Translational reprogramming following UVB irradiation is mediated by DNA-PKcs and allows selective recruitment to the polysomes of mRNAs encoding DNA repair enzymes

Abstract

UVB-induced lesions in mammalian cellular DNA can, through the process of mutagenesis, lead to carcinogenesis. However, eukaryotic cells have evolved complex mechanisms of genomic surveillance and DNA damage repair to counteract the effects of UVB radiation. We show that following UVB DNA damage, there is an overall inhibition of protein synthesis and translational reprogramming. This reprogramming allows selective synthesis of DDR proteins, such as ERCC1, ERCC5, DDB1, XPA, XPD, and OGG1 and relies on upstream ORFs in the 5' untranslated region of these mRNAs. Experiments with DNA-PKcs-deficient cell lines and a specific DNA-PKcs inhibitor demonstrate that both the general repression of mRNA translation and the preferential translation of specific mRNAs depend on DNA-PKcs activity, and therefore our data establish a link between a key DNA damage signaling component and protein synthesis.

Figure 1.

Inhibition of protein synthesis following UVB-induced DNA damage. HeLa cells were mock- or UVB-irradiated (275 J/m²) and harvested at the times shown following exposure. (A) The relative amount of cyclo-pyrimidine dimers (CPDs) produced by this treatment were measured by ELISA using a monoclonal antibody specific to CPDs. Measurements are the mean of four independent experiments, and error bars represent 1 standard deviation from the mean. (B) Protein synthesis rates were determined by liquid scintillation counting of newly incorporated [³⁵S]-methionine at the time points indicated. Measurements are the mean of three independent experiments normalized to that of the unirradiated cells. Error bars represent 1 standard deviation from the mean. (C) Cell lysates derived from control and UVB-irradiated cells were separated by SDS-PAGE, immunoblotted, and probed with antibodies against eIF2α. There was a significant change in the phosphorylation status of eIF2α but no other eIFs (see Supplemental Fig. S1C). (D) Cell lysates were separated on (10%–50%) sucrose density gradients and the absorbance across the gradient read at 254 nm. Positions of the 40S, 60S ribosomal subunits and polysomal fractions are indicated.

Figure 2.

Post-transcriptional changes in gene expression and changes in the rate of protein synthesis of a number of polypeptides confirm the polysome profiling data. (A) Untreated HeLa cells or HeLa cells exposed to UVB light were harvested after 4 h and applied to 10%–50% ([w/v] sucrose density gradients). mRNA was generated from the individual fractions from the gradients, and Northern analysis was performed to determine the polysome/subpolysome distribution of the mRNAs shown. Representative blots are shown of paired experiments carried out at the same time, but analysis was performed on three independent occasions. These data are in agreement with that obtained from polysome profiling. For example, the data would predict that DDB1 becomes more polysomally associated following UVB exposure while APE1 becomes less polysomally associated following UVB exposure. (B, panel i) To measure the rates of protein synthesis of DDB1, ERCC1, ERCC5, OGG1, ATM, c-Myc, RAD52, and APE1, untreated HeLa cells and HeLa cells 2 and 4 h following UVB exposure were incubated with [³⁵S]-methionine for 10 min prior to harvesting. Total cell lysates were prepared from these samples, and the proteins shown were immunoprecipitated using specific antibodies. Immunoprecipitated proteins were resolved by SDS-PAGE and detected using PhosphorImager analysis. These experiments were performed on three independent occasions. (Panel ii) To measure total RNA levels, parallel samples were taken, and mRNA was extracted. These samples were subjected to Northern analysis, and DDB1, ERCC1, ERCC5, OGG1, ATM, c-Myc, RAD52, and APE1 mRNAs were detected using specific probes as indicated. These experiments were performed on three independent occasions. (C) The data obtained from the experiments shown in A and B (carried out in parallel) were analyzed using ImageQuant software to measure the fold-change for each of the mRNAs and proteins following UVB exposure. This allowed the calculation of the relative translational efficiency. The data show that there is an increase in translational efficiency for proteins that were found to be more polysomally associated (e.g., DDB1) but a marked decrease in translational efficiency in the proteins that were less polysomally associated (e.g., RAD52).

Figure 3.

