Resolution of R-loops by topoisomerase III-β (TOP3B) in coordination with the DEAD-box helicase DDX5

SUMMARY

The present study demonstrates how TOP3B is involved in resolving R-loops. We observed elevated R-loops in TOP3B knockout cells (TOP3BKO), which are suppressed by TOP3B transfection. R-loop-inducing agents, the topoisomerase I inhibitor camptothecin, and the splicing inhibitor pladienolide-B also induce higher R-loops in TOP3BKO cells. Camptothecin- and pladienolide-B-induced R-loops are concurrent with the induction of TOP3B cleavage complexes (TOP3Bccs). RNA/DNA hybrid IP-western blotting show that TOP3B is physically associated with R-loops. Biochemical assays using recombinant TOP3B and oligonucleotides mimicking R-loops show that TOP3B cleaves the single-stranded DNA displaced by the R-loop RNA-DNA duplex. IP-mass spectrometry and IP-western experiments reveal that TOP3B interacts with the R-loop helicase DDX5 independently of TDRD3. Finally, we demonstrate that DDX5 and TOP3B are epistatic in resolving R-loops in a pathway parallel with senataxin. We propose a decatenation model for R-loop resolution by TOP3B-DDX5 protecting cells from R-loop-induced damage.

Graphical abstract

In brief

Saha et al. show that the topoisomerase TOP3B is recruited to DNA:RNA hybrids known as R-loops in response to the R-loop-inducing agents camptothecin and pladienolide B and can resolve R-loop structures in human cells in coordination with the helicase DDX5.

INTRODUCTION

R-loops are prevalent (5% of the human genome) and represent dynamic non-B DNA structures consisting of RNA-DNA hybrids with displaced single-stranded DNA segments (Sanz et al., 2016). They are found at promoters and terminators, rDNA, tRNA-coding genes, Ty elements, centromeres, and telomeres (Chan et al., 2014; Ginno et al., 2012; Wahba et al., 2016). R-loops (regulatory or unscheduled/aberrant) can form co-transcriptionally in cis when nascent RNA reanneals with the template DNA strand or in trans when the RNA originating from a distal locus (viz., long non-coding RNAs) or in homologous chromosome invades duplex DNA (Ariel et al., 2020; Barroso et al., 2019; Chedin and Benham, 2020; Cloutier et al., 2016; Gomez-Gonzalez and Aguilera, 2021; Niehrs and Luke, 2020; Wahba et al., 2013).

Regulatory R-loops play important roles in cellular metabolism, including class switch recombination (Yu et al., 2003), mitochondrial replication (Pohjoismaki et al., 2010; Xu and Clayton, 1996), protection against promoter methylation (Ginno et al., 2012; Grunseich et al., 2018; Niehrs and Luke, 2020), transcription termination (Skourti-Stathaki et al., 2011), and chromosome segregation (Kabeche et al., 2018). Unscheduled or aberrant R-loops formed during transcription (Gomez-Gonzalez and Aguilera, 2021) can impair transcription elongation, leading to replication stress, transcription-replication conflicts, DNA breaks, chromosome fragility, and genomic instability (Aguilera and Garcia-Muse, 2012; Aguilera and Gomez-Gonzalez, 2017; Cristini et al., 2019; Crossley et al., 2019; Gan et al., 2011; Hamperl et al., 2017; Hamperl and Cimprich, 2016; Helmrich et al., 2011; Huertas and Aguilera, 2003; Sollier and Cimprich, 2015). R-loops are also associated with a growing number of human diseases, including neurological and autoimmune disorders and cancer (Garcia-Muse and Aguilera, 2019; Perego et al., 2019; Richard and Manley, 2017; Wells et al., 2019).

Cells deploy at least two different mechanisms, prevention of R-loop formation and resolution of existing R-loops, to protect the genome from excessive R-loop accumulation. Nucleases such as RNase H1/RNase H2 degrade the RNA of RNA-DNA hybrids, and RNA/DNA helicases like senataxin (SETX), Aquarius (AQR), DHX9, and members of the DEAD box RNA helicase family unwind the RNA-DNA hybrids and directly remove the excessive R-loops (Amon and Koshland, 2016; Chakraborty et al., 2018; Cristini et al., 2018; Hodroj et al., 2017; Li et al., 2016; Lockhart et al., 2019; Mersaoui et al., 2019; Okamoto et al., 2019; Paulsen et al., 2009; Skourti-Stathaki et al., 2011; Sollier et al., 2014; Song et al., 2017; Zhao et al., 2018). Alternatively, mRNA biogenesis and processing factors such as the THO complex, THSC/TREX-2 (Bhatia et al., 2014; Castellano-Pozo et al., 2012; Dominguez-Sanchez et al., 2011; Gomez-Gonzalez et al., 2011; Huertas and Aguilera, 2003), FIP1L (Stirling et al., 2012), and splicing factors such as SRSF1 (Li and Manley, 2005) sequester the nascent RNA and facilitate its dissociation from the DNA template. In addition, topoisomerases act by relaxing co-transcriptionally generated hypernegative supercoiling behind transcription forks and prevent further R-loop formation (El Hage et al., 2010; Huang et al., 2018; Manzo et al., 2018; Pommier et al., 2016; Sordet et al., 2009; Tuduri et al., 2009; Wilson-Sali and Hsieh, 2002; Yang et al., 2014; Zhang et al., 2019).

Topoisomerases are essential enzymes that solve nucleic acid topological constraints and permit vital metabolic processes, including replication, transcription, recombination, chromosome segregation, and chromatin remodeling (Pommier et al., 2016, 2022). Human topoisomerase 3β (TOP3B) is unique among the six vertebrate topoisomerases as it is the only topoisomerase that can act on both DNA and RNA (Ahmad et al., 2017a, 2017b; Pommier et al., 2022; Saha et al., 2020; Stoll et al., 2013; Xu et al., 2013). Loss of TOP3B is associated with various human diseases, including schizophrenia, autism, intellectual disability, genomic instability, and breast/renal cancer (Ahmad et al., 2017a; Daghsni et al., 2018; Kaufman et al., 2016; Oliveira-Costa et al., 2010; Stoll et al., 2013; Xu et al., 2013; Zhang et al., 2019). TOP3B forms cellular complexes with its auxiliary factor Tudor domain-containing protein 3 (TDRD3), which also binds fragile X mental retardation protein (FMRP) and participates in DNA and RNA metabolic processes, namely transcription and translation in the nucleus and cytoplasm, respectively (Stoll et al., 2013; Xu et al., 2013).

Loss of TOP3B has been linked with an increased burden of cellular R-loops in different organism and model systems (Huang et al., 2018; Wilson-Sali and Hsieh, 2002; Yang et al., 2014; Zhang et al., 2019). Hypernegative DNA supercoils, which open the DNA duplex, are preferential substrates for TOP3B, and it is plausible that the relaxation of these single-stranded regions by TOP3B prevents and destabilizes R-loops (Huang et al., 2018; Pommier et al., 2016; Yang et al., 2014; Zhang et al., 2019). Experimental evidence from a few recent studies supports the role of TOP3B in R-loop metabolism. TDRD3, the TOP3B scaffolding protein, which reads asymmetric di-methylated arginine marks in core histones and RNA polymerase II, can target TOP3B to transcriptionally active CpG island (CGI) promoters to resolve R-loops (Yang et al., 2014). Consistently, the methylation of two arginine residues, R833 and R835, present in the C-terminal RGG motif of TOP3B, has been shown to be important for R-loop resolution both in vitro and in vivo (Huang et al., 2018). Furthermore, TOP3B genetic alterations leading to persistent and excessive TOP3B cleavage complexes (TOP3Bccs) cause excessive R-loop accumulation both in human colon carcinoma HCT116 and human embryonic kidney HEK293 cells (Saha et al., 2020). Finally, lymphoblast cells derived from a patient with bilateral renal cancer and homozygous deletion of the TOP3B gene showed higher R-loop signals (Zhang et al., 2019).

Although TOP3B has been implicated in R-loop homeostasis, the molecular mechanisms by which it potentially resolves R-loops had not been elucidated. Using cellular, pharmacological, and biochemical approaches, we establish that TOP3B is recruited to R-loops and that TOP3B interacts with the DEAD box RNA helicase DDX5 independently of TDRD3 and SETX to resolve R-loops.

RESULTS

Elevated R-loop levels in TOP3B knockout (KO) cells and in response to camptothecin (CPT) and pladienolide B (Plad-B)

We generated TOP3BKO HCT116 and K562 cells by CRISPR-Cas9 (Figure S1A). Those TOP3B-null clones (clone E2–4 HCT116 and clones E5–11 and E5–13 K562) were confirmed by western blotting (Figures S1B and S1C). We isolated genomic DNA and determined R-loops in wild-type (WT) and TOP3BKO HCT116 and K562 cells by dot blot analysis using the S9.6 anti-body specific for RNA-DNA hybrids (Boguslawski et al., 1986). Consistent with independent reports (Huang et al., 2018; Zhang et al., 2019), TOP3B depletion caused a significant increase in R-loop levels both in HCT116 (Figures 1A–1D) and K562 cells (Figures S1D and S1E). Disappearance of signals after treatment of the genomic DNA samples with RNase H before slot-blot analysis confirmed the specificity of the R-loop signals detected with S9.6 antibody (Figures 1A–1D, S1D, and S1E). We also confirmed the absence of double-stranded RNA (dsRNA) contamination in the genomic DNA samples used for R-loop detection by performing slot blot and probing with J2 antibody specific for dsRNA (Figure S1H). Ectopic expression of WT TOP3B in TOP3BKO cells (Figures S1F and S1G) suppressed R-loops, whereas the active site mutant Y336F TOP3B failed to reduce R-loops to baseline level, implicating a causal role of TOP3B in suppressing R-loops (Figures 1A–1D).

