This odd marine worm, Xenoturbella bocki, is in the news right now, and I had to look it up in Pechenik's Biology of the Invertebrates(amzn/b&n/abe/pwll) to remind myself of what it was. Here's the complete entry:
Xenoturbella bocki
This marine worm, first described in 1949 as an acoel flatworm and later claimed as either an early metazoan offshoot or a primitive deuterostome, has recently been affiliated with primitive bivalve molluscs, based upon a study of gamete development (oogenesis) and an analysis of sequence data from both 18S rRNA and mitochondrial genes. Little is known about its reproductive mode, and developmental studies that might help to resolve the phylogenetic issues are just starting to be reported. A second species was described in 1999.
The animals are up to 4 cm long, vermiform (worm-shaped), and covered by locomotory cilia. They have no digestive tract, and indeed no organs at all. Their only conspicuous morphological feature, other than their cilia, is a statocyst for determining orientation. To date, they have been collected only off the coasts of Sweden and Scotland, in sediments at depths of 20 m to 100 m.
That's it. Part of that is now known to be wrong: the data showing an affinity to the molluscs is an artifact, caused by the fact that it somehow eats bivalves, and partly digested clam material contaminated the samples. Otherwise, not much is known; I've found papers describing the presence of oocytes inside the animal, but no one as far as I know has actually observed its development. It's a strange, mysterious blob of a worm.
We now know a little more about it, though. A rather thorough molecular study of the organism has accumulated sufficient sequence information to place it in a phylogenetic tree, and it falls into the deuterostome clade, along with us chordates, echinoderms, and hemichordates. At the same time, it's different enough that the investigators believe it warrants its very own phylum, the Xenoturbellida, bringing the number of extant deuterostome phyla up to four. If only there were a parliament of metazoans, our party would now have an additional vote, although I suspect the Xenoturbellida would always be siding with the flatworms, anyway.
Here's the summary tree from their data:
Bayesian analysis of nuclear data after the addition of asteroid, hemichordate and xenoturbellid data. The new sequences join the branch to the echinoderm, and the cephalochordates now join the chordate branch. This indicates that the previous result is due to systematic error. Xenoturbella is the sister group of the Ambulacraria (echinoderms plus hemichordates).
You may notice that there are two major branches of that tree. The authors also propose two new clades to represent those groups: the Olfactores, consisting of the vertebrates and urochordates, and the Ambulacria (hemichordates and echinoderms).
Another purpose of the paper was to resolve a recent issue, the possibility suggested in some previous work that the cephalochordates were more closely related to the echinoderms than to the other chordates. That was a strange result, and would imply that echinoderms had to have lost a great many characters common to cephalochordates and vertebrates, such as a notochord. This seemed most improbable, and this new paper suggests it was an artifact of the analysis. They threw in more data (more data is always good) and carried out a few manipulations to remove details known to skew results, and got more detailed trees that fit the expectations from morphology much better, and, of course, are stronger and more consistent in themselves.
Bayesian analysis of mitochondrial data with the amino acids M, I, N and K recoded as missing data places the cephalochordates with vertebrates; Xenoturbella is the sister group of Ambulacraria. Support for critical nodes is shown in the order BPP (%), TF (%) and MLBP (%). Long branches leading to the urochordates have been shortened as indicated by a break (/) in the branch.
I do have two reservations about the work on this animal, though.
One is that it is all well and good to have more solid molecular analyses, but I have this little bias…it bothers me to see an animal called a deuterostome when no one has looked at a gastrula. I'd say at this point that the deuterostomy of Xenoturbellida is a hypothesis, one likely to be found correct, but I'd sure like to see some developmental biology. I'd like to see more biology of this organism, period. We've got its place in the tree of history marked, but we still don't know anything about its life history, its physiology, its development—and given its place in the family tree, the embryo and larva could be very informative.
Another problem I have is more with the press reports than in the work itself. For example, look at this press release, Enigmatic worm identified as mankind's long lost relative:
Scientists have discovered one of mankind's closest invertebrate relatives in the shape of a rare 3cm worm that resides in mud at the bottom of a Swedish fjord.
The research, published in today's edition of Nature (21 August), is the first conclusive proof that humans and the 'Xenoturbella' worm, whose Latin name means strange flatworm, derive from a common ancestor, thereby placing Xenoturbella in the same division of the animal kingdom as man.
Ick, ick, ick. Everything is about this animal's relationship to human beings, and it implies so much that is wrong. Look at the trees. This creature is closer to the lines of starfish and acorn worms than to the vertebrates. The appeal of Xenoturbella is that it represents a group that branched off within the deuterostome clade long, long ago. It's the distance of this relative that's interesting, but it's not unique. Fruit flies are also invertebrate relatives of humans, and are even more distant than this strange worm; shouldn't the media be just as excited about the relationship of its readers to sea urchins, flies, clams, bacteria, and trees?
Bourlat SJ, Juliusdottir T, Lowe CJ, Freeman R, Aronowicz J, Kirschner M, Lander ES, Thorndyke M, Nakano H, Kohn AB, Heyland A, Moroz LL, Copley RR, Telford MJ. (2006) Deuterostome phylogeny reveals monophyletic chordates and the new phylum Xenoturbellida. Nature. 444(7115):85-8.
