Thursday, October 04, 2007

Is "Prokaryote" a Useful Term?

Coturnix (Bora Zivkovic) is the Online Community Manager at PLoS-ONE (Public Library of Science). Part of his job is to get people to post comments on the PLoS websites. [New in Science Publishing, etc.]

So when Bora suggested we get involved in a debate on "Is "prokaryotic" an outdated term?" I hopped on over to the PLoS website and read the comments. I discovered that you have to register on PLoS in order to comment so I went ahead and did that and posted a response to the question.

I don't like registering on websites, it's a painful process, especially in this case 'cause you have to answer a lot of questions. It took me about ten minutes to figure out what to do and to convince the program to let me register even though I didn't want to receive email spam from PLoS. I also had to make up a user ID—Larry_Moran, in this case—because, apparently your name isn't good enough. This is not a very open process.


The Three Domain Hypothesis
Anyway, the question is important. If you think the Three Domain Hypothesis is well established, then you believe there are two non-eukaryotic domains (Bacteria, Archaea). Furthermore, the eukaryotes cluster with the Archaea according to this hypothesis. Thus, the word "prokaryote" encompasses a paraphyletic group and becomes useless.

But we wouldn't be having this discussion if the Three Domain Hypothesis is incorrect. In that case, the root of the tree might well be a split between eukaryotes and prokaryotes. The point is that the discussion about usefulness of "prokaryote" is really a debate about the validity of the Three Domain Hypothesis and we shouldn't forget that. It's wrong to assume that your side has won that debate and then start to solidify your apparent victory by defining your opponent's point of view out of existence!


  1. Thus, the word "prokaryote" encompasses a polyphyletic group and becomes useless

    paraphyletic, not polyphyletic

  2. Right, if you accept a single common ancestor and sister-group status for Archaea and eukaryotes, "prokaryotes" would be paraphyletic.

    But so what? There are a lot of named but paraphyletic (and even a few polyphyletic) groups that are actually quite useful, informally, for efficient communication as "organizational grades."
    The classic example is Reptiles (paraphyletic if birds are excluded) for ectothermic amniotes, but there are many more: jawless fishes, bony fishes, lophophorates, sponges, flukes, polychaete worms, gymnosperms, algae.
    These are useful groups from a functional perspective. As long as everybody involved is aware that the name is not being used to identify a clade, it doesn't bother me at all.
    I do not advocate retention of such groups in formal taxonomy, which I agree should reflect phylogeny as closely as possible and should only name monophyletic groups. But for teaching, or informal conversation, as long as there is full disclosure, it's no big deal.
    In the case of prokaryotes, I still think it is a useful term when talking about organizational grades of cells. If you are uncomfortable with naming a (possibly!) paraphyletic group, then just make sure to refer to "prokaryotic cells" instead.

  3. I agree with Sven Dimilo. There are many such paraphyletic groups in primatology, despite the influence of cladists in that field. The group "monkeys" is paraphyletic; Old World monkeys share a clade with apes (and humans) that excludes New World monkeys. Prosimians (strepsirhines + tarsiers) are paraphyletic, as is the pongid group (chimps, gorillas, orangutans).

    Grades reflect shared traits, broad ecological niches, levels of complexity; in other words, types - even if they are sometimes paraphyletic. Some innovations - specialized multicellularity, bipedality, language etc. - are so consequential that they demand the recognition of a new "kind" and place members of the same clade lacking that innovation in a paraphyletic grade partly defined by primitive traits.

    There is nothing wrong in principle with accepting a certain cladistic scheme while using terms like "prokaryote." Grades are real.

    (Note that I am not advocating the recognition of polyphyetic groups, except in a very limited way for comparison only - for example, homeotherms, arboreal mammals, unicellular organisms.)


  4. Sven,

    You have a valid point, but I'd like to argue against it.

    Most, if not all, of the taxonomically incorrect terms (those that are poly- or paraphyletic) were not at one time incorrect in the sense that they were thought to be correct taxonomically. Thus, the modern use of the older terms, while retaining some usefulness, remain as uncorrected terms in common use. That does not smell right to me.

    More importantly, it is not really useful in teaching. Sure, we can use those terms, say, through tenth grade, but then in AP Biology or College biology we have to explain to the students that what they learned was technically wrong, not because we have learned more since they originall were taught a certain set of terms, but because we were simply telling them the wrong thing. You can explain all you want to the befuddled students that we were actually acting in their own interest, but you should then not be surprised if these same individuals are later easily caught up in the "Scientists change what they say every few minutes, so who can you believe? way of thinking that is so common (and counterproductive) these days.

    Finally, these differences in terms are actually great teaching tools because they demonstrate not only how science advances, but how evolution actually works. Simplification (using the para-polyphyletic terms because it is easier) is counterproductive if it obviates fantastic learning opportunities.

  5. Another argument, besides the reality of grades, against cladistic anti-gradist purism: compromised clades due to hybridization, lateral transfer, and symbiogenesis.

    Let's be careful not to overly reify clades. After all, some of today's widely recognized clades will meet the same fate of cladistic invalidation as some of the paraphyletic groups we've discussed.


  6. GL: I see what you mean, and it's well taken. Note that I did advocate full disclosure; if I'm going to use a grade name my students know it's not a monophyletic group. I always go back to the current clado-phylogeny for emphasis.

