WRN exonuclease imparts high fidelity on translesion synthesis by Y family DNA polymerases

Abstract

Purified translesion synthesis (TLS) DNA polymerases (Pols) replicate through DNA lesions with a low fidelity; however, TLS operates in a predominantly error-free manner in normal human cells. To explain this incongruity, here we determine whether Y family Pols, which play an eminent role in replication through a diversity of DNA lesions, are incorporated into a multiprotein ensemble and whether the intrinsically high error rate of the TLS Pol is ameliorated by the components in the ensemble. To this end, we provide evidence for an indispensable role of Werner syndrome protein (WRN) and WRN-interacting protein 1 (WRNIP1) in Rev1-dependent TLS by Y family Polη, Polι, or Polκ and show that WRN, WRNIP1, and Rev1 assemble together with Y family Pols in response to DNA damage. Importantly, we identify a crucial role of WRN's 3' → 5' exonuclease activity in imparting high fidelity on TLS by Y family Pols in human cells, as the Y family Pols that accomplish TLS in an error-free manner manifest high mutagenicity in the absence of WRN's exonuclease function. Thus, by enforcing high fidelity on TLS Pols, TLS mechanisms have been adapted to safeguard against genome instability and tumorigenesis.

Keywords: DNA lesions; DNA repair; UV damage; Werner syndrome protein WRN; Y family DNA polymerases; fidelity of translesion synthesis.

Figure 1.

Requirement of WRN for replication through UV lesions in conjunction with WRNIP1, Rev1, and Y family Pols. (A) RF progression in unirradiated WT HFs depleted for WRN, WRNIP1, or Rev1. (B) RF progression in UV-irradiated WT HFs depleted for WRN, WRNIP1, or Rev1. (C) RF progression in unirradiated WRN^−/− HFs depleted for WRNIP1 or TLS Pols. (D) RF progression in UV-irradiated WRN^−/− HFs depleted for WRNIP1 or TLS Pols. In A–D, quantitative determination of RF progression (mean CldU:IdU ratio) was based on ∼400 DNA fibers analyzed from four independent experiments. Error bars indicate the SD. P-values were derived using Student's two-tailed t-test. (ns) Not significant, (*) P < 0.05, (***) P < 0.001.

Figure 2.

UV-induced assembly of WRN, WRNIP1, and Rev1 with Y family Pols. (A) Requirement of WRNIP1 and Rev1 for the assembly of WRN into UV-induced foci. (B) Requirement of WRN and Rev1 for the assembly of WRNIP1 into UV-induced foci. (C,D) Requirement of WRN and WRNIP1 for the assembly of Rev1 (C) and Polη (D) into UV-induced foci. In A–D, error bars indicate the SD. P-values were derived using Student's two tailed t-test. (****) P < 0.0001. For each analysis, ∼200–300 cells were analyzed. Scale bars: A, left, 10 µm; A, right, 2 µm. (E) Coimmunoprecipitation (co-IP) of FLAG-Rev1 with WRN, WRNIP1, and Polη in chromatin fractions from UV-irradiated cells. GM00637 HFs expressing FLAG-Rev1 were UV-irradiated and incubated for 4 h. Chromatin extracts from unirradiated and UV-irradiated cells were immunoprecipitated with FLAG M2 agarose (Sigma). Co-IPs of FLAG-Rev1 with WRN, WRNIP1, and Polη were determined by Western blot analysis.

Figure 3.

