Genotoxic stress causes the accumulation of DNA-dependent protein kinase catalytic subunit phosphorylated at serine 2056 at nuclear speckles and alters pre-mRNA alternative splicing

Abstract

RNA splicing has emerged as a critical player in the DNA damage response (DDR). However, the underlying mechanism(s) by which pre-mRNA splicing is coordinately regulated by genotoxic stress has remained largely unclear. Here, we show that a DDR factor, DNA-dependent protein kinase (DNA-PK), participates in the modulation of pre-mRNA splicing in the presence of DNA double-strand break (DSB)-induced genotoxic stress. Through indirect immunostaining, we made the surprising discovery that DNA-PK catalytic subunits (DNA-PKcs) autophosphorylated at serine 2056 (S2056) accumulate at nuclear speckles (dynamic nuclear structures that are enriched with splicing factors), following their dissociation from DSB lesions. Inactivation of DNA-PKcs, either using a small molecule inhibitor or by RNA interference, alters alternative splicing of a set of pre-mRNAs in A549 cells treated with the topoisomerase II inhibitor mitoxantrone, indicative of an involvement of DNA-PKcs in modulating pre-mRNA splicing following genotoxic stress. These findings indicate a novel physical and functional connection between the DNA damage response and pre-mRNA splicing, and enhance our understanding of how mRNA splicing is involved in the cellular response to DSB lesions.

Keywords: DNA damage response; DNA‐PKcs; genotoxic stress; nuclear speckles; pre‐mRNA splicing.

Figure 1

S2056‐phosphorylated DNA‐PKcs accumulates at discrete nuclear foci outside of DNA damage sites in MTX‐treated A549 cells. (A) A549 cells were treated with 10 μm MTX for 60 min and subjected to indirect immunofluorescence staining for phosphorylated DNA‐PKcs (p‐DNA‐PKcs) (S2056) and 53BP1 as indicated. The images in the upper panel are representative of cells in which S2056‐phosphorylated DNA‐PKcs foci colocalized with 53BP1 foci following 10 μm MTX treatment for 60 min. The images in the lower panel are representative of cells in which S2056‐phosphorylated DNA‐PKcs accumulated at discrete nuclear areas that had an absence of 53BP1 signals following 10 μm MTX treatment for 60 min. Typical p‐DNA‐PKcs (S2056) foci outside of DSBs are indicated by arrows. Scale bar, 5 μm. (B) The phosphorylation status of DNA‐PKcs at S2056 was monitored by western blot. (C) A549 cells were transduced with control (shNTC) or shDNA‐PKcs virion for 48 h followed by a 4 h 0.5% DMSO or 10 μm MTX treatment. Protein levels of DNA‐PKcs, S2056‐phosphorylated DNA‐PKcs and GAPDH were monitored by western blot as indicated. (D) Cell lysates from MTX‐treated A549 cells were incubated with calf intestinal alkaline phosphatases (CIAP) for different times. The phosphorylation status of DNA‐PKcs at S2056 was monitored by western blot. (E) A549 cells were transduced with control (shNTC) or shDNA‐PKcs virion for 48 h followed by 10 μm MTX treatment for the indicated times. Representative images of indirect immunostaining of p‐DNA‐PKcs (S2056) are presented. Scale bar, 25 μm.

Figure 2

S2056‐phosphorylated DNA‐PKcs accumulates at nuclear areas that lack 53BP1 in MTX‐treated A549 cells. (A) Colocalization of phosphorylated DNA‐PKcs (p‐DNA‐PKcs (S2056)) with nuclear foci localizing 53BP1 was analyzed by indirect immunostaining in A549 cells treated with 10 μm MTX for the indicated times. Representative images are presented. Scale bar, 5 μm. (B) Quantitative analyses of cells with > 90% p‐DNA‐PKcs (S2056) foci localized within or outside of 53BP1 foci. Only cells that carried more than five discrete p‐DNA‐PKcs (S2056) nuclear foci were included in the quantitative analyses. The number of cells scored for each condition is indicated in the graphs. Error bars represent standard deviations from three independent experiments.

