A paper appeared very recently in PLoS One by Philippe et al. on the phylogenetic position of Acoela in the animal evolutionary tree (freely available at the link). I promised last week that I'd comment on the paper when I'd read it, and that would be now.
Acoela are a smallish group (only a few hundred described species) of marine "worms", superficially similar to flatworms (Platyhelminthes) in appearance. Acoela lack a proper gut - the mouth in the diagram above (from Answers.com) opens into a pharynx leading to a shapeless mass of digestive cells, where individual cells take up food particles and digest them via phagocytosis. In the past they were included in Platyhelminthes, and still appear as such in textbooks (which are always out of date), the entire complex being regarded as the basalmost of all bilaterian phyla. However, while Acoela resemble flatworms in features such as absence of a coelom or through-gut (the features previously regarded as primitive for bilaterians), recent molecular investigations have positioned them well away from the Platyhelminthes. Meanwhile, the Platyhelminthes cluster within the Protostomia, implying that their supposed primitive features instead represent derivations from more complex organisms. I am not aware of any attempts so far to explain exactly why and how Platyhelminthes came to dispense seemingly integral features like an anus, but they seem to have done exactly that.
The Acoela are a different story, though. Molecular analyses have placed Acoela at the very base of the bilaterian tree, below the divergence of deuterostomes and protostomes. With this topography, the primitive features of Acoela might indeed be primitive, making this a potentially significant taxon in understanding how the triploblastic Bilateria arose from their diploblastic, radial ancestors (probably similar to modern Cnidaria). Another small marine worm group*, the Nemertodermatida, shares a number of features with Acoela and the two are usually united as Acoelomorpha. However, the monophyly of Acoelomorpha is not certain (Ruiz-Trillo et al., 2002), and could do with further investigation. The paper I'm dealing with today only deals directly with Acoela.
*If you're wondering if "small marine worm group" means "small group of marine worms" or "group of small marine worms", both are equally applicable.
The supposed position of Acoelomorpha as basalmost bilaterians is particularly interesting because Acoelomorpha are not unlike in appearance to a cnidarian planula larva (shown above in a photo from the site of Prof. Fiorenza Accordi). As such, they give credence to the idea that Bilateria may have originated from the development of sexual maturity in a planuloid larva (Baguñà & Riutort, 2004) [As an aside, the Wikipedia page for "planula" previously stated that planulae are incapable of feeding - this only appears to be true for Medusozoa, as Anthozoa have feeding planulae.]
Enter Philippe et al., who used a positively huge molecular data set derived from 68 protein-coding genes to test the position of Acoela, using the exemplar Convoluta pulchra. Because Acoela show a rapid evolutionary speed, making long-branch attraction potentially a significant problem, Philippe et al. used an analytical model that is supposedly more resistant to long-branch attraction. Their final tree agreed with previous studies that Acoela were not related to Platyhelminthes, but disagreed about their position outside the deuterostome + protostome clade. Instead, Convoluta appeared as the sister group to Deuterostomia, a previously unsuspected position.
Under an alternative model, Convoluta appeared in its more traditional position with Platyhelminthes, but this is almost certainly due to long-branch attraction. Platyhelminthes had the longest branches in the analysis (other than Convoluta itself), and their removal caused Convoluta to madly leap away and latch itself onto the chordate Oikopleura, possessor of the next-longest branch. Also, Philippe et al. were able to identify the presence in Convoluta of the gene for guanidinoacetate N-methyltransferase. This gene is found in all animals other than protostomes, but has been lost in the latter. Strangely enough, when the analysis was run without including non-bilaterians, Convoluta became the sister in Deuterostomia of the worm-like Xenoturbella, which is the sister of the Ambulacraria (hemichordates and echinoderms). Xenoturbella resembles Acoela in lacking a through-gut and in the ultrastructure of the epidermal ciliary rootlets, so this is not outside the bounds of possibility.
Nevertheless, I remain somewhat skeptical of the overall result. Bootstrap support for the result was low (only 34%), but a position for Convoluta outside the deuterostome + protostome clade (Eubilateria) got even less support (only 7%). In fact, bootstrap support was pretty low overall within the Deuterostomia (in contrast, Protostomia [including Chaetognatha as basalmost branch] had a bootstrap result of 99%). I feel compelled to chant the usual mantra of all studies - "we need more data!" (though the authors of a study that did, after all, compare 11,959 positions might reply with "for the love of - how much bloody data do you want?"). One particular comment in the paper got my attention:
In more conceptual terms, the position of acoels out of the Platyhelminthes should warn us against the naive view that considers some features as ‘lost’, ‘absent’, or ‘reduced’ in clades (e.g. acoels) than might never have had them in the first place.
However, if Acoela are indeed related to Deuterostomia, then I would feel that they are almost certainly secondarily reduced. It is much easier to regard characters shared between eubilaterians, such as fully formed brain ganglia and the through-gut (Baguñà & Riutort, 2004), as lost in Acoelomorpha rather than gained independently in deuterostomes and protostomes (or even protostomes, chordates and ambulacrarians, if Acoelomorpha are together with Xenoturbella[which also lost these features] the sister to Ambulacraria).
I was going to mention Hox genes here too. Baguñà & Riutort (2004) had a nice tree to support the basal position of acoelomorphs that showed Acoelomorpha with only one central Hox gene and one posterior Hox gene, while the basal number for Eubilateria was four central and at least two posterior Hox genes. However, I then noticed that the author list for the current paper included the same two authors, Baguñà and Riutort. Now, the ideal rule in science is to "believe the data, not the reporter", but I couldn't help wondering if they now knew something that I didn't (quite possible, after all). A quick search doesn't reveal much change for Acoela - Nemertodermatida do have two central Hox genes, but as the second cannot be readily attached to any of the missing eubilaterian genes, it may represent an independent duplication (Jiménez-Guri et al., 2006). However, an apparently not-yet-published study available online by Fritzsch et al. found that Xenoturbella also had only two central Hox genes (though sequence analysis of said genes united them with other deuterostomes rather than acoelomorphs). As a result, if acoelomorphs are closer to deuterostomes than protostomes, I feel that the position in the restricted analysis as sister to Xenoturbella seems more likely than the position as sister to all Deuterostomia in the total analysis. At least that only requires one loss of the characters mentioned.