m5C modification of mRNA serves a DNA damage code to promote homologous recombination

Abstract

Recruitment of DNA repair proteins to DNA damage sites is a critical step for DNA repair. Post-translational modifications of proteins at DNA damage sites serve as DNA damage codes to recruit specific DNA repair factors. Here, we show that mRNA is locally modified by m⁵C at sites of DNA damage. The RNA methyltransferase TRDMT1 is recruited to DNA damage sites to promote m⁵C induction. Loss of TRDMT1 compromises homologous recombination (HR) and increases cellular sensitivity to DNA double-strand breaks (DSBs). In the absence of TRDMT1, RAD51 and RAD52 fail to localize to sites of reactive oxygen species (ROS)-induced DNA damage. In vitro, RAD52 displays an increased affinity for DNA:RNA hybrids containing m⁵C-modified RNA. Loss of TRDMT1 in cancer cells confers sensitivity to PARP inhibitors in vitro and in vivo. These results reveal an unexpected TRDMT1-m⁵C axis that promotes HR, suggesting that post-transcriptional modifications of RNA can also serve as DNA damage codes to regulate DNA repair.

Fig. 1. m⁵C mRNA methylation is enriched at transcriptionally active sites with DNA damage.

a U2OS-TRE cells transfected with TA-KR/TA-Cherry/tetR-KR/tetR-Cherry plasmids were exposed to light for 30 min for KR activation and allowed to recover for 1 h before harvest (scale bar: 10 μm). Quantification of frequency of cells in 500 cells with m⁵C foci from three independent experiments, mean ± SD (upper right). Fold increase of m⁵C mean intensity = mean intensity of m⁵C at TA-KR/mean intensity of background (n = 20, mean ± SD) (lower right). b U2OS-TRE cells were transfected with TA-KR/TA-Cherry to induce local oxidative damage or for the control condition. Cells were then stained for m⁵C with four different anti-m⁵C antibodies. Frequency of m⁵C-positive cells in 500 cells was quantified (n = 3, mean ± SD). c U2OS-TRE cells transfected with TA-KR were digested with RNaseH1, RNaseA, or DNase I and stained for m⁵C quantification (scale bar: 10 μm). d The mRNA from Flp-in 293 cells treated with or without 2 mM H₂O₂ for 40 min was used for m⁵C measurement via dot blot. Quantification of m⁵C levels (mean ± SD) from three independent experiments normalized with Ctrl and methylene blue is shown. e ³²P-labeled mRNA monophosphate nucleosides were run on 2D gels for 2D-TLC analysis. In vitro-transcribed 4B mRNA with or without m⁵C was run in parallel. Representative images from three sets of independent experiments are shown with arrows showing the directions of each solvent run. Position of each nucleotide and m⁵C are labeled (Left). f ³²P-labeled mRNA monophosphate nucleosides from U2OS cells with or without 2 mM H₂O₂ for 40 min were run on 2D gels for 2D-TLC analysis. Representative images from three sets of independent experiments. Associated quantification of relative increase in m⁵C in peroxide-treated cells compared to control, normalized to nucleotide C (right). Statistical analysis was performed with the unpaired two tailed Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig. 2. TRDMT1 mediates m⁵C mRNA methylation at DNA damage sites.

a U2OS-TRE cells pre-treated with the indicated siRNA were transfected with TA-KR to induce local oxidative damage and stained for m⁵C. The frequency of cells in 500 cells with m⁵C foci (n = 3, mean ± SD) is shown. b The U2OS-TRE WT and TRDMT1 KO cells were transfected with TA-KR, stained for m⁵C, and quantified (scale bar: 10 μm) (n = 3, mean ± SD). c WT U2OS and TRDMT1 KO cells were treated with 2 mM H₂O₂ for 40 min. The mRNA was then extracted from the cell lysates for m⁵C measurement via dot blot. Quantification of m⁵C levels (mean ± SD) from three independent experiments normalized with Ctrl and methylene blue is shown. d ³²P-labeled monophosphate nucleosides from mRNA from WT U2OS and TRDMT1 KO cells with or without 2 mM H₂O₂ for 40 min were run on 2D gels for 2D-TLC analysis. Representative images from three sets of independent experiments. Associated quantification of relative increases in m⁵C from H₂O₂-treated cells were compared to the control and normalized to nucleotide C (n = 3, mean ± SD). e The TRDMT1 KO and stably rescued cell lines expressing WT TRDMT1 or the C79A mutant were transfected with TA-KR, stained for m⁵C, and quantified (scale bar: 10 μm) (n = 3, mean ± SD). f U2OS-TRE cells transfected with TA-KR/TA-Cherry/tetR-KR/tetR-Cherry and GFP-TRDMT1/GFP-NSUN2 plasmids were light-irradiated and allowed to recover for 1 h before fixation and were quantified (scale bar: 10 μm) (n = 10, mean ± SD). Statistical analysis was performed with the unpaired two tailed Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig. 3. TRDMT1 is required for damage removal and HR.

