney rank sum test, and the parametric paired t test: *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Note added in proof: After the manuscript was accepted for publication, the authors alerted the Editorial Office to an additional paper related to their data discussion: K. B. Ynag, A. Rasouly, V. Epshtein, C. Martinez, T. Nguyen, I. Shamovsky, E. Nudler, Persistence of backtracking by human RNA polymerase II. Mol Cell. 84, 897–909 (2024). Supplementary Materials This PDF file includes: Figs.S1 to S9 legend for tableS1 tableS2 References Other Supplementary Material for this manuscript includes the following: tableS1 REFERENCES AND NOTES 1. Y. Pommier, Y. Sun, S.- Y. n. huang, J. l. nitiss, Roles of eukaryotic topoisomerases in transcription, replication and genomic stability. Nat. Rev. Mol. Cell Biol. 17, 703–721 (2016). 2. G. capranico, J. Marinello, G. chillemi, type i DnA topoisomerases. J. Med. Chem. 60, 2169–2192 (2017). 3. Y. Pommier, topoisomerase i inhibitors: camptothecins and beyond. Nat. Rev. Cancer 6, 789–802 (2006). 4. Y. Pommier, A. nussenzweig, S. takeda, c. Austin, human topoisomerases and their roles in genome stability and organization. Nat. Rev. Mol. Cell Biol. 23, 407–427 (2022). 5. J. Marinello, G. chillemi, S. Bueno, S. G. Manzo, G. capranico, Antisense transcripts enhanced by camptothecin at divergent cpG- island promoters associated with bursts of topoisomerase i- DnA cleavage complex and R- loop formation. Nucleic Acids Res. 41, 10110–10123 (2013). 6. l. Baranello, D. Wojtowicz, K. cui, B. n. Devaiah, h. J. chung, K. Y. chan- Salis, R. Guha, K. Wilson, X. Zhang, h. Zhang, J. Piotrowski, c. J. thomas, D. S. Singer, B. F. Pugh, Y. Pommier, t. M. Przytycka, F. Kouzine, B. A. lewis, K. Zhao, D. levens, RnA polymerase ii regulates topoisomerase 1 activity to favor efficient transcription. Cell 165, 357–371 (2016). 7. l. Baranello, D. Bertozzi, M. v. Fogli, Y. Pommier, G. capranico, DnA topoisomerase i inhibition by camptothecin induces escape of RnA polymerase ii from promoterproximal pause site, antisense transcription and histone acetylation at the human hiF- 1alpha gene locus. Nucleic Acids Res. 38, 159–171 (2010). 8. A. Khobta, F. Ferri, l. lotito, A. Montecucco, R. Rossi, G. capranico, early effects of topoisomerase i inhibition on RnA polymerase ii along transcribed genes in human cells. J. Mol. Biol. 357, 127–138 (2006). 9. P. Duann, M. Sun, c. t. lin, h. Zhang, l. F. liu, Plasmid linking number change induced by topoisomerase imediated DnA damage. Nucleic Acids Res. 27, 2905–2911 (1999). 10. F. Kouzine, A. Gupta, l. Baranello, D. Wojtowicz, K. Ben- Aissa, J. liu, t. M. Przytycka, D. levens, transcriptiondependent dynamic supercoiling is a shortrange genomic force. Nat. Struct. Mol. Biol. 20, 396–403 (2013). 11. A. A. cohen, n. Geva- Zatorsky, e. eden, M. Frenkel- Morgenstern, i. issaeva, A. Sigal, R. Milo, c. cohen- Saidon, Y. liron, Z. Kam, l. cohen, t. Danon, n. Perzov, U. Alon, Dynamic proteomics of individual cancer cells in response to a drug. Science 322, 1511–1516 (2008). 12. S. D. Desai, t.