Computational Characterization of ncRNA Fragments in Various Tissues of the Brassica rapa Plant

Abstract

Recently, a novel type of non-coding RNA (ncRNA), known as ncRNA fragments or ncRFs, has been characterised in various organisms, including plants. The biogenesis mechanism, function and abundance of ncRFs stemming from various ncRNAs are poorly understood, especially in plants. In this work, we have computationally analysed the composition of ncRNAs and the fragments that derive from them in various tissues of Brassica rapa plants, including leaves, meristem tissue, pollen, unfertilized and fertilized ova, embryo and endosperm. Detailed analysis of transfer RNA (tRNA) fragments (tRFs), ribosomal RNA (rRNA) fragments (rRFs), small nucleolar RNA (snoRNA) fragments (snoRFs) and small nuclear RNA (snRNA) fragments (snRFs) showed a predominance of tRFs, with the 26 nucleotides (nt) fraction being the largest. Mapping ncRF reads to full-length mature ncRNAs showed a strong bias for one or both termini. tRFs mapped predominantly to the 5' end, whereas snRFs mapped to the 3' end, suggesting that there may be specific biogenesis and retention mechanisms. In the case of tRFs, specific isoacceptors were enriched, including tRNA^Gly(UCC) and tRF^Asp(GUC). The analysis showed that the processing of 26-nt tRF5' occurred by cleavage at the last unpaired nucleotide of the loop between the D arm and the anticodon arm. Further support for the functionality of ncRFs comes from the analysis of binding between ncRFs and their potential targets. A higher average percentage of binding at the first half of fragments was observed, with the highest percentage being at 2-6 nt. To summarise, our analysis showed that ncRFs in B. rapa are abundantly produced in a tissue-specific manner, with bias toward a terminus, the bias toward the size of generated fragments and the bias toward the targeting of specific biological processes.

Keywords: Brassica rapa; apical meristem; embryo; endosperm; leaves; ncRFs; ncRNA; ncRNA fragments; non-coding RNA; pollen; pollinated ovules; rRF; snRF; snoRF; tRF; unpollinated ovules.

Figure 1

Characterisation of read number. The total number of reads (in thousands) in different tissues (A); and the percentage of reads mapped to genomic regions encoding ncRNAs (B), organellar genomes (C), the nuclear genome (excluding ncRNA regions) (D), genic versus intergenic regions (F), exonic versus intronic and unmapped reads (E). Percentages above the bars indicate the percentage of reads mapped to a specific region in a given tissue, as compared to all reads. The data shows the average calculated from two biological replicates (with standard deviation (SD)). Asterisks show significant differences in a particular tissue type as compared to other tissues (one—p < 0.05 and two—p < 0.01); the diamond symbol shows significant difference at p < 0.1.

Figure 2

The size of non-coding RNA (ncRNA) reads mapped to known ncRNA sequences in the Brassica rapa genome. The Y-axis shows the number of reads of a certain size. The mean (with standard deviation, SD) size of reads is also shown. “UnpollOvules” = unpollinated ovules; “PollOvules” = pollinated ovules.

Figure 3

The relative read frequency of various ncRNAs in various tissues of Brassica rapa. The relative read frequency was calculated by dividing the read number of specific ncRNAs by the total read number. In each figure, the data points not connected by the same letters are significantly different from each other. tRNA = transfer RNA; rRNA = ribosomal RNA; snoRNA = small nucleolar RNA; snRNA = small nuclear RNA; miRNA = micro RNA; ApicMeristem = apical mersitem; UnpollOvules = unpollinated ovules; PollOvules = pollinated ovules.

Figure 4

Size distribution of reads mapping to different ncRNAs in different tissues. The y-axis shows the size of reads. The bottom and top of the rectangle indicate the first and third quartiles, respectively. The lower and upper ends of the vertical line extending outside the rectangle represent the minimum and maximum, respectively. The thick horizontal line inside the rectangle is the median, and the circle beyond the rectangle displays an outlier.

Figure 5

Read distribution across the whole length of the corresponding ncRNA. The entire length of each ncRNA was taken to be 100% and then divided into bins (each of 10%), with 0–10% representing the first 10% of the ncRNA length (starting from the 5′ end). The relative frequency of reads in a specific bin was calculated by dividing the frequency of reads in a specific bin by all reads mapping to the entire length of the ncRNA.

Figure 6

The frequency of the occurrence of reads mapping to various tRNAs (A); and tRFs (B) in different tissues of Brassica rapa. The y-axis shows the frequency (as a percentage of the total) of the occurrence of reads mapping to a specific tRNA, relative to reads mapping to all tRNAs.

Figure 7

A correlation between transfer RNA (tRNA) and tRNA fragments (tRFs) in leaves (A), apical meristems (B), pollen (C), unpollinated ovules (D), pollinated ovules (E), embryos (F) and endosperms (G). The y-axis shows the tRF to tRNA ratio (tRF5′ in all cases, unless specifically labelled). The tRF to tRNA ratio was calculated by dividing the number of tRF reads in the indicated type by the total number of tRNA reads. A summary of all tissues is shown in panel H.

Figure 8

The structure of tRNAs to which the most abundant tRF reads match. (A,C,D,F,G) The secondary structure of tRNAs, with a thick red line indicating tRF and the arrow indicating the potential cleavage site; (B,E,H) the number of reads (y-axis) mapping to a specific tRNA, and their relative size (x-axis).

Figure 9

The percentage of cleavage and translational inhibition of targets by ncRFs. (A–D) data for tRF, rRF, snRF and snoRF, respectively; (E,F) summarise data for cleavage and inhibition, respectively. “ApiMer” = apical meristem; “Unpol_Ov” = unpollinated ovule; “Pol_Ov” = pollinated ovules.