PCNA ubiquitination is important, but not essential for translesion DNA synthesis in mammalian cells

Abstract

Translesion DNA synthesis (TLS) is a DNA damage tolerance mechanism in which specialized low-fidelity DNA polymerases bypass replication-blocking lesions, and it is usually associated with mutagenesis. In Saccharomyces cerevisiae a key event in TLS is the monoubiquitination of PCNA, which enables recruitment of the specialized polymerases to the damaged site through their ubiquitin-binding domain. In mammals, however, there is a debate on the requirement for ubiquitinated PCNA (PCNA-Ub) in TLS. We show that UV-induced Rpa foci, indicative of single-stranded DNA (ssDNA) regions caused by UV, accumulate faster and disappear more slowly in Pcna(K164R/K164R) cells, which are resistant to PCNA ubiquitination, compared to Pcna(+/+) cells, consistent with a TLS defect. Direct analysis of TLS in these cells, using gapped plasmids with site-specific lesions, showed that TLS is strongly reduced across UV lesions and the cisplatin-induced intrastrand GG crosslink. A similar effect was obtained in cells lacking Rad18, the E3 ubiquitin ligase which monoubiquitinates PCNA. Consistently, cells lacking Usp1, the enzyme that de-ubiquitinates PCNA exhibited increased TLS across a UV lesion and the cisplatin adduct. In contrast, cells lacking the Rad5-homologs Shprh and Hltf, which polyubiquitinate PCNA, exhibited normal TLS. Knocking down the expression of the TLS genes Rev3L, PolH, or Rev1 in Pcna(K164R/K164R) mouse embryo fibroblasts caused each an increased sensitivity to UV radiation, indicating the existence of TLS pathways that are independent of PCNA-Ub. Taken together these results indicate that PCNA-Ub is required for maximal TLS. However, TLS polymerases can be recruited to damaged DNA also in the absence of PCNA-Ub, and perform TLS, albeit at a significantly lower efficiency and altered mutagenic specificity.

Figure 1. UV sensitivity of mouse embryo fibroblasts carrying the Pcna^K164R/K164R mutation.

(A) Outline of ubiquitination and deubiquitination reactions of PCNA. Treatment with DNA-damaging agents, such as UV light, induces monoubiquitination of PCNA at Lys164 by the Rad18 E3 ligase. A common model suggests that the monoubiquitinated PCNA (PCNA-mUb) directly recruits TLS polymerases enabling translesion DNA synthesis. The deubiquitinating enzyme Usp1 removes the ubiquitin from PCNA-mUb, thereby negatively regulating the level of PCNA-Ub. The Rad5 homologs Shprh and Hltf, and an unidentified additional E3 ligase (marked x?) can extend the ubiquitin chain of PCNA-mUb, and the polyubiquitinated PCNA (PCNA-polyUb) thus formed promotes homology-dependent (template switch) error-free damage tolerance. (B) Pcna^+/+ and Pcna^K164R/K164R MEFs were irradiated at the indicated UV doses, and assayed for viability after 48 hours by measuring the level of cellular ATP. Each point represents the average of 3 independent experiments.

Figure 2. Time course of accumulation and clearance of UV-induced Rpa foci.

Pcna^+/+ and Pcna^K164R/K164R MEFs were collected at the G1/S boundary by centrifugal elutriation, and allowed to attach to microscope slides. Two hours later they were UV irradiated at a dose of 8 J/m². At the indicated time points cells were pre-extracted and then fixed, immuno-stained with anti-Rpa antibody and DAPI-stained. (A) Representative images of cells stained with DAPI or antibodies against Rpa, either with or without UV irradiation. The timeline of the experiment is shown at the top. (B) Quantification of the percentage of cells with Rpa foci. For each cell line at each time point at least 100 cells were counted and the percentage of cells with Rpa foci was determined. The results are the average of two independent experiments. Error bars represent standard deviation. Full symbols, Pcna^K164R/K164R MEFs; Empty symbols, wild-type MEFs; Squares, UV irradiated cells; Triangles, unirradiated cells.

Figure 3. TLS in Pcna^K164R/K164R mouse embryo fibroblasts.

