N6-methyladenosine RNA modification suppresses antiviral innate sensing pathways via reshaping double-stranded RNA

Abstract

Double-stranded RNA (dsRNA) is a virus-encoded signature capable of triggering intracellular Rig-like receptors (RLR) to activate antiviral signaling, but whether intercellular dsRNA structural reshaping mediated by the N⁶-methyladenosine (m⁶A) modification modulates this process remains largely unknown. Here, we show that, in response to infection by the RNA virus Vesicular Stomatitis Virus (VSV), the m⁶A methyltransferase METTL3 translocates into the cytoplasm to increase m⁶A modification on virus-derived transcripts and decrease viral dsRNA formation, thereby reducing virus-sensing efficacy by RLRs such as RIG-I and MDA5 and dampening antiviral immune signaling. Meanwhile, the genetic ablation of METTL3 in monocyte or hepatocyte causes enhanced type I IFN expression and accelerates VSV clearance. Our findings thus implicate METTL3-mediated m⁶A RNA modification on viral RNAs as a negative regulator for innate sensing pathways of dsRNA, and also hint METTL3 as a potential therapeutic target for the modulation of anti-viral immunity.

Fig. 1. METTL3 globally inhibits innate immune signaling cascade.

a qPCR analysis of Ifnb1 expression following a 12 h treatment with different triggers as indicated before overexpression of Mettl3 in RAW264.7 cells (Left). qPCR analysis of IFNB1 expression following a 12 h treatment with HBV before knockdown of METTL3 in AC12 cells (Right). n = 3 biologically independent experiments. b IFN-β promoter activity in HEK293T cells transfected with METTL3 vector upon different treatments for 12 h (Left). IFN-β promoter activity in AC12 cells transfected with siMETTL3 upon HBV treatment for 12 h (Right). n = 3 biologically independent experiments. c Western blot analysis of the proteins in innate immune signaling isolated from peritoneal macrophage of WT or Mettl3-deficient mice. The peritoneal macrophage was treated with PBS or VSV infection for 12 h before lysis. d Western blot analysis of the proteins from A549 cell line before treated with siScramble or siMETTL3 for 2 days. e–h Western blot analysis of the proteins from RAW264.7 (e), HeLa (f), Huh7 (g), and LO2 (h) cell lines before overexpression of METTL3 for 2 days. i Immunofluorescence analysis of IRF3 translocation and activation in HEK293T cells after VSV infection for 6 h or PBS treatment. The arrowed cells indicated the METTL3 overexpressed cells which inhibited the nuclear translocation and activation of IRF3. The “*” labeled cells indicated control cells. α-Tubulin indicated the cytosolic part. n = 2 biologically independent experiments. j GO and pathway enrichment analysis of upregulated genes upon Mettl3 knockout in RAW264.7 cells after VSV infection (two-tailed hypergeometric test with p < 0.05). k Two-tailed Gene Set Enrichment Analysis of the Ifnb1 signaling transcriptional signature in Mettl3 knockout relative to control groups. NES, normalized enrichment score. l Scatter plot showing the alteration of genes expression by comparing Mettl3 knockout and WT RAW264.7 cells after VSV infection for 12 h (yellow, upregulated; blue, downregulated; gray, no significant change; red triangle, ISGs). Data are representative of 2–3 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, as determined by two-tailed unpaired Student’s t test (Fig. 1a, b, i). Error bars represent mean ± SEM.