a WT or TRDMT1 KO U2OS-TRE cells were transfected with TA-KR and stained for γH2AX at the indicated timepoint after damage (n = 3, mean ± SD). b Stable expression of Myc-tagged TRDMT1 in TRDMT1 KO U2OS-TRE cell lines shown in Western blots. Expression of β-actin is shown as a positive control. c U2OS-TRE WT, TRDMT1 KO, and TRDMT1 stably expressing cells were stained for γH2AX after damage caused by TA-KR at the indicated timepoint (n = 3, mean ± SD). d DR-GFP cells were pre-treated with siTRDMT1, siBRCA1, or control siRNA and then transfected with the NLS-I-SceI plasmid to induce DSBs. The GFP-positive population was analyzed by flow cytometry (n = 3, mean ± SD). e WT or TRDMT1 KO U2OS cells were co-transfected with CRISPR/Cas9-sgLMNA, LMNA-mClover, and mCherry plasmids. The fraction of mClover-positive cells in the mCherry-positive population was analyzed by flow cytometry (n = 3, mean ± SD). f EJ5-GFP cells were pre-treated with siTRDMT1 or control siRNA and then transfected with the NLS-I-SceI plasmid to induce DSBs. The GFP-positive population was analyzed by flow cytometry (n = 3, mean ± SD). Statistical analysis was performed with the unpaired two tailed Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig. 4. RAD52 is a m⁵C reader.

a U2OS-TRE WT and TRDMT1 KO cells were transfected with TA-KR and stained for RAD51 1 h after light irradiation. Representative figures were shown (scale bar: 10 μm). b U2OS-TRE WT, TRDMT1 KO, and TRDMT1 stably rescued cell lines were transfected with TA-KR. and stained for RAD51 1 h after light irradiation. Fold increase of RAD51 foci intensity was calculated (n = 25, mean ± SD). c U2OS-TRE WT and TRDMT1 KO cells were transfected with TA-KR and GFP-RAD52 and fixed 1 h after light irradiation. Representative figures were shown (scale bar: 10 μm). d U2OS-TRE WT, TRDMT1 KO, and TRDMT1 stably rescued cell lines were co-transfected with GFP-RAD52 and TA-KR and quantified for RAD52 foci frequency (n = 3, mean ± SD). e RNA oligos with or without m⁵C modification and their complementary DNA oligos were chemically synthesized for RAD52 binding experiments. f RAD52 protein in the cell lysate was pulled down by biotin-labeled DNA:RNA hybrids, with or without m⁵C RNA modification, coated on streptavidin magnetic beads. g The binding of RAD52 protein to DNA:RNA hybrids with or without m⁵C RNA modification was measured in electrophoretic mobility shift assays and quantified (n = 3, mean ± SD). Statistical analysis was performed with the unpaired two tailed Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Fig. 5. TRDMT1 inhibition sensitizes cells to radiation and PARPi.

a Clinical and RNA-seq gene expression data were downloaded from TCGA. RSI was calculated by a rank-based linear regression algorithm. The median value of BRCA1 and TRDMT1 TPM was defined as the cutoff to classify patients into high expression and low expression groups (n = 1091, mean ±SD). b WT, TRDMT1 KO U2OS, or siBRCA1 cells were treated with Olaparib at the indicated dose. The survival rate was measured via the colony formation assay (n = 3, mean ± SEM). c Tumor volume in mice injected with LV-shTRDMT1 or LV-NC pre-treated MDA-MB-231 cells at the indicated time (n = 3, mean ± SEM). One week after injection, 50 mg/kg Olaparib or saline was intraperitoneally injected into the xenograft tumors once every other day. d Tumor volume in mice injected with LV-shTRDMT1 or LV-NC pre-treated MDA-MB-231 cells at day 12 after PARPi treatment (n = 3, mean ± SEM). Statistical analysis was performed with the unpaired two tailed Student’s t-test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.