Trouble in the Fourth Domain?

In March I wrote about two studies that raised the tantalizing possibility that the tree of life, which till now has appeared to have three main branches, turns out to have a fourth.

Some of the evidence for the fourth branch (or “domain,” as taxonomists would call it) came from a newly discovered and very strange group of viruses. They’re known as giant viruses, because they’re about a hundred times bigger than typical viruses and can have over a thousand genes. If there was indeed a fourth domain , it meant that giant viruses were part of one of the oldest lineages on Earth. By studying them we might learn about the earliest stages in life’s evolution.

Since then, there have been a couple developments that merit a follow-up. In April, Didier Raoult of Mediterranean University in Marseille and his colleagues published a new study on another species of giant virus. Their previous studies on the fourth domain involved giant viruses that were first discovered in the water in air conditioners, infecting amoebae called Acanthamoeba. But now scientists are finding giant viruses all over the world, in lots of different single-celled hosts. One of the newest of these discoveries is a giant virus that infects an ocean-dwelling amoeba called Cafeteria roenbergensis. which lives inside amoebae in the ocean.

Raoult and his colleagues took a close look at the Cafeteria virus’s genes. A number of its genes don’t exist inside other known giant viruses. The ancestors of Cafeteria may have picked up them from hosts, or from unknown viruses. (This shuttling of genes from species to species is called horizontal gene transfer.)

But Raoult and his colleagues found some genes shared by the Cafeteria virus and by other giant viruses. They compared some of these shared genes to versions found in other forms of life, like bacteria and eukaryotes (we are eukaryotes, as are plants, fungi, and amoebae). Raoult and his colleagues found that the most compelling evolutionary tree joining these genes together had a four-branch structure. They concluded that the Cafeteria virus “unabiguously” points to the existence of a fourth domain.

But in June a new study came out that raises some serious doubts about the fourth domain.

The new study is the work of Tom Williams, Martin Embley, and Eva Heinz of the University of Newcastle. They felt that Raoult and his colleagues might have been tricked by the slippery nature of evolution. One of the big challenges in drawing evolutionary trees based on DNA is that similarities can be deceiving. Just because two species have stretches of DNA that look alike doesn’t necessarily mean that they inherited that DNA from a recent ancestor. The DNA may have independently evolved into a similar state in each lineage.

This is no big secret. Everybody in the business of reconstructing evolutionary history from DNA knows they have to contend with this kind of mirage, called homoplasy. Scientists can reduce homoplasy’s confusion by steering clear of genes that are prone to homoplasy. They can look at particular kinds of elements of DNA that are particularly unlikely to suffer from homoplasy. They can also take into account experiments on living organisms that show how likely different kinds of mutations are. Knowing those probabilities can help scientists figure out how likely an evolutionary tree is to be accurate.

Embley and his colleagues performed some new tests on the DNA that was used in the fourth-domain studies. They concluded that Raoult and his colleagues had used an evolutionary model that did a bad job of avoiding homoplasy. Embley and his colleagues then carried out an analysis of their own, using other models that took into account some details of biology that Raoult and his colleagues didn’t think were important. There are twenty amino acids that proteins can be built from, for example. But that doesn’t mean that during evolution a pariticular amino acid can switch to any of the other nineteen. Some switches are simply impossible. So Embley and his colleagues built these constraints into their model.

When Embley and his colleagues redrew the tree of life, the support for a fourth domain “effectively disappeared,” they write. They could not reject the possibility that the supposedly ancient genes in giant viruses are not ancient at all. Instead, the viruses picked the genes up more recently from their amoebae hosts. Once inside the viruses, these genes quickly evolved so that they ended up looking different from their original forms.

I got in touch with Jonathan Eisen of the University of California at Davis to get his comment on the new paper. Eisen, as I wrote in March, was an author on a study on microbial genes scooped up from the ocean; he and his co-authors suggested that the genes might be pointing to a fourth domain, although they were a lot more tentative in their conclusions than Raoult and his colleagues. Eisen thinks Embley’s probably right. “Paper looks pretty sound,” he wrote in an email to me.

