Preferential loss and gain of introns in 3′ portions of genes suggests a reverse-transcription mechanism of intron insertion

Abstract

In an attempt to gain insight into the dynamics of intron evolution in eukaryotic protein-coding genes, the distributions of old introns, that are conserved between distant phylogenetic lineages, and new, lineage-specific introns along the gene length, were examined. A significant excess of old introns in 5′-regions of genes was detected. New introns, when analyzed in bulk, showed a nearly flat distribution from the 5′- to the 3′-end. However, analysis of new intron distributions in individual genomes revealed notable lineage-specific features. While in intron-poor genomes, particularly yeast Schizosaccharomyces pombe (Sp), the 5′-portions of genes contain a significantly greater number of new introns than the 3′-portions, the intron-rich genomes of humans and Arabidopsis show the opposite trend. These observations seem to be compatible with the view that introns are both lost and inserted in 3′-terminal portions of genes more often than in 5′-portions. Overrepresentation of 3′-terminal sequences among cDNAs that mediate intron loss appears to be the most likely explanation for the apparent preferential loss of introns in the distal parts of genes. Preferential insertion of introns in the 3′-portions suggests that introns might be inserted via a reverse-transcription-mediated pathway similar to that implicated in intron loss. This mechanism could involve duplication of a portion of the coding region during reverse transcription followed by homologous recombination and subsequent rapid sequence divergence in the copy that becomes a new intron.