mRNA-Mediated intron losses: Evidence from extraordinarily large exons

Multicellular eukaryotes that have high intron density have their introns almost evenly distributed within genes, but unicellular eukaryotes that are generally intron poor have their introns asymmetrically distributed toward the 5' ends of genes. This was explained by homologous recombination of genomic DNA with the cDNA reverse transcribed from the 3' polyadenylated tail of spliced mRNA. This paper is to study whether mRNA-mediated intron losses have ever occurred in multicellular eukaryotes. If intron losses were mRNA-mediated, adjacent introns should be commonly lost together. A direct result is fusion of several previously adjacent exons and producing a large exon. We found that extraordinarily large exons (ELEs) are common not only in unicellular eukaryotes but also in multicellular eukaryotes. The percentage of genes having ELEs is negatively correlated with intron abundance. In addition, the number of lost introns estimated from the relative lengths of ELEs is negatively correlated with the number of extant introns. These results support mRNA-mediated intron losses in all eukaryotes. Moreover, we found that the ELEs of intron-common eukaryotes (with more than 0.5 intron per gene on average) are not only located at 3' ends but also at 5' ends and the middle of genes. This is contrary to what would be expected if the involved cDNAs were reverse transcribed from the 3' polyadenosine ends. A remarkable difference in intron distribution was revealed between intron-rare eukaryotes and intron-common eukaryotes. The intron-rare eukaryotes show very strong 5'-biased intron distribution, whereas the intron-common eukaryotes display even intron distribution or only weak 5'-biased distribution. We suspected that intron losses from 3' end of genes may be limited in intron-rare eukaryotes. The intron losses from intron-common eukaryotes should have other priming mechanism, like self-primed reverse transcription.