Figure 1. R-loops accumulate in TOP3BKO cells, and R-loop-inducing agents produce elevated R-loops in TOP3BKO cells and TOP3Bccs in TOP3B-proficient cells.

(A and B) Representative slot blots showing the increased accumulation of R-loops in TOP3BKO HCT116 cells (wild type [WT]), R-loop induction by camptothecin (CPT; 20 μM, 10 min) and pladienolide-B (Plad-B; 5 μM, 2 h), and the suppression of R-loops by transfection of TOP3B. Y336F TOP3B fails to suppress R-loops in TOP3BKO cells. Genomic DNA from HCT116 WT, TOP3BKO, TOP3B, and Y336F TOP3B-transfected TOP3BKO cells was probed with S9.6 antibody. (C and D) Quantitation of R-loops from 3 independent experiments. Data are plotted as means ± standard deviations (SDs). Statistical significance was calculated using 2-tailed unpaired t test. *p < 0.01; **p < 0.001; n.s., not significant.

(E–H) Time-dependent induction of R-loops by CPT (20 μM) and Plad-B (5 μM) in the indicated HCT116 cell lines. Representative blots are shown in (E) and (G). Quantitations from 3 independent experiments are shown in (F) and (H) (means ± SDs, and statistical significance calculated using 2-tailed unpaired t test; *p < 0.01; **p < 0.001).

(I–L) Time-dependent induction of TOP3Bccs by CPT (20 μM) and Plad-B (5 μM) as detected by RADAR assay in the indicated HCT116 cell lines. Representative blots are shown in (I) and (K). Quantitations from 3 independent experiments are shown in (J) and (L) (means ± SDs), and statistical significance calculated using 2-tailed unpaired t test. *p < 0.01; **p < 0.001.

See also Figures S1 and S2.

Next, we measured R-loops in WT and TOP3BKO cells under three previously reported R-loop-inducing conditions acting by different mechanisms: topoisomerase I (TOP1) inhibition by CPT, splicing inhibition by Plad-B, and downregulation of the R-loop-resolving helicase SETX (Chakraborty et al., 2018; Cristini et al., 2018; Grunseich et al., 2018; Marinello et al., 2013, 2016; Nguyen et al., 2017, 2018; Skourti-Stathaki et al., 2011; Yeo et al., 2014; Yuce and West, 2013). Treatments with CPT (20 μM, 10 min) and Plad-B (5 μM, 2 h) or transfection with small interfering RNA (siRNA) against SETX caused greater accumulation of R-loops in TOP3BKO cells compared to WT HCT116 cells (Figures 1A–1D and S2A–S2C). Rescue experiments by transfecting WT TOP3B in TOP3BKO cells reduced R-loop burden after treatments with CPT or Plad-B, whereas transfection of the active site mutant Y336F TOP3B failed to reduce R-loops in TOP3BKO cells (Figures 1A–1D).

These results show that TOP3B is an important regulator that keeps R-loops at baseline levels (both in HCT116 and K562 cells) and under R-loop-inducing conditions (TOP1 inhibition by CPT, splicing inhibition with Plad-B, or downregulation of SETX).

Trapping of TOP3B by CPT and Plad-B coincides with R-loop formation

TOP3B has been proposed to prevent R-loop formation by relaxing DNA hypernegative supercoiling behind transcription forks (Huang et al., 2018; Pommier et al., 2016; Yang et al., 2014; Zhang et al., 2019). To find out whether TOP3B is catalytically activated by cellular R-loops, we performed RADAR (rapid approach to DNA adduct recovery) assays (Kiianitsa and Maizels, 2013; Saha et al., 2020) to determine whether TOP3B-nucleic acids covalent intermediates (i.e., TOP3Bccs) (Saha et al., 2020) were induced after R-loop induction by CPT and Plad-B (Figures 1E–1L).

Experiments were carried out to determine the timing of CPT- and Plad-B-induced R-loops in WT, TOP3BKO, and TOP3BKO HCT116 cells transfected with TOP3B (Figures 1E–1H). While CPT induced R-loops rapidly (at the first measurable 5-min time point) and transiently (Figures 1E and 1F) (Marinello et al., 2016), Plad-B induced R-loops with much slower kinetics and persistence over 4 h (Figures 1G and 1H). In both cases and all time points, significantly more R-loops were induced in the TOP3BKO cells, and transfection of TOP3B in the TOP3BKO cells reduced the R-loops to the levels observed in the WT cells (Figures 1E–1H).

Having determined the kinetics and TOP3B-dependence of R-loop formation, we hypothesized that TOP3B may be recruited to the R-loops induced by CPT and Plad-B and performed TOP3B RADAR assays under the same conditions as those used to induced R-loops. Figures 1I–1L and S1I–S1L show that both CPT and Plad-B induced TOP3Bccs at the peak time of formation of the R-loops: 5–10 min for CPT and 2–4 h for Plad-B. The TOP3Bcc signals were not detected in untreated conditions, abolished in the TOP3BKO cells, and enhanced in the TOP3B-transfected cells (Figures 1I–1L and S1I–S1L), demonstrating that both CPT- and Plad-B-R-loops induce the formation of TOP3Bccs coincidently with the formation of R-loops.

To further establish that R-loop formation induces TOP3Bccs, we tested whether genetic inactivation of the R-loop resolvase SETX (Skourti-Stathaki et al., 2011) also induced the formation of TOP3Bccs. Consistent with the association of TOP3B with R-loops, knocking down SETX induced significantly higher levels of R-loops in TOP3BKO than WT HCT116 cells (Figures S2A–S2C) as well as TOP3Bccs (Figures S2D and S2E).

Together, these results demonstrate that TOP3B suppresses cellular R-loops and that it is catalytically engaged by forming TOP3cc upon R-loop formation.

R-loop formation induces both DNA and RNA TOP3Bccs

To determine whether TOP3Bccs formed both on DNA and RNA after R-loop formation, we used our previous technical approach (Saha et al., 2020) and examined the TOP3Bccs induced by CPT (20 μM, 10 min) and Plad-B (5 μM, 2 h) in TOP3B-complemented TOP3BKO HCT1116 cells. After isolating nucleic acids containing TOP3Bccs by RADAR assay, equal amounts of RADAR assay samples were digested either with a mix of RNase A and RNase T1 to remove RNA, or with DNase 1 to digest DNA, or with benzonase to remove total nucleic acids (Figure 2A).

Figure 2. R-loop formation enhances both DNA and RNA TOP3Bccs, which are recruited to R-loops.

(A) Outline of the experiment for detection of TOP3Bccs in DNA and RNA after treatments with CPT (20 μM, 10 min) and Plad-B (5 μM, 2 h).

(B–D) Representative slot blots of TOP3Bccs induced by CPT and Plad-B. Protein-nucleic acid adducts were isolated by RADAR assay and samples were digested either with excess RNase A (200 μg/mL) and RNase T1 (200 units/μL) mix or with DNase 1 (10 U) or benzonase. Samples were ethanol precipitated, resuspended, and slot blotted. TOP3Bccs were detected with anti-TOP3B antibodies.

(C–E) Quantitation of TOP3Bccs from 3 independent experiments as shown in (B) and (D). Data are plotted as means ± SDs. *p ≤ 0.01 and **p ≤ 0.001 (2-tailed unpaired t test).

(F) Outline of the S9.6 IP-western blot experiment performed in HCT116 TOP3BKO cells and FLAG TOP3B-transfected TOP3BKO cells after treatments with the R-loop-inducing agents CPT (20 μM, 10 min) and Plad-B (5 μM, 2 h).

(G and H) TOP3B interacts with R-loops after CPT and Plad-B treatments as determined by S9.6 IP-western blotting. See also Figures S1 and S2.

Slot blotting showed that TOP3Bccs signals decreased significantly after digestions with RNase A and RNase T1 mix, indicating the presence of RNA TOP3Bcc induced by CPT as well as by Plad-B (Figures 2B–2E). DNase 1 partially but significantly reduced the levels of TOP3Bccs, indicating the presence of DNA TOP3Bccs (Figures 2B–2E). Benzonase totally eliminated the TOP3Bcc signals, which further showed that TOP3B works both on DNA and RNA as R-loops form inside cells.

Recruitment of TOP3B to R-loops

Next, we performed S9.6 immunoprecipitation (IP)-western blot experiments (Cristini et al., 2018) to determine whether TOP3B interacts with R-loop structures in human cells. FLAG-tagged TOP3B-complemented TOP3BKO HCT116 cells were treated with CPT (20 μM, 10 min) or Plad-B (5 μM, 2 h), nuclear fractions were separated, and IP was performed using S9.6 antibody after treatment of the nuclear fraction with RNase A and RNase III mix to eliminate the RNA-RNA hybrids and selectively retain the R-loops. Western blotting of the IP samples with anti-FLAG anti-body was used to detect TOP3B in the R-loop samples (Figure 2F).