- K. li, A. Rodriguez- Bauman, e. h. Rubin, l. F. liu, Ubiquitin/26S proteasomemediated degradation of topoisomerase i as a resistance mechanism to camptothecin in tumor cells. Cancer Res. 61, 5926–5932 (2001). 13. B. Gómez- González, A. Aguilera, transcriptionmediated replication hindrance: A major driver of genome instability. Genes Dev. 33, 1008–1026 (2019). 14. M. P. crossley, M. Bocek, K. A. cimprich, R- loops as cellular regulators and genomic threats. Mol. Cell 73, 398–411 (2019). 15. t. García- Muse, A. Aguilera, R loops: From physiological to pathological roles. Cell 179, 604–618 (2019). 16. S. hamperl, M. J. Bocek, J. c. Saldivar, t. Swigut, K. A. cimprich, transcriptionreplication conflict orientation modulates rloop levels and activates distinct DnA damage responses. Cell 170, 774–786.e19 (2017). 17. M. holmström, v. Winters, Micronucleus induction by camptothecin and amsacrine in bone marrow of male and female cD- 1 mice. Mutagenesis 7, 189–193 (1992). 18. J. Marinello, A. Arleo, M. Russo, M. Delcuratolo, F. ciccarelli, Y. Pommier, G. capranico, topoisomerase i poisontriggered immune gene activation is markedly reduced in human smallcell lung cancers by impairment of the cGAS/StinG pathway. Br. J. Cancer 127, 1214–1225 (2022). 19. Y. h. hsiang, M. G. lihou, l. F. liu, Arrest of replication forks by drugstabilized topoisomerase i- DnA cleavable complexes as a mechanism of cell killing by camptothecin. Cancer Res. 49, 5077–5082 (1989). 20. D. Strumberg, A. A. Pilon, M. Smith, R. hickey, l. Malkas, Y. Pommier, conversion of topoisomerase i cleavage complexes on the leading strand of ribosomal DnA into 5′- phosphorylated DnA doublestrand breaks by replication runoff. Mol. Cell. Biol. 20, 3977–3987 (2000). 21. M. Regairaz, Y.- W. Zhang, h. Fu, K. K. Agama, n. tata, S. Agrawal, M. i. Aladjem, Y. Pommier, Mus81- mediated DnA cleavage resolves replication forks stalled by topoisomerase i- DnA complexes. J. Cell Biol. 195, 739–749 (2011). 22. h. Fu, M. M. Martin, M. Regairaz, l. huang, Y. You, c.- M. lin, M. Ryan, R. Kim, t. Shimura, Y. Pommier, M. i. Aladjem, the DnA repair endonuclease Mus81 facilitates fast DnA replication in the absence of exogenous damage. Nat. Commun. 6, 6746 (2015). 23. S. Di Marco, Z. hasanova, R. Kanagaraj, n. chappidi, v. Altmannova, S. Menon, h. Sedlackova, J. langhoff, K. Surendranath, D. hühn, R. Bhowmick, v. Marini, S. Ferrari, i. D. hickson, l. Krejci, P. Janscak, RecQ5 helicase cooperates with mus81 endonuclease in processing stalled replication forks at common fragile sites during mitosis. Mol. Cell. 66, 658–671.e8 (2017). 24. B. Pardo, M. Morielcarretero, t. vicat, A. Aguilera, P. Pasero, homologous recombination and Mus81 promote replication completion in response to replication fork blockage. EMBO Rep. 21, e49367 (2020). 25. n. chappidi, Z. nascakova, B. Boleslavska, R. Zellweger, e. isik, M. Andrs, S. Menon, J. Dobrovolna, c. Balbo Pogliano, J. Matos, A. Porro, M. lopes, P. Janscak, Fork cleavagereligation cycle and active transcription mediate replication restart after fork stalling at cotranscriptional R- loops. Mol. Cell 77, 528–541.e8 (2020). 26. P. Pourquier, l. M. Ueng, J. Fertala, D. Wang, h. J. Park, J. M. essigmann, M. A. Bjornsti, Y. Pommier, induction of reversible complexes between eukaryotic DnA topoisomerase i and DnA- containing oxidative base damages. 