N6-methyladenosine promotes induction of ADAR1-mediated A-to-I RNA editing to suppress aberrant antiviral innate immune responses

Fig 1. m6A methylation on ADAR mRNAs. (A) m6A-seq data of the ADAR1 and ADAR2 transcripts in HepG2 cells showing m6A modification on both transcripts, modified from Dominissini and colleagues [9]. An arrow indicates the transcription direction. (B) RIP RT-qPCR showing binding of YTHDF1 and YTHDF2 to ADAR1p110 and ADAR1p150 mRNAs in A172 cells. The signals were normalized to input samples. HPRT1 and GAPDH were used as negative controls. mTOL and MX1 were used as positive controls. (C) Immunoblot analysis showing knockdown effects of METTL3, METTL14, and YTHDF1 on IFN-induced ADAR1p150 protein expression in A172 cells. (D) m6A-RIP RT-qPCR showing m6A modification on ADAR1p110 and ADAR1p150 mRNAs. HPRT1 and GAPDH were used as negative controls. mTOL and MX1 were used as positive controls. (E) Immunoblot analysis showing time course of ADAR1p150 protein expression after IFN-α stimulation. Immunoblot images are representative of 3 biological replicates. (B, D) n = 3 for all experiments. Data are presented as the mean ± SEM. The numerical values for this figure are available in S1 Data. ADAR, adenosine deaminase acting on RNA; IFN, interferon; IgG, immunoglobulin G; IP, immunoprecipitation;…

Fig 2. YTHDF1 knockdown reduces IFN-induced A-to-I RNA editing. (A) Predicted secondary structure of TNKS and EIF2AK2 transcripts by the mfold web server. Red circles indicate A-to-I RNA editing sites. (B) Direct sequencing chromatogram showing from RT-PCR products of TNKS, EIF2AK2, and PSMB2 mRNA following IFN-α stimulation. (C) Effect of YTHDF1 knockdown on A-to-I RNA editing of a few selected transcripts following IFN-α stimulation. Reduced IFN-induced A-to-I RNA editing was observed with YTHDF1 knockdown. Two-tailed Student t tests were performed to assess the statistical significance of differences between groups, *p < 0.05, **p < 0.01, ***p < 0.001, n.s. p ≧ 0.05, N.D. means not detected. n = 3 for all experiments. Data are presented as the mean ± SEM. (D) A heatmap of over editing levels from 3 biological replicates at 488 selected editing sites. Editing level changes were shown in each row covering all editing sites, ranked by the editing level. High editing levels were displayed in yellow and low in blue. An overall IFN-induced A-to-I RNA editing levels were reduced with YTHDF1 knockdown. (E) A flowchart for identification of IFN-induced editing sites. The numerical values for…

Fig 3. YTHDF1 knockdown induces IFN responses. Cells were treated with mock (−) or IFN-α (+). (A) RT-qPCR showing significantly elevated expression of IFN genes upon YTHDF1 knockdown and IFN stimulation. The signals were normalized to control siRNA samples. (B) Spontaneous IFN-β secretion by cells transfected with control or YTHDF1-specifc siRNA as quantified by ELISA after stimulation of IFN-α. (C) Immunoblot analyses of TBK1 and STAT1 phosphorylation levels. Immunoblot images are representative of 3 biological replicates. (D) RT-qPCR showing significantly elevated expression of ISGs upon YTHDF1 knockdown and IFN stimulation. The signals were normalized to mock treatment samples. (E, F) Heatmaps of expression levels of ISGs (E) and NF-κB–inducible genes (F) from RNA-seq data. (G) RT-qPCR showing significantly elevated expression of NF-κB–inducible genes upon YTHDF1 knockdown and IFN stimulation. The signals were normalized to mock treatment samples except for CXCL11. (H) Immunoblot analyzing of IκBα and NF-κB p65 phosphorylation levels. Immunoblot images are representative of 3 biological replicates. (A, D, G) The signals were normalized to GAPDH. (A, B, D, G) Two-tailed Student t…

Fig 4. ADAR1p150 A-to-I RNA editing activity is required for the YTHDF1-dependent IFN responses. (A) Immunoblot analysis in stable cell lines with lentivirus expressing control EGFP, wild-type ADAR1p150, or catalytically inactive mutant of ADAR1p150E912A, respectively. (B) RT-qPCR showing knockdown effect of YTHDF1 on IFN genes, ISGs, and NF-κB–inducible genes in stable cell lines following IFN-α stimulation. The enhanced IFN responses were attenuated by overexpression of ADAR1p150, but not by catalytic inactive mutant ADAR1p150E912A. (C) RT-qPCR showing knockdown effect of YTHDF1 on IFN genes, ISGs, and NF-κB–inducible genes in A172 cells that were pretreated with the BX759 inhibitor for 1 h and then treated with IFN-α. The enhanced IFN responses were attenuated by the BX759. (D) RT-qPCR showing that concomitant knockdown of YTHDF1 and MDA5 attenuated the enhanced expression of IFN genes, ISGs, and NF-κB–inducible genes upon IFN-α stimulation. (B–D) The signals were normalized to GAPDH and then normalized to control siRNA samples. Two-tailed Student t tests were performed to assess the statistical significance of differences between groups, *p < 0.05, **p < 0.01, ***p < 0.001. n = 3…

Fig 5. YTHDF1 knockdown inhibits viral replication in cells. (A) RT-qPCR showing significant decrease in the expression of VSV genome RNA and antigenome RNA upon YTHDF1 knockdown at 12 and 20 h after rVSV-GFP infection. (B) Viral titers in culture medium at 12, 20, and 24 h after rVSV-GFP infection were determined by plaque assay. YTHDF1 knockdown inhibited the release of infectious virus. (C) RT-qPCR showing elevated expression levels of IFN genes, ISGs, and NF-κB–inducible genes upon YTHDF1 knockdown at 4 and 8 h after rVSV-GFP infection. The signals were normalized to GAPDH. (D) RT-qPCR showing knockdown effect of YTHDF1 on the expression of VSV genome RNA and antigenome RNA at 12 and 20 h after rVSV-GFP infection in stable cell lines. (E) Viral titers in culture medium at 12, 20, and 24 h after rVSV-GFP infection were determined by plaque assay. (F) RT-qPCR showing knockdown effect of YTHDF1 on IFN genes, ISGs, and NF-κB–inducible genes in stable cell lines. The signals were normalized to GAPDH and then normalized to control siRNA samples. (A–F) Two-tailed Student t tests were performed to assess the statistical significance of differences between groups, *p < 0.05, **p < 0.01, ***p…