Abstract
Long intergenic non-coding RNAs (lincRNAs) represent a large fraction of transcribed loci in eukaryotic genomes. Although classified as non-coding, most lincRNAs contain putative open reading frames (pORFs), and it is unclear what hinders translation of cytoplasmic lincRNAs.
Here, we analysed lincRNA sequences from six eukaryotes, including human and fission yeast, and identified signatures of translational repression. In species under strong selective pressure, i.e. yeast and mostly worm, we detected fewer and shorter pORFs in lincRNAs than in intronic and non-transcribed control regions, and the sequence context around their putative start codons appeared suboptimal for translation initiation. Moreover, trinucleotides (“codons”) in lincRNA pORFs tended to be decoded by tRNAs with fewer gene copies than codons in mRNAs or control regions, likely to impede translation elongation.
In order to better understand how in multicellular species varying tRNA and cytoplasmic lincRNA expression levels between cell types affect lincRNA translation regulation, we analysed data from human cell lines. In three out of five cases, we detected signatures of translational repression of cytoplasmic lincRNAs, i.e. a codon usage that corresponded to lower expressed tRNAs than control trinucleotides. Importantly, the degree of codon-tRNA correspondence was in agreement with the ribosome-association strengths of cytoplasmic lincRNAs across cell lines.
Together, the codon usage of lincRNA pORFs is biased, and it appears to function to repress ribosome binding. This mechanism is capable of counteracting lincRNA translation in a cell type-specific manner.
Significance statement Although long intergenic non-coding RNAs (lincRNAs) are classified as non-coding, mechanistic insight into what hinders translation of cytoplasmic lincRNAs is still lacking. We investigated signatures in lincRNA sequences counteracting translation in six eukaryotes and five human cell lines. The detected signatures were more pronounced in species evolving under stronger selective pressures, i.e. worm and yeast, and for lincRNAs with higher cytoplasmic expression. In human, signatures were cell type-specific. Our observations suggest that lincRNA sequences have been shaped to counteract translation. Mechanistically, we provide evidence that a poor correspondence between lincRNA tri-nucleotide and tRNA frequencies, which is less than for coding RNAs and non-coding controls, reduces ribosome-binding to lincRNAs. The identified signatures may assist distinguishing peptide-from real non-coding lincRNAs.
Competing Interest Statement
The authors have declared no competing interest.