7, 8- dihydro- 8- oxoguanine and 5- hydroxycytosine. J. Biol. Chem. 274, 8516–8523 (1999). 27. S. Katyal, Y. lee, K. c. nitiss, S. M. Downing, Y. li, M. Shimada, J. Zhao, h. R. Russell, J. h. J. Petrini, J. l. nitiss, P. J. McKinnon, Aberrant topoisomerase- 1 DnA lesions are pathogenic in neurodegenerative genome instability syndromes. Nat. Neurosci. 17, 813–821 (2014). 28. h. takashima, c. F. Boerkoel, J. John, G. M. Saifi, M. A. M. M. Salih, D. Armstrong, Y. Mao, F. A. Quiocho, B. B. Roa, M. nakagawa, D. W. Stockton, J. R. lupski, Mutation of tDP1, encoding a topoisomerase idependent DnA damage repair enzyme, in spinocerebellar ataxia with axonal neuropathy. Nat. Genet. 32, 267–272 (2002). 29. R. hirano, h. interthal, c. huang, t. nakamura, K. Deguchi, K. choi, M. B. Bhattacharjee, K. Arimura, F. Umehara, S. izumo, J. l. northrop, M. A. M. Salih, K. inoue, D. l. Armstrong, J. J. champoux, h. takashima, c. F. Boerkoel, Spinocerebellar ataxia with axonal neuropathy: consequence of a tdp1 recessive neomorphic mutation? EMBO J. 26, 4732–4743 (2007). Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 20 of 21 30. n. Kim, S. n. huang, J. S. Williams, Y. c. li, A. B. clark, J.- e. cho, t. A. Kunkel, Y. Pommier, S. Jinks- Robertson, Mutagenic processing of ribonucleotides in DnA by yeast topoisomerase i. Science 332, 1561–1564 (2011). 31. S. M. cerritelli, J. iranzo, S. Sharma, A. chabes, R. J. crouch, D. tollervey, A. el hage, high density of unrepaired genomic ribonucleotides leads to topoisomerase 1- mediated severe growth defects in absence of ribonucleotide reductase. Nucleic Acids Res. 48, 4274–4297 (2020). 32. A. Aguilera, B. Gómezgonzález, DnA- RnA hybrids: the risks of DnA breakage during transcription. Nat. Struct. Mol. Biol. 24, 439–443 (2017). 33. J. Marinello, S. Bertoncini, i. Aloisi, A. cristini, G. M. tagliazucchi, M. Forcato, O. Sordet, G. capranico, M. M. tagliazucchi, G. Forcato, Dynamic effects of topoisomerase i inhibition on R- loops and short transcripts at active promoters. PLOS ONE 11, e0147053 (2016). 34. A. cristini, G. Ricci, S. Britton, S. Salimbeni, S. huang, J. Marinello, P. calsou, Y. Pommier, G. Favre, G. capranico, n. Gromak, O. Sordet, Dual processing of R- loops and topoisomerase i induces transcriptiondependent DnA doublestrand breaks. Cell Rep. 28, 3167–3181.e6 (2019). 35. K. Bartsch, K. Knittler, c. Borowski, S. Rudnik, M. Damme, K. Aden, M. e. Spehlmann, n. Frey, P. Saftig, A. chalaris, B. Rabe, Absence of Rnase h2 triggers generation of immunogenic micronuclei removed by autophagy. Hum. Mol. Genet. 26, 3960–3972 (2017). 36. A. M. S. Giordano, M. luciani, F. Gatto, M. A. Alezz, c. Beghè, l. Della volpe, A. Migliara, S. valsoni, M. Genua, M. Dzieciatkowska, G. Frati, J. tahraoui- Bories, S. c. Giliani, S. Orcesi, e. Fazzi, R. Ostuni, A. D’Alessandro, R. Di Micco, i. Merelli, A. lombardo, M. A. M. Reijns, n. Gromak, A. Gritti, A. Kajaste- Rudnitski, DnA damage contributes to neurotoxic inflammation in Aicardi- Goutières syndrome astrocytes. J. Exp. Med. 219, e20211121 (2022). 