uORFs allow continued synthesis of DNA repair enzymes. (A) The plasmid constructs that contained the uORFs derived from DDB1, ERCC1, and ERCC5 (and the mutated versions) (S3A) were transfected into HeLa cells, which were then exposed to UVB light. After the times shown post-irradiation, the cells were lysed, and luciferase activity was determined. Since translation of the transfection control was also reduced following UVB exposure, mRNA was isolated from cell lysates, and the amount of firefly luciferase mRNA present in the cells was measured by quantitative PCR. The luciferase expression was then calculated relative to the level of mRNA present. The error bars represent one standard deviation from the mean. The data show that the presence of the 5′ UTRs derived from DDB1 (panel ii), ERCC1 (panel iii), or ERCC5 (panel iv) confer resistance to translational repression, and luciferase is still produced following UVB exposure. However, mutating the AUG codons is sufficient to inhibit the production of firefly luciferase following UVB exposure. More of the 5′ UTR derived from ATF4 was insufficient to increase the expression of luciferase. (B, panel iii) Cells were transfected with the constructs shown and then treated with salburinal for 12 h. (Panel i) This compound is a selective inhibitor of complexes that dephosphorylate eIF2α (Boyce et al. 2005), and in the presence of this compound, there was an increase in the phosphorylation of the α subunit of eIF2 as expected. However, the data show that in contrast to ATF4 (panel ii) the 5′ UTRs derived from ERCC5 (panel iv), ERCC1 (panel v), or DDB1 (panel vi) were not resistant to the effects of this compound, and the decrease in translation following treatment was similar to that found with the pGL3 reporter vector. Taken together, these data suggest that a reduction in ternary complex levels alone is insufficient to elicit a response.

Figure 4.

DNA-PKcs is required for both translational inhibition and reprogramming of protein synthesis. (A) HeLa cells were pretreated with inhibitors of DNA-PKcs (Nu-7441), ATM (Ku-55933), or DMSO control and then exposed to UVB light. Protein synthesis rates determined by liquid scintillation counting of newly incorporated [³⁵S]-methionine 4 h following exposure. Measurements are the mean of three independent experiments, and the values were normalized to that of the nonirradiated cells. Error bars represent one standard deviation from the mean. (B) To assess the effect that the ATM (Ku-55933) or DNA-PKcs (Nu-7441) inhibitors had on the phosphorylation status of eIF2α and the eIF2α kinases shown (GCN2, PRK, and PERK), cells were preincubated with these compounds and exposed to UVB light, and samples were harvested after 4 h. Cell lysates generated from these samples were separated by SDS-PAGE and immunoblotted with the antibodies shown. In the presence of the ATM inhibitor (Ku-55933), there is an increase in the degree of phosphorylation of eIF2α in nonirradiated cells; however, following UVB exposure, there is further phosphorylation of this protein. In contrast, in the presence of the DNA-PKcs inhibitor (Nu-7441), there is no increase in the phosphorylation of eIF2α following UVB exposure. The kinases that are known to phosphorylate eIF2α, GCN2, PKR, and PERK were also tested, and the DNA-PKcs inhibitor blocks the phosphorylation of GCN2 but not the other eIF2α kinases. (C) To assess the effect that UVB light and the DNA-PKcs inhibitor had on GCN2, cells were transfected with a plasmid that encodes GCN2 containing a Flag tag. Following exposure with UVB in the presence/absence of the DNA-PKcs inhibitor GCN2, protein was immunoprecipitated and separated by SDS-PAGE and immunoblotted for total GCN2. The data show that in the presence of the DNA-PKcs inhibitor GCN2 phosphorylation is inhibited. (D) HeLa cells were pretreated with the ATM inhibitor (Ku-55933) (panel ii) or the DNA-PKcs (Nu-7441) (panel iii) or untreated and were then either exposed to UVB light. Cell lysates were prepared 4 h following exposure and were applied to 10%–50% (w/v) sucrose density gradients. These were fractionated and the absorbance across the gradient measured at 254 nm. (E) Cell lysates from HeLa cells ±Ku-55933 or ±Nu-7441 and treated with ±UVB light were fractionated using sucrose density gradient analysis on 10%–50% (w/v) sucrose gradients. Northern analysis was performed on the mRNA isolated from the individual fractions to determine the polysome/subpolysome distribution of DDB1, ERCC1, and ERCC5.

Figure 5.

(A, panel i) Cell lysates from M059K or M059J cells were prepared separated by SDS-PAGE and immunoblotted for DNA-PKcs. As expected, no DNA-PKcs was detected in M059J cells. (panel ii) M059K or M059J cells were exposed to UVB light, and protein synthesis rates were determined by liquid scintillation counting of newly incorporated [³⁵S]-methionine. Measurements are the mean of three independent experiments normalized to that of the nonirradiated cells. Error bars represent one standard deviation from the mean. (B) Cell lysates from M059K (panel i) orM059J (panel ii) cells were exposed to UVB light and immunoblotted for eIF2α, GCN2, and PERK. The data show that there in no change in the degree of phosphorylation of eIF2α or GCN2 in cells that lack DNA-PKcs. (C) M059K (panel i) orM059J (panel ii) cells were exposed to UVB light and, after 4 h, cell lysates were prepared and fractionated using sucrose density gradient analysis on 10%–50% (w/v) sucrose gradients. (D) Cell lysates were generated from M059K (panel i) orM059J (panel ii) cells ± exposure to UVB light and applied to 10%–50% (w/v) sucrose gradients. mRNA was isolated from individual fractions, and Northern analysis was performed to determine the polysome/subpolysome distribution of DDB1, ERCC1, and ERCC5.