Molecular phylogeny of the subgenus Drosophila (Diptera, Drosophilidae) with an emphasis on Neotropical species and groups: a nuclear versus mitochondrial gene approach

The genus Drosophila has played an essential role in many biological studies during the last 100 years but much controversy and many incompletely addressed issues still remain to be elucidated regarding the phylogeny of this genus. Because information on the Neotropical species contained in the subgenus Drosophila is particularly incomplete, with this taxonomic group being underrepresented in many studies, we designed a study to answer some evolutionary questions related to these species. We subjected at least 41 Drosophilidae taxa to a phylogenetic analysis using a 516-base pair (bp) fragment of the alpha-methyldopa (Amd) nuclear gene and a 672 bp fragment of the mitochondrial cytochrome oxidase subunit II (COII) gene both individually and in combination. We found that the subgenus Drosophila is paraphyletic and subdivided into two main clusters: the first containing species traditionally placed in the virilis-repleta radiation and the second assembling species of the immigrans-Hirtodrosophila radiation. Inside the first of these clusters we could detect the monophyly of both the flavopilosa (the sister-clade of the annulimana group) and the mesophragmatica (closely related to the repleta group) species groups. Concerning the immigrans-Hirtodrosophila lineage, Zaprionus, Liodrosophila, Samoaia, and Hirtodrosophila were the early offshoots, followed by the immigrans, quinaria, testacea, and funebris species groups. The tripunctata radiation appears to be a derived clade, composed of a paraphyletic tripunctata group, intimately interposed with members of the cardini, guarani, and guaramunu species groups. Overall, the COII gene yielded a poor phylogenetic performance when compared to the Amd gene, the evolutionary hypothesis of which agreed with the total evidence tree. This phenomenon can be explained by the fast saturation of transitional substitutions in COII, due to strong biases in both base composition and substitution patterns, as also by its great among-site rate variation heterogeneity.