Genome-wide analysis of long noncoding RNAs, 24-nt siRNAs, DNA methylation and H3K27me3 marks in Brassica rapa.

ABSTRACT

Long noncoding RNAs (lncRNAs) are RNA fragments that generally do not code for a protein but are involved in epigenetic gene regulation. In this study, lncRNAs of Brassica rapa were classified into long intergenic noncoding RNAs, natural antisense RNAs, and intronic noncoding RNAs and their expression analyzed in relation to genome-wide 24-nt small interfering RNAs (siRNAs), DNA methylation, and histone H3 lysine 27 trimethylation marks (H3K27me3). More than 65% of the lncRNAs analyzed consisted of one exon, and more than 55% overlapped with inverted repeat regions (IRRs). Overlap of lncRNAs with IRRs or genomic regions encoding for 24-nt siRNAs resulted in increased DNA methylation levels when both were present. LncRNA did not overlap greatly with H3K27me3 marks, but the expression level of intronic noncoding RNAs that did coincide with H3K27me3 marks was higher than without H3K27me3 marks. The Brassica genus comprises important vegetables and oil seed crops grown across the world. B. rapa is a diploid (AA genome) thought to be one of the ancestral species of both B. juncea (AABB genome) and B. napus (AACC) through genome merging (allotetrapolyploidization). Complex genome restructuring and epigenetic alterations are thought to be involved in these allotetrapolyploidization events. Comparison of lncRNAs between B. rapa and B. nigra, B. oleracea, B. juncea, and B. napus showed the highest conservation with B. oleracea. This study presents a comprehensive analysis of the epigenome structure of B. rapa at multi-epigenetic levels (siRNAs, DNA methylation, H3K27me3, and lncRNAs) and identified a suite of candidate lncRNAs that may be epigenetically regulated in the Brassica genus.