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

Abstract

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.

Keywords: CP: Molecular biology; DDX5; DNA-RNA hybrids; DNA/RNA topoisomerase 3B; R-loop; TDRD3; senataxin; topoisomerase cleavage complexes.

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.

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.

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.

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.

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.

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).

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.