37. A. cristini, M. tellier, F. constantinescu, c. Accalai, l. O. Albulescu, R. heiringhoff, n. Bery, O. Sordet, S. Murphy, n. Gromak, Rnase h2, mutated in Aicardi- Goutières syndrome, resolves cotranscriptional R- loops to prevent DnA breaks and inflammation. Nat. Commun. 13, 2961 (2022). 38. Z. Andrysik, M. D. Galbraith, A. l. Guarnieri, S. Zaccara, K. D. Sullivan, A. Pandey, M. MacBeth, A. inga, J. M. espinosa, identification of a core tP53 transcriptional program with highly distributed tumor suppressive activity. Genome Res. 27, 1645–1657 (2017). 39. Y. luo, B. c. hitz, i. Gabdank, J. A. hilton, M. S. Kagda, B. lam, Z. Myers, P. Sud, J. Jou, K. lin, U. K. Baymuradov, K. Graham, c. litton, S. R. Miyasato, J. S. Strattan, O. Jolanki, J. W. lee, F. Y. tanaka, P. Adenekan, e. O’neill, J. M. cherry, new developments on the encyclopedia of DnA elements (encODe) data portal. Nucleic Acids Res. 48, D882–D889 (2020). 40. l. lotito, A. Russo, G. chillemi, S. Bueno, D. cavalieri, G. capranico, Global transcription regulation by DnA topoisomerase i in exponentially growing Saccharomyces cerevisiae cells: Activation of telomereproximal genes by tOP1 deletion. J. Mol. Biol. 377, 311–322 (2008). 41. J. Marinello, G. capranico, in Methods in Molecular Biology (humana Press inc., 2022), vol. 2528, pp. 203–213. 42. D. Roy, Z. Zhang, Z. lu, c.- l. hsieh, M. R. lieber, competition between the RnA transcript and the nontemplate DnA strand during R- loop formation invitro: A nick can serve as a strong R- loop initiation site. Mol. Cell. Biol. 30, 146–159 (2010). 43. l. chen, J.- Y. chen, X. Zhang, Y. Gu, R. Xiao, c. Shao, P. tang, h. Qian, D. luo, h. li, Y. Zhou, D.- e. Zhang, X.- D. Fu, R- chiP using inactive Rnase h reveals dynamic coupling of R- loops with transcriptional pausing at gene promoters. Mol. Cell 68, 745–757.e5 (2017). 44. S. K. Das, v. Kuzin, D. P. cameron, S. Sanford, R. K. Jha, Z. nie, M. t. Rosello, R. holewinski, t. Andresson, J. Wisniewski, t. natsume, D. h. Price, B. A. lewis, F. Kouzine, D. levens, l. Baranello, MYc assembles and stimulates topoisomerases 1 and 2 in a “topoisome”. Mol. Cell 82, 140–158.e12 (2022). 45. D. castillo- Guzman, F. chédin, Defining R- loop classes and their contributions to genome instability. DNA Repair 106, 103182 (2021). 46. D. Zatreanu, Z. han, R. Mitter, e. tumini, h. Williams, l. Gregersen, A. B. Dirac- Svejstrup, S. Roma, A. Stewart, A. Aguilera, J. Q. Svejstrup, elongation factor tFiiS prevents transcription stress and R- loop accumulation to maintain genome stability. Mol. Cell 76, 57–69.e9 (2019). 47. R. M. Sheridan, n. Fong, A. D’Alessandro, D. l. Bently, Widespread backtracking by RnA Pol ii is a major effector of gene activation, 5′ pause release, termination, and transcription elongation rate. Mol. Cell 73, 107–118.e4 (2019). 48. A. canela, S. Sridharan, n. Sciascia, A. tubbs, P. Meltzer, B. P. Sleckman, A. nussenzweig, DnA breaks and end resection measured genomewide by end sequencing. Mol. Cell 63, 898–911 (2016). 49. l. halász, Z. Karányi, B. Boros- Oláh, t. Kuik- Rózsa, É. Sipos, É. nagy, Á. Mosolygól, A. Mázló, É. Rajnavölgyi, G. halmos, l. Székvölgyi, RnA- DnA hybrid (R- loop) immunoprecipitation mapping: An analytical workflow to evaluate inherent biases. Genome Res. 27, 1063–1073 (2017). 