Reevaluation of the cox1 group I intron in Araceae and angiosperms indicates a history dominated by loss rather than horizontal transfer.

The origin and modes of transmission of introns remain matters of much debate. Previous studies of the group I intron in the angiosperm cox1 gene inferred frequent angiosperm-to-angiosperm horizontal transmission of the intron from apparent incongruence between intron phylogenies and angiosperm phylogenies, patchy distribution of the intron among angiosperms, and differences between cox1 exonic coconversion tracts (the first 22 nt downstream of where the intron inserted). We analyzed the cox1 gene in 179 angiosperms, 110 of them containing the intron (intron(+)) and 69 lacking it (intron(-)). Our taxon sampling in Araceae is especially dense to test hypotheses about vertical and horizontal intron transmission put forward by Cho and Palmer (1999. Multiple acquisitions via horizontal transfer of a group I intron in the mitochondrial coxl gene during evolution of the Araceae family. Mol Biol Evol. 16:1155-1165). Maximum likelihood trees of Araceae cox1 introns, and also of all angiosperm cox1 introns, are largely congruent with known phylogenetic relationships in these taxa. The exceptions can be explained by low signal in the intron and long-branch attraction among a few taxa with high mitochondrial substitution rates. Analysis of the 179 coconversion tracts reveals 20 types of tracts (11 of them only found in single species, all involving silent substitutions). The distribution of these tracts on the angiosperm phylogeny shows a common ancestral type, characterizing most intron(+) and some intron(-) angiosperms, and several derivative tract types arising from gradual back mutation of the coconverted nucleotides. Molecular clock dating of small intron(+) and intron(-) sister clades suggests that coconversion tracts have persisted for 70 Myr in Araceae, whose cox1 sequences evolve comparatively slowly. Sequence similarity among the 110 introns ranges from 91% to identical, whereas putative homologs from fungi are highly different, but sampling in fungi is still sparse. Together, these results suggest that the cox1 intron entered angiosperms once, has largely or entirely been transmitted vertically, and has been lost numerous times, with coconversion tract footprints providing unreliable signal of former intron presence.