Are there really three domains of life?

[Ygggdrasil]

Most biology textbooks state that life can be classified into three domains: bacteria, eukarotes, and archaea. This classification began from early studies looking at the evolutionary relationship between these three groups of organisms that concluded that all archaea are more similar to themselves than to bacteria and eukaryotes, all eukaryotes are more similar to themselves than to bacteria or archaea, and all bacteria are more similar to themselves than to archaea or eukaryotes. These studies led to the phylogenetic tree shown in panel a below, which shows the evolutionary relationships between bacteria, eukaryotes (eukaryota), and the main branches of archaea (highlighted in blue):

(http://www.nature.com/nature/journal/v504/n7479/full/nature12779.html)

Inferring evolutionary relationships from billions of years ago, however, is a tricky business. The studies that led to these conclusions relied on some assumptions – for example that base compositions remain constant across different lineages and that evolutionary rates remain constant across the different DNA sequences being analyzed – that simplified the analysis, but were not necessarily correct. Newer methods that try to address these concerns are beginning to suggest a different evolutionary history that includes only two domains of life (see panel b above).

These studies suggest that all archaea are not more similar to themselves than to eukaryotes (for example, eukaryotes and eocytes are more similar to each other than eocytes to euryarchaeota). Thus, these studies conclude that eukaryotes are best classified within archaea rather than being a separate branch of the evolutionary tree.

In evolutionary terms, the three domain model hypothesized that the last common ancestor of all life first split into two different populations: bacteria and the archaea/eukaryote ancestor. The archaea/eukaryote ancestor then branched off into two populations, the archaea and eukaryotes giving the three domains of life. The newer model instead suggest that archaea and eukaryotes did not evolve in parallel. Rather, archaea evolved first and then the eukaryotes later evolved from an ancient archaeal lineage.

Of course, the art of constructing evolutionary relationships from extant sequencing data is by no means an easy process, so the question of a two or three domains tree of life is by no means settled. Newer methods or better analyses may certainly provide more clarity in the future. While this information has little practical implications for biology (other than better understanding how eukaryotes evolved), it's always interesting to see facts in textbooks being challenged by new data. For a more in depth discussion of the topic, see the following review article from this week's issue of Nature:

Williams et al. 2013. An archaeal origin of eukaryotes supports only two primary domains of life. Nature 504: 231. http://dx.doi.org/10.1038/nature12779[/URL]

 

[Pythagorean]

The narrative I perceived when I was last reading about this was that there were archaea and bacteria, then "one day" an archaea ate a bacteria and, instead of digesting it, turned it into something useful (i.e. endosymbiotic theory).

Is that the general idea? Or would eukaryotes have split off sometime before endosymbiosis occurred?

edit:
(I guess when I say prokaryotes, I mean bacteria)

 

[Ygggdrasil]

The narrative I perceived when I was last reading about this was that there were archaea and bacteria, then "one day" an archaea ate a bacteria and, instead of digesting it, turned it into something useful (i.e. endosymbiotic theory).

Is that the general idea? Or would eukaryotes have split off sometime before endosymbiosis occurred?

edit:
(I guess when I say prokaryotes, I mean bacteria)

Yes, that's an accurate view of what the evidence shows. According to the two domain model, some archaea acquired mitochondria through endosymbiosis and gave rise to the ancestor to all eukaryotes. Because membrane-enclosed organelles are a defining feature of eukaryotes (and these arose thorough endosymbiosis), eukaryotes probably could not have split off from archaea prior to the endosymbiosis occurring.

What's in question with the two-domain vs three-domain model is when the endosymbiotic event leading to the origin of eukaryotes took place. If this event took place very early in the evolution of archaea (before they diversified into different kingdoms), then the three domain model is correct. If eukaryotes evolved much later after diversification of the archaea and descended from a specific branch of the archaeal tree, the two domain model is more correct.

 

[Pythagorean]

So... How are the endosymbiotic organelles constructed?
Is there DNA in the nucleus that codes for them?
Does bacterial genetic promiscuity explain how that would be possible (an endosymbiote getting its DNA into the nucleus)?
Would the nucleus, having a membrane, have originally been a bacteria itself or part of the "mother cell" (presumably an archaea)?

 

[jim mcnamara]

Mitochondria obviously kept DNA and it has plasmid-like structure. As in bacteria. This one example at least lends support to what you are saying. Minus the genetic promiscuity issue.

 

[atyy]

Well, hopefully textbooks are not teaching the 3 domains hypothesis as fact!

 

[Ygggdrasil]

So... How are the endosymbiotic organelles constructed?
Is there DNA in the nucleus that codes for them?
Does bacterial genetic promiscuity explain how that would be possible (an endosymbiote getting its DNA into the nucleus)?
Would the nucleus, having a membrane, have originally been a bacteria itself or part of the "mother cell" (presumably an archaea)?