Defects in WRN's 3′ → 5′ exonuclease activity confer high error-proneness on TLS by Y family Pols opposite UV lesions. (A) UV-induced (5 J/m²) mutation frequencies in the cII gene in BBMEFs expressing WT WRN or E84A WRN. UV mutations resulting from TLS opposite CPDs (top) were examined in a BBMEF cell line expressing a (6-4) PP photolyase gene, and UV mutations resulting from TLS opposite (6-4) PPs (bottom) were examined in a BBMEF cell line expressing a CPD photolyase gene. Mutation frequencies and SEM were calculated from four independent experiments. (a) UV-induced mutation frequencies resulting from TLS through CPDs or (6-4) photoproducts were calculated by subtracting the corresponding spontaneous mutation frequency from the mutation frequency in UV-irradiated cells. (B) Diagrammatic representation of error-proneness imposed by WRN's exonuclease deficiency upon error-free TLS by Polη opposite CPDs. The figure depicts the error-proneness of Polθ in WRN-depleted BBMEFs expressing WT WRN (mutation frequency 27 × 10⁻⁵) and the elevation in error-proneness (mutation frequency 54.2 × 10⁻⁵) that occurs in BBMEFs expressing E84A WRN due to the added error-proneness of Polη. In BBMEFs codepleted for WRN and Polθ and expressing WT WRN, TLS by Polη operates in an error-free manner, but in BBMEFs expressing E84A WRN, TLS by Polη exhibits high error-proneness (mutation frequency 26 × 10⁻⁵). (C) Diagrammatic representation of elevated error-proneness conferred by WRN's exonuclease deficiency on TLS by Polη and Polι opposite (6-4) PPs. The figure depicts the increase in error-proneness of TLS by Polη and Polι in WRN-depleted BBMEFs expressing E84A WRN (mutation frequency 29.1 × 10⁻⁵) from that in BBMEFs expressing WT WRN (mutation frequency 11.9 × 10⁻⁵). In B and C, the corresponding spontaneous mutation frequency has been subtracted from the UV-induced mutation frequency; thus, the figures depict mutation frequencies derived from TLS through CPDs (B) or (6-4) PPs (C).

Figure 4.

UV-induced (5 J/m²) mutational spectra in the cII gene in BBMEFs resulting from TLS through CPDs or (6-4) PPs. (A) Mutational spectra resulting from TLS through CPDs in WRN-depleted BBMEFs expressing WT WRN are shown above the sequence, and those expressing E84A WRN are shown below the sequence. (B) Mutational spectra resulting from TLS through CPDs in BBMEFs codepleted for WRN and Polθ and expressing WT WRN are shown above the sequence, and those expressing E84A WRN are shown below the sequence. (C) Mutational spectra resulting from TLS through (6-4) PPs in WRN-depleted BBMEFs expressing WT WRN are shown above the sequence, and those expressing E84A WRN are shown below the sequence. The designations for the other mutational changes are deletion (X), addition (+), and tandem mutations (underlined).

Figure 5.

Defects in WRN's 3′ → 5′ exonuclease activity confer high error-proneness on error-free TLS by Polκ opposite Tg and on error-free TLS by Polι opposite εdA. (A) Effects of E84A WRN on mutation frequencies and nucleotides inserted opposite the Tg (top) or the εdA (bottom) lesions carried on the leading strand DNA template of a duplex plasmid in WRN^−/− HFs. (a) In WRN^−/− HFs treated with control (NC) or Polθ siRNA and expressing E84A WRN, mutations occur at the 5′ template residue next to the Tg lesion and involve a change of 5′T > C. Thus, the sequence 5′-CAATTgG-3′ is changed to 5′-CAACTG-3′ (the corresponding 5′ residues are underlined). (b) Numbers in parentheses indicate the total number of mutations. (c) FLAG-DA-Rev1 refers to catalytically inactive D570A E571A Rev1. (B) Diagrammatic representation of error-proneness imposed by WRN's exonuclease deficiency upon error-free TLS by Polκ opposite Tg. In WRN^−/− HFs depleted for Polθ and expressing E84A WRN, error-free TLS by Polκ exhibits a mutational pattern in which mutations generated by Polκ occur primarily by the change of 5′TTg > 5′CT and less frequently by a change at the Tg lesion site. (C) Diagrammatic representation of highly elevated error-proneness imposed by WRN's exonuclease deficiency upon error-free TLS by Polι opposite εdA. In WRN^−/− HFs expressing WT WRN, mutations (∼19%) result from error-prone TLS by Rev1 and Polθ. In WRN^−/− HFs inactivated for both the Rev1 and Polθ error-prone TLS pathways and expressing WT WRN, TLS by Polι operates in an error-free manner; in striking contrast, expression of E84A WRN in these cells confers a remarkably high error-proneness on Polι (∼33% mutations). Mutation frequency analyses in WRN^−/− HFs expressing E84A WRN and treated with Polθ siRNA indicate that Rev1's error-proneness is not significantly elevated in the absence of WRN's exonuclease activity.