WRN helicase unwinds Okazaki fragment-like hybrids in a reaction stimulated by the human DHX9 helicase

Abstract

Mutations in the Werner gene promote the segmental progeroid Werner syndrome (WS) with increased genomic instability and cancer. The Werner gene encodes a DNA helicase (WRN) that can engage in direct protein-protein interactions with DHX9, also known as RNA helicase A or nuclear DNA helicase II, which represents an essential enzyme involved in transcription and DNA repair. By using several synthetic nucleic acid substrates we demonstrate that WRN preferably unwinds RNA-containing Okazaki fragment-like substrates suggesting a role in lagging strand maturation of DNA replication. In contrast, DHX9 preferably unwinds RNA-RNA and RNA-DNA substrates, but fails to unwind Okazaki fragment-like hybrids. We further show that the preferential unwinding of RNA-containing substrates by WRN is stimulated by DHX9 in vitro, both on Okazaki fragment-like hybrids and on RNA-containing 'chicken-foot' structures. Collectively, our results suggest that WRN and DHX9 may also cooperate in vivo, e.g. at ongoing and stalled replication forks. In the latter case, the cooperation between both helicases may serve to form and to dissolve Holliday junction-like intermediates of regressed replication forks.

Figure 1.

Unwinding of short fully hybridized strands. Time course for the unwinding of M2:M1 by WRN (A) or DHX9 (B), unwinding of R2:R1 by WRN (C) and DHX9 (D), and melting of M2:R1 by WRN (E) or DHX9 (F) 28 nM of WRN or of DHX9 and 1 nM of substrates were used for each experiment. Triangle represents heat-denatured DNA substrate control, NE no enzyme. The positions of the substrates and reaction products are indicated; double lines depict RNA containing strands. The asterisk represents the 5′-labeled nucleotide. Quantifications of duplex unwinding by WRN (G) and DHX9 (H) are also depicted. Error bars indicate SD as derived from three independent experiments.

Figure 2.

Unwinding of Okazaki fragment-like structures. (A) Time course of unwinding of R8D:M2. (B) Time course of unwinding of R4D:M2. (C) Time course of unwinding of R1D:M2. Unwinding of either construct by 28 nM WRN (left) or DHX9 (right) and 1 nM of substrate is shown in a 10-min time scale. Triangle represents, heat-denatured control; NE, no enzyme. 8R, 4R and 1R indicate the numbers of ribonucleotides at the 5′ ends. (D) Quantification of unwinding of R8D (diamonds), R4D (squares) and R1D (triangles) by WRN. Error bars indicate SD as derived from three independent experiments.

Figure 3.

DHX9 stimulates the unwinding of Okazaki fragment-like structures by WRN. (A) Unwinding of 1 nM R8D:M2 catalyzed by 14 nM WRN (lanes 3 and 6), 14 nM DHX9 (lanes 4 and 7) or 14 nM each of WRN and DHX9 (lanes 5 and 8) after 3 min (lanes 3–5) or 5 min (lanes 6–8) incubation at 37°C. (B) The same experiment as shown in panel A except that 1 nM R1:M2 was used as substrate. (C) The same experiment as shown in panel A except that 1 nM M1:M2 was used. (D) Quantification of the unwinding of R8D:M2 by 14 nM WRN in the presence of 14 nM DHX9. (E) Unwinding of 1 nM R8D:M2 catalyzed by 21 nM poisoned WRN (w, lane 3), 21 nM poisoned DHX9 (d, lane 4), and 21 nM each of NEM-poisoned WRN and DHX9 (wd, lane 5). Poisoning could be rescued by adding 0.1 M DTT together with NEM (lanes 6–8). (F) Unwinding of 1 nM R8D:M2 catalyzed by 21 nM WRN (lane 3), DHX9 (lane 4), WRN plus DHX9 (lane 5), NEM-poisoned WRN plus untreated DHX9 (w, lane 6), NEM-poisoned DHX9 plus untreated WRN (Wd, lane 7), and poisoned WRN plus poisoned DHX9 (wd, lane 8). Error bars indicate SD as derived from three independent experiments.

Figure 4.

WRN binds to R8D:M2, M2:R1, and M2:M1 in an ATPγS-dependent manner. (A) Substrate (50 fmol) was incubated without (NE, lane 1) and with increasing amounts of WRN (as indicated) in the absence of any nucleotide. (B) The same experiment as shown in panel A, but this time in the presence of 1 mM ATPγS. The radiolabeled DNA species were analyzed on a native 4% PAGE and visualized by PhosphorImaging. The bands in the slots on top of the gel were also visible in many no enzyme controls and probably represent aggregated substrate.

Figure 5.

DHX9 stimulates the WRN-catalyzed branch migration of an RNA-primer containing HJ. (A) WRN-catalyzed (28 nM) unwinding of 1 nM of the synthetic RNA-primer HJ substrate R8M4:M2*:S1:S2. WRN unwound about half of the substrate within 30 min at 37°C (lane 3) while DHX9 failed to do so (lane 4). With 28 nM each of DHX9 and WRN complete unwinding was achieved (lane 5). The reaction products were resolved on a native 10% polyacrylamide gel and visualized by PhosphorImaging. (B) 60 fmol radiolabeled HJ were incubated with the indicated amounts of WRN and/or DHX9. After 20 min preincubation at 37°C in the presence of 1 mM ATPγS, the products were separated by electrophoresis through a 4% non-denaturing polyacrylamide gel at 4°C and visualized using a PhosphorImager. The bands in the slots on top of the gel were also visible in the no enzyme control (lane 1) and probably represent aggregated substrate.

Figure 6.

Model for a cooperation of WRN and DHX9 at stalled replication forks. The arrowheads indicate 3′-ends, the triangle indicates a damaged site, the double line shows primer RNA, and the gray lines display proposed intermediate structures. See text for more details.