Figures 2G and 2H show the presence of TOP3B in the samples obtained from cells in which R-loops were induced by CPT (Figure 2G) or Plad-B (Figure 2H). We conclude that TOP3B is recruited to R-loops.

Purified TOP3B cleaves the single-stranded regions of R-loops

To elucidate how TOP3B may act on R-loops from a mechanistic viewpoint, we performed biochemical experiments with recombinant human TOP3B and different R-loop substrates. Electrophoretic mobility shift assays (EMSAs) with purified TOP3B were first performed to compare the affinity of TOP3B for single-stranded DNA (ssDNA), ssRNA, RNA-DNA, DNA-DNA, and RNA-RNA hybrid substrates (Figure S2A). We designed a ssDNA substrate (adopted and modified from the R-loop-prone regions in the gene body of DDX5, which had been previously identified as a TDRD3/TOP3B target; Yang et al., 2014) and ssRNA oligonucleotides having similar sequences. These oligonucleotides were annealed with cDNA or RNA sequences to generate RNA-DNA, DNA-DNA, and RNA-RNA hybrid substrates (Figure S2A). TOP3B showed higher binding affinity for ssDNA over ssRNA substrate (Figure S2B). By contrast, TOP3B displayed low affinity for RNA-DNA, DNA-DNA and RNA-RNA hybrid substrates (Figure S2B).

TOP3B-mediated cleavage assays using these same ssDNA, ssRNA, RNA-DNA, DNA-DNA, and RNA-RNA substrates (Figure S2A) and catalytically active TOP3B showed that, consistent with the binding assays, TOP3B was only able to cleave the ssDNA and ssRNA substrates (Figure S2C). These results demonstrate that TOP3B selectively binds and cleaves single-stranded nucleic acids but not the RNA-DNA hybrid segment of R-loop structures.

Next, we designed D-loop/R-loop-mimicking or mismatch-bubble substrate oligonucleotides based on the DDX5 R-loop prone regions identified as a TDRD3/TOP3B target (Yang et al., 2014). The top strand was annealed with a partially complementary bottom strands, resulting in an 11-base mismatch bubble (Figures 3A–3C) or with a fully complementary strand-forming duplex DNA (Figure 3D). We also generated corresponding D-loop and R-loop substrates (Figures 3A and 3B). TOP3B was only able to cleave the unpaired single-stranded region of the D-loop, R-loop, or mismatch-bubble substrates (Figure 3E, lanes 2, 3, and 4), consistent with the fact that TOP3B can only cleave single-stranded nucleic acids.

Figure 3. TOP3B cleaves the unpaired DNA strand of R-loops.

(A–D) TOP3B substrates derived from the DDX5 R-loop locus. The 41-nt top strand containing a strong TOP3B cleavage site (arrowhead) was labeled with γ-³²P at the 5′-end, and annealed with a partially complementary bottom strand, resulting in an 11-base mismatch bubble (C) or with a fully complementary strand forming double-stranded DNA (dsDNA) (D). Annealing construct C with an additional 19-nt DNA/RNA fragment partially complementary to the unpaired region of the bottom strand leads to the formation of a D-loop (A) or an R-loop (B).

(E) TOP3B cleavage assay using the DNA constructs described in (A)–(D). TOP3B cleavage products are indicated by the arrowhead.

(F) R-loop substrate (Nguyen et al., 2017) containing 2 extended peripheral duplex DNA arms (30 bp) and a centrally located DNA bubble (31 nt) with an RNA-DNA hybrid region (25 bp). Top and bottom DNA strands were radiolabeled with γ-³²P at their 5′-end.

(G) TOP3B cleavage sites in the centrally located unpaired region (opposite to the RNA-DNA hybrid) with 3 different major cleavage sites (37, 43, and 60 nucleotides from the 5′-end).

See also Figure S3.

To confirm these findings, we tested a previously reported “R-loop substrate” (Figure 3F) containing two extended peripheral dsDNA arms (30 bp) and a centrally located bubble of ssDNA (31 nt) in which we could form a 25-bp RNA-DNA hybrid region (Nguyen et al., 2017). We annealed the oligonucleotides with either the top or the bottom strand radiolabeled (at the 5′ end) or with both strands radiolabeled before annealing (Figure 3F). TOP3B only cleaved the top DNA strand of this R-loop substrate in the centrally located unpaired region (opposite to RNA-DNA hybrid) at 3 major sites generating 37-, 43-, and 60-nt products (Figure 3G).

Together, these experiments show that TOP3B can selectively bind and cleave the displaced DNA strand of R-loop structures.

The DEAD-box helicase 5 (DDX5) interacts with TOP3B

Our results thus far show that R-loops recruit TOP3B and that TOP3B engages both DNA and RNA in the R-loop structures. Because we also found that TOP3B only cleaves the single-stranded regions of R-loops, we hypothesized that some DNA/RNA helicase(s) may unwind the R-loops to generate substrates for TOP3B-mediated cleavage and strand passage activities that would contribute to the resolution of R-loops.

To search for TOP3B-interacting helicases, we immunoprecipitated endogenous TOP3B and performed IP-liquid chromatography-tandem mass spectrometry (LC-MS/MS) in human embryonic kidney HEK293 cells. Endogenous TOP3B pulldown in HEK293 cells retrieved 755 significantly enriched proteins (with peptide-spectrum match [PSM] value > 10), which were ranked according to their PSM values (Table S1). DDX5, along with several other DNA/RNA helicases (e.g., SNRNP200, HELZ2, DHX9, DDX3X, DDX17, DDX21, DDX6, DHX15, DHX19A, AQR) were found to be strong interactors of human TOP3B (Figure 4A; Table S1). Because DDX5 (p68) had been reported to be a part of the Top3β-TDRD3 complex associated with heterochromatin formation in Drosophila (Lee et al., 2018), we set out to determine whether human TOP3B works with DDX5 to resolve R-loops. To confirm the first set of results, we repeated the IP-LC-MS/MS analysis in HEK293 cells transfected with hemagglutinin (HA)-tagged TOP3B (Table S2). Pulling down HA-TOP3B retrieved 471 proteins with PSM values >10. Again, we found DDX5 to be an interaction partner of ectopically expressed TOP3B (Figure 4B; Table S2). Similarly, DDX5 was found in HCT116 cells as an interactor of ectopically expressed TOP3B (Figure 4C, left; Table S3).

Figure 4. TOP3B interacts with the R-loop-associated helicase DDX5.

(A) TOP3B pull-down-LC-MS/MS showing that endogenous TOP3B interacts with DDX5 in HEK293 cells. Shown are the number of peptide-spectral matches (PSMs) identified.

(B) HA-tagged pull-down-LC-MS/MS showing TOP3B interaction with DDX5 in HEK293 cells. After transfection of HA-TOP3B, HEK293 cells were subjected to HA-tagged pulldown followed by LC-MS.

(C) FLAG-tagged pull-down-LC-MS/MS showing TOP3B interaction with DDX5 in HCT116 cells both before and after treatments with Plad-B. After transfection of FLAG-TOP3B, TOP3BKO HCT116 cells were subjected to FLAG-tagged pull-down, followed by LC-MS.

(D and E) FLAG-tagged pull-down-western blotting experiment showing that TOP3B interacts with DDX5 both before and after treatments with CPT or Plad-B. SETX was included as a negative control. Following transfection with FLAG-TOP3B, TOP3BKO HCT116 cells were treated with Plad-B (5 μM, 2 h) or CPT (20 μM, 10 min) and subjected to FLAG IP and western blotting. (D) is a representative experiment and (E) displays quantitation of pulled down DDX5 as shown in (D). Data are plotted as means ± SDs. **p ≤ 0.001 and ***p ≤ 0.0001 (2-tailed unpaired t test).

See also Tables S1, S2, and S3.

To determine whether R-loops affect the interaction of human TOP3B with DDX5, we tested their interaction in cells undergoing R-loop induction in response to Plad-B (Figure 4C; Table S3). After pulling down ectopically expressed FLAG-tagged TOP3B (using FLAG ab) from TOP3BKO HCT116 cells (before and after Plad-B treatment to induce R-loop formation), we repeated the mass spectrometry (IP-LC-MS/MS) analyses. Pull down of FLAG-tagged TOP3B before and after R-loop induction (i.e., Plad-B treatment) retrieved 270 and 314 proteins with PSM values >10, respectively (Table S3). DDX5 interaction with TOP3B appeared comparable in both conditions. These results demonstrate that the DDX5-TOP3B interaction is independent of the nature of affinity tag of ectopically expressed TOP3B (FLAG/HA) and that it is also not cell line specific (HCT116/HEK293). In addition, during sample preparation for IP-MS, cell lysates were treated with benzonase, suggesting that the TOP3B-DDX5 interaction is not mediated by nucleic acids.

To validate the TOP3B-DDX5 interaction identified by IP-MS experiments, we performed TOP3B immunoprecipitation (IP; using FLAG antibody) followed by western blotting in HCT116 cells transfected with FLAG-tagged TOP3B before and after treatments with Plad-B as well as with CPT. Consistent with the IP-MS results, DDX5 was detected as an interacting partner of TOP3B, and the interaction was independent of CPT or Plad-B treatments (Figures 4D and 4E). These results suggested that TOP3B forms a prominent complex with DDX5 even before R-loop induction.