50. Y. hu, B. Stillman, Origins of DnA replication in eukaryotes. Mol. Cell 83, 352–372 (2023). 51. D. J. emerson, P. A. Zhao, A. l. cook, R. J. Barnett, K. n. Klein, D. Saulebekova, c. Ge, l. Zhou, Z. Simandi, M. K. Minsk, K. R. titus, W. Wang, W. Gong, D. Zhang, l. Yang, S. v. venev, J. h. Gibcus, h. Yang, t. Sasaki, M. t. Kanemaki, F. Yue, J. Dekker, c. l. chen, D. M. Gilbert, J. e. Phillipscremins, cohesinmediated loop anchors confine the locations of human replication origins. Nature 606, 812–819 (2022). 52. M. M. Martin, M. Ryan, R. G. Kim, A. l. Zakas, h. Fu, c. M. lin, W. c. Reinhold, S. R. Davis, S. Bilke, h. liu, J. h. Doroshow, M. A. Reimers, M. S. valenzuela, Y. Pommier, P. S. Meltzer, M. i. Aladjem, Genomewide depletion of replication initiation events in highly transcribed regions. Genome Res. 21, 1822–1832 (2011). 53. t. Sasaki, S. Ramanathan, Y. Okuno, c. Kumagai, S. S. Shaikh, D. M. Gilbert, the chinese hamster dihydrofolate reductase replication origin decision point follows activation of transcription and suppresses initiation of replication within transcription units. Mol. Cell. Biol. 26, 1051–1062 (2006). 54. O. Brison, S. elhilali, D. Azar, S. Koundrioukoff, M. Schmidt, v. nähse, Y. Jaszczyszyn, A. M. lachages, B. Dutrillaux, c. thermes, M. Debatisse, c.- l. chen, transcriptionmediated organization of the replication initiation program across large genes sets common fragile sites genomewide. Nat. Commun. 10, 5693 (2019). 55. M. J. Scherr, S. A. Wahab, D. Remus, K. e. Duderstadt, Mobile originlicensing factors confer resistance to conflicts with RnA polymerase. Cell Rep. 38, 110531 (2022). 56. B. D. Pope, t. Ryba, v. Dileep, F. Yue, W. Wu, O. Denas, D. l. vera, Y. Wang, R. S. hansen, t. K. canfield, R. e. thurman, Y. cheng, G. Gülsoy, J. h. Dennis, M. P. Snyder, J. A. Stamatoyannopoulos, J. taylor, R. c. hardison, t. Kahveci, B. Ren, D. M. Gilbert, topologically associating domains are stable units of replicationtiming regulation. Nature 515, 402–405 (2014). 57. G. Miglietta, M. Russo, G. capranico, G- quadruplex–R- loop interactions and the mechanism of anticancer G- quadruplex binders. Nucleic Acids Res. 48, 11942–11957 (2020). 58. R. c. Duardo, F. Guerra, S. Pepe, G. capranico, non- B DnA structures as a booster of genome instability. Biochimie 214, 176–192 (2023). 59. D. A. Koster, K. Palle, e. S. M. Bot, M. A. Bjornsti, n. h. Dekker, Antitumour drugs impede DnA uncoiling by topoisomerase i. Nature 448, 213–217 (2007). 60. S. G. Manzo, S. R. hartono, l. A. Sanz, J. Marinello, S. De Biasi, A. cossarizza, G. capranico, F. chedin, DnA topoisomerase i differentially modulates R- loops across the human genome. Genome Biol. 19, 100 (2018). 61. t. García- Muse, A. Aguilera, transcription–replication conflicts: how they occur and how they are resolved. Nat. Rev. Mol. Cell Biol. 17, 553–563 (2016). 62. D. Dutta, K. Shatalin, v. epshtein, M. e. Gottesman, e. nudler, linking RnA polymerase backtracking to genome instability in e.coli. Cell 146, 533–543 (2011). 