Convergent intron gains in hymenopteran elongation factor-1α

The eukaryotic translation elongation factor-1α gene (eEF1A) has been used extensively in higher level phylogenetics of insects and other groups, despite being present in two or more copies in several taxa. Orthology assessment has relied heavily on the position of introns, but the basic assumption of low rates of intron loss and absence of convergent intron gains has not been tested thoroughly. Here, we study the evolution of eEF1A based on a broad sample of taxa in the insect order Hymenoptera. The gene is universally present in two copies - F1 and F2 - both of which apparently originated before the emergence of the order. An elevated ratio of non-synonymous versus synonymous substitutions and differences in rates of amino acid replacements between the copies suggest that they evolve independently, and phylogenetic methods clearly cluster the copies separately. The F2 copy appears to be ancient; it is orthologous with the copy known as F1 in Diptera, and is likely present in most insect orders. The hymenopteran F1 copy, which may or may not be unique to this order, apparently originated through retroposition and was originally intron free. During the evolution of the Hymenoptera, it has successively accumulated introns, at least three of which have appeared at the same position as introns in the F2 copy or in eEF1A copies in other insects. The sites of convergent intron gain are characterized by highly conserved nucleotides that strongly resemble specific intron-associated sequence motifs, so-called proto-splice sites. The significant rate of convergent intron gain renders intron-exon structure unreliable as an indicator of orthology in eEF1A, and probably also in other protein-coding genes.

Highlights► The eEF1A gene is present in two copies throughout the insect order Hymenoptera. ► The F1 copy was originally intron free and later successively accumulated introns. ► At least three introns appeared at the same position as introns in other copies, mostly outside Hymenoptera. ► All convergent intron gains are associated with highly conserved sequence motives, the proto-splice sites. ► Because of a high rate of parallel intron gain, intron-exon structure is unreliable as an indicator of orthology in eEF1A.