DisA Limits RecG Activities at Stalled or Reversed Replication Forks

Abstract

The DNA damage checkpoint protein DisA and the branch migration translocase RecG are implicated in the preservation of genome integrity in reviving haploid Bacillus subtilis spores. DisA synthesizes the essential cyclic 3', 5'-diadenosine monophosphate (c‑di-AMP) second messenger and such synthesis is suppressed upon replication perturbation. In vitro, c-di-AMP synthesis is suppressed when DisA binds DNA structures that mimic stalled or reversed forks (gapped forks or Holliday junctions [HJ]). RecG, which does not form a stable complex with DisA, unwinds branched intermediates, and in the presence of a limiting ATP concentration and HJ DNA, it blocks DisA-mediated c-di-AMP synthesis. DisA pre-bound to a stalled or reversed fork limits RecG-mediated ATP hydrolysis and DNA unwinding, but not if RecG is pre-bound to stalled or reversed forks. We propose that RecG-mediated fork remodeling is a genuine in vivo activity, and that DisA, as a molecular switch, limits RecG-mediated fork reversal and fork restoration. DisA and RecG might provide more time to process perturbed forks, avoiding genome breakage.

Keywords: DNA repair; DisA; RecG; c-di-AMP; stalled fork; template switching.

Figure 1

DisA may coexist with RecG on HJ DNA. (A) For EMSA assays [γ³²P] HJ-J3 DNA (0.1 nM in molecules) was incubated with increasing DisA (6–25 nM, lanes 2–4), RecG (7.5–30 nM, lanes 5–7) or increasing DisA and fixed RecG (lanes 14 to 16) concentrations, or pre-incubated with a fixed RecG (lanes 8–10) or increasing DisA concentrations (lanes 11–13) and then the second protein added for 15 min in buffer B containing 5 mM ATPγS at 37 °C. The protein-DNA complexes were separated using 6% PAGE in TAE buffer and visualized by autoradiography. The RecG-HJ (G-I and G-II) and DisA-HJ (A, retained in the well) complexes are indicated. (B) DNAse I footprinting assays. The [γ³²P] HJ-J3 DNA was pre-incubated with increasing RecG (15–60 nM) or DisA concentrations (30–120 nM) or increasing DisA concentrations and a fixed RecG concentration (60 nM) in buffer C containing 5 mM ATPγS for 15 min at 37 °C, and then DNAse I was added. Samples were separated in 15% dPAGE and autoradiographed. The region protected by DisA in the presence of RecG is marked with rectangles. C denotes a HJ DNA control without DNase I treatment. In (A,B) the assay was repeated >3 times with similar results, a representative gel is shown. (C) RecG bound to HJ DNA inhibits DisA-mediated c-di-AMP synthesis. DisA (25 nM) was incubated in buffer E containing 100 µM ATP in the presence of HJ-J3 DNA (125 nM) and increasing RecG concentrations (7 to 30 nM) for 30 min at 37 °C. Samples were separated by TLC, dried and visualized by phosphor imaging. Spots corresponding to the ATP and c-di-AMP molecules were quantified. The position of ATP, the linear pppA-pA intermediate, c-di-dAMP and the origin are indicated. A representative gel and the mean ± SD of >3 independent experiments are shown. The radiolabeled strand is indicated with an asterisk.

Figure 2

RecG unwinds HJs and nascent DNA strands from partial and complete replication fork substrates. The different radiolabeled DNA substrates (HJ-J4, Y fork, 5´-fork, 3´-fork or flayed DNA) were incubated with increasing RecG (15 and 30 nM) concentrations in buffer C containing 5 mM ATP (15 min at 37 °C). After deproteinization, substrates and products were separated by 6% PAGE in TAE buffer and analyzed by auto-radiography. The radiolabeled strands are indicated with an asterisk. In lanes 16 and 17 the [γ³²P]-labelled oligonucleotides are loaded as controls. A representative gel and below the mean ± SD of 3 independent experiments is shown.

Figure 3

DisA inhibits RecG-mediated fork reversal and HJ regression. (A,C,E) [γ³²P] forked-Lead (A), [γ³²P] forked-Lag DNA (C), or [γ³²P] HJ-J4 DNA (0.1 nM) (E) was pre-incubated with increasing DisA concentrations (doubling from 12–48 nM) or a fixed amount of RecG (15 nM) in buffer B containing 10 mM MgCl₂ (5 min at 37 °C). Then, the second protein (variable DisA [RecG → DisA], a constant amount of RecG [DisA → RecG] or no protein) and 5 mM ATP were added and the reaction was further incubated for 15 min at 37 °C. The reaction was stopped, deproteinized and separated by 6% native PAGE. Gels were dried and visualized by phosphor imaging. -, no protein added; B, boiled product. (B,D,F) The relative amount of flayed DNA product in three independent experiments such as those shown in A, C and E was quantified, and the mean ± SD are represented. The radiolabeled strand is indicated with an asterisk.

Figure 4

DisA inhibits the ATPase activity of RecG. (A) RecG (5 nM) was incubated with or without 3199-nt cssDNA or HJ-3 DNA (10 μM in nt) in buffer D containing 5 mM ATP and the ATP regeneration system, and the ATPase activity was measured for 30 min. Insert, RecG (15 nM) was incubated with cssDNA in buffer D containing 5 mM ATP and 2x the ATP regeneration system, and the ATPase activity was measured for 30 min. (B) HJ-J3 DNA was incubated with RecG (15 nM), DisA (24–36 nM) or both in buffer D containing 5 mM ATP and the ATP regeneration system and the ATPase activity measured for 30 min. (C) cssDNA was incubated with RecG (15 nM), DisA (6–36 nM) or both proteins and the ATPase activity was measured. (D) cssDNA was pre-incubated with RecG (15 nM) or DisA (24 nM) (5 min at 37 °C), then DisA (RecG + ssDNA → DisA) or RecG (DisA + ssDNA → RecG) was added and the ATPase activity was measured. (E to H) cssDNA (E,G) or HJ-3 DNA (F,H) was incubated with RecG (15 nM) and DisA ∆C290 (6–36 nM) (E,F) or with RecG (15 nM) and DisA D77N (6–36 nM) (G,H) and the ATPase activity was measured. All reactions were repeated three or more times with similar results, and a representative graph is shown here.

Figure 5

Model for the control of fork remodeling by DisA. DisA moves onto dsDNA synthesizing c-di-AMP until an unrepaired DNA lesion (represented here by the green square on the leading strand template) causes blockage of replication fork movement. The fork remodeler RecG (or RecA) helps DisA to bind the stalled or reversed fork and then DisA-mediated c-di-AMP synthesis is suppressed. Low levels of c-di-AMP indirectly increase in turn (p)ppGpp synthesis which that inhibits DNA replication. In the upper panel, DisA suppresses RecG-mediated fork reversal to avoid fork breakage (A). If fork reversal occurs, RecU in concert with RuvAB may catalyze HJ cleavage (B), leading to the formation of a one-ended DSB that is lethal during spore revival. In the lower panel, (C) RecA-mediated template switching followed by DNA synthesis is also modulated by DisA to allow time for specific enzymes to repair the damage, and then replication re-starts.