Figure 1.
TDG-Deficient Mouse Cells Are Resistant to 5-FU
(A) Western blot analysis of whole-cell protein extracts derived from SV40 immortalized MEFs (left), spontaneously immortalized MEFs (middle), and ES cell lines (right) used. Tdg genotypes are as indicated. A highly specific polyclonal anti-mouse TDG antibody (TDGab) was used to detect TDG; beta-actin staining served as loading control (β-actab). TDG is undetectable in extracts from Tdg−/− cells and reduced in heterozygous MEFs. Stable transfection of a Tdg expression construct (pTdg) restores TDG levels in knockout MEFs.
(B) TDG-deficient MEF and ES cells exhibit increased 5-FU resistance, and ectopic expression of wild-type Tdg in knockout MEFs restores 5-FU sensitivity. The sensitivity to increasing amounts of 5-FU was measured for the different cell lines after a continuous treatment of 48 h. Shown are survival curves as percentages of mock-treated cells.
(C) TDG-deficient MEFs are not generally resistant to induced DNA base damage. Sensitivity to MMS was measured after a treatment for 1 h with increasing concentrations of MMS. Shown are survival curves as percentages of mock-treated cells.
Shown are means ± standard error of the mean (SEM) from at least three independent experiments. pC, vector control; pTdg, Tdg-expressing vector.
Figure 2.
TDG-Induced Cell Death upon 5-FU Treatment
(A) Exposure to 5-FU results in death of TDG-proficient cells (Tdg+/+), but slow growth of TDG-deficient (Tdg−/−) cells. The growth of TDG wild-type and knockout MEF cultures was monitored and recorded in real time during treatment with indicated concentrations of 5-FU. Shown are mean percentages of confluence at the respective time points with error bars representing standard deviations (SD) of three independent cultures.
(B) HeLa cells overexpressing TDG show increased sensitivity towards 5-FU, whereas siRNA-mediated TDG knockdown results in increased 5-FU resistance. Shown is the clonogenic survival of HeLa cells with different TDG expression levels following treatment with 5-FU for 72 h (left). Data points represent percentages of colony-forming units (5-FU/mock; mean ± SEM) from at least three independent experiments. Corresponding TDG protein levels 72 h after plating of the cells are shown in the right panel. TDG was detected with a monoclonal anti-human TDG antibody (TDGmab) in extracts of cells treated with siRNA directed against TDG, or a corresponding siRNA control.
pC, vector control; pTdg, Tdg expressing vector; T↓, Tdg siRNA; C↓, control siRNA.
Figure 3.
Involvement of TDG in Processing of Uracil and 5-FU
(A) Base release activities of purified recombinant mouse TDG (mTDG) and nuclear protein extracts of TDG wild-type (Tdg+/+), heterozygous (Tdg+/−), and knockout (Tdg−/−) MEFs on uracil, 5-FU, and G•T containing synthetic 60-mer DNA duplexes. Shown are representative results of base release assays with the intact substrate DNA strands (S) and the cleaved products (P) resolved on denaturing polyacrylamide gels. All reactions were performed in the presence of the UNG inhibitory UGI peptide. Purified TDG processes thymine, uracil, and 5-FU when opposite guanine as well as 5-FU paired with adenine, but only inefficiently uracil opposite adenine.
(B) Quantitation of base release activities in nuclear extracts. G•T processing activity is reduced in protein extracts of heterozygous cells and absent from knockout extracts. Tdg knockout extracts also show a significant reduction of A•5-FU processing. All other uracil- and 5-FU–containing substrates were processed with similar efficiencies by all three nuclear extracts. Data are presented as means ± SD from three independent experiments.
An asterisk (*) indicates the 5'-fluorescein-labeled strand.
doi:10.1371/journal.pbio.1000091.g003
Figure 4.
5-FU–Induced DNA Strand Breaks Are Reduced in TDG-Deficient Cells whereas Overall Repair Activity Is Increased.
(A) Complementation of Tdg knockout MEFs with wild-type and catalytically deficient TDG. Stable transfectants of Tdg−/− MEFs ectopically expressing either TDG variant from the native promoter show TDG levels about the same as endogenous, as detected by western blotting.
(B) Reduced levels of 5-FU–induced DNA strand breaks in cells lacking active TDG. Steady-state levels of DNA single- and double-strand breaks in the cell lines indicated were assessed by the alkaline Comet assay using automated comet tail moment analysis. 5-FU treatment results in a significant tail moment increase in wild-type, but not in Tdg knockout MEFs. The generation of 5-FU–specific DNA strand breaks in Tdg knockout cells is restored by complementation with wild-type Tdg, but not with the catalytically inactive mutant. Shown are box plots with individual tail moments per cell, medians, interquartile ranges (boxes), 2.5%–97.5% percentiles (whiskers) and outliers (dots) of pooled data (600 to 900 cells) obtained from three independent experiments.