DDX5-TOP3B interaction is independent of TDRD3 and DDX5 localizes to chromatin independently of TDRD3

In the fruit fly, TDRD3 has been reported to be a scaffold linking Top3β to p68 (DDX5) (Lee et al., 2018). To determine whether the TOP3B-DDX5 interaction is TDRD3 dependent in human cells, we pulled down endogenous TOP3B from WT and TDRD3KO HCT116 cells (Shuaikun et al., 2022) using TOP3B antibody and performed IP-LC-MS/MS. A total of 198 and 409 proteins came out as TOP3B interaction partners (having PSM values >10) in WT and TDRD3KO HCT116 cells, respectively (Table S4). Whereas DDX5 was detected in the interactome of TOP3B both in WT and TDRD3KO HCT116 cells, as expected, TDRD3 was detected only in WT cells (Figure 5A; Table S4). Consistent with a recent publication (Yuan et al., 2021), an additional DNA-RNA helicase, DHX9, was detected as a TOP3B interaction partner in both WT and TDRD3KO HCT116 cells (Figure 5A).

Figure 5. DDX5 interacts with TOP3B independently of TDRD3.

(A and B) TOP3B and FLAG-TOP3B pull-down-LC-MS/MS showing that TOP3B interacts similarly with DDX5 and DHX9 both in WT and TDRD3KO HCT116 cells. In (B), cells were transfected for 48 h with FLAG-TOP3B before FLAG IP and LC-MS/MS.

(C and D) Pull-down-western blot experiments showing that TOP3B interacts with DDX5 both in WT and siTDRD3-transfected HCT116 cells.

(E and F) In vitro pull-down experiment showing that purified recombinant DDX5 interacts with purified recombinant TOP3B. (E) is a representative experiment, and (F) displays quantitation of pulled down TOP3B as shown in (E). Data are plotted as means ± SDs. **p ≤ 0.001 (2-tailed unpaired t test).

See also Figure S4 and Tables S4 and S5.

Next, we repeated the IP-MS experiment with ectopically expressed TOP3B in WT and TDRD3KO HCT116 cells transfected with FLAG-tagged TOP3B. A total of 1,048 and 806 proteins were found as interaction partners (having PSM values >10) of ectopically expressed TOP3B in WT and TDRD3KO HCT116 cells, respectively (Table S5). DDX5 along with DHX9 were again detected as interaction partners of TOP3B in both WT and TDRD3KO HCT116 cells, and TDRD3, as expected, was only present in WT cells (Figure 5B; Table S5). To confirm the IP-MS results, we performed TOP3B IP (with TOP3B antibody), followed by western blotting in WT and siTDRD3-transfected HCT116 cells. Again, DDX5 was detected as an interacting partner of TOP3B both in WT and TDRD3-knockdown cells (Figure 5C). Reciprocal IP using DDX5 antibody followed by western blotting in WT and siTDRD3-transfected HCT116 cells confirmed that TOP3B was detected in DDX5 immunoprecipitates both in WT and TDRD3 downregulated cells (Figure 5D). These results demonstrate that TOP3B interacts with DDX5 independently of TDRD3 in human cells.

To further establish that DDX5 interacts with TOP3B independently of TDRD3, we performed biochemical experiments with recombinant DDX5 and TOP3B proteins. As outlined in Figure 5E (left), purified DDX5 was incubated with DDX5 antibody, protein-antibody conjugates were incubated with protein A/G magnetic beads, and after washing to remove unbound DDX5, the protein A/G beads containing the DDX5 protein-antibody conjugate were incubated with purified TOP3B. After washing the protein A/G beads to remove unbound TOP3B, samples were analyzed by SDS-PAGE and immunoblotting with TOP3B antibody. Purified DDX5 (bait) directly pulled down TOP3B (prey) (Figures 5E and 5F).

To further establish the lack of effect of TDRD3 on DDX5, we checked the subcellular distribution of DDX5 in WT and siTDRD3-transfected HCT116 and HEK293 cells (Figure S4). Both DDX5 and TDRD3 were found in the cytosolic and nuclear fractions and knocking down TDRD3 had no detectable effect on the subcellular distribution of DDX5. Together, these results indicate a direct interaction of TOP3B with DDX5 independent of TDRD3 and R-loop formation.

DDX5 and TOP3B work in an epistatic manner to resolve R-loops

DDX5 is known to participate in R-loop resolution (Cloutier et al., 2012; Mersaoui et al., 2019; Villarreal et al., 2020; Yu et al., 2020). As we found that TOP3B is in a complex with DDX5, we tested whether they work together to resolve R-loops in cells. To answer this question, we measured R-loop levels in TOP3BKO, DDX5-depleted, and DDX5-depleted TOP3BKO HCT116 cells (Figures 6A–6C). Consistent with previous reports and previous results in the present study, both DDX5 and TOP3B depletion increased cellular R-loop levels (Figures 6B and 6C). Depletion of both DDX5 and TOP3B produced no further increase in R-loop levels compared to only DDX5 or TOP3B depleted cells (Figures 6B and 6C).

Figure 6. DDX5 and TOP3B are epistatic and independent of SETX for R-loop resolution.

(A) Control representative immunoblots showing expression levels of TOP3B and DDX5 in WT (control), TOP3BKO, and siDDX5-transfected WT and TOP3BKO HCT116 cells.

(B and C) Slot-blot analysis of R-loop formation. Genomic DNA was isolated from the indicated cells, slot blotted, crosslinked, and probed with S9.6 antibody. (B) Representative slot blot, and (C) displays the quantitation of R-loop formation from 3 independent experiments. Data are plotted as means ± SDs. n.s., not significant, *p ≤ 0.01, and **p ≤ 0.001 (2-tailed unpaired t test).

(D) Control representative immunoblots showing expression levels of TOP3B and SETX in WT (control), TOP3BKO, and siSETX-transfected WT and TOP3BKO HCT116 cells.

(E and F) Slot-blot analysis of R-loop formation. Genomic DNA was isolated from the indicated cells, slot blotted, crosslinked, and probed with S9.6 antibody. (E) A representative slot blot, and (F) is the quantitation of R-loop formation from 3 independent experiments. Data are plotted as means ± SDs. n.s., not significant, *p ≤ 0.01, and **p ≤ 0.001 (2-tailed unpaired t test).

As a control experiment, we also depleted SETX, which was not present in our TOP3B interactome IP-MS (Figure 6D) and measured R-loop levels (Figures 6E and 6F). As expected, SETX inactivation increased R-loops; but unlike DDX5-TOP3B co-depletion, SETX depletion in TOP3BKO cells further increased cellular R-loop levels compared to only TOP3B- or SETX-depleted cells (Figures 6E and 6F). These results indicate that SETX works independently of TOP3B to resolve R-loops, whereas DDX5 and TOP3B work in an epistatic manner to regulate cellular R-loop levels.

DISCUSSION

How the dual DNA-RNA topoisomerase TOP3B (Ahmad et al., 2016) is implicated in genome stability in eukaryotic cells has remained underexplored (Huang et al., 2018; Mohanty et al., 2008; Yang et al., 2014; Zhang et al., 2019). The present study focuses on the molecular mechanisms by which TOP3B protects the genome from the threat of excessive R-loops. We show that human TOP3B (1) is recruited to cellular R-loops in human cells, (2) can cleave R-loop structures in vitro in their single-stranded region, (3) suppresses cellular R-loops that accumulate both in its absence and after treatments with R-loop-inducing agents (CPT and Plad-B), (4) interacts with DDX5 independent of its scaffolding protein TDRD3, and (5) works in an epistatic manner with DDX5 and in parallel to SETX to resolve cellular R-loops. In addition, we show that CPT and Plad-B are the first small molecules that induce TOP3Bccs indirectly following R-loop accumulation.

R-loops form preferentially in hypernegatively supercoiled DNA, which favors the formation of RNA-DNA heteroduplexes (Barroso et al., 2019; Belotserkovskii and Hanawalt, 2022; Chedin and Benham, 2020) (Figure 7). Topoisomerases can prevent excessive R-loop formation by relaxing hypernegative DNA supercoils. Bacterial type IA topoisomerase, Topo I, was first reported to function in R-loop homeostasis (Drolet et al., 1995; Masse and Drolet, 1999; Phoenix et al., 1997; Usongo et al., 2008), and both Topo I and Topo III can prevent co-transcriptional R-loop formation (Broccoli et al., 2000; Brochu et al., 2018). Loss of the eukaryotic type IB topoisomerase TOP1 and inhibition of TOP1 activity by camptothecin and its derivatives cause R-loop accumulation associated with replication stress, transcriptional block during ribosomal RNA synthesis, and transcription-mediated DNA breaks (El Hage et al., 2010; Marinello et al., 2013, 2016; Sordet et al., 2009; Tuduri et al., 2009). Hence, it is plausible that TOP3B prevents R-loop formation by relaxing hypernegatively supercoiling in the ssDNA segments associated with R-loops (Huang et al., 2018; Yang et al., 2014; Zhang et al., 2019) (Figure 7B, left).