Despite being derived from bacteria, mitochondria contain orders of magnitude less DNA than most bacteria. The majority of proteins that mitochondria require to function are encoded in the nucleus and transported into the mitochondria after synthesis in the cytoplasm (a small number of mitochondrial proteins are encoded by DNA within the mitochondria and translated by ribosomes inside the mitochondria). The explanation for this situation is that after endosymbiosis, gene transfer occurred between the mitochondria and the cell nucleus. This process appears to be ongoing as mitochondria from different branches of the eukaryotic tree encode different numbers of genes in their mitochondria. For ideas about how this gene transfer occured, see the following review article: http://www.nature.com/nrg/journal/v5/n2/full/nrg1271.html

While we're fairly sure about the origins of mitochondria (endosymbiosis of an α-proteobacterium), how the eukaryotic nucleus evolved is still very much an open question. The follwing news piece from Science discusses some theories (although it is from 2004, so some of the thinking about this issue may have changed since then): http://www.sciencemag.org/content/305/5685/766.full

 

[Pythagorean]

Thanks for the comments and literature; this has always been an interesting subject.

 

[atyy]

Actually, the main difference seems to be the placement of the eukaroytic branch between the 3 domains and eocyte proposals. But even in the 3 domains hypothesis, archaea and eukaryotes are don't seem obviously equal as domains to bacteria - naively, wouldn't the common ancestor to archaea and bacteria be more equivalent to bacteria as a domain (assuming the 3 domains tree)? If so, what is the raesoning for 3 domains in if one assumes the "3 domains" tree, instead of 2 domains?

Also, if I understand correctly, the main advance in the discovery of archaea was that prokaryotes were done away with and split into bacteria and archaea - that seems to still hold?

 

[Ygggdrasil]

Actually, the main difference seems to be the placement of the eukaroytic branch between the 3 domains and eocyte proposals. But even in the 3 domains hypothesis, archaea and eukaryotes are don't seem obviously equal as domains to bacteria - naively, wouldn't the common ancestor to archaea and bacteria be more equivalent to bacteria as a domain (assuming the 3 domains tree)? If so, what is the raesoning for 3 domains in if one assumes the "3 domains" tree, instead of 2 domains?

Also, if I understand correctly, the main advance in the discovery of archaea was that prokaryotes were done away with and split into bacteria and archaea - that seems to still hold?

The three domain tree gives three domains because each of bacteria, archaea, and eukaryota form mutually exclusive monophyletic groups. That is, all species in that group descend from a common ancestor and the group includes all descendents of that common ancestor. The two domain tree does not support eukaryota as a separate domain from archaea because archaea would not form a monophyletic group if you excluded eukarya from the group. Again, this has little practical implications for most biologists but it does change the number of domains we can define because of the rules of taxonomy and systematics.

The discovery of archaea did something similar. Previously, life was classified into five kingdoms (monera, protista, animalia, plantae, and fungi) with all prokaryotes falling into kingdom monera. The discovery of archaea, prokaryotes that were more closely related to eukaryotes than to bacteria, revealed that monera was no longer a monophyletic group. Thus, biologists (led by Woese) reclassified life into (what they thought at the time) three new monophyletic groupings (bacteria, archaea, and eukarya) which they termed domains.

Of course, given the extensive amount of horizontal gene transfer in evolution, it's not even clear that a classical phylogenetic tree best represents some of these early evolutionary events. Eukaroyotes clearly arose by combining genes from archaea and bacteria, but typical phylogenetic trees have no way of representing that evolutionary history.

 

[atyy]

The three domain tree gives three domains because each of bacteria, archaea, and eukaryota form mutually exclusive monophyletic groups. That is, all species in that group descend from a common ancestor and the group includes all descendents of that common ancestor. The two domain tree does not support eukaryota as a separate domain from archaea because archaea would not form a monophyletic group if you excluded eukarya from the group. Again, this has little practical implications for most biologists but it does change the number of domains we can define because of the rules of taxonomy and systematics.

The discovery of archaea did something similar. Previously, life was classified into five kingdoms (monera, protista, animalia, plantae, and fungi) with all prokaryotes falling into kingdom monera. The discovery of archaea, prokaryotes that were more closely related to eukaryotes than to bacteria, revealed that monera was no longer a monophyletic group. Thus, biologists (led by Woese) reclassified life into (what they thought at the time) three new monophyletic groupings (bacteria, archaea, and eukarya) which they termed domains.

Of course, given the extensive amount of horizontal gene transfer in evolution, it's not even clear that a classical phylogenetic tree best represents some of these early evolutionary events. Eukaroyotes clearly arose by combining genes from archaea and bacteria, but typical phylogenetic trees have no way of representing that evolutionary history.

