Mapping 3 ' transcript ends in the bank vole (Clethrionomys glareolus) mitochondrial genome with RNA-Seq
Abstrakt
Poly(A)-enriched RNA-Seq reads collected for two wild-caught bank voles (Clethrionomys glareolus) were mapped to the complete mitochondrial genome of that species. Transcript polyadenylation was detected as unmapped adenine residues at the ends of the mapped reads.
Where the tRNA punctuation model applied, there was the expected polyadenylation, except for the nad5 transcript, whose polyadenylated 3' end is at an intergenic sequence/cytochrome b boundary. As in human, two pairs of bank vole genes, nad4l/nad4 and atp8/atp6, are expressed from bicistronic transcripts.
TAA stop codons of four bank vole protein-coding genes (nad1, atp6, cox3 and nad4) are incompletely encoded in the DNA and are completed by polyadenylation. This is three genes (nad2, nad3 and cob) less than in human.
The bank vole nad2 gene encodes a full stop codon (TAA in one vole and TAG in the other), which is followed by a 2 bp UTR and the gene conforms to the tRNA punctuation model. In contrast, the annotations of the reference mouse and some other rodent mt genomes in GenBank include complete TAG stop codons in both nad1 and nad2, which overlap downstream trnI and trnW, respectively.
Thus the RNA-Seq data of bank voles provides a model for stop codons of mt-encoded genes in mammals comparable to humans, but at odds with some of the interpretation based purely on genomic data in mouse and other rodents. This work demonstrates how RNA-Seq data were useful to recover mtDNA transcriptome data in a non-model rodent and to shed more light on mammalian mtDNA transcriptome and post-transcriptional modification.
Even though gene content and organisation of mtDNA are strongly conserved among mammals, annotations that neglect the transcriptome may be prone to errors in relation to the stop codons