    But of course Tupaia is right--the current clado-phylogeny is a hypothesis, and, well, scientists do change what they say about a lot of these phylogenies pretty often these days. I am using a textbook (one of the Big Ones) that shows two different phylogenies for amniotes in different chapters and a third in the powerpoint slide corresponding to one of those print figures (they differ only in the placement of turtles, but still). I don't see that it would be helpful to subject middle-schoolers to this kind of flux, nor to insist on using terms like "non-avian dinosaurs" or "non-hominoidean cattarhines" instead of New World monkeys.

    I guess it depends on context, too. Having presented a current phylogeny, I don't feel too bad about referring to "pseudocoelomates" later on in the course (even if nematodes are ecdysozoans and rotifers are lophotrochozoans), if the point I'm trying to make is about hydrostatic skeletons instead of evolutionary relationships.

    And I still talk about prokaryotes too, even though I'm not as ready as Dr. Moran is to jettison the 3-domains system.

  7. Thank you for registering and commenting on PLoS ONE, Larry. Sorry for the complex registration process - but we have to do it. After all, it is not just another blog, but a permanent record of scientific research and the ongoing debate about it (and as attractive to spammers as can be!). Now that you are registered, I hope to see you there often ;-)

  8. Certainly in describing a phylogentic group, the word prokaryote is useless, not only because of the three domain system, but also as some "prokaryotes" have a nucleus (e.g. Brocardia anammoxidans , see Lindsay et al (2001) Cell compartmentalisation in planctomycetes: novel types of structural organisation for the bacterial cell. Arch Microbiol 175 413-429).
    However, when describing (or lecturing about) mixed communities of microbes, saying "Bacteria and Archaea" frequently is a real pain! Its a shame that the term Eubacteria was not adopted so that "bacteria" could be used in a similar way to Prokaryote for describing common properties of the Bacteria & Archaea.

  9. Given what we now know about all the horizontal gene transfers, endosymbioses and genome fusions that have gone on, does it really make sense even in principle at this point to try to root the tree of life? In a sense it seems as though both the three domains and the traditional classification of prokaryotes have been refuted- both are over-simple compared to the actual history which is more net- than tree-like. (What one does about that in terms of nomenclature, I have no idea.)

  10. Heleen,

    Thanks for catching the slip-up. I fixed it on the blog.

  11. oops--just noticed my error above; I meant Old World monkeys.

    I agree that the prudent position at present is to keep the Tree unrooted. Coincidentally, yesterday I ran across a review in Science from May 2006 (abstract here) that I thought made a pretty good case for the three domains based not only on rRNA but also genome comparisons and protein-folding apparatus; IIRC there are correlated differences in membrane phospholipids (some are branched in Archaea) and cell-wall composition that distinguish Archaea from both bacteria and eukaryotes. To me (and I am no expert; I study vertebrate physiology) it seems parsimonious to regard similarities between Archaea and bacteria as homoplastic and similarities between Archaea and eukaryotes as results from horizontal gene transfer, and therefore to retain three domains but refuse to root them. I'd be interested in contrasting perspectives.

  12. ...of course similarities between bacteria and Archaea could also result from later HGT as well as homoplasy.
    My point is that, regardless of probably unknowable origins, the extant Archaea seem to be different enough, and evolving (pretty much) separately for long enough, that reports of the death of the 3-domain hypothesis are exaggerated.

  13. Unless life itself is polyphyletic (I doubt it), "prokaryote" seems to me a useful paraphyletic term. I agree "full diclosure" must be made, that we are not referring to a group of taxa, but to a certain "set" of plesiomorphic traits..."pelycosauria" ,"crossopterygia"...still handy. Maybe its good to keep them in lower case or between commas

  14. Tupaia, most well-accepted clades are not going to become obsolete. There is much more consensus now with cladistics than there was before (when they "aided" their phylogenetic hypotheses with adaptationist speculations) .
    Even without cladistics, the reality of some clades is evident enough, that they have never been seriously challenged. Many aspects of the tree of life are simply a matter of fact.
    Phylogentic systematics is to me one of the most serious and respectable fields of evolutionary biology; I predict we will continue to see progress and agreement (and many bad stuff that now makes it into the great journals will be biting the dust)

  15. a review in Science

    Cool, first author "Chuck" Kurland was responsible for the molecular biology course I could fit into my studies. (He wanted to draw over civil engineers and physicists to help with models, equipment et cetera, so made one among other "specialties" courses.) I see in his CV he is still professor at Uppsala U, and still engaged in outreach. And still very jovial, I'm sure.

    So now you know who may share some of the blame that I display the audacity to comment here on stuff I certainly don't know anything about. :-P

  16. I think biologists could learn about terminology from mathematicians, since math is the queen of utterly precise and rigorous definitions. Mathematicians would say "by an abuse of language..." and then refer to T. Rex as a dinosaur, and not as a "non-avian dinosaur" without the slightest twinge of unease.

  17. I'm pretty sure math has its share of ... unsystematic ... naming. (Pi is a Greek letter, e is a Latin letter, to just pick a well-known example.)

    However, cladistics as such is just applied set theory, after all - and just with finite sets at that, the boring kiddie version.

    There's a reason I preferred career in IT to one in math, even though I was rather good with math in school. Being able to have proofs is one thing, being expected to create them all the time is something entirely different.