| Summary: | <p>Abstract</p> <p>Background</p> <p>The most frequent case of horizontal transfer in plants involves a group I intron in the mitochondrial gene <it>cox1</it>, which has been acquired via some 80 separate plant-to-plant transfer events among 833 diverse angiosperms examined. This homing intron encodes an endonuclease thought to promote the intron's promiscuous behavior. A promising experimental approach to study endonuclease activity and intron transmission involves somatic cell hybridization, which in plants leads to mitochondrial fusion and genome recombination. However, the <it>cox1 </it>intron has not yet been found in the ideal group for plant somatic genetics - the Solanaceae. We therefore undertook an extensive survey of this family to find members with the intron and to learn more about the evolutionary history of this exceptionally mobile genetic element.</p> <p>Results</p> <p>Although 409 of the 426 species of Solanaceae examined lack the <it>cox1 </it>intron, it is uniformly present in three phylogenetically disjunct clades. Despite strong overall incongruence of <it>cox1 </it>intron phylogeny with angiosperm phylogeny, two of these clades possess nearly identical intron sequences and are monophyletic in intron phylogeny. These two clades, and possibly the third also, contain a co-conversion tract (CCT) downstream of the intron that is extended relative to all previously recognized CCTs in angiosperm <it>cox1</it>. Re-examination of all published <it>cox1 </it>genes uncovered additional cases of extended co-conversion and identified a rare case of putative intron loss, accompanied by full retention of the CCT.</p> <p>Conclusions</p> <p>We infer that the <it>cox1 </it>intron was separately and recently acquired by at least three different lineages of Solanaceae. The striking identity of the intron and CCT from two of these lineages suggests that one of these three intron captures may have occurred by a within-family transfer event. This is consistent with previous evidence that horizontal transfer in plants is biased towards phylogenetically local events. The discovery of extended co-conversion suggests that other <it>cox1 </it>conversions may be longer than realized but obscured by the exceptional conservation of plant mitochondrial sequences. Our findings provide further support for the rampant-transfer model of <it>cox1 </it>intron evolution and recommend the Solanaceae as a model system for the experimental analysis of <it>cox1 </it>intron transfer in plants.</p>
|
|---|