Figure 7. Proposed model for R-loop resolution by TOP3B.

(A) Transcription generates positive supercoiling (+Sc) in front of the translocating RNA polymerase complex (POL) and negative supercoiling (−Sc), behind which are normally suppressed by topoisomerase I and II (TOP1 and TOP2) in duplex DNA.

(B–D) R-loops unwinding by DDX5 helicase generates intertwined DNA-RNA molecules and hypernegative supercoiling (−−Sc) behind the melted RNA-DNA duplex. We propose that TOP3B resolves R-loops by 2 possible mechanisms: TOP3B-mediated DNA relaxation of hypernegative Sc (upper left) and TOP3B-mediated decatenation by cutting the single-stranded DNA and passing RNA (C) or by cutting single-stranded RNA and passing DNA (D).

(E) R-loops are suppressed in the presence of TOP3B and DDX5.

Our results support a DNA-RNA decatenation mechanism by which TOP3B could resolve cellular R-loops (Figures 7C and 7D). Indeed, we observed that TOP3B is recruited to R-loops in cells and that the single-stranded regions of DNA and RNA in R-loops catalytically engage TOP3B as indicated by the induction of DNA and RNA TOP3B cleavage complexes (TOP3Bccs) in R-loops. Our biochemical experiments further show that purified human TOP3B can bind and cleave the single-stranded regions of R-loops, which could decatenate R-loops (Figures 7C and 7D). Although it has been reported that Drosophila melanogaster Top3B can cleave the unpaired strand of R-loops in biochemical constructs (Wilson-Sali and Hsieh, 2002), our report demonstrates this activity for human TOP3B in cells. In addition, our study suggests that TOP3B decatenation is coupled with DNA/RNA duplex unwinding by the DDX5 helicase that could promote TOP3B-mediated cleavage and strand passage activity on the ssDNA and RNA regions of R-loops to promote their resolution (Figures 7B–7D).

Although R-loops form throughout the human genome, these non-B DNA structures are mainly distributed at the transcription start sites (TSSs) of gene promoters and at transcription termination sites (TTSs) (Ginno et al., 2012; Manzo et al., 2018). In the present study, we used three different mechanisms for R-loop induction (i.e., CPT and Plad-B treatment and siSETX transfection) to demonstrate that TOP3B is recruited to and resolves R-loops. CPT is known to induce TOP1cc rapidly within 5–10 min in cells and short treatment with CPT (a highly specific TOP1cc poison; Pommier and Marchand, 2011) is known to rapidly induce R-loops (Cristini et al., 2019; Manzo et al., 2018; Marinello et al., 2016). CPT-induced TOP1 sequestration in TOP1ccs blocks transcription (Ljungman and Hanawalt, 1996; Pommier et al., 2016), increases negative supercoil density in cells (Koster et al., 2007; Pommier et al., 2022), and induces promoter-associated (mostly CGI promoters) R-loops at active TSSs (Cristini et al., 2018; Marinello et al., 2013, 2016). This mechanism of R-loop induction is different from Plad-B, which inhibits the splicing factor SF3B1, leading to aberrant R-loops with relatively slower kinetics (Chakraborty et al., 2018; Nguyen et al., 2017, 2018) that map to intergenic regions, extending downstream of RNA polymerase II (RNAPII) transcribing genes (Castillo-Guzman et al., 2020). Plad-B does not affect all splicing events, and sequestration of nascent RNA by the spliceosome is not likely to be solely controlled by SF3B1 activity alone (Effenberger et al., 2014). This may be the reason we see slower R-loop for-mation kinetics (compared to CPT treatment) in cells after Plad-B-mediated SF3B1 inhibition. By contrast, SETX resolves R-loops formed at transcription pause sites, as well as at double-strand break (DSB) sites (Alzu et al., 2012; Cohen et al., 2018; Mischo et al., 2011; Skourti-Stathaki et al., 2011). As treatments with CPT and Plad-B and the downregulation of SETX recruit TOP3B to newly formed R-loops, it appears that TOP3B can be recruited to and resolves R-loops formed by different mechanisms at diverse genomic locations (i.e., in TSSs of active CGI promoters, transcription pause sites, DSBs, and in intergenic regions, extending downstream of RNAPII transcribing genes). Further mapping studies are warranted to compare the distribution of TOP3B over different R-loop forming loci.

Before our study, it was known that the ATP-dependent RNA helicase DDX5, a member of the DEAD box (Asp-Glu-Ala-Asp [DEAD]) protein family, unwinds RNA-RNA and RNA-DNA duplexes (Hirling et al., 1989; Rossler et al., 2001; Xing et al., 2017) and is an important player in RNA metabolic processes involving structured RNAs (Buszczak and Spradling, 2006; Cloutier et al., 2012; Fuller-Pace, 2013; Xing et al., 2017). Although DDX5 can resolve G-quadruplex structures (Wu et al., 2019), it is primarily known as a key factor in R-loop homeostasis (Gomez-Gonzalez et al., 2021; Kang et al., 2021; Kim et al., 2020; Mersaoui et al., 2019; Sessa et al., 2021; Villarreal et al., 2020; Yu et al., 2020). Dbp2, the yeast homolog of DDX5, resolves the RNA-DNA hybrids of R-loops (Cloutier et al., 2012) and its human counterpart DDX5 also resolves R-loops both in vitro and in vivo (Mersaoui et al., 2019). DDX5 associates with 5′-3′ exoribonuclease 2 (XRN2) to resolve R-loops at TTSs (Mersaoui et al., 2019). In addition, DDX5-depleted cells accumulate R-loops at promoters, near the TSSs and TTSs (Villarreal et al., 2020). DDX5 can also resolve R-loops at DNA DSBs (Yu et al., 2020). Although human DDX5 is known to work with several proteins, including ATAD5, BRCA2, Thrap3 (Gomez-Gonzalez and Aguilera, 2021; Kang et al., 2021; Kim et al., 2020; Sessa et al., 2021), it was not known to associate with human topoisomerases in the context of R-loops, except for a recent study showing that Drosophila p68 (homolog of DDX5) interacts with Top3β-TDRD3 as a part of RNAi machinery and promotes heterochromatin formation (Lee et al., 2018). Our study provides evidence for an epistatic role of TOP3B and DDX5 in R-loop resolution.

Type IA topoisomerases are commonly coupled with helicases, as in the case of reverse gyrase (Yang et al., 2020), yeast Top3, and human TOP3A (Bizard and Hickson, 2020). Human TOP3A, the paralog of TOP3B, works with the BLM helicase (in the context of Holiday junction resolution), FANCM (to suppress sister chromatid exchanges and helps in replication restart), and the SNF2 family helicase/translocase PICH (to induce positive DNA supercoiling, which helps ultra-fine bridge resolution during mitosis) (Ahmad et al., 2017b; Bizard and Hickson, 2020; Zhao et al., 2018). Here, we provide evidence for the coupling of human TOP3B with the DDX5 helicase in the context of R-loop resolution. By contrast to p68, the Drosophila ortholog of DDX5, whose interaction with TOP3B has been reported to be TDRD3 mediated in the context of heterochromatin regulation (Lee et al., 2018), we found that human TOP3B interacts with DDX5 independently of TDRD3. Based on the epistatic relationship between DDX5 and TOP3B, we propose that DDX5-mediated unwinding of R-loops may generate topological tensions and single-stranded nucleic acids regions where TOP3B-mediated DNA/RNA strand cleavage/strand passage could release topological tension and resolve R-loops (Figure 7).

As we were looking for TOP3B interacting helicases by IP-MS experiments, we identified many other helicases in addition to DDX5, including members of DEAD box and DExH box families as interactors with both endogenous and ectopically expressed TOP3B protein (e.g., SNRNP200, HELZ2, DHX9, DDX3X, DDX17, DDX21, DDX6, DHX15, DHX19A, RECQL, AQR) (Tables S1, S2, and S3). One may ask, why so many? Consistent with the role of TOP3B in suppressing R-loops, many of these helicases have known roles in R-loop metabolism, namely, AQR, DDX21, DHX9, DHX19A (Chakraborty et al., 2018; Cristini et al., 2018; Hodroj et al., 2017; Paulsen et al., 2009; Sollier et al., 2014; Song et al., 2017). We also found that the helicase DHX9 interacts with TOP3B, which is consistent with a recent study reporting a role for DHX9 in R-loop resolution (Yuan et al., 2021). However, by contrast to our results showing that DHX9 interacts with TOP3B independently of TDRD3, that study found that the TOP3B interaction was TDRD3 dependent (in the MCF7 cell line) and that the TDRD3-DHX9 complex worked independently of TOP3B to resolve promoter-associated R-loops (Yuan et al., 2021). Further studies are therefore warranted to decipher how DHX9 and other TOP3B interacting helicases couple with TOP3B for the resolution of different R-loops. In addition, DEAD box and DExH box family helicases often work together in complexes, and an interaction between DDX5 and DHX9 has recently been reported (Kim et al., 2020). It is therefore not excluded that TOP3B could form complexes with more than one DEAD box and DExH box helicases for R-loop resolution. Also, TOP3B may work with different helicases to resolve R-loops at different locations and under diverse circumstances (e.g., promoter-associated R-loops, transcription pause site-associated R-loops, ribosomal R-loop, DSB-associated R-loops).

