Like a lot of evolutionary biologists not studying the root of the tree of life, I assumed the three domain hypothesis was fairly well supported by the research community. This model posits that the tree of life can be broken up into three taxa at its most basal level: eukaryotes, bacteria, and archaea. It's the hypothesized evolutionary relationships of these taxa that caused researchers to break the tree of life into three domains.
It turns out the three domain model isn't as supported as many of us assume. Larry Moran put me in my place, pointing out arguments against the three domain model. Not being familiar with this area of research and not having enough time to research it thoroughly, I asked if Dr. Moran would write up a review for those of still clinging to the three domain model. This was before he had a blog. Now he's got a blog, and he's started his review. Let's see if he can convince us 3Dists that we're wrong.
- Log in to post comments
For the most part objections to the three domains arise from a "Canadian School" of microbiology comprised of Doolittle, Golding, Moran, and Cavalier-Smith (who used to be in Canada) and people who were students or postdocs of these people. With the possible exception of Doolittle these are not people who were converted away from the 3-domains -- they were people who never accepted them in the first place, because they were supporters of phylogenies based on genes like HSP70 which don't support the three domains.
What genomics has shown us is that information processing genes for the most part strongly support the three domains, while many genes involved in metabolism don't. Which makes perfect sense; picking up a few genes through HGT to degrade a new sugar seems pretty easy; changing from sigma-factor based transcription as in bacteria to the TATA-based system such as in eukaryotes and Archaea would be considerbly harder. If someone isolated a relative of E. coli that used archaeal transcription I'd agree that the three domains had been falsified, but so far all the "triumphs" of the Canadian school have been bacteria with some archaeal metabolic genes and vice versa.
It is also worth stressing that while like all theories, criticism of the 3 domains is certainly welcome, it is important to realize that critics are a small minority -- and I say that not just as a former student of Gary Olsen (a member of "Woese's Army" to use Margulis' amusing phrase). The fact that even after Doolittle's 1999 Science paper (the first one in years to give serious credibility to the possibility that the 3-domains were wrong), Woese has received the National Medal of Science, the Crafoord Prize, and most recently, has been inducted into the Royal Society of London, indicates that most people in the field aren't convinced that the 3-domains are in any serious danger.
Again, it is possible that this will change in the future, but as with critics of string theory, it is important to distinguish the loudness of the critics from their small numbers.
Thanks for that clarifying comment, JB. The shared TATA-based transcription system of eukaryotes & archaeans does seem a critical point.
To follow up on a point in the original Fri. phylogeny post, even though "prokaryotes" is paraphyletic, there is nothing wrong with the adjective "prokaryotic" to describe the shared cell type of bacteria and Archaea. (This is a very useful thing to be able to do when teaching Intro Bio.)
Yeah, it seems half the debate about the 3-domains isn't really about phylogeny but about terminology...
What about the differences in membrane composition? Isn't that an additional basis for the three-domain model? Or does he think that archaeal membrane differences are comparable to gram-positive and gram-negative differences in eubacteria?
Moran cites the book that Sapp put together, which is perhaps most up to date, but the article that I'm most familiar with is Woese's 2000 article in PNAS. As I don't have a copy of Sapp's book though, I hope that he lists the lines of evidence presented therein that counters the 3 domain model - it'd be interesting to see.
I agree that this is good additional support for the monophyly of the Archaea, but if that were the only evidence I wouldn't be convinced. A common argument by opponents of the three domains is to argue "Bacterial group X also has membranes of a composition not seen by other groups; why don't you consider them the fourth domain then?".
Although historically the distinct nature of the Archaea was first determined by 16S rRNA and secondarily confirmed by membrane composition, it is really the later discovery of the eukaryotic nature of the archaeal transcription system that really clinched the three domains (and which can't be dismissed by technical arguments about long branch attraction and rooting methods).
Or does he think that archaeal membrane differences are comparable to gram-positive and gram-negative differences in eubacteria?
In a long review article from 2002, Cavalier-Smith reclassifies Archaea as a phylum, Archaebacteria, in a new subkingdom, Unibacteria, based on the presence of a single membrane. In the same year the archaeon Ignicoccus was found to have a double membrane...
In addition, a lot of the arguments in that review suggest that many archaeal-specific features are just adaptations to thermophily. At the time most of the available archaeal genome data was from thermophiles. However, we now have genome sequence for a range of archaea that grow between 0-100 °C and the evidence for features specific to archaea is mounting. They really are more like Eukarya than Bacteria in many ways - JB mentioned the example of transcription and there are other cases too.
Comparative genomics is the key to resolving this, not old-fashioned morphology or microbiology.
There is also the fact that eukaryotes, eubacteria and archaea all use different enzymes to make Gln-tRNA(Gln) used in protein synthesis. Archaea and most bacteria use an indirect pathway in which Glu is attached to tRNA(Gln) which is then amidated by a tRNA-dependent amidotransferase (AdT). Archaea use GatDE whereas bacteria utilize GatCAB. Eukaryotes use a glutaminyl-tRNA synthetase (GlnRS)in their cytoplasm as do some bacteria like E. coli. The bacterial GlnRS is related to the eukaryotic GlxRS family. Chloroplasts utilize the indirect route with GatCAB. Mitocondria the jury is still out (the cytoplasmic GlnRS and tRNA(Gln) is imported into the mitochondria but the GatB homolog, Pet112 in yeast, has been shown to be important for mitocondrial function). I am not sure why Moran keeps bringing up the aminoacyl-tRNA synthetases (AARSs). The evidence is that in the last common ancestral community not all twenty AARSs were present. Some were later to evolve (like GlnRS, AsnRS, etc). What I think Moran is really providing evidence against is the idea of a last common ancestor and for the idea of a last common community.