A gatekeeping function of the replicative polymerase controls pathway choice in the resolution of lesion-stalled replisomes

Abstract

DNA lesions stall the replisome and proper resolution of these obstructions is critical for genome stability. Replisomes can directly replicate past a lesion by error-prone translesion synthesis. Alternatively, replisomes can reprime DNA synthesis downstream of the lesion, creating a single-stranded DNA gap that is repaired primarily in an error-free, homology-directed manner. Here we demonstrate how structural changes within the Escherichia coli replisome determine the resolution pathway of lesion-stalled replisomes. This pathway selection is controlled by a dynamic interaction between the proofreading subunit of the replicative polymerase and the processivity clamp, which sets a kinetic barrier to restrict access of translesion synthesis (TLS) polymerases to the primer/template junction. Failure of TLS polymerases to overcome this barrier leads to repriming, which competes kinetically with TLS. Our results demonstrate that independent of its exonuclease activity, the proofreading subunit of the replisome acts as a gatekeeper and influences replication fidelity during the resolution of lesion-stalled replisomes.

Keywords: DNA replication; damage avoidance; replication stalling; repriming; translesion synthesis.

Fig. 1.

In vitro reconstitution of Pol IV-mediated TLS. (A) A rolling-circle replication-based TLS assay. Replication reactions were performed through the indicated steps. Newly synthesized DNAs were labeled with incorporated [α-³²P]-dATP during replication, separated on a denaturing gel and visualized by autoradiography. (B) A single N²-FFdG in the leading-strand template potently inhibits the formation of RL leading-strand products by the E. coli replisome. Control, a lesion-free template; N²-FFdG, a N²-FFdG–containing template. (C) Pol IV promotes replication of the N²-FFdG–containing template. (Left) Replication of the N²-FFdG–containing template in the presence of indicated amounts of Pol IV. Reactions were quenched 12 min after initiation. (Right) Magnified view of the same gel; −, no Pol IV; +, 156 nM Pol IV; numbers (n), number of passages through a N²-FFdG lesion.

Fig. 2.

ε–Cleft mediates polymerase exchange. (A) Interactions between Pol III core (αεθ) or Pol IV and the β₂ clamp via the CBM–cleft interaction. (Upper) α–Cleft, the cleft occupied by the α subunit; ε–cleft, the cleft occupied by the ε subunit. Pol IV interacts with a cleft via the C-terminal CBM. For simplicity, the θ subunit of Pol III core is not depicted. (Lower) Mutations on CBMs to strengthen the interaction with a cleft. (B) Strengthening the ε–cleft interaction suppresses in vitro TLS. (Upper) RL leading-strand replication products. (Lower) Relative band intensities were calculated with respect to the band intensity in the absence of Pol IV for the lesion-free template and the maximal band intensity for the lesion-containing templates in the presence of Pol IV, respectively; αεθ, wild-type ε-containing replisome; αε_Lθ, ε_L-containing replisome. (Inset, Left) Relative replication (RR) corresponds to replication of the control template normalized by replication of the αεθ-containing replisome at [Pol IV] = 156 nM. (Inset, Right) Relative TLS (RT) was calculated by first normalizing replication of the lesion-containing template with replication of the control and then calculating the ratio with respect to the αεθ replisome. Reported values correspond to [Pol IV] = 156 nM (mean ± SD, n = 3). (C) Enhancing the α–cleft contact has little effect on in vitro replication or TLS. (Upper) RL leading-strand replication products. (Lower) Relative band intensities of leading-strand replication products; αεθ, wild-type ε-containing replisome; α^M3εθ, α^M3-containing replisome. (Inset, Left) RR, replication of the control template at [Pol IV] = 156 nM. (Inset, Right) replication-normalized RT at [Pol IV] = 156 nM (mean ± SD, n > 2). (D) Strengthening the ε–cleft interaction sensitizes cells to NFZ and MMS. Cultures of indicated strains were serially diluted and spotted on LB-agar plates containing indicated concentrations of either NFZ or MMS. dnaQ and dnaE, the genes encoding the ε and the α subunits, respectively.

Fig. 3.