(A) MEFs were assayed for TLS as described under Materials and Methods, using the indicated site-specific lesions. TLS extents were given as percentage relative to TLS assayed with isogenic wild type MEFs. Average results of at least three independent experiments are presented. Error bars represent standard deviations. The detailed data are presented in Table S1. (B) Percentage of accurate and mutagenic TLS across a TT CPD in Pcna^+/+ and Pcna^K164R/K164R MEFs. (C) Percentage of accurate and mutagenic TLS across TT 6-4 PP. The mutation types shown are: targeted (opposite the lesion), semi-targeted (at the nucleotides flanking the lesion), and mixed (both targeted and semi-targeted). (D) Percentage of accurate and mutagenic TLS across a cisPt-GG adduct. The percentage of events is calculated out of all TLS events. The detailed mutational spectra are presented in Table S2.

Figure 4. TLS in Rad8^−/− mouse embryo fibroblasts.

(A) MEFs were assayed for TLS as described in the legend to Figure 3. The detailed data are presented in Table S3. (B) Percentage of accurate and mutagenic TLS across a TT CPD in Rad18^+/+ and Rad18^−/− MEFs. (C) Percentage of accurate and mutagenic TLS across TT 6-4 PP. The mutation types shown are: targeted (opposite the lesion), semi-targeted (at the nucleotides flanking the lesion), and mixed (both targeted and semi-targeted). (D) Percentage of accurate and mutagenic TLS across a cisPt-GG adduct. The percentage of events is calculated out of all TLS events. The detailed mutational spectra are presented in Table S4.

Figure 5. TLS in Shprh^−/−Hltf^−/− mouse embryo fibroblasts.

(A) MEFs were assayed for TLS as described in the legend to Figure 3. The detailed data are presented in Table S5. (B) Percentage of accurate and mutagenic TLS across a TT CPD in Shprh^+/+Hltf^+/+ and Shprh^−/−Hltf^−/− MEFs. (C) Percentage of accurate and mutagenic TLS across TT 6-4 PP. The mutation types shown are: targeted (opposite the lesion), semi-targeted (at the nucleotides flanking the lesion), and mixed (both targeted and semi-targeted). (D) Frequency of accurate and mutagenic TLS across a cisPt-GG adduct. The percentage of events is calculated out of all TLS events. The detailed mutational spectra are presented in Table S6.

Figure 6. TLS in Usp1^−/− mouse embryo fibroblasts.

(A) MEFs were assayed for TLS as described in the legend to Figure 3. The detailed data are presented in Table S7. (B) Percentage of accurate and mutagenic TLS across a TT CPD in Usp1^+/+ and Usp1^−/− MEFs. (C) Percentage of accurate and mutagenic TLS across TT 6-4 PP. The mutation types shown are: targeted (opposite the lesion), semi-targeted (at the nucleotides flanking the lesion), and mixed (both targeted and semi-targeted). (D) Frequency of accurate and mutagenic TLS across a cisPt-GG adduct. The percentage of events is calculated out of all TLS events. The detailed mutational spectra are presented in Table S8.

Figure 7. Epistasis analysis of the contributions of ubiquitinated PCNA and TLS DNA polymerases to UV sensitivity.

Pcna^+/+ and Pcna^K164R/K164R MEFs were transfected with siRNA against the TLS genes Rev3L (the catalytic subunit of polζ; A) PolH (encoding polη; B) or Rev1 (C), and after 48 h they were UV irradiated at the indicated doses. Sensitivity was determined 10 days after UV irradiation by measuring colony forming ability. Each point represents the mean of three independent experiments.

Figure 8. Model describing interactions that stabilize a TLS DNA polymerase during lesion bypass.

The binding of a TLS DNA polymerase to a primer-template-lesion involves at least 7 known stabilizing interactions, 4 of which involve PCNA (top drawing). Two of these interactions, which involve the ubiquitin moiety, are lost in cells carrying the PcnaK164R mutation (lower drawing). The remaining 5 interactions (lower drawing) are sufficient to promote TLS, albeit with lower efficiency and altered mutagenic specificity. See text for details.