What do the textbooks currently teach? The reason I wonder whether they really teach the 3 domains as fact, rather than hypothesis, is that when I was in high school (1990) , my textbooks did not teach the 5 kingdoms as fact, and mentioned alternatives like the 4 kingdoms. What was taught to me as fact was that the modern aim of a classification was to reflect evolutionary history, but it was also taught that the reconstructions could not be certain. I was taught Margulis's endosymbiont hypothesis, which is now considered strongly supported, and if I understand you correctly, represents the more serious challenge to the definition of either 3 or 2 monophyletic domains. Woese's archaea was not in my textbooks, but my high school teachers did teach it. To be honest, I don't remember exactly which of the above was in my textbooks (but I definitely remember that some had 5 kingdoms and others 4, and that Margulis's book was one of our references), perhaps it was just my excellent teachers that avoided dogma.

Edit: Reading about, it seems that Woese and Fox originally divided prokaryotes into archaea and bacteria in 1977, but Woese's 3 domains proposal is very late - 1990. So what I must have learned about in high school in 1990 was the 1977 version (which I wrongly assumed was the 3 domains proposal, when I later heard about 3 domains containing archaea). So it's really the 1977 proposal that should have made it into textbooks by now, not the 1990 3 domains.

http://tolweb.org/Life_on_Earth/1 does state that the monophyly of archaea is uncertain. So I'm not sure any textbook facts are being challenged.

 

[Pythagorean]

How are clades clustered in the tree of evolution? From Yggg's link is it something to do with how synapomorphies are defined and how much weight you give to each in a comparison? Doesn't that allow for a lot of subjectivity?

edit:

I guess Yggg based the difference between the 2-domain and 3-domain as a matter of time:

What's in question with the two-domain vs three-domain model is when the endosymbiotic event leading to the origin of eukaryotes took place. If this event took place very early in the evolution of archaea (before they diversified into different kingdoms), then the three domain model is correct. If eukaryotes evolved much later after diversification of the archaea and descended from a specific branch of the archaeal tree, the two domain model is more correct.

So I guess you infer closeness of time from the closeness of characteristics?

 

[Ygggdrasil]

How are clades clustered in the tree of evolution? From Yggg's link is it something to do with how synapomorphies are defined and how much weight you give to each in a comparison? Doesn't that allow for a lot of subjectivity?

Most evolutionary relationships are now inferred from comparison of sequence data. In reconstructing the phylogeny, researchers gather highly conserved sequences from a wide variety of organisms (e.g. ribosomal RNA sequences, genes encoding essential proteins like RNA and DNA polymerase that are present in all organisms) then use a model for the rate of mutation of these sequences to compute the most likely series of evolutionary events that led to the set of modern day sequences. Of course, exactly how you model evolution, how you implement the computation, and how you chose the sequences leaves room for much subjectivity in the analysis, so phylogenetics remains much more an art than an exact science. The review article I cited in the OP discusses some of the differing conclusions from various studies. While many studies support the two domain model advanced by the authors, there are still a number of modern studies that find support for the three domain model.

What do the textbooks currently teach? The reason I wonder whether they really teach the 3 domains as fact, rather than hypothesis, is that when I was in high school (1990) , my textbooks did not teach the 5 kingdoms as fact, and mentioned alternatives like the 4 kingdoms. What was taught to me as fact was that the modern aim of a classification was to reflect evolutionary history, but it was also taught that the reconstructions could not be certain. I was taught Margulis's endosymbiont hypothesis, which is now considered strongly supported, and if I understand you correctly, represents the more serious challenge to the definition of either 3 or 2 monophyletic domains. Woese's archaea was not in my textbooks, but my high school teachers did teach it. To be honest, I don't remember exactly which of the above was in my textbooks (but I definitely remember that some had 5 kingdoms and others 4, and that Margulis's book was one of our references), perhaps it was just my excellent teachers that avoided dogma.

Edit: Reading about, it seems that Woese and Fox originally divided prokaryotes into archaea and bacteria in 1977, but Woese's 3 domains proposal is very late - 1990. So what I must have learned about in high school in 1990 was the 1977 version (which I wrongly assumed was the 3 domains proposal, when I later heard about 3 domains containing archaea). So it's really the 1977 proposal that should have made it into textbooks by now, not the 1990 3 domains.

http://tolweb.org/Life_on_Earth/1 does state that the monophyly of archaea is uncertain. So I'm not sure any textbook facts are being challenged.