Fig. 2. Deficiency in METTL3 protects mice against VSV infection in vivo.

a Survival of Mettl3^f/f Lyz2-Cre and Mettl3^f/f littermate female mice (n = 5 per group) at various times (horizontal axes) after intraperitoneal infection with VSV (1 × 10⁹ PFU per mouse). b, c Microscopy of hematoxylin-and-eosin-stained lung (b) or liver (c) sections from female mice treated with PBS or VSV. d qRT-PCR analysis of Ifnb1 and VSV mRNA in the spleens (left), livers (middle), and lung (right) of Mettl3^f/f Lyz2-Cre and Mettl3^f/f littermate female mice (n = 5 per group) given intraperitoneal injection of PBS or infected for 24 h by intraperitoneal injection of VSV (2.5 × 10⁸ PFU per mouse); results are presented relative to those of actin. e Western blot analysis of VSV-G in the spleens, livers, and lungs of infected mice. f ELISA analysis of IFN-βin serum. n = 6 biologically independent animals. g Isolated peritoneal macrophages from VSV-infected mice, and then performed Western blot by indicated antibodies. h qRT-PCR analysis of Mettl3 conditionally knocked out in hepatocytes. n = 3 biologically independent experiments. i Survival of Mettl3^f/fAlb-Cre and Mettl3^f/f littermate female mice (n = 5 per group) at various times (horizontal axes) after intraperitoneal infection with VSV (1 × 10⁹ PFU per mouse). j, k qRT-PCR analysis of Ifnb1 (i) and VSV (j) mRNA in the spleens, livers, and lung of Mettl3^f/f Alb-Cre and Mettl3^f/f littermate female mice (n = 5 per group) given intraperitoneal injection of PBS or infected for 24 h by intraperitoneal injection of VSV (2.5 × 10⁸ PFU per mouse); results are presented relative to those of actin. l ELISA analysis of IFN-β in serum. n = 5 biologically independent animals. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by two-tailed unpaired Student’s t test (d, f, h, j–l). Log-rank (Mantel–Cox) test (a, i). Error bars represent mean ± SEM.

Fig. 3. VSV infection induces METTL3 cytoplasmic translocation and dampens type I IFNs.

a Western blot analysis of Mettl3 protein along VSV infection of RAW264.7 cells. The upper graph indicates statistic result. b Cellular fractionation and Western blot analysis of METTL3 protein localization before or post VSV infection in HEK293T cells. c Immunofluorescence analysis of endogenous METTL3 protein subcellular localization before or post VSV infection for 6 h in HeLa cells. d Immunofluorescence analysis of endogenous METTL3 protein subcellular localization before or post HBV (DNA virus) infection as indicated timepoints in AC12 cells. e The domain organization and dissection of the human METTL3 protein. NLS-mut: nuclear localization signal-mutation. f Immunofluorescence analysis of METTL3 protein subcellular localization after NLS-mutation in HeLa cells. g Dot blot analysis of mRNA-m⁶A level after overexpression of NLS-mut or WT METTL3 in HeLa cells. Methylene blue staining indicated the loading control. h qRT-PCR analysis of IFNB1 expression following a 12 h treatment with PBS or VSV infection before transfection of Vector/ NLS-mut/ WT METTL3, respectively, in HeLa cells. i IFN-β promoter activity in HEK293T cells transfected with METTL3 vectors upon VSV infection. j RAW264.7 cells overexpressed catalytic-mutated Mettl3 (Mettl3-mut) or WT Mettl3 in mock (PBS) or VSV infection for 12 h. The supernatants were collected to perform plaque assay, which indicated the VSV titer. k, l qRT-PCR and ELISA analysis of Ifnb1 expression in RAW264.7 cells overexpressed Mettl3-mut or WT Mettl3 in mock (PBS) or VSV infection for 12 h. Data are representative of 3 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, as determined by two-tailed unpaired Student’s t test. Error bars represent mean ± SEM.