Bourlat SJ, Nielsen C, Lockyer AE, Timothy D, Littlewood J, Telford MJ (2003). Xenoturbella is a deuterostome that eats molluscs. Nature 424: 925-928.
- Log in to post comments
Wow!
NO Digestive tract or organs?
No specialization at all?
MYOB'
.
Does it live in sulfidic sediments? Maybe it has sulfide-oxidizing bacterial symbionts.
I guess if you have no other morphology, it's important to have a statocyst, to know whether you're going forwards or backwards. Does a statocyst look like one of those arrows on packing boxes, that say, "this side up." :-)
In phylogenetic analysis using molecular data, I would disagree that more data is ALWAYS better. It's more important that you get the right gene(s) for your organisms and type of study rather than just sequencing a whole bunch of stuff.
But I definitely agree with your doubts about this paper. It would have been far better for them to produce a paper on the life history and development of this worm and THEN produce a paper on the systematics. And why even pick Bayes...did they say? I think Bayes was a poor choice for this paper when they haven't even examined the life history/development of this worm.
If I remember right from invertebrate zoology, a statocyst is a cell that has a little grain of...stuff...in it that bumps against the cell wall.
That thing EATS CLAMS? When someone finally figures out how it does that, I'd pay good money for a video.
Apparently this Israelsson dude who did his PhD on Xenoturbella still disagrees with the ingested molluscan material angle:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&li…
Interesting. I don't have access to the article right now so I don't know what gene they amplified here.
"the data showing an affinity to the molluscs is an artifact"
PZ! Please help hold the line here..."Data" is plural! Always!!
As for eating clams, not quite. According to the abstract from the original report( http://www.nature.com/nature/journal/v424/n6951/abs/nature01851.html ), the thing "eats" (how with "no organs"???) bivalve embryos and larvae, and that's what contaminated earlier DNA samples.
Bayesian analysis was reasonable because details of nerve net, epidermis, and cilia morphology already suggested an affinity with deuterostomes.
I think it's silly to claim that the auhors should have published a detailed life history before publishing on molecular systematics--they had samples of adults, and nothing else is known!
And I totally agree with PZ that it would be oh-so-cool to know what the embryos and larvae of these things look like...of course it'll be hard to find that out if they in fact lack gonads!!!
That Israelsson paper has some interesting things in it. First, from the introduction,
...that said, here's some controversy about phylogeny (specifically, about the mollusk-type embryos...food or babies?):
http://www.springerlink.com/content/j332nx6l1r5v6211/
http://links.jstor.org/sici?sici=0962-8452%2819990422%29266%3A1421%3C83…
...and here's a poster presenting independent (immonocytological) evidence for deuterostome affinities:
http://www.senckenberg.de/odes/Stach.pdf
...and here's how they eat with no organs:
http://www.springerlink.com/content/jjctccaanypunrrl/
I love Google.
That second cladogram looks odd to me.
Maybe I'm reading it wrong, but it looks like the Urochordates have ended up as far away from the rest of the deuterostomes as the protostomes.
I'm a bit concerned about the fact that the hagfishes Myxine and Eptatretus are grouping with the snakes, Homo is splitting the reptiles, and Lepidosiren the South American Lungfish is off with the ray-fin Protopterus.
Apologies for duplicate links. And the older articles seem now to have been disproved.
The new study was not designed to resolve relationships within vertebrates (which do group as monophyletic, enough for the point of the article)...the troublesome nodes are probably poorly supported, and long-branch attraction probably explains the hagfish thing.
But tuinicates outside of protostomes + other deuterostomes? Weird indeed.
"Bayesian analysis of mitochondrial data with the amino acids M, I, N and K recoded as missing data"
What is the rationale for recoding those particular amino acids as missing?
But tuinicates outside of protostomes + other deuterostomes? Weird indeed.
I don't see a root on that tree so maybe tunicates can't really be said to be the 'outermost' taxon there, and they have a long branch thing going on too. But you're right that they are still in the 'wrong place'; maybe they offer some explanations in the article.
But a more pressing question is: if these animals brood their young, but have no organs, how do they "do it"? ;)
Only when they use mitochondrial DNA, you'll notice; in the previous cladogram, which uses nuclear DNA, the Urochordates group in with the rest of the Chordates. According to both the Bourlat papers cited above, the rapid evolution of the Urochordates produces a lot of long-branch attraction, so they tend to misroot. (I don't know enough about either cladistic theory or these organisms to tell you why that makes nuclear DNA more likely than mtDNA to root them correctly.)
As I understand it, it's not unusual for a given character suite to sort out larger taxa well but smaller taxa poorly. In this case, the authors are only using the mitochondrial tree to back up their claims of cephalochordate/vertebrate and Ambulacraria/Xenoturbellida monophyly.
