Single-base mapping of m6A by an antibody-independent method

Abstract

N ⁶-methyladenosine (m⁶A) is one of the most abundant messenger RNA modifications in eukaryotes involved in various pivotal processes of RNA metabolism. The most popular high-throughput m⁶A identification method depends on the anti-m⁶A antibody but suffers from poor reproducibility and limited resolution. Exact location information is of great value for understanding the dynamics, machinery, and functions of m⁶A. Here, we developed a precise and high-throughput antibody-independent m⁶A identification method based on the m⁶A-sensitive RNA endoribonuclease recognizing ACA motif (m⁶A-sensitive RNA-Endoribonuclease-Facilitated sequencing or m⁶A-REF-seq). Whole-transcriptomic, single-base m⁶A maps generated by m⁶A-REF-seq quantitatively displayed an explicit distribution pattern with enrichment near stop codons. We used independent methods to validate methylation status and abundance of individual m⁶A sites, confirming the high reliability and accuracy of m⁶A-REF-seq. We applied this method on five tissues from human, mouse, and rat, showing that m⁶A sites are conserved with single-nucleotide specificity and tend to cluster among species.

Fig. 1. m⁶A identification method based on m⁶A sensitive RNA endoribonuclease.

(A) Validation for the methylation sensitivity of endoribonuclease MazF by synthetic m⁶A-containing RNA oligonucleotide. (B) Validation for the methylation sensitivity of endoribonuclease ChpBK. (C) Various ratios of m⁶A-containing oligo mixed with unmethylated oligo digested by endoribonuclease. The m⁶A/A ratios of synthetic oligo are indicated to imitate practical RNAs with different m⁶A ratios. (D) Relative grayscale of proportionally mixed RNA oligo digested by endoribonuclease. (E) The schematic diagram of m⁶A-REF-seq.

Fig. 2. Transcriptome-wide distribution of m⁶A revealed by m⁶A-REF-seq.

(A) The analysis pipeline of m⁶A-REF-seq data. (B) The snapshot of sequencing reads shows a known m⁶A site in 18S rRNA. (C) Overlap sites among three replicates after removing RNA secondary structure. (D) The overall shift of methylation ratio after FTO treatment. The value of methylation change indicates the methylation ratio of each m⁶A site in MazF minus that after FTO treatment. (E) Transcriptome-wide distribution of m⁶A. Pie chart shows the percentages of m⁶A sites located within CDS, 5′UTR, and 3′UTR. (F) Single-base m⁶A sites from m⁶A-REF-seq show a typical transcriptome-wide distribution pattern of m⁶A. (G) Overlap of m⁶A sites to the m⁶A peaks identified by antibody-based method. (H) The proportion of motifs containing m⁶A sites. The red square includes the RRACA motif and the green square includes the DRACA motif.

Fig. 3. Single-base validation using ligation-amplification method and qPCR.

(A) Schematic diagram of ligation-amplification method for single-base m⁶A validation. (B) Validation results of six individual sites. Five of the sites are validated to be m⁶A sites, whereas the other one is confirmed to be not modified. Data for m⁶A-CLIP, miCLIP-CITS, miCLIP-CIMS, and MeRIP-seq are downloaded from published literatures. (C) Schematic diagram of qPCR to quantify the methylation level of a specific m⁶A site. The undigested mRNA sample is treated as control. (D) qPCR results and methylation ratios of six m⁶A sites. H1 to H3 represent the highly methylated sites (>0.75), while L1 to L3 represent the weakly methylated sites (<0.35). The left y axis represents the ΔΔCt values, and the right y axis represents the methylation ratio determined by m⁶A-REF-seq. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Fig. 4. Single-base method reveals high resolution features of m⁶A.

(A) The distance of individual m⁶A sites to the stop codons. DRACA and DRACH background motifs are extracted from the transcriptome. (B) The relative position of m⁶A sites within exons. (C) The length distribution of m⁶A-containing exons versus all exons in m⁶A-modified genes. (D) The distances between two m⁶A sites are significantly enriched within 200 bp regions compared to random sampling (Fisher’s exact test, **P < 0.01).

Fig. 5. Conservation of m⁶A in mammals.

(A) Metagene plots of m⁶A in the brain of human, mouse, and rat. (B) Shared m⁶A-modified genes among three species. (C) Diagram showing the m⁶A sites conserved in the corresponding short regions from different species. (D) Frequency of distances for pairwise m⁶A in brain. Randomly picked ACA motifs are assigned for the same analysis as control. (E) Conservation scores of all m⁶A sites, methylation sites in ortholog genes, and conserved m⁶A sites are compared to that for all A sites in ACA motifs (P < 9 × 10⁻¹⁰, Wilcoxon test).