Tobamovirus evolution: gene overlaps, recombination, and taxonomic implications

Tobamoviruses, mostly isolated from solanaceous plants, may representancient virus lineages that have codiverged with their hosts. Recentlycompleted nucleotide sequences of six nonsolanaceous tobamoviruses allowedassessment of the codivergence hypothesis and support a third subgroupwithin tobamoviruses. The genomic sequences of 12 tobamoviruses and thepartial sequences of 11 others have been analyzed. Comparisons of thepredicted protein sequences revealed three clusters of tobamoviruses,corresponding to those infecting solanaceous species (subgroup 1), thoseinfecting cucurbits and legumes (subgroup 2), and those infectingcrucifers. The orchid-infecting odontoglossum ringspot tobamovirus wasassociated with subgroup 1 genomes by its coat and movement proteinsequences, but with the crucifer-pathogenic tobamoviruses by the remainderof its genome, suggesting that it is the progeny of a recombinant. For fourof five genomic regions, subgroup 1 and 3 genomes were equidistant from asubgroup 2 genome chosen for comparison, suggesting uniform rates ofevolution. A phylogenetic tree of plant families based on the tobamovirusesthey harbor was congruent with that based on rubisco sequences but had adifferent root, suggesting that codivergence was tempered by rare events ofviruses of one family colonizing another family. The proposed subgroup 3viruses probably have an origin of virion assembly in the movement proteingene, a large (25-codon) overlap of movement and coat protein open readingframes, and a comparably shorter genome. Codon-position- dependent basecompositions and codon prevalences suggested that the coat protein frame ofthe overlap region was ancestral. Bootstrapped parsimony analysis of thenucleotides in the overlap region and of the sequences translated from the-1 frame (the subgroup 3 movement protein frame) of this region producedtrees inconsistent with those deduced from other regions. The results areconsistent with a model in which a no or short overlap organization wasancestral. Despite encoding of subgroup 2 and 3 movement protein C-terminiby nonhomologous nucleotides, weak similarities between their amino acidsequences suggested convergent sequence evolution.