(C) 5-FU treatment triggers DNA SSB repair in TDG wild-type and knockout cells. The top panel shows nuclei of Tdg-proficient and -deficient cells stained with a polyclonal anti-XRCC1 antibody (XRCC1ab) after 5-FU treatment. The statistical analysis of XRCC1 foci per cell across the populations analyzed (n ≥ 100 cells per population) is shown as a scatter plot with medians and the interquartile ranges.
pC, empty vector; pTdg, vector expressing TDG; pTdgcat, vector expressing a catalytic dead variant.
Figure 5.
5-FU Treatment Induces a TDG-Dependent S-Phase Delay
The histograms show the effect of the 5-FU treatment on the relative cell-cycle distribution (% cells) of TDG-proficient and -deficient MEFs (A), and of TDG knockout cell lines stably transfected with a plasmid expressing Tdg from its authentic promoter (B). 5-FU treatment of TDG-proficient cells results in a significant accumulation cells in S-phase at the expense of G1 cells, whereas TDG-deficient cells show only insignificant changes in cell-cycle distribution. Expression of wild-type Tdg in knockout MEFs partially restored the 5-FU–dependent S-phase delay. The data shown represent averages of three independent experiments with fold changes upon 5-FU treatment.
pC, empty vector; pTdg, vector expressing TDG.
Figure 6.
TDG-Dependent Activation of DNA Damage Responses upon 5-FU Treatment
(A–C) TDG mediates late Chk1 activation following 5-FU treatment. Activation of Chk1 in TDG-proficient and -deficient MEFs (A) as well as in complemented knockout cells (B) was determined by western blotting with a S345 phospho-specific antibody against Chk1 (Chk1-Pab). After treatment with 10 μM 5-FU, wild-type but not TDG-deficient MEFs show a strong accumulation of S345 phosphorylated Chk1. Total Chk1 protein is the same in both MEF lines before and after 5-FU treatment (Chk1ab). TDG levels in wild-type cells, detected with a specific anti-mTDG antibody (TDGab), are reduced in 5-FU–exposed cells, reflecting an accumulation of cells in S-phase, where TDG is absent. Tdg knockout MEFs stably expressing an ectopic copy of Tdg (B) contain low levels of TDG, which is sufficient to induce Chk1 activation upon 5-FU treatment.
(C) Dynamics of Chk1 activation in TDG-proficient and -deficient MEFs during and after exposure to 5-FU or HU. The 5-FU–containing (10 μM) or HU-containing (2.5 mM) medium was replaced with drug-free medium after 24 or 16 h, respectively. Samples were taken at the time points indicated and analyzed for Chk1 S345 phosphorylation by western blotting. After 16 h into treatment, activated Chk1 appears equally in extracts from 5-FU– and HU-treated cells; at 24 h, the Chk1-p signal is undetectable in the HU-treated samples and significantly reduced in 5-FU–treated cells; at 40 h, significant levels of phosphorylated Chk1 reappear in 5-FU–exposed TDG-proficient MEFs but not in TDG-deficient MEFs.
(D) The induction of γH2AX foci by 5-FU treatment is significantly reduced in TDG-deficient MEFs. The top panels show examples of MEFs immunostained with a monoclonal antibody against γH2AX (γH2AX mab) after treatment with 5 μM 5-FU. The statistical analysis of γH2AX foci per cell across the populations analyzed (n > 95 cells per population) is depicted in the lower panel as scatter plot with medians and the interquartile ranges.
pC, empty vector; pTdg, vector expressing TDG; TDG-S, TDG modified with SUMO. An asterisk (*) indicates an unspecific cross-reaction of the secondary antibody
Figure 7.
TDG-Dependent 5-FU Cytotoxicity
Illustrated are the cell-cycle distributions of the three relevant UDGs, TDG, UNG2, and Smug1 (top), together with expected levels of genomic 5-FU, uracil, AP-sites, and the observed Chk1 activation following 5-FU treatment (bottom). TDG is present during the G2/M and G1 phases but is degraded prior to and absent from S-phase. UNG2 shows a strictly inverse regulation whereas Smug1 is expressed throughout the entire cell cycle. Treatment with 5-FU for 24 h gives rise to misincorporation of appreciable levels of 5-FU and uracil during S-phase (S1), resulting in Chk1 activation by ongoing replication-associated UNG2 and Smug1-dependent BER. Although Smug1 and UNG2 will initiate faithful repair of uracil and 5-FU bases directly after DNA synthesis, these pathways will become saturated under 5-FU exposure and, in addition, are relatively inefficient in processing the 5-FU•A base pairs. Hence, some of them will persist in the DNA into the subsequent G2 and G1 phases of the cell cycle. There, TDG will initiate repair, but turnover with a low rate, leading to an accumulation of AP-sites and/or DNA SSBs. During the subsequent S-phase, these repair intermediates will interfere with DNA replication, causing replication fork stalling, fork collapse, DNA double-strand breaks, and a second round of Chk1 activation. Due to genome fragmentation, cells will then induce apoptosis.