ε Subunit suppresses TLS through a gatekeeping role. (A) Strengthening the ε–cleft interaction with the dnaQ(ε_L) mutation suppresses TLS at the fork for both wild-type and catalytically defective ε subunits. (Upper) RL leading-strand replication products. (Lower) Relative band intensities of these leading-strand replication products; ε^CD, catalytically defective ε (ε^D12A,E14A); ε_L^CD, ε^CD with the dnaQ(ε_L) mutation. (Inset, Right) Replication-normalized RT (relative TLS) at [Pol IV] = 78 nM (mean ± SD, n = 2). (B) Weakening the ε–cleft interaction promotes Pol IV-mediated TLS at the fork. RL leading-strand replication products resulting from replication of a lesion-free control template (Upper, Left) or a N²-FFdG–containing template (Upper, Right) by the indicated Pol III complexes. (Lower) Relative band intensities of these leading-strand replication products. (Inset, Left) Inhibition of replication by Pol IV; replication is normalized to replication in the absence of Pol IV (NR, normalized replication). (Inset, Right) Replication-normalized RT at [Pol IV] = 156 nM (mean ± SD, n > 2); αεθ, wild-type replisome; αε_Qθ, ε_Q-containing replisome; α, ε-free replisome.

Fig. 4.

Pol IV diverts lesion-stalled replisomes from repriming to TLS. (A) Southern blot probes used in this study. A lesion-stalled replisome on the leading-strand template reprimes downstream of the lesion, leaving an ssDNA gap between the lesion and a newly synthesized downstream primer. (B) Strengthening of the ε–cleft interaction suppresses TLS at the fork resulting in persistent repriming. (Left) Lesion-stalled replisomes reprime downstream of N²-FFdG. Repriming and leading-strand replication products resulting from replication of the N²-FFdG–containing template by wild-type (αεθ) or ε_L-containing (αε_Lθ) replisomes. Replication products were separated in a denaturing agarose gel and detected by Southern blot with a mixture of leading-strand probes (900 and 1,901 nts) shown in Fig. 4A. (Right) Integrated intensities of repriming replication products (mean ± SD, n = 3). (C) Pol IV-mediated TLS at the fork kinetically competes with repriming. (Right, Upper) Replication of the N²-FFdG–containing template was initiated in the absence of Pol IV, and Pol IV (100 nM final) was added to the reaction 30 or 270 s after initiation. (Left) Reactions were quenched at the indicated time after initiation, and replication products were separated in a denaturing agarose gel and detected by Southern blot with a leading-strand probe (900 nts) shown in Fig. 4A. (Right, Lower) Quantitation of repriming products. Amounts of repriming products were plotted relative to the amount of repriming products detected at the final time point (720 s) in the absence of Pol IV. The lines are linear connections of immediate time points (mean ± SD, n = 3).

Fig. 5.

Weakening the ε–cleft interaction increases utilization of TLS in cells. (A) Extent of Pol IV-mediated TLS across G-BaP(-) (N²-(-)-transanti-benzo(a)pyrene dG) in indicated strains. (B) Extent of Pol V-mediated TLS across TT-CPD and TT(6–4) in indicated strains. (C) Lesion tolerance at a G-AAF adduct within the NarI site. Scheme of Pol II- or Pol V-mediated TLS across G-AAF at NarI recognition site. The 3 possible outcomes of lesion-tolerance pathways are illustrated: DA using homologous recombination results in copying the local sister chromatid sequence in green, TLS0 mediated by Pol V and TLS-2 (with −2 frameshift slippage intermediate) mediated by Pol II. (D) Extent of Pol II- and Pol V-mediated TLS across G-AAF (N²-acetylaminofluorene dG) in indicated strains. To monitor Pol II-mediated TLS with −2 frameshift, the lesion-containing sequence was inserted into the genome with +2 frameshift so that the lacZ⁺ sequence is synthesized only when −2 frameshift took place (see details in SI Appendix, Supplemental Methods). Results shown here are the average and the SD of at least 4 independent experiments.

Fig. 6.

Conformational basis of pathway choice of lesion-stalled replisomes. (A) Conformational transitions of Pol III core complex during TLS. (a) Pol III core (αεθ) occupies both α– and ε–clefts during processive replication. (b and c) Upon encountering a lesion on the leading strand, Pol III core swings away from the P/T junction while remaining bound to the β₂ clamp through the α-cleft interaction. This conformational transition is either facilitated or suppressed by the dnaQ(ε_Q) or the dnaQ(ε_L) mutations, respectively. (d and e) Pol IV binds to the ε–cleft and take over the P/T junction and performs TLS past the lesion. (f) After synthesizing a short patch, Pol IV swings away from the P/T junction and Pol III core reestablishes the ε–cleft interaction and resumes processive replication. Within closed states, states a and b, the ε-cleft is occupied by the ε subunit. Within open states, states c though f, the ε–cleft is not occupied by the ε subunit. The θ subunit of Pol III core is not depicted for simplicity. (B) Pathway choice for lesion-stalled replisomes during the SOS response: Repriming followed by gap-filling TLS vs. TLS at the fork.