Is RNA editing implicated in group II intron survival in the angiosperm mitochondrial genome?

Introns may be considered as optional because they are removed from mRNA molecules, but introns are fairly preserved for unknown reasons. Previously, the mitochondrial rps3 gene of sugar beet ( Beta vulgaris L., Caryophyllales) was shown to represent a unique example of an intron loss. We have determined the distribution of the rps3 intron in 19 Caryophyllalean species. The intron was absent from the Amaranthaceae and the Achatocarpaceae. In the Caryophyllaceae, Dianthus japonicus rps3 was pseudogenized, but the intronic sequence was retained. Intact intron-bearing rps3 copies were cloned from Portulaca grandiflora and Myrtillocactus geometrizans , members of the sister clade of the Amaranthaceae-Achatocarpaceae-Caryophyllaceae clade. Most of the C-to-U RNA-editing sites in P. grandiflora and M. geometrizans rps3 transcripts were homologous in the two species, as well as in the sugar beet rps3, which, unlike the other 12 rps3 transcripts, lacks editing in the exonic regions around the intron. Provided that the loss of editing preceded the loss of rps3 intron, it appears conceivable that a requirement for editing could have prevented the loss of group II introns retained in angiosperm mitochondrial genomes. This interpretation is an alternative to the conventional one that views the loss of editing as a mere trace of RNA-mediated gene conversion.