Limitations of the study

One of the potential caveats of the study is the use of TOP3B overexpression to show that TOP3B suppresses R-loops. However, the fact that overexpression of the active site mutant (catalytically inactive) Y336F TOP3B could not suppress cellular R-loops in TOP3BKO cells demonstrates the catalytic role of TOP3B in R-loop homeostasis.

Although we demonstrate that endogenous TOP3Bccs form on R-loops, to detect the binding to RNA and DNA in the R-loops, we overexpressed TOP3B to maximize the signal.

Finally, this study does not include direct biochemical experiments with purified TOP3B, DDX5, and R-loop substrates, and further experiments are warranted.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by lead contact Yves Pommier.

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed Materials Transfer Agreement.

Data and code availability

• This mass spectroscopy data generated in this study is available as Tables S1, S2, S3, S4 and S5. Original imaging data supporting the current study have been deposited at Mendeley Data (https://doi.org/10.17632/htjp2m4sdb.1) and are publicly available as of the date of publication. The DOI is listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Monoclonal ANTI-FLAG® M2 Mouse Monoclonal antibody, Sigma-Aldrich	Sigma-Aldrich	Cat# F1804; RRID: AB_262044
Anti-GAPDH Rabbit Monoclonal Antibody, Unconjugated, Clone 14C10	Cell Signaling Technology	Cat# 2118; RRID: AB_561053
Anti-TOP3B Rabbit Monoclonal antibody [EP7779] - C-terminal (ab183520)	Abcam	RRID: AB_183520
Rabbit Anti-HA-Tag Monoclonal Antibody, Unconjugated, Clone C29F4	Cell Signaling Technology	Cat# 3724; RRID: AB_1549585
Sheep Anti-Mouse IgG ECL Antibody, HRP Conjugated, GE Healthcare	GE Healthcare	Cat# NA9310–1mL; RRID: AB_772193
Donkey Anti-Rabbit IgG ECL Antibody, HRP Conjugated, GE Healthcare	GE Healthcare	Cat# NA9340–1mL; RRID: AB_772191
Anti-dsRNA monoclonal antibody J2 (mouse, IgG2a, kappa chain)	Jena Bioscience	Cat# RNT-SCI-10010200; RRID: AB_2651015
Anti-DNA-RNA Hybrid Antibody, clone S9.6	Millipore Sigma	Cat# MABE1095; RRID: AB_2861387
Recombinant Anti-DDX5 antibody [EPR7239] (ab126730)	Abcam	RRID: AB_126730
Recombinant Anti-Senataxin antibody [BLR050F] - BSA free (ab243904)	Abcam	Cat# ab243904; RRID: AB_2893218
TDRD3 (D3O2G) Rabbit mAb #5942	Cell Signaling Technology	Cat# 5942; RRID: AB_2797626
HDAC1 (D5C6U) XP® Rabbit mAb #34589	Cell Signaling Technology	Cat# 34589S; RRID: AB_2756821
Histone H3 Rabbit Antibody #9715	Cell Signaling Technology	Cat# 9715S; RRID: AB_331563
Bacterial and virus strains
NEB® 5-alpha Competent E. coli (High Efficiency)	NEW ENGLAND BioLabs Inc.	Cat# C2987H
MAX Efficiency™ DH10B Competent Cells	ThermoFisher Scientific	Cat# 18297010
DE77, a DH10Bac-derived strain	Bac-to-Bac system, Thermo Fisher	N/A
Chemicals, peptides, and recombinant proteins
DMEM - Dulbecco’s Modified Eagle Medium	ThermoFisher Scientific	Cat# 11965–092
Fetal Bovine Serum	Gemini	Cat# 100–106
Penicillin-Streptomycin	ThermoFisher Scientific	Cat# 15140–122
Trypsin-EDTA (0.05%)	ThermoFisher Scientific	Cat# 25300054
cOmplete Mini, EDTA-free (protease inhibitor cocktail)	Roche	Cat# 11836170001
Recombinant Human TOP3B	(Saha et al., 2020)	N/A
PEI	Polysciences	Cat# 23966
Lipofectamine 3000 Reagent	ThermoFisher Scientific	Cat# L3000015
Lipofectamine® RNAiMAX transfection reagent	ThermoFisher Scientific	Cat# 13778150
Benzonase	Sigma-Aldrich	Cat# E8263
Pierce™ ChIP-grade Protein A/G Magnetic Beads	ThermoFisher Scientific	Cat# 26162
Tris-glycine SDS sample buffer	Novex	Cat# LC2676
SuperSignal™ West Femto Maximum Sensitivity Substrate	ThermoFisher Scientific	Cat# 34095
TRIzol™ Reagent	ThermoFisher Scientific	Cat# 15596026
Micrococcal Nuclease Solution (≥100 U/μL)	ThermoFisher Scientific	Cat# 88216
RNase A	ThermoFisher Scientific	Cat# EN0531
RNase T1	ThermoFisher Scientific	Cat# EN0542
RNase H	New England Biolabs	Cat# M0297L
ShortCut® RNase III	New England BioLabs	Cat# M0245L
Invitrogen™ TURBO™ DNase (2 U/μL)	ThermoFisher Scientific	Cat# AM2238
N-Ethylmaleimide	Millipore Sigma	Cat# E3876–25G
DNAzol	ThermoFisher Scientific	Cat# 10503027
Glycogen, RNA grade	ThermoFisher Scientific	Cat# R0551
Invitrogen™ Proteinase K Solution	ThermoFisher Scientific	Cat# 4333793
Invitrogen™ Proteinase K Solution (20 mg/mL), RNA grade	ThermoFisher Scientific	Cat# 25530049
N-Lauroylsarcosine sodium salt	Millipore Sigma	Cat# L9150
Camptothecin (CPT)	Developmental Therapeutics Program (NCI/NIH)	NSC#94600
Pladienolide B, 0.5 mg (CAS 445493–23-2)	Santa Cruz Biotechnology	Cat# sc-391691
Gibco™ Hygromycin B (50mg/mL)	ThermoFisher Scientific	Cat# 10–687-010
DDX5 (NM_004396) Human Recombinant Protein	OriGene	Cat# TP300371
Mini Quick Spin Oligo Columns	Millipore Sigma	Cat# 11814397001
Critical commercial assays
MycoAlert kit - 100 tests	Lonza	Cat# LT07–318
Deposited data
Original imaging data	This paper: Mendeley Data	https://data.mendeley.com/datasets/htjp2m4sdb/draft?a=f39bb027-f618-4d44-a18b-2c36fae27413
Experimental models: Cell lines
HEK293	ATCC	CRL-1573
HCT116	Developmental Therapeutics Program (NCI/NIH)	N/A
K562	ATCC	CCL-243™
TDRD3KO HCT116	(Shuaikun et al., 2022)	N/A
TOP3BKO HCT116	This study	N/A
TOP3BKO K562	This study	N/A
Oligonucleotides
TOP3B forward primer for cloning into pcDNA3-HA: 5′-GCTTGGATCC AAG ACT GTG CTC ATG GTT-3′	IDT oligo	N/A
TOP3B reverse primer for cloning into pcDNA3-HA: 3′- CCAAGAATTCTCATA CAAAGTAGGCGGC-5′	IDT oligo	N/A
TOP3B forward primer for cloning into Gateway entry vector pENTR3C: 5′-CGGGGTACCATGAAG ACTGTGCTCATGG-3′	IDT oligo	N/A
TOP3B reverse primer for cloning into Gateway entry vector pENTR3C: 5′-AGGCTACATCAGCTACG TACGGACAGAGACCACC-3′	IDT oligo	N/A
ON-TARGETplus SMARTpool siRNA targeting DDX5 GCAAAUGUCAUGGAUGUUA CAACCUACCUUGUCCUUGA GCAUGUCGCUUGAAGUCUA CCAAAUAUGCACAAUGGUA	Dharmacon	CAT#: L-003774–00-0005
SMARTpool: ON-TARGETplus SETX siRNA GCACGUCAGUCAUGCGUAA UAGCACAGGUUGUUAAUCA AAAGAGUACUUCACGAAUU GGACAAAGAGUUCGAUAGA	Dharmacon	CAT#: L-021420–00-0005
ON-TARGETplus SMARTpool siRNA targeting TDRD3; CUGAAGCACAUAACGGAAA CAAAUGUGAUAGACCGUAU UAGCAGAGAACUUGAUCGA CCGAAAUAGGGAAGUUUUA	Dharmacon	CAT#: L-014655–01-0010
Recombinant DNA
Human TOP3B-Myc-FLAG	OriGene	Cat# RC223204
pcDNA3-TOP3B-HA	(Saha et al., 2020)	N/A
pENTR3C-TOP3B	(Saha et al., 2020)	N/A
Software and algorithms
GraphPad Prism 9 (software for drawing graphs and statistics analysis)	GraphPad	N/A
Fiji	(Schindelin et al., 2012)	http://fiji.sc

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

HEK293 (ATCC, Manassas, VA), K562 (ATCC, Manassas, VA) and HCT116 (Developmental Therapeutics Program, National Cancer Institute) cell lines were grown in Dulbecco’s modified Eagle’s medium (Life Technologies, Carlsbad, CA) supplemented with 10% Fetal Bovine Serum (Gemini, West Sacramento, CA, 100–106) and 1% penicillin-streptomycin (ThermoFisher Scientific, 15140122) at 37°C in humidified 5% CO₂ chamber. Cell lines were tested for mycoplasmas (Lonza, LT07–318). Tni-FNL cells were cultured in Gibco Express 5 medium with 18mM glucose. Cell line authentication was carried out using short tandem repeat analysis at Frederick National Laboratory, NCI-NIH.