From the paper:
"We were aware, however, that the mitochondrial genetic code changes in the lineage leading to Ambulacraria and an additional change in the echinoderms13 might have led to compositional bias that could artefactually distance them from Xenoturbella. Indeed, the proportions of amino acids that were affected by the changes of genetic code (methionine and isoleucine, asparagine and lysine (M, I, N, K)) in the echinoderms differ from the average seen in deuterostomes and metazoans."
It looks like you have an older edition of Biology of the Invertebrates. The 5th edition that I have has the following differences (p. 449):
In the first paragraph, the sentence that beings "Little is known..." was moved to the end of the second paragraph, and was replaced with this:
The end of that final sentence was changed from "are just starting to be reported" to "are just now underway".
The asterisk next to "Now" is referencing a footnote citation of the Bourlat, Nielsen (et al.) paper from Nature, 2003.
Oh, and I was wondering if Dr. Pechenik used the cover image for the book just to please cephalopodophiles:
http://ec2.images-amazon.com/images/P/0072348992.01._SCLZZZZZZZ_V107411…
Protopterus is the African lungfish genus.
mtDNA evolves faster than most nuclear genes and is therefore more likely to produce long-branch attraction artefacts.
BTW, one datum, two data... just the Latin word for "given".
Oh, and cladistic theory: more data is always better, except in very few cases when a bit more bad data makes things worse; the solution to the latter problem is to add yet more data. The signal adds up, the noise cancels itself out.
This little critter could end up being very interesting once its genome is fully sequenced. I suspect that the modern representatives of this new phylum are degenerative descendents of a more complex organism (hence its lose of identifiable organs) that could have been very close morphologically to the deuterostome last common ancestor (LCA). Perhaps many of the genes of that LCA are preserved in the genome of Xenoturbella as inactivated pseudogenes. With continued advances in genetic engineering, who knows, someday we may be able to resurect the deuterostome LCA based on this sort of evidence.
Errata:
I need to learn how to write more gooder.
Protopterus is the African lungfish genus.
D'oh. Of course it is! I'm always muddling it with Polypterus (Bichirs)! I really shouldn't post when I'm tired after work. ;)
I though long-branch attraction would be the explanation for some of these things. I'll have to read the paper and find out the author's rational for the unexpected internal stuff, but it probably all works out ok.
CCP:
Data, I think, is largely used as a collective, at least in American usage.
Hm. As a nonscientist, I can only say, "Frikkin' cool! A new phylum!"
CCP is right, despite the common mis-usage. "Data" is always always always plural. I'm not one of these "language is immutable" types and welcome new usages or words that contribute new shades of meaning or dispense with arbitray detritus of language history. But "data...is" is just wrong. Phyla/phylum, strata/stratum... no one has trouble with those but suddenly some people think you're a grammar snob for insisting on singular datum.
"media" I give up on, but "data" becomes a collective over my dead, pedantic body.
Cool animal, though, huh?
"And why even pick Bayes...did they say? I think Bayes was a poor choice for this paper when they haven't even examined the life history/development of this worm", followed by "Bayesian analysis was reasonable because details of nerve net, epidermis, and cilia morphology already suggested an affinity with deuterostomes"
Bayesian phylogenetic analysis is one way to allow researchers to analyze large data sets using model-based methods, which are more robust (but not immune) to various types of systematic error. Barring some arguments regarding the use of flat priors (and whether the priors really are flat and, if not, what that means), Bayesian analysis is as appropriate for analysis of sequence data as maximum likelihood analysis, and it's faster. It really doesn't have anything to do with prior observations that suggested Xenoturbella is a deuterostome. That information could theoretically have been used to modify the topology prior for the analysis (as you might expect in a typical Bayesian analysis), but it wasn't -- the topology and parameter priors were flat, which is the default in the software they used (MrBayes) (and FWIW, I don't think you can use unequal topology priors in the current version of MrBayes). The current standard in Bayesian phylogenetic analysis is to minimize the influence of priors as much as possible and let the likelihood function govern the outcome.
And yes, the position of urochordates on the mitochondrial protein-coding gene tree is weird. I thought it was just drawn with an odd root, but I've checked the supplementary information about their analyses, and they included various non-bilaterian outgroups (cnidarians, sponges, ctenophores and others) and got a strange rooting even with Bayesian analysis. There are forms of systematic bias (e.g., base/amino acid composition bias) that can mislead just about any phylogenetic method now in general use, including Bayesian methods.
Very cool worms, though!
I don't know why you object to this, the resemblance to Salvador T. Cordova is inescapable.
Two things come to mind upon reading the summary of this critters eating style. The prey, bivalve molluscs and the eversion of the gastrodermis reminds me a lot of starfish feeding.
I've read Bourlat et al. (2003) where they reported that Xenoturbella is not a bivalve mollusc but that the studied samples (by Noren&Jondelius in 1997)were contamined with bivalves embryos eaten by X. In this paper, they do mention that X. has a gut ! "we dissected away as much of the GUT as possible from our specimen". BTW, there's a new paper on X. by Perseke&al (2007) The mitochondrial DNA of Xenoturbella bocki: genomic architecture and phylogenetic analysis. Theory Biosci. 126(1): 35-42.