Figure 3

S2056‐phosphorylated DNA‐PKcs (p‐DNA‐PKcs) accumulates at nuclear areas that lack γ‐H2AX in MTX‐treated A549 cells. (A) Colocalization of p‐DNA‐PKcs (S2056) with nuclear foci localizing γ‐H2AX was analyzed by indirect immunostaining in A549 cells treated with 10 μm MTX for the indicated times. Representative images are presented. Scale bar, 5 μm. (B) Quantitative analyses of cells with > 90% p‐DNA‐PKcs (S2056) foci localized within or outside of γ‐H2AX foci. Only cells that carried more than five discrete p‐DNA‐PKcs (S2056) nuclear foci were included in the quantitative analyses. The number of cells scored for each condition is indicated in the graphs. Error bars represent standard deviations from three independent experiments.

Figure 4

S2056‐phosphorylated DNA‐PKcs is redistributed to nuclear speckles at the late stage of DDR. (A) Colocalization of phosphorylated DNA‐PKcs (p‐DNA‐PKcs (S2056)) with nuclear speckles (SRSF2) was analyzed via indirect immunostaining in A549 cells treated with 10 μm MTX for the indicated times. Representative images are presented. (B–D) Colocalization of p‐DNA‐PKcs (S2056) (B), 53BP1 (C), and p‐ATM (S1981) (D) foci with SRSF2 was analyzed by indirect immunostaining in A549 cells treated with DMSO, 10 μm MTX, 34 μm Etop, 2 μm CPT and 625 ng·mL⁻¹ NCS for 4 h, respectively. Representative images are presented. Scale bar, 5 μm. (E) Quantitative analyses of cells with > 90% DDR factor foci colocalized with speckles in (B–D). Only cells that carried more than five discrete DDR factor nuclear foci were included in the quantitative analyses. The number of cells scored for each condition is indicated in the graphs. Error bars represent standard deviations from three independent experiments.

Figure 5

DNA‐PKcs is involved in the control of pre‐mRNA alternative splicing in MTX‐treated A549 cells. (A) Schematic representation of human SRSF1 pre‐mRNA alternative splicing pattern. Whether the isoform contains a premature termination codon (PTC) was indicated on the right. (B) A549 cells were pretreated with 0.5% DMSO or 20 μm DNA‐PK inhibitor NU7026 for 30 min, followed by a 4 h treatment of DMSO or 10 μm MTX. Regular RT‐PCR was performed to monitor the splicing efficiency of an AS intron of the SRSF1 pre‐mRNA. Primer locations are indicated in (A) by the red arrows. A representative agarose gel electrophoresis image is presented. The ratio of the signal of the spliced RNA to that of the unspliced isoform is indicated. (C) A549 cells were treated as indicated in (B). qRT‐PCR was applied to monitor the abundance of the indicated spliced exon junctions of the SRSF1 mRNA using junction specific primers. qRT‐PCR signal of each junction was normalized to that of SRSF1 exon 1 in parallel assays. Ratios of the normalized junction abundances in the NU7026‐treated cells to that in the DMSO‐treated cells are shown in the graph. (D) A549 cells were transduced with control (shNTC) or shDNA‐PKcs virion for 48 h followed by a 4 h treatment of DMSO or 10 μm MTX. Relative abundance of the indicated exon junction was analyzed by qRT‐PCR as indicated in (C). Error bars represent standard deviations from three independent experiments. Significance of changes in splicing efficiency was assessed using Student's two‐tailed t test with significant changes indicated by **P < 0.01; NS, not significant. (E) A549 cells were treated as described in (B). Regular RT‐PCR was performed to monitor splicing of the indicated transcripts. Representative agarose gel electrophoresis images are presented. The ratio of the signal of the RNA carrying the alternative spliced exon (inclusion) to that of the RNA skipping the corresponding exon (skipping) is indicated.