N 6-methyladenosine of chromosome-associated regulatory RNA regulates chromatin state and transcription

Abstract

N ⁶-methyladenosine (m⁶A) regulates stability and translation of messenger RNA (mRNA) in various biological processes. In this work, we show that knockout of the m⁶A writer Mettl3 or the nuclear reader Ythdc1 in mouse embryonic stem cells increases chromatin accessibility and activates transcription in an m⁶A-dependent manner. We found that METTL3 deposits m⁶A modifications on chromosome-associated regulatory RNAs (carRNAs), including promoter-associated RNAs, enhancer RNAs, and repeat RNAs. YTHDC1 facilitates the decay of a subset of these m⁶A-modified RNAs, especially elements of the long interspersed element-1 family, through the nuclear exosome targeting-mediated nuclear degradation. Reducing m⁶A methylation by METTL3 depletion or site-specific m⁶A demethylation of selected carRNAs elevates the levels of carRNAs and promotes open chromatin state and downstream transcription. Collectively, our results reveal that m⁶A on carRNAs can globally tune chromatin state and transcription.

Fig. 1.. Mettl3 KO in mESCs leads to increased nascent RNA transcription and chromatin accessibility.

(A and B) Analysis of nascent RNA synthesis in WT or Mettl3^−/− mESCs [(A), Mettl3^−/−−1 and −2 are two independently generated KO lines], and Mettl3^−/− mESCs rescued with WT or an inactive mutant Mettl3 (B). Nascent RNA synthesis was detected by using a click-it RNA Alexa fluor 488 imaging kit. EU, 5-ethynyl uridine; DAPI, 4′,6-diamidino-2-phenylindole. (C and D) Analysis of chromatin accessibility in WT or Mettl3^−/− mESCs (C), and Mettl3^−/− mESCs rescued with WT or mutant Mettl3 (D). DNase I–treated TUNEL assay was performed. For (A) to (D), the nucleus is counterstained by DAPI. EV (empty vector) refers to Mettl3^−/− mESCs when transfected with empty vector plasmid. dsDNA, double-stranded DNA.

Fig. 2.. Transcript turnover of carRNAs is regulated by m⁶A.

(A) LC-MS/MS quantification of the m⁶A/A ratio in nonribosomal (non-Rib) caRNAs (including pre-mRNA) extracted from WT or Mettl3^−/− mESCs. n = 3 biological replicates; error bars indicate means ± SEM. (B) m⁶A level changes on carRNAs were quantified through normalizing m⁶A sequencing results with spike-in between WT and Mettl3 KO mESCs. n = 2 biological replicates. (C) carRNAs were divided into methylated (m⁶A) or nonmethylated (non-m⁶A) groups. The boxplot shows greater increases in transcript abundance fold changes of the m⁶A group versus the non-m⁶A group upon Mettl3 KO over WT mESCs. For (A) and (C), P values were determined by two-tailed t test. (D) Cumulative distribution and boxplots (inside) of nuclear carRNA half lifetime changes in CKO Ythdc1 and control mESCs. (E) Cumulative distributions and boxplots (inside) of the half lifetime changes of carRNAs upon Ythdc1 CKO. carRNAs were divided into methylated (m⁶A) or nonmethylated (non-m⁶A) groups. Depletion of YTHDC1 led to greater half lifetime increases of m⁶A-marked carRNAs than non-m⁶A–marked ones. For (D) and (E), P values were calculated by a nonparametric Wilcoxon-Mann-Whitney test. h, hours.

Fig. 3.. The m⁶A level of carRNAs affects downstream gene expression and transcription rate.

(A and B) Volcano plot of genes with differential expression levels in Mettl3^−/−−1 (A) or Mettl3^−/−−2 (B) versus WT mESCs [P < 0.05 and |log₂FC| > 1 (FC, fold change)]. Genes with upstream, m⁶A-marked carRNAs are shown with orange circles. Gene expression level was normalized to ERCC spike-in with linear regression method. DEG, differentially expressed gene. (C and D) Cumulative distribution and boxplot (inside) of gene transcription rate in Mettl3^−/−−1 (C) or Mettl3^−/−−2 (D) versus WT mESCs. (E and F) Cumulative distribution and boxplot (inside) of transcription rate difference between Mettl3^−/−−1 (E) or Mettl3^−/−−2 (F) versus WT mESCs. Genes were categorized into two subgroups according to whether their upstream carRNAs contain m⁶A (m⁶A) or not (non-m⁶A). For (C) to (F), P values were calculated by a nonparametric Wilcoxon-Mann-Whitney test. (G) Heatmap showing the m⁶A level fold changes (log₂FC < −1) on carRNAs and downstream gene transcription rate difference between Mettl3 KO and WT mESCs. (H) Venn diagram showing the overlap between the m⁶A peaks and super-enhancer peaks in mESCs. (I) Boxplot showing fold changes of the abundance of m⁶A-marked and non-m⁶A–marked seRNAs between Mettl3^−/− and WT mESCs. For (A), (B), and (I), P values were determined by two-tailed t test. (J) Heatmap showing fold change (log₂FC < −0.38) of m⁶A level of seRNAs and transcription rate difference of their downstream genes between Mettl3 KO and WT mESCs.

Fig. 4.. The m⁶A level of carRNAs affects local chromatin state and downstream transcription.

(A) Profiles of H3K27ac (top) and H3K4me3 (bottom) level changes on gene body together with 2.5 kb upstream of the TSS (transcription start site) and 2.5 kb downstream of the TTS (transcription termination site) in WT and Mettl3 KO mESCs. Genes were categorized into two groups according to whether they harbor upstream m⁶A-marked carRNAs (m⁶A) or not (non-m⁶A). (B and C) Profiles of EP300 (B) or YY1 (C) DNA binding at their peak center and flanking 2.5 kb regions in WT and Mettl3 KO mESCs. (D and E) Profiles of EP300 (D) and YY1 (E) DNA binding at the center of m⁶A peaks overlapped with carRNAs and its flanking 2.5 kb regions in WT mESCs. m⁶A peaks were categorized into highly (high), moderately (medium), or lowly (low) methylated groups, according to their m⁶A levels in WT mESCs. (F and G) The correlation between changes in m⁶A level of the carRNAs and changes in EP300 (F) or YY1 (G) DNA binding at genomic regions that show m⁶A differences with Mettl3 KO. The genomic regions were categorized into 100 bins on the basis of fold change rank of m⁶A level upon Mettl3 KO. (H) Barplots showing H3K27ac level changes at genomic regions that are m⁶A methylated (m⁶A only, without EP300 and YY1 binding), bound by EP300 or YY1 (EP300/YY1 only, without m⁶A carRNA), and m⁶A methylated with EP300 and YY1 binding (m⁶A + EP300/YY1). The last group showed the highest increase upon Mettl3 KO. (I) Analysis of chromatin accessibility in Mettl3 KO mESCs treated with control or LINE1 antisense oligos (ASOs). DNase I–treated TUNEL assay was performed. (J) A dCas13b-FTO (WT or inactive mutant) construct with gRNA targeting the seRNA of Arntl was used to reduce the m⁶A level of Arntl seRNA. After treatment, increased half lifetime of the target seRNA, elevated local H3K27ac and H3K4me3 levels, and increased Arntl transcription rate were observed, accompanied by the decreased seRNA m⁶A level. (K) A schematic model showing how m⁶A affects transcription by regulating the decay of upstream carRNAs stability and chromatin state.