The birth of the Epitranscriptome: deciphering the function of RNA modifications

Abstract

Recent studies have found methyl-6-adenosine in thousands of mammalian genes, and this modification is most pronounced near the beginning of the 3' UTR. We present a perspective on current work and new single-molecule sequencing methods for detecting RNA base modifications.

Figure 1

Peak distribution. (a) We plotted the distribution of the peaks reported across gene bodies by the MeRIP-seq and m⁶A-seq studies. Note the very well defined enrichment for peaks near the stop codon and in the 3' UTR. The m⁶A-seq HepG2 peaks also show a peak in the 5' UTR. (b) The distribution of peaks across the transcriptome using the BWA-based MeRIPPeR pipeline [35] on the data from both groups. Data from [19] and [20]. CDS, coding sequence.

Figure 2

Peak distribution across the transcriptome. The peak distribution depicted is the average across the entire transcriptome. Peaks are mapped to transcripts and assigned to the following transcriptomic features: 1 kB upstream from the TSS and downstream from the transcription end site, 5' and 3' UTRs, coding segments (CDS), and exon and intron segments. In the bottom row, peaks mapping to transcripts with four or more exons are shown, with the first, penultimate and last exons separated into individual boxes, as are their neighboring introns. The remaining exons and introns are shown in the middle boxes as a contiguous segment. Genes with only two or three exons are shown in the middle row and single exon genes are shown in the top row. Data from [19] and [20].

Figure 3

Distribution of [AG]ACU motif sites. The [AG]ACU motif was used to find potential m⁶A sites within peaks, and the distribution of these potential sites across gene bodies plotted. Data from [19] and [20].

Figure 4

Single-molecule sequencing of RNA to detect epitranscriptomic changes. SMRT sequencing with the Pacific Biosciences RS shows longer times (inter-pulse distances) to incorporate m⁶A versus standard adenosines. (a) Experimental design for using a DNA primer in a reverse transcription reaction. Sequencing of the unmodified template shows, in a single-molecule sequencing trace, base incorporation via a reverse transcriptase-mediated cDNA synthesis reaction. (b) Shows sequencing as with (a), but using an RNA template with m⁶A instead of normal adenosines. Incorporation of thymines (T) show significant delay (longer inter-pulse distances). A.U. stands for normalized arbitrary units in fluorescence measurement. (c) Exponential fit of experimentally observed inter-pulse distances (IPDs). (d) Shows the difference between the average IPDs for native As and m⁶As. The average IPD in each case is the reverse of the exponential decay rate. The error bars indicate the range around each average IPD that includes 83% of the observed IPDs (that is, ±½ of standard deviation of the exponential fit). We used an Ansari-Bradley test in Matlab to confirm that the distribution functions were different (P = 0.0043).

Figure 5

Known types of RNA modifications. Known modifications to RNA bases are grouped by RNA type, base and species: (a) archaea; (b) bacteria; (c) eukarya; (d) all species. Data are compiled from the RNA Modification Database [13].