More Recent Comments

Wednesday, November 22, 2006

The Three Domain Hypothesis (part 2)

Jan Sapp sets the tone by outlining the history of bacterial classification and phylogenetic analysis. We’re mostly concerned with the fourth era—the one that begins in the 1990's with the publication of the first bacterial genomes.
By the late 1990's, just when the three-domain proposal and the outlines of a “universal phylogenetic tree” were becoming well established, the microbial order based on rRNA was challenged by data from complete genome analysis of bacteria. Phylogenies based on genes other than those for rRNA often indicated different genealogies, and indeed a somewhat chaotic order. The new genomic data also indicated that archaebacteria and bacteria had many genes in common: perhaps they were not that different after all.
Sapp then goes on to discuss the attack on the Three Domain Hypothesis by Ernst Mayr in an oft-quoted PNAS paper (Mayr, 1998). Mayr’s objections have more to do with classification and taxonomy than with any real dispute over the validity of the molecular data. It’s about the fact that Mayr doesn’t like cladistics. He doesn’t want molecular phylogenies to trump visible phenotypes and “common sense” (Mayr’s, of course). Mayr argues that archaebacteria and bacteria both look like bacteria so they should be lumped together in a single prokaryotic empire.

I’m not interested in that debate. If the gene trees say that archaebacteria form a separate domain then that’s good enough for me no matter how much they resemble other prokaryotes. Woese (1998) has published an adequate reply to Mayr.

The real arguments are based on conflicting gene trees and the increasingly obvious similarity between bacteria and archaebacteria at the molecular level. How do we resolve the conflicts between the ribosomal RNA trees and examples of equally well-supported trees from proteins? The first thing that comes to mind is that some of the gene phylogenies are just wrong. They are artifacts of some sort and don’t really represent the history of the genes. Most of the debate on this topic concerns the validity of the SSU trees since they are based on nucleotide sequences. It’s well-known that ribosomal RNA trees are prone to long branch attraction artifacts to a greater extend than trees based on amino acid sequences. It’s also well-known that there are some famous mistakes in rRNA trees.

For the time being, let’s assume that all genes trees are accurate representations of the gene history, bearing in mind that the opponents of the Three Domain Hypothesis are not prepared to concede that point.

Conflicting gene trees then have to be artifacts of a different sort. Some of them will accurately represent the evolution of the species while others will not. The ones that don’t follow the phylogeny of the species will deviate because the genes have a different history. Either they have been transferred singly from one species to another or they have been transferred en masse by some sort of fusion event. Sapp discusses both these possibilities.

Lateral gene transfer (LTG)—also called horizontal gene transfer (HGT)—is the latest fad in microbial evolution. You can explain away all the conflicting gene phylogenies by invoking interspecies transfer. But here’s the problem: which genes were transferred and which ones represent the “true” species phylogeny? Several papers in the book address this problem and we’ll cover them in separate postings.

Keep in mind that LGT can get you out of a messy situation but there’s a price to pay. If you envisage a time when cells were frequently swapping lots of genes to form a “net” of life, then that, in and of itself, is enough to refute the standard version of the Three Domain Hypothesis. What you’re left with is a hypothesis about the phylogeny of “some” genes and a different phylogeny for others. This gets us into playground fights about “my gene is better than your gene.” Supporters of the Three Domain Hypothesis are willing to go there in order to save the hypothesis. Do their arguments hold up?

The other way of explaining the conflict is to invoke whole genome fusions followed by selective loss of half the genes. There are several models to explain the origin of eukaryotic cells by fusion of a primitive archaebacterium with a primitive bacterium. Such an event would account for the data, which shows that most eukaryotic genes are more closely related to bacteria but some are closer to archaebacteria. There are other interesting models, for example one model postulates fusion of a primitive eukaryotic cell with a primitive bacterial cell to form the first archaebacterium! This also accounts for the data but it pretty much wipes out one of the three domains!

Most people take these fusion models seriously. If one of the fusion models is correct, then the original Three Domain Hypothesis is refuted. (One of the complications is the transfer of genes from mitochondria to the eukaryotic nucleus. We’re not talking about those genes. Those ones are relatively easy to recognize.)

Jan Sapp closes his introduction with a summary of the problems that will be addressed in the rest of the book.
... with the development of genomics, the hitherto unappreciated ubiquity of LGT was postulated to explain many gene histories other than those for rRNA. The species concept was again considered to be inapplicable to bacteria, not because of the absence of genetic recombination, as long thought, but because there seemed to be so little barrier to it. Doubts about the inability to construct bacterial genealogies arose anew because of the scrambling of the genetic record from LGT. While debates continue over which (if any) provide the most reliable phylogenetic guide, so too do debates over the origin of the eukaryotic cell nucleus and over the inheritance of acquired bacterial genomes.