63. e. nudler, RnA polymerase backtracking in gene regulation and genome instability. Cell 149, 1438–1445 (2012). 64. J. Brochu, É. vlachos- Breton, D. irsenco, M. Drolet, characterization of a pathway of genomic instability induced by R- loops and its regulation by topoisomerases in E. coli. PLOS Genet. 19, e1010754 (2023). 65. G. Abdurashidova, S. Radulescu, O. Sandoval, S. Zahariev, M. B. Danailov, A. Demidovich, l. Santamaria, G. Biamonti, S. Riva, A. Falaschi, Functional interactions of DnA topoisomerases with a human replication origin. EMBO J. 26, 998–1009 (2007). 66. h. Fu, c. e. Redon, B. l. thakur, K. Utani, R. Sebastian, S.- M. M. Jang, J. M. Gross, S. Mosavarpour, A. B. Marks, S. Z. Zhuang, S. B. lazar, M. Rao, S. t. Mencer, A. M. Baris, l. S. Pongor, M. i. Aladjem, Dynamics of replication origin overactivation. Nat. Commun. 12, 3448 (2021). 67. B. l. thakur, A. Ray, c. e. Redon, M. i. Aladjem, Preventing excess replication origin activation to ensure genome stability. Trends Genet. 38, 169–181 (2022). 68. Y.- W. Zhang, M. Regairaz, J. A. Seiler, K. K. Agama, J. h. Doroshow, Y. Pommier, Poly(ADP- ribose) polymerase and XPF- eRcc1 participate in distinct pathways for the repair of topoisomerase iinduced DnA damage in mammalian cells. Nucleic Acids Res. 39, 3607–3620 (2011). 69. A. De Magis, S. G. Manzo, M. Russo, J. Marinello, R. Morigi, O. Sordet, G. capranico, DnA damage and genome instability by G- quadruplex ligands are mediated by R loops in human cancer cells. Proc. Natl. Acad. Sci. U.S.A. 116, 816–825 (2019). 70. M. García- Rubio, S. i. Barroso, A. Aguilera, in Methods in Molecular Biology (humana Press inc., 2018), vol. 1672, pp. 347–361. 71. n. Wong, S. John, A. nussenzweig, A. canela, in Methods in Molecular Biology (humana Press inc., 2021), vol. 2153, pp. 9–31. 72. A. collins, P. Møller, G. Gajski, S. vodenková, A. Abdulwahed, D. Anderson, e. e. Bankoglu, S. Bonassi, e. Boutet- Robinet, G. Brunborg, c. chao, M. S. cooke, c. costa, S. costa, A. Dhawan, J. de lapuente, c. Del Bo', J. Dubus, M. Dusinska, S. J. Duthie, n. e. Yamani, B. engelward, i. Gaivão, l. Giovannelli, R. Godschalk, S. Guilherme, K. B. Gutzkow, K. habas, A. hernández, O. herrero, M. isidori, A. n. Jha, S. Knasmüller, i. M. Kooter, G. Koppen, M. Kruszewski, c. ladeira, B. laffon, M. larramendy, l. l. hégarat, A. lewies, A. lewinska, G. e. liwszyc, A. l. de cerain, M. Manjanatha, R. Marcos, M. Milić, v. M. de Andrade, M. Moretti, D. Muruzabal, M. novak, R. Oliveira, A.- K. Olsen, n. Owiti, M. Pacheco, A. K. Pandey, S. Pfuhler, B. Pourrut, K. Reisinger, e. Rojas, e. Rundén- Pran, J. Sanz- Serrano, S. Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024 Duardo et al., Sci. Adv. 10, eadm8196 (2024) 24 May 2024 Science ADvAnceS | ReSeARch ARticle 21 of 21 Shaposhnikov, v. Sipinen, K. Smeets, h. Stopper, J. P. teixeira, v. valdiglesias, M. valverde, F. van Acker, F.- J. van Schooten, M. vasquez, J. F. Wentzel, M. Wnuk, A. Wouters, B. Žegura, t. Zikmund, S. A. S. langie, A. Azqueta, Measuring DnA modifications with the comet assay: A compendium of protocols. Nat. Protoc. 18, 929–989 (2023). 73. M. Martin, cutadapt removes adapter sequences from highthroughput sequencing reads. EMBnet 17, 10–12 (2011). 