To generate TOP3BKO cells, the sequence TGCTCAGTCCACGAGTACAC targeting the second exon of TOP3B or the sequence TGAATGCACAGCCGATTCGC targeting the fifth exon of TOP3B was cloned into px330. A Hygromycin B-resistance gene flanked by homology arms (of ~1 kb) upstream and downstream of either target site was cloned into pCR2.1, serving as the HDR sequence. The modified pX330 plasmid and the corresponding HDR sequence was co-transfected into HCT116 cells with Lipofectamine 3000 following the manufacturer’s instructions or into K562 cells using Neon Transfection System (Pulse voltage = 1000 v, Pulse width = 50 ms, Pulse number = 1). 48 h post transfection, transfected cells were selected with 400 μg/mL of Hygromycin B in HCT116 cells and 200 μg/mL of Hygromycin B in K562 cells. Single clones of either cell lines were established from the Hygromycin-resistant population and subsequently screened by immunoblotting analysis to identify TOP3BKO cells. TDRD3KO HCT116 cells were generated as described (Shuaikun et al., 2022).

METHODS DETAILS

Mammalian expression constructs and transient expression in mammalian cell

Human TOP3B-Myc-FLAG cDNA ORF (CAT#: RC223204) Clone were purchased from OriGene. The full-length cDNAs of TOP3B was PCR-amplified from TOP3B-Myc-FLAG cDNA ORF Clone (CAT#: RC223204) respectively using cloning primers (TOP3B forward primer 5′-GCTTGGATCCAAGACTGTGCTCATGGTT-3′; TOP3B reverse primer 3′- CCAAGAATTC TCATACAAAGTAGGCGGC-5′) and subcloned into pcDNA3-HA with BamHI and EcoRI sites. Plasmids were transfected in HCT116 and HEK293 cells using Lipofectamine 3000 Reagent (CAT#: L3000015, ThermoFisher Scientific) according to the manufacturer’s protocol for 48–72 h.

Recombinant human TOP3B production

Recombinant TOP3B was purified from baculovirus system as described previously (Saha et al., 2020). Briefly, TOP3B was initially PCR amplified from Human TOP3B-Myc-FLAG cDNA ORF (CAT#: RC223204) using forward primer: 5′-CGGGGTACCATGAAG ACTGTGCTCATGG-3′ and reverse primer: 5′-CCGCTCGAGTCATACAAAGTAGGCGGCCAG-3′ and cloned into Gateway entry vector pENTR3C (Invitrogen, CAT#: A10464). TOP3B was then subcloned by Gateway LR recombination (Thermo Fisher) into pDest-635 (22876-X01–635) for insect cell expression which includes an N-terminal His6X tag. Bacmid was prepared in DE77, a DH10Bac-derived strain (Bac-to-Bac system, Thermo Fisher) and after purification, bacmid DNA was verified by PCR amplification across the bacmid junctions. Bacmids were transfected in SF-9 cells using PEI (1 mg/mL with 5% glucose; Polysciences, CAT#: 23966), recombinant baculovirus stock was collected and titrated using ViroCyt (Beckamn). Two liters of Tni-FNL cells were set in a baffled 5-liter Thomson Optimum Growth Flask in Gibco Express 5 medium with 18mM glucose at a cell density of 1 × 10⁶ cells/mL at 27°C and 24 hrs later infected at a MOI (multiplicity of infection) of 3. After 3 days of incubation at 21°C, cell pellets were collected by centrifugation at 2000 rpm for 11 min and flash frozen on dry ice. Cell pellet was thawed by the addition of 200 mL of lysis buffer (20 mM HEPES, 300 mM NaCl, 1 mM TCEP and 1:100 v/v of Sigma protease inhibitor P8849) and homogenized by vortexing. The cells were lysed by performing two passes on an M-110EH-30 microfluidizer (Microfluidics) at 7000 psi, clarified at 100K x g for 30 minutes at 4°C using an optima L-90K ultracentrifuge (Beckman), filtered (0.45 micron) and applied to a f20 mL IMAC HP column (GE Scientific) that was pre-equilibrated with lysis buffer containing 50 mM imidazole on a Bio-Rad NGC. Column was washed with lysis buffer containing 50 mM imidazole and proteins were eluted with lysis buffer containing 500 mM imidazole. After SDS-PAGE/Coomassie staining, positive fractions were pooled, dialyzed to 20 mM HEPES, 50 mM NaCl, 1 mM TCEP, 0.5 mM PMSF, 1:1000 v/v of PI, 50% glycerol, pH 7.2. Protein concentration was determined (0.88 mg/mL) and stored at −80°C for future use.

TOP3B mediated cleavage of oligonucleotide substrates

Oligonucleotide substrates were labeled on the 5′ end with [γ-³²P] ATP and T4 Polynucleotide Kinase before passing through mini Quick Spin Oligo Columns (Millipore Sigma) and annealed after heating at 95°C for 5 minutes. 6 nM of labeled substrate was incubated with 18 nM of recombinant TOP3B in 5 mM Tris-HCl (pH 7.5), 100 mM potassium glutamate (pH 7.0), 0.1 mM MgCl2, 0.02% v/v Tween-20 and 2mM DTT, incubated at 37 C for 30 minutes before addition of 0.2% SDS and 1 volume of gel loading buffer [96% (v/v) formamide, 10 mM ethylenediaminetetraacetic acid (EDTA), 1%(w/v) xylene cyanol and 1%(w/v) bromophenol blue]. Samples were analyzed by 20% denaturing polyacrylamide gel (containing 7 M Urea) electrophoresis gels, which were dried and exposed on Phosphor Imager screens. Imaging was done using a Typhoon 8600 and ImageQuant software (GE Healthcare, UK).

siRNA transfection

Silencing of DDX5, SETX and TDRD3 were done using ON-TARGETplus SMARTpool siRNA targeting DDX5 (CAT#: L-003774–00-0005, Dharmacon), SMARTpool: ON-TARGETplus SETX siRNA (CAT#: L-021420–00-0005, Dharmacon), ON-TARGETplus SMARTpool siRNA targeting TDRD3 (CAT#: L-014655–01-0010, Dharmacon), respectively. All siRNAs were used at a final concentration of 25 nM and transfected using Lipofectamine^® RNAiMAX transfection reagent (CAT#: 13778150, ThermoFisher Scientific) following the manufacturer’s protocol for 48–72 h.

RADAR assay and detection of DNA and RNA TOP3Bccs

RADAR assay was performed for detection of TOP3Bccs as described previously (Saha et al., 2020). Wild type HCT116 cells, HA or FLAG-tagged WT TOP3B transfected HCT116 cells (1 × 10⁶) were treated with CPT or Plad-B, washed with PBS and lysed by adding 1 mL DNAzol (ThermoFisher Scientific, CAT#:10503027). Nucleic acids were precipitated following addition of 0.5 mL of 100% ethanol, incubation at −20°C for 5 min and centrifugation (12,000 3 g for 10 min). Precipitates were washed twice in 75% ethanol, resuspended in 200 μL TE buffer, heated at 65°C for 15 minutes, followed by shearing with sonication (40% power for 15 s pulse and 30 s rest 5 times). Samples were centrifuged at 15,000 rpm for 5 min and the supernatant containing nucleic acids with covalently bound proteins were collected. Nucleic acid containing protein adducts were quantitated, slot-blotted and TOP3Bccs were detected with either rabbit monoclonal Anti-TOP3B antibody [EP7779] - C-terminal (Abcam, CAT#: ab183520) or mouse monoclonal anti-FLAG M2 antibody (Millipore Sigma, St. Louis, MO, CAT#: F1804) or rabbit Anti-HA-Tag Monoclonal Antibody, Unconjugated, Clone C29F4 (Cell Signaling Technology, CAT#: 3724).

For detection of DNA and RNA TOP3Bccs, 10 μg RADAR assay samples were digested either with excess RNase A (200 μg/mL; ThermoFisher Scientific Cat# EN0531) and RNase T1 (200 units/mL; ThermoFisher Scientific, Cat# EN0542) mix, or with DNase 1 (10 units; ThermoFisher Scientific, Cat# AM2238) or Benzonase (250 units; Sigma-Aldrich, Cat# E8263). Samples were ethanol-precipitated, resuspended, quantitated, slot-blotted and TOP3Bccs were detected with Anti-TOP3B antibody [EP7779] - C-terminal (Abcam, CAT#: ab183520).