Microbobial Phylogeny and Evolution: Concepts and Controversies Jan Sapp, ed., Oxford University Press, Oxford UK (2005)
Jan Sapp The Bacterium’s Place in Nature
Norman Pace The Large-Scale Structure of the Tree of Life.
Woflgang Ludwig and Karl-Heinz Schleifer The Molecular Phylogeny of Bacteria Based on Conserved Genes.
Carl Woese Evolving Biological Organization.
W. Ford Doolittle If the Tree of Life Fell, Would it Make a Sound?.
William Martin Woe Is the Tree of Life.
Radhey Gupta Molecular Sequences and the Early History of Life.
C. G. Kurland Paradigm Lost.


Mayr, E. (1998) Two Empires or Three? Proc. Natl. Acad. Sci. USA 95:9720-0823.

Woese, C. R. (1998) Default taxonomy: Ernst Mayr’s view of the microbial world. Proc. Natl. Adad. Sci. USA 95:11043-11046.


6 comments :

Anonymous said...

Such promiscuous little beasties, exchanging nucleotides willy nilly like that. No self-control, I tell ya. It's a little like trying to draw yer family tree in Arkansas...

(Okay. Yes, family trees with branches that grow back together, I guess it's not really the same thing as species trees and gene trees which don't reconcile because they're not sketching the same history. Bad metaphor just to get the easy line. Mea culpe.)

Anyway, thanks for this. I'm one of those who'd heard there was some ferment in the area, didn't really know what the current thinking covered. Nice to see some possibilities laid out.

John S. Wilkins said...

Keep 'em coming. I note that some have said that LGT doesn't affect the core genomes of these lineages. So I'm keen to see more. More I say! [Even if I am on the wrong team this week.]

Jonathan Badger said...

If you envisage a time when cells were frequently swapping lots of genes to form a “net” of life, then that, in and of itself, is enough to refute the standard version of the Three Domain Hypothesis [...] If one of the fusion models is correct, then the original Three Domain Hypothesis is refuted

Well, if you want to say, "we know more now than Woese did in 1990; the evolutionary history of life is likely to have been more complicated than he thought" then few people would disagree with you. In much the same sense, we now know that gene regulation is more complicated than Monod and Jacob thought in 1961, but the operator/repressor model isn't "refuted" by that.

It doesn't make any sense to return to thinking of Archaea as bacteria because their similarity to eukaroytes is no longer merely a untested hypothesis of phylogenetic trees; one of the main reasons why people work on Archaea today is the stunningly eukaryotic nature of their information processing systems -- for which we now have solid biochemical and genetic evidence. This evidence won't disappear no matter what advances in phylogenetics occur.

Larry Moran said...

jonathan badger says,
one of the main reasons why people work on Archaea today is the stunningly eukaryotic nature of their information processing systems -- for which we now have solid biochemical and genetic evidence

Maybe people work on archaebacteria because some of their genes are derived from primitive eukaryotes via an ancient fusion between a eukaryote and a bacterium. Did you ever think of that? :-)

For every "stunning" relationship between archaebacteria and eukaryotes, there's an equally "stunning" relationship between archaebacteria and other bacteria, and an equally "stunning" relationship between eukaryotes and bacteria.

Do you think that's a problem?

How does the Three Domain Hypothesis account for that data?

Jonathan Badger said...

Maybe people work on archaebacteria because some of their genes are derived from primitive eukaryotes via an ancient fusion between a eukaryote and a bacterium. Did you ever think of that? :-)

Yes, I've heard of that idea. It does seem to be rather lacking on the plausibility scale (it's a bit like the suggestion that fungi evolved from animals and not vice versa), but, yes, it could be *possible*. Still, I don't see how this would threaten the three domains any more than endosymbiotic events between bacteria and eukaryotes do.

For every "stunning" relationship between archaebacteria and eukaryotes, there's an equally "stunning" relationship between archaebacteria and other bacteria, and an equally "stunning" relationship between eukaryotes and bacteria.

Do you think that's a problem?


No, because it isn't true that the other shared relationships are particularly "stunning". The bacterial/eukaryotic and bacterial/archaeal similarities are for the most part in rather dull metabolic and structural genes.

Why are those genes dull? Because evolution (like reproduction) is ultimately a matter of information flow. In the end metabolism is just a way of making the ATP needed for copying of information. So it's not particularly surprising that the black boxes involved in making the ATP can be shuffled around -- even across domains.

Now, a relative of E.coli that used TATA-based transcription -- now *that* would be the discovery of the decade and *would* be a serious problem for the three domains.

Anonymous said...

Hey, I liked you blog a lot, you got some cool stuff here. Do you ever use anything other than blogging...I just started switching! I switched to US Free Ads and Squidoo because I think they are easier instead of blogging. Here's some of my US Free as examples:

Peel Away Ads
Beats 365
000 Relationships
All Types Of Wedding Speeches
Bring Back A Lost Love

Well come back and check them out sometime, keep up the good work also.

Gm