74. h. li, R. Durbin, Fast and accurate short read alignment with Burrows- Wheeler transform. Bioinformatics 25, 1754–1760 (2009). 75. S. G. landt, G. K. Marinov, A. Kundaje, P. Kheradpour, F. Pauli, S. Batzoglou, B. e. Bernstein, P. Bickel, J. B. Brown, P. cayting, Y. chen, G. DeSalvo, c. epstein, K. i. Fisher- Aylor, G. euskirchen, M. Gerstein, J. Gertz, A. J. hartemink, M. M. hoffman, v. R. iyer, Y. l. Jung, S. Karmakar, M. Kellis, P. v. Kharchenko, Q. li, t. liu, X. S. liu, l. Ma, A. Milosavljevic, R. M. Myers, P. J. Park, M. J. Pazin, M. D. Perry, D. Raha, t. e. Reddy, J. Rozowsky, n. Shoresh, A. Sidow, M. Slattery, J. A. Stamatoyannopoulos, M. Y. tolstorukov, K. P. White, S. Xi, P. J. Farnham, J. D. lieb, B. J. Wold, M. Snyder, chiP- seq guidelines and practices of the encODe and modencODe consortia. Genome Res. 22, 1813–1831 (2012). 76. h. li, B. handsaker, A. Wysoker, t. Fennell, J. Ruan, n. homer, G. Marth, G. Abecasis, R. Durbin; 1000 Genome Project Data Processing Subgroup, the sequence alignment/ map format and SAMtools. Bioinformatics 25, 2078–2079 (2009). 77. Y. Zhang, t. liu, c. A. Meyer, J. eeckhoute, D. S. Johnson, B. e. Bernstein, c. nussbaum, R. M. Myers, M. Brown, W. li, X. S. liu, Modelbased analysis of chiP- seq (MAcS). Genome Biol. 9, R137 (2008). 78. A. R. Quinlan, i. M. hall, BeDtools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010). 79. M. e. Ritchie, B. Phipson, D. Wu, Y. hu, c. W. law, W. Shi, G. K. Smyth, limma powers differential expression analyses for RnA- sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015). 80. M. Russo, B. De lucca, t. Flati, S. Gioiosa, G. chillemi, G. capranico, DROPA: DRiP- seq optimized peak annotator. BMC Bioinformatics. 20, 414 (2019). 81. M. lawrence, R. Gentleman, v. carey, rtracklayer: An R package for interfacing with genome browsers. Bioinformatics 25, 1841–1842 (2009). 82. F. Ramírez, D. P. Ryan, B. Grüning, v. Bhardwaj, F. Kilpert, A. S. Richter, S. heyne, F. Dündar, t. Manke, deeptools2: A next generation web server for deepsequencing data analysis. Nucleic Acids Res. 44, W160–W165 (2016). 83. P. Danecek, J. K. Bonfield, J. liddle, J. Marshall, v. Ohan, M. O. Pollard, A. Whitwham, t. Keane, S. A. Mccarthy, R. M. Davies, twelve years of SAMtools and BcFtools. Gigascience 10, giab008 (2021). 84. A. canela, Y. Maman, S. n. huang, G. Wutz, W. tang, G. Zagnolivieira, e. callen, n. Wong, A. Day, J.- M. Peters, K. W. caldecott, Y. Pommier, A. nussenzweig, topoisomerase iiinduced chromosome breakage and translocation is determined by chromosome architecture and transcriptional activity. Mol. Cell 75, 252–266.e8 (2019). 85. the encODe Project consortium, An integrated encyclopedia of DnA elements in the human genome. Nature 489, 57–74 (2012). 86. J. t. Robinson, h. thorvaldsdóttir, W. Winckler, M. Guttman, e. S. lander, G. Getz, J. P. Mesirov, integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011). 