R-loop detection by dot-blot method using s9.6 ab

For R-loop detection by slot-blot, genomic DNA was extracted from HCT116 cells using DRIP protocol as described previously (Sanz et al., 2021; Sanz and Chedin, 2019). Briefly, cells were lysed in TE buffer containing SDS and proteinase K (at 37°C overnight), phase separated using phenol/chloroform/isoamyl alcohol (25:24:1), ethanol precipitated and resuspended in TE buffer. Genomic DNA was digested using cocktail of restriction enzymes (HindIII, SspI, EcoRI, BsrGI and XbaI; 30 U each), treated with RNase A (10 μg/mL; ThermoFisher Scientific Cat# EN0531) and shortcut RNase III (2 units; New England Biolabs; Cat# M0245L) and again purified by phenol/chloroform/isoamyl alcohol (25:24:1) extraction. Increasing concentrations of genomic DNA were spotted on a nitrocellulose membrane, crosslinked with UV light (120 mJ/cm2)), blocked with PBS-Tween (0.1%) buffer and 5% non-fat milk (Room temperature for 1hr) and incubated with mouse S9.6 antibody (1:500 dilution, overnight at 4°C, Millipore Sigma, Cat# MABE1095). After washing with PBS-Tween (0.1%), membrane was incubated with HRP-conjugated anti-mouse secondary antibody, washed, and developed with ECL techniques. In case of RNase H treated control, 10 μg genomic DNA was pre-incubated with 20 U of RNase H for three hours at 37°C.

RNA/DNA hybrid IP

RNA/DNA hybrid immunoprecipitation experiment was performed as described previously with minor modifications (Cristini et al., 2018). Briefly, cells were lysed in 85 mM KCl, 5 mM PIPES (pH 8.0), and 0.5% NP-40 for 10 min on ice and centrifuged to isolate nuclei. Pelleted nuclei were resuspended in RSB buffer (10 mM Tris-HCl pH 7.5, 200 mM NaCl, 2.5 mM MgCl₂) with 0.2% sodium deoxycholate, 0.1% SDS, 0.05% sodium lauroyl sarcosinate and 0.5% Triton X-100, and extracts were sonicated for 10 min. Extracts were then diluted 1:4 in RSB with 0.5% Triton X-100 and subjected to IP with the S9.6 antibody, bound to protein A magnetic beads (ThermoFisher Scientific/Pierce; Cat#: 88845). RNase A (10 μg/mL; ThermoFisher Scientific Cat# EN0531) and shortcut RNase III (2 units; New England Biolabs; Cat# M0245L) were added during IP. Beads were washed (4x with RSB with 0.5% Triton X-100; 2x with RSB), separated using magnetic racks and eluted in Tris-Glycine SDS Sample Buffer (2X) (ThermoFisher Scientific Novex^™; Cat#: LC2676) for SDS-PAGE analysis.

Immunoprecipitation (IP)

Cells were washed with PBS and incubated on a shaker with IP lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.2% Triton X-100, 5% glycerol, 1 mM DTT, 20 mM N-ethylmaleimide and protease inhibitor cocktail) supplemented with 2 μL benzonase (250 units/μL, Sigma-Aldrich, CAT#: E8263) followed by 30 minutes incubation in ice. After brief sonication (40% power for 30 s pulse and 1 min rest 3 times), samples were centrifuged at 15,000 X g at 4°C for 30 min and the supernatant was collected and an aliquot (20 μL) of the supernatant was saved as input. The rest of the supernatant was diluted in IP lysis buffer containing desired antibody (3 μg/tube) and rotated overnight at 4°C. Next day, 30 μL Protein A/G magnetic beads (ThermoFisher Scientific Cat#: 26162) were added and incubated with the lysates for anther 4 hrs. After magnetic separation, beads were washed with RIPA buffer 2 times. Finally beads were resuspended in Tris-Glycine SDS Sample Buffer (2X) (ThermoFisher Scientific Novex^™; Cat#: LC2676) for SDS-PAGE and immunoblotted with different antibodies as indicated or further processed for MS analysis.

Mass spectrometry analysis

MS samples were either separated by SDS-PAGE for in-gel trypsin digestion or in-solution digested with trypsin using S-traps following the manufacturer’s protocol. (Wisniewski et al., 2009). Dried peptides were solubilized in 2% acetonitrile, 0.5% acetic acid, 97.5% water for mass spectrometry analysis. They were trapped on a trapping column and separated on a 75 μm 3 15 cm, 2 μm Acclaim PepMap reverse phase column (Thermo Scientific) using an UltiMate 3000 RSLCnano HPLC (Thermo Scientific). Peptides were separated at a flow rate of 300 nL/min followed by online analysis by tandem mass spectrometry using either a Thermo Orbitrap Fusion mass spectrometer or a Thermo Orbitrap Exploris 480 mass spectrometer. Peptides were eluted into the mass spectrometer using a linear gradient from 96% mobile phase A (0.1% formic acid in water) to 55% mobile phase B (0.1% formic acid in acetonitrile). Proteome Discoverer 2.4 (Thermo) was used to search the data against human proteins from the UniProt database using SequestHT. The search was limited to tryptic peptides, with maximally two missed cleavages allowed. Cysteine carbamidomethylation was set as a fixed modification, and methionine oxidation was set as a variable modification. The precursor mass tolerance was 10 ppm, and the fragment mass tolerance was 0.6 Da or 0.02 Da for Fusion and Exploris 480 data, respectively. The Percolator node was used to score and rank peptide matches using a 1% false discovery rate.

Western blotting and antibodies

To prepare whole cell lysates for western blotting, cells were resuspended with RIPA buffer (150 mM NaCl, 1% NP- 40, 0.5% Sodium deoxycholate, 0.1% SDS, 50 mM Tris pH 7.5, 1 mM DTT and protease inhibitor cocktail). After thorough mixing, samples were agitated at 4°C for 30 min, sonicated for 30 seconds with 40% pulse (3 times), centrifuged at 15,000 X g at 4°C for 15 min, and supernatant was collected.

For detection of total cellular TOP3B (both free and nucleic acid bound), cells were washed with PBS and incubated on a shaker with IP lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.2% Triton X-100, 5% glycerol, 1 mM DTT, 20 mM N-ethylmaleimide and protease inhibitor cocktail) supplemented with 2 μL benzonase (250 units/μL, Sigma-Aldrich, CAT#: E8263) followed by 30 minutes incubation in ice. After brief sonication (40% power for 30 s pulse and 1 min rest 4 times), samples were centrifuged at 15,000 X g at 4°C for 30 min and the supernatant was collected.

Lysed samples were mixed with tris-glycine SDS sample buffer (Novex, LC2676) and loaded onto Novex tris-glycine gels (Novex). Blotted membranes were blocked with 5% non-fat dry milk in PBS with 0.1% Tween-20 (PBST). The primary antibodies were diluted in 5% milk in PBST by 1:1000 for Mouse monoclonal anti-FLAG M2 (Millipore Sigma, St. Louis, MO, CAT#: F1804), 1:10000 for Rabbit monoclonal anti-GAPDH (Cell Signaling Technology, Danvers, MA, CAT#: 2118S), 1:1000 for Rabbit monoclonal anti-HA (Cell Signaling Technology, Danvers, MA, CAT#: 3724S), 1:1000 for Rabbit monoclonal anti-TDRD3 (Cell Signaling Technology, Danvers, MA, CAT#: 5942S), 1:1000 for Rabbit monoclonal anti-DDX5 antibody (abcam, Waltham, MA, CAT#: ab126730), 1:1000 for Rabbit monoclonal anti-TOP3B antibody (abcam, Waltham, MA, CAT#: ab183520), 1:1000 for Rabbit monoclonal anti-Senataxin antibody (abcam, Waltham, MA, CAT#: ab243904), 1:10000 for Rabbit polyclonal anti-H3 antibody (Cell Signaling Technology, Danvers, MA, CAT#: 9715S), 1:1000 for Rabbit monoclonal anti-HDAC1 (Cell Signaling Technology, Danvers, MA, CAT#: 34589S). Secondary antibodies were diluted (1:10000) in 5% non-fat milk in PBST and signal was detected by ECL chemiluminescence reaction (Thermo Scientific, Waltham, MA).

Subcellular fractionation

Subcellular fractionation of HEK293 cells was performed using Thermo Scientific^™ subcellular protein fractionation kit for cultured cells (Thermo Scientific, Waltham, MA, cat# 78840) following the manufacturer’s instructions. The cytoplasmic, nucleoplasmic and the chromatin fractions were subjected to Western blotting using indicated antibodies.

In vitro pull-down experiment

Purified DDX5 protein was incubated with DDX5 antibody and protein-antibody conjugate was further incubated with protein A/G magnetic beads. After subsequent washing to remove unbound DDX5 protein, protein A/G beads containing DDX5 protein-antibody conjugate was incubated with purified TOP3B. Finally, Protein A/G beads were washed to remove any unbound TOP3B, analyzed by SDS-PAGE and western blotting using TOP3B antibody.

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantifications were carried out using the Fiji software. Data are provided as means ± standard deviations (SD) from the number of independent experiments performed. Statistical analyses and graphical representation were carried out using GraphPad prism 7 software. Statistical test methods are described in each figure legend. Statistical significance is represented by * and ** and indicate a computed p value <0.01 and <0.001 respectively. n.s. = not significant.

Highlights.

• The topoisomerase TOP3B forms DNA and RNA TOP3B cleavage complexes in cellular R-loops

• TOP3B works in an epistatic manner with the helicase DDX5 to resolve cellular R-loops

• Recombinant human TOP3B cleaves R-loop structures in their single-stranded region

• TOP3B can resolve cellular R-loops by a DNA-RNA decatenation mechanism