Rev1 is essential for DNA damage tolerance and non-templated immunoglobulin gene mutation in a vertebrate cell line

Abstract

The majority of DNA damage-induced mutagenesis in the yeast Saccharomyces cerevisiae arises as a result of translesion replication. This process is critically dependent on the deoxycytidyl transferase Rev1p, which forms a complex with the subunits of DNA polymerase zeta, Rev3p and Rev7p. To examine the role of Rev1 in vertebrate mutagenesis and the DNA damage response, we disrupted the gene in DT40 cells. Rev1-deficient DT40 grow slowly and are sensitive to a wide range of DNA-damaging agents. Homologous recombination repair is likely to be intact as basal and damage induced sister chromatid exchange and immunoglobulin gene conversion are unaffected. How ever, the mutant cells show a markedly reduced level of non-templated immunoglobulin gene mutation, indicating a defect in translesion bypass. Furthermore, ultraviolet exposure results in marked chromosome breakage, suggesting that replication gaps created in the absence of Rev1 cannot be efficiently repaired by recombination. Thus, Rev1-dependent translesion bypass and mutagenesis is likely to be a trade-off for the ability to complete replication of a damaged template and thereby maintain genome integrity.

Fig. 1. Generation of ΔRev1-DT40. (A) Schematic of the human Rev1 cDNA. The grey boxes represent sequence features: BRCT, homology to BRCA-1 C-terminus. Boxes I–VIII also show significant homology between species and contain motifs common to translesion polymerases. ‘Rev7 interaction’ denotes the region mapped in human Rev1 to interact with Rev7 (Murakumo et al., 2001). The white boxes show the positions of exons 5 and 6 with the amino acid identity between chicken and human indicated (see also Supplementary figure 1). The black bar shows the region deleted in the ΔRev1-DT40 in this study. (B) Restriction map of the Rev1 genomic locus in DT40 containing exons homologous to human exons 5 and 6. The targeting construct is derived from the 4.5 kb EcoRI fragment defined by the bold RI sites. P, PstI; R, EcoRI. (C) Southern blot of genomic DNA digested with PstI and probed with the PstI–EcoRI fragment indicated in (B). (D) Northern blot for Rev1. The left-hand two lanes are probed with a chicken Rev1 cDNA fragment spanning polymerase motifs VII and VIII. The right hand two lanes shows the same blot hybridized with chicken MMS2 cDNA.

Fig. 2. Growth and cell cycle characteristics of ΔRev1-DT40. (A) Growth characteristics of DT40 (open circles) and its ΔRev1 derivative (filled symbols). (B) Cloning efficiency. Data were derived from eight plates for DT40 and 13 for ΔRev1 from two separate sorting sessions. (C) Annexin V staining. The upper right quadrant represents dead apoptotic cells [Annexin V positive, propidium iodide (PI) positive]. The lower right quadrant incipiently apoptotic cells (Annexin V positive, PI negative). (D) 2D cell cycle analysis in respsonse to cisplatin. Representative plots of fixed cells pulsed with BrDU counterstained with PI. The cell population is gated to exclude clumps of two or more cells and the proportion of cells in sub-G₁, G₁, S and G₂/M is indicated by the appropriate gate. The three panels from left to right are untreated and 6 and 24 h after a 1 h pulse of 10 µM cisplatin.

Fig. 3. Sensitivity of ΔRev1-DT40 to DNA-damaging agents. (A) UV light (254 nm); (B) hydrogen peroxide; (C) NQO; (D) X-rays; and (E) cisplatin. Curves are derived from at least three separate experiments. Wild-type DT40 open circles, ΔRev1-DT40 filled diamonds. For UV curves data are shown for two additional, independently targeted ΔRev1-DT40 (filled squares and triangles). Error bars are 1 SD.

Fig. 4. (A) Induction of SCE following DNA damage. The histograms represent the percentage of metaphases examined (y-axis) containing a given number of SCE (x-axis) with wild type in grey and ΔRev1-DT40 in black. From top to bottom the three panels show SCE in metaphases from untreated cells, NQO-treated and hydrogen peroxide-treated cells, respectively. The mean number of SCE per metaphase is indicated in the top right of each panel. Each database contains at least 100 metaphases. (B) Chromosomal aberrations following UV irradiation. For each time point aberrations are shown as stacked histogram bars with chromatid breaks in black, chromosome (isochromatid) breaks in grey and exchanges (radial structures) in white. More than 50 metaphases were scored blind for each time point.

Fig. 5. Diversification of the immunoglobulin light chain variable region in ΔRev1-DT40. (A) Fluctuation analysis of IgM-loss variants in wild-type DT40 and ΔRev1. The median Ig loss population for DT40 in these experiments was 0.27% and for ΔRev1-DT40 0.31%. (B) Graphic representation of consecutive sequences from sorted Ig negative cells from DT40 and ΔRev1-DT40. Each line represents an individual sequence. Black ‘lollipops’ are non-templated mutations and bars gene conversion tracts. The pseudogene donor for each tract is indicated above it (M, multiple possible donors; dup, duplication). Deletions are indicated by gaps in the sequence line. (C) Analysis of the whole database of DT40 and Rev1-DT40 immunoglobulin light chain sequence variation in selected surface Ig loss-variants. The pie charts represent the proportion of sequences with the indicated number of changes. The total number of sequences analysed is indicated in the centre of the pies. PCR error was calculated for 30 cycles of PCR assuming an error rate for PfuTurbo of 1.3 × 10^–6/bp/cycle. (D) Ratio of non-templated to templated changes in DT40 IgM-loss variants. The bars show the number of changes per mutated sequence: black bars non-templated (PM); white bars templated (T). The figures above the bars represent the ratio of non-templated point mutations to templated changes (gene conversions).

Fig. 6. Simplified model illustrating lesion channelling between homologous recombination and translesion synthesis. TLS, translesion synthesis; HR, homologous recombination. The effect of deleting (Δ) each gene is given under the gene name. Damage stalling a replication fork can be dealt with by either homologous recombination or post- replication repair. Ablation of genes involved in the ‘early’ stages of either of these pathways (e.g. XRCC2/3 or RAD18) results in lesions being channelled into the other pathway with a consequent increase in the detectable endpoints of those pathways (SCE for homologous recombination, Ig point mutation for translesion synthesis). However, loss of genes in the ‘later’ stages of either pathway (e.g. RAD54 and Rev1) is proposed to result in an intermediate that cannot be subsequently processed by the other pathway.