Fig. 4. METTL3 mediates m⁶A modification on VSV RNA.

a A schematic representation of the experimental procedure used in b. b Dot blot analysis of VSV RNA m⁶A level treated with immunoprecipitated (IP)-IgG or IP-METTL3, respectively. Methylene blue staining indicated the loading control. c MeRIP-qPCR analysis of VSV RNA m⁶A level following treatment with overexpression of WT or Mettl3-mut, respectively. n = 2 biologically independent experiments. d RNA-FISH analysis of the co-localization of VSV RNA and m⁶A modification in the cytosol. e The conserved sequence motif of m⁶A residues in CIMS-based miCLIP-seq. f Integrative genomics viewer (IGV) plots of the METTL3-binding regions and m⁶A modification on VSV negative sense (−) genomic RNA (Upper graph) and VSV positive sense (+) RNAs (Lower graph). m⁶A sites are indicated by red triangles. g MeRIP-qPCR analysis of specific m⁶A sites on VSV RNA in WT or Mettl3 KO RAW264.7 cells. n = 2 biologically independent experiments. *p < 0.05, **p < 0.01, as determined by two-tailed unpaired Student’s t test (c, g). Error bars represent mean ± SEM.

Fig. 5. METTL3-mediated m⁶A modification reshapes viral dsRNA.

a Immunofluorescent analysis of dsRNA level after PBS treatment or VSV infection for 12 h in shNC and shMETTL3 HeLa cells. n = 10 cells examined over 2 independent experiments. b Dot blot analysis of dsRNA level after VSV infection for 12 h in HeLa cell. Methylene blue staining indicates equal RNA loading. The bar graph shows the statistics (right). n = 3 independent experiments. c, d Immunoblot (c) and dot blot (d) shows the immunoprecipitated (IP)-IgG or IP-dsRNA, and the m⁶A level in the IP-dsRNA in shScr and shMETTL3 HeLa cells. Methylene blue staining indicates equal RNA loading. The bar graph shows the statistics (right). e Anti-dsRNA-RIP-qPCR analysis of the dsRNA level in VSV RNA in shScr and shMETTL3 HeLa cells. The bar graph (Right panel) shows the statistics from the mean value of each m⁶A site (Left panel). Data are representative of 2 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as determined by two-tailed unpaired Student’s t test (a, b, d, e-left) or two-tailed paired Student’s t test (e-right). Error bars represent mean ± SEM.

Fig. 6. m⁶A modification impairs viral RNA sensing efficacy by RLRs.

a Integrative genomics viewer (IGV) plots the m⁶A sites and RIG-I and MDA5-binding regions on VSV (+) RNA. RNA-seq data were used as input control. b Venn diagram showing the overlap between high-confidence MDA5 and RIG-I binding clusters. The number of clusters in each category is shown in parenthesis. c Biotin-labeled RNA pull-down and Western blot analysis of Rig-I and Mda5 bindings to RNA oligo with or without single m⁶A modification (Lower graph). The upper graph indicates predicted structure of VSV (+) RNA: 2912–2974. Three times each experiment was repeated independently with similar results. Histograms show mean relative RNA content pulled down from 3 independent replicates. d Immunofluorescence analysis of HeLa cells increased the co-localization between RIG-I/ MDA5 and dsRNA induced by VSV infection for 8 h. n = 20 cells examined over 2 independent experiments with similar results. e RIP-qPCR analysis of increase of RIG-I and MDA5 binding to VSV (+) RNA (region: 1129–1329 nt, referred to Fig. 5a PAR-CLIP result) after deficient for METTL3 in HeLa cells. Data are representative of 2 independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, as determined by two-tailed unpaired Student’s t test (c–e). Error bars represent mean ± SEM.

Fig. 7. Model for METTL3-mediated m⁶A modification dampening viral RNA secondary structure to avoid innate immunity sensing.

In the proposed model, VSV RNA contains dsRNA structures to initiate innate immune sensing. During the VSV infection, METTL3 can be attracted from nucleus to cytoplasm to contact and modify VSV RNA. This m⁶A modification impairs the conformation of duplex structures in VSV RNA and interferes the sensing by dsRNA sensors involving RIG-I and MDA5, which attenuates innate immune response and helps virus invasion. When the host is deficient for METTL3, there are more dsRNA structures recognizing by RLRs to drive the expression of type I IFNs, following enhances in anti-viral function. However, the DNA virus HBV generates cccDNA in nuclear with sufficient m⁶A modification to produce less dsRNA.