I always remember being taught about the three domains (I graduated from high school in 2003, so this is well after Woese's proposal would have had time to be integrated into textbooks), with the five kingdom model presented as an earlier version that had been superseded by the three domain model. Indeed, in Albert's Molecular Biology of the Cell (2002 edition, though my copy of the 2008 edition says essentially the same thing), a widely used cell biology textbook, the authors write:

It now appears that the procaryotes comprise two distinct groups that diverged early in the history of life on Earth, either before the ancestors of the eucaryotes diverged as a separate group or at about the same time. The two groups of procaryotes are called the bacteria (or eubacteria) and the archaea (or archaebacteria). The living world therefore has three major divisions or domains: bacteria, archaea, and eucaryotes (Figure 1-21).

(https://www.ncbi.nlm.nih.gov/books/NBK26866/#A40)

The accompanying figure shows the three domain tree. Of course, the textbook (which focuses on eukaryotic cell biology) does not discuss evolution in depth, so more general biology textbooks that discuss evolution in much greater depth might have a more nuanced discussion.

 

[eloheim]

There has been some buzz over the last couple of years about a possible forth domain of life, composed of mega viruses. I don't know how it will all pan out, but it's certainly intriguing!

Giant viruses open Pandora's box (Nature News blurb with links to published research.)

But these viruses, described today in Science1, are more than mere record-breakers — they also hint at unknown parts of the tree of life. Just 7% of their genes match those in existing databases...The researchers are now trying to determine the viruses’ origins by characterizing the unknown genes and the proteins they encode. They have long suspected that giant viruses evolved from cells; if they are right, the ancestors of Pandoraviruses must have been very different from the bacteria, archaea and eukaryotes we have today. “We think that at some point, the dynasty on Earth was much bigger than those three domains,†says Abergel. Some cells gave rise to modern life, and others survived by parasitizing them and evolving into viruses.

Pandoraviruses: Amoeba Viruses with Genomes Up to 2.5 Mb Reaching That of Parasitic Eukaryotes.

Genomics of Megavirus and the elusive fourth domain of Life

I think it's safe to say that the future of such genomic analyses will reveal more than a couple surprises we can't even see coming!

 

[Pythagorean]

Wow, ok, considering viruses as a domain of life made me interested in the origin of viruses, for anyone interested:

http://www.nature.com/scitable/topicpage/the-origins-of-viruses-14398218

From this comes an interesting suggestion that the eukaryote nucleus could have once been a virus:

"Others have argued that precursors of today's NCLDVs led to the emergence of eukaryotic cells. Villarreal and DeFilippis (2000) and Bell (2001) described models explaining this proposal. Perhaps, both groups postulate, the current nucleus in eukaryotic cells arose from an endosymbiotic-like event in which a complex, enveloped DNA virus became a permanent resident of an emerging eukaryotic cell."

 

[eloheim]

Wow, ok, considering viruses as a domain of life made me interested in the origin of viruses, for anyone interested...

Thanks for that. I remember spending a few days on a similar path when I was learning about the mega viruses too. IMHO, the most interesting of the three possibilities given for virus origins (and the one relevant to the thread at hand):

 

[Ygggdrasil]

A new paper in Nature pretty much nails down the two domain hypothesis. See https://www.physicsforums.com/threa...he-evolution-of-complex-cellular-life.812833/

 

[Torbjorn_L]

Spang works @ my alma mater!

But of course the "eocyte" hypothesis was slightly wrong too, eukaryotes sits within TACKs as sisters to Lokiarchaeota, and their closest common ancestor is with Korarchaeota and not Crenarchaeota.

While Petrov et al don't show it in their phylogenetic trees [Fig 1], the PTC of the ribosome on the basis of insertion fingerprints suggests that Archaea/Eukaryote clade within Bacteria instead of the A/E clade being sisters [Fig 3]. ["Evolution of the ribosome at atomic resolution", Petrov et al, PNAS]. Anyway, that is another paper that (weakly) support the Spang et al find.

I'm no expert, but while the ML bootstrap support is 80 (modest, I think) the real hot stuff is with the genes of Lokiarchaeota actins et cetera related to basal actins. IMHO.

 

[Ygggdrasil]

I'm no expert, but while the ML bootstrap support is 80 (modest, I think) the real hot stuff is with the genes of Lokiarchaeota actins et cetera related to basal actins. IMHO.

Yes, I'd agree. I would not be surprised to see the phylogeny change as new related archaea are discovered, but the really cool part of the paper was the sign that these archaea contained genes related to the actin cytoskeleton and membrane-remodeling machinery (like the ESCRT complex) which helps clarify the evolutionary origin of these features. One probably could have guessed that these would have had to been present in the eukaryotic ancestor as they would seem to be necessary for endosymbiosis to occur, but it's nice to see data confirming this.

I'll have to take a look at that Petrov paper. Thanks for pointing it out.