87. O. Pedro, J. A. Méridacerro, A. G. Rondón, B. Gómez- González, A. Aguilera, DnA- RnA hybrids at DSBs interfere with repair by homologous recombination. eLife 10, e69881 ( 2021). Acknowledgments: We thank the Genomic Unit of cABiMeR for DnA sequencing and e. S. Oliva for technical assistance. Funding: this work was funded by AiRc (Associazione italiana per la Ricerca sul cancro, Milan, italy) (iG 2019 –iD 23032 project to G.c.); by the Ministry of University and Research (Rome, italy) under project PRin 2017–KSZZJW_004 to J.M.; and by projects i+D+i PiD2019- 104270GB- i00/BMc, funded by Mcin/Aei/10.13039/501100011033/, and eRc2014 AdG669898 tARlOOP, funded by the european Research council to A.A. the research leading to these results has also received funding from the european Union–nextGenerationeU through the italian Ministry of University and Research under PnRR- M4c2- i1.3 Project Pe_00000019 “heAl itAliA” to G.c., cUP (J33c22002920006). the views and opinions expressed are those of the authors only and do not necessarily reflect those of the european Union or the european commission. neither the european Union nor the european commission can be held responsible for them. R.c.D. is a recipient of a FiRc- AiRc postdoc fellowship for italy (26701). Author contributions: G.c. conceived the study. G.c., J.M., B.G.- G., and A.A. supervised the research. R.c.D., J.M., S.P., and F.G. conducted the wet laboratory experiments. M.R. and S.M. analyzed the genomic data. G.c., J.M., and A.A secured funding. R.c.D., J.M., S.P., M.R., and S.M. prepared the figures. G.c. wrote the manuscript with input from all the authors. All authors read, discussed, and approved the last version of the manuscript. Competing interests: the authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Genomic DRiP- seq and enD- seq sequencing data and normalized bigwigs are available at GeO under accession GSe239328. RnAPii chiP- seq, tOP1 chiP- seq, and tOP1- seq data (6) were obtained from GeO under accession number GSe57628. GRO- seq data (38) were obtained from GeO under accession number GSe86165. SnS- seq (22) data were obtained from GeO under accession number GSe28911. tAD and replication iZs data (51) were obtained from the Supplementary Materials. Bigwigs of histone markers chiP- seq and Dnaseseq in hct116 cells for the hg19 assembly were obtained from encODe under accession the following accession numbers: encFF401XSl (h3K27ac), encFF679MXB (h3K9me2), encFF885ZMJ (h3K4me1), encFF736AWh (h3K4me2), encFF176nSX (h3K4me3), and encFF039Glc (Dnaseseq). BeD files for histone markers chiP- seq and in hct116 cells for the hg19 assembly were obtained from encODe under accession the following accession numbers: encFF782UBh (h2AFZ), encFF349lKU (h3K27ac), encFF590YiB (h3K27ac_1), encFF846MnK (h3K4me2), encFF699vYO (h3K4me3), encFF982eZU (h3K4me3_1), encFF348RGO (h3K9ac), and encFF723YDW (h3K9me3). Scripts used to perform reads processing, alignment on reference genome, filtering and peak calling of DRiP- seq and enD- seq reads, as well as bed files of DRiP- seq and enD- seq regions used in this work are available here on Zenodo (DOi: 10.5281/ zenodo.10907219) or github.com (marcrusso/Duardo_et_al_2024). Submitted 7 november 2023 Accepted 18 April 2024 Published 24 May 2024 10.1126/sciadv.adm8196 Downloaded from https://www.science.org at Universidad de Sevilla on October 11, 2024