Background: Inferring the evolutionary history of phylogenetically isolated, deep-branching groups of taxa-in particular determining the root-is often extraordinarily difficult because their close relatives are unavailable as suitable outgroups. One of these taxonomic groups is the phylum Parabasalia, which comprises morphologically diverse species of flagellated protists of ecological, medical, and evolutionary significance. Indeed, previous molecular phylogenetic analyses of members of this phylum have yielded conflicting and possibly erroneous inferences. Furthermore, many species of Parabasalia are symbionts in the gut of termites and cockroaches or parasites and therefore formidably difficult to cultivate, rendering available data insufficient. Increasing the numbers of examined taxa and informative characters (e.g., genes) is likely to produce more reliable inferences.
Principal findings: Actin and elongation factor-1α genes were identified newly from 22 species of termite-gut symbionts through careful manipulations and seven cultured species, which covered major lineages of Parabasalia. Their protein sequences were concatenated and analyzed with sequences of previously and newly identified glyceraldehyde-3-phosphate dehydrogenase and the small-subunit rRNA gene. This concatenated dataset provided more robust phylogenetic relationships among major groups of Parabasalia and a more plausible new root position than those previously reported.
Conclusions/significance: We conclude that increasing the number of sampled taxa as well as the addition of new sequences greatly improves the accuracy and robustness of the phylogenetic inference. A morphologically simple cell is likely the ancient form in Parabasalia as opposed to a cell with elaborate flagellar and cytoskeletal structures, which was defined as most basal in previous inferences. Nevertheless, the evolution of Parabasalia is complex owing to several independent multiplication and simplification events in these structures. Therefore, systematics based solely on morphology does not reflect the evolutionary history of parabasalids.