For now, Eisen’s undecided on where giant viruses fit into the tree of life. They could have branched off early, he says, “but I think it is equally plausible that even the big viruses have stolen their cellular-organism-like genes from hosts of some kind.”

One way to cut down on the uncertainty would be to fill in more branches on the tree of life. It’s easy to forget that for all the millions of species scientists have discovered, there are millions–maybe tens of millions–more that are waiting to be found. Right now, scientists are forced reconstruct the tree of life by comparing species that are separated by hundreds of millions or billions of years of evolution. The more species scientists add to the tree of life, the closer those comparisons will become. And there’s no telling what sort of strange branches the tree will turn out to have.

“There very well may be some weird stuff out there,” Eisen said.

PS: I also contacted Raoult to comment on Embley’s paper and have yet to hear back. I’ll add his responses when I do.

[Update: revised post to clarify that the host is named Cafeteria. ]

8 thoughts on “Trouble in the Fourth Domain?

  1. Two things:

    1. The first instance of the word ‘homoplasy’ has a typo. (If it weren’t the first, I wouldn’t bring it up.)

    2. How is “homoplasy” pronounced, please?

    [CZ: Thanks–fixed the typo. Prononciation is hoe-moe-play-zee. You can listen here.]

  2. Nicely done, Carl, and it’s wonderful you’re doing a follow-up semi-refutation — or more like a tap on the brakes. It seems like a pullback from “Tantalizing possibility” to something closer to “Interesting if true.” This is one of the lovely things about having a blog: You’d have a hard time selling this to anyplace that got as wide a readership. Now you can just do it.

    Your post did disappoint me in one way, however: Having read that the first giant virus was found only in the water in air conditioners (WTF?), I was really hoping the new one, Cafeteria roenbergensis, would be found exclusively in that slightly scary water trapped between cafeteria trays after washing. I was betting on it when I saw the name. Alas, no.

    Science has a way of generating expectations … and then disappointing them.

    Oh — did I mention the writing? I particularly admire the passage on homoplasy, the most slippery part of this thing: clear as a bell and phrased in terms of human endeavor.

  3. Seeing as Cafeteria roenbergensis is a unicellular protist, I presume that you’re referring to a virus from this species. And David, Cafeteria got its name from its ability to feed on a wide range of foods.

    [CZ: Thanks, Chris. I’ve fixed the post.]

  4. I note I am still reserving judgement on the origin of these viruses. They are very strange relative to “normal” viruses (if there is a such thing) in that they are (1) very large (2) have very large genomes (3) encode some pretty standard “housekeeping” genes and (4) even apparently have their own viruses.

    I note, in our phylogenetic analysis related to finding unusual sequences in environmental samples, we were/are a bit worried about phylogenetic artifacts. We did use some of the methods Embley used in their paper and found that the sequences we found to be novel still appeared to be phylogenetically novel. However, issues like homoplasy are hard to rule out when one has very long branches in a tree – and this is one of the reasons we took a toned down approach to our claims.

  5. I tend to think that what they got in their fourth domain tree is no more than a artifact caused by long branch attraction, their homoplasy due to the high rate of evolution of laterally acquired genes in the giant virus. I’ve never been too akin to these premature strong claims.

  6. I agree with you Sergio on the fact that the fourth domain is nothing more than a artifact.”Just because two species have stretches of DNA that look alike doesn’t necessarily mean that they inherited that DNA from a recent ancestor”…this means that somethings are happening that we might get tricked by and not realize it because they look alike. I honestly think it’s smart for the scientists to be forced to rebuild the tree of life. Maybe then the comparisons will be more clearer and the results will be easier to see. The scientists are doing everything they can to see where every virus that comes from and how.

Leave a Reply

Your email address will not be published. Required fields are marked *