A Blog by Ed Yong

Tree or ring: the origin of complex cells

Tree_ringThe natural world is full of great partnerships. Bacteria give animals the guts to digest all manner of otherwise inedible foods. Algae allow corals to harness the power of the sun and construct mighty reefs.  Ants cooperate to become mighty superorganisms. But the greatest partnership of all is far more ancient. It’s so old that we can only infer that it took place by looking for signals of history, embedded into the genomes of modern species. The details of how and when it happened are still the source of fierce debate but this was undoubtedly the most important merger in the history of life on Earth: a partnership between two simple cells that would underlie the rise of every living animal, plant, fungus and alga.

All complex life belongs to a single group called the eukaryotes, whose members, from humans to amoebas, share a common ancestry. Their cells are distinguished by having several internal compartments, including the nucleus, which shelters their precious DNA, and the mitochondria, which provide them with power.

This internal organisation sets the eukaryotes apart from the two other domains of life: the prolific bacteria; and the archaea, masters of extreme environments. These two groups are very different in their biochemistry but they are look superficially similar. Both are comprised of solitary cells that lack mitochondria or any other internal compartments; their DNA is unfettered by a nucleus. A chasm of organisation separates these simple cells from the complex eukaryotes, and the crossing of this chasm, several billion years ago, is one of the most important events in the planet’s history.

An event so deep in time was always going to be difficult to piece together. Those ancestral cells are hardly going to leave fossils behind, but there are clues to their ancient prehistory in the genomes of living species. Some genes are so important that they’re shared between all living things – neither bacterium nor human could do without them. These core genes have diverged over the course of evolution, but they’re similar enough to reflect a shared history. By comparing them across modern species, biologists can look at which are more closely related than others.

In the 1970s, the great biologist Carl Woese used one such gene to construct a grand tree of life, where eukaryotes and archaea are sister lineages, both descended from bacterial ancestors (left image above). This is the traditional view of the origin of eukaryotes, but it was based on just one gene. As others were analysed, a more confusing picture emerged. While many eukaryotic genes are indeed closely related to those of archaea, as predicted by Woese’s tree, others have more similar counterparts in bacteria.

These two classes also tend to have different roles. The archaea-like genes tend to be “informational” genes, which are involved in DNA: decoding it, using it to produce proteins, and making new copies of it. The bacteria-like genes tend to be “operational” – they’re involved in dogsbody jobs like making amino acids, fat molecules and so on.

Now, in a new study, James Cotton and James McInerney have found that the archaea-like genes also seem to be more important than the bacteria-like ones, even though they’re less common. The duo looked at every one of the 6,700 gene in the baker’s yeast Saccharomyces cerevisiae. They found that around 2,000 are most closely related to bacterial genes and just 500 or so are most closely related to archaeal ones.

But not all genes are equal. Some are so indispensible that if they don’t work, the yeast dies. And these vital genes are more than twice as likely to have archaeal counterparts as bacterial ones. There are other indicators of importance: the archaea-like genes are also about twice as active as the bacteria-like ones; they’re more likely to interact with other genes; and they’re present in fewer copies. So in the yeast genome, the archaea-like genes are a small group but an elite one.

Cotton and McInerney think that their results support a model for the origin of eukaryotes that’s very different to Woese’s tree. In this scenario – the so-called “ring of life” – eukaryotes arose from a fusion of the other two domains of life. In a fateful encounter, an archaeon swallowed a bacterium, setting up an alliance that would allow both of them to escape the restraints of their simple structures.

The bacterium was eventually domesticated, becoming the mitochondria of today. It transferred many of its genes into the host genome, producing the chimeric mash-up of archaeal and bacterial DNA that we see in modern eukaryotes. Meanwhile, it retained some of its own genes and indeed, mitochondria still have a small genome of their own.

Cotton and McInerney’s results certainly fit with this idea. The genes of the archaeal host have been interacting with each other for longer, well before any bacterial genes were added into the mix. Once that happened, some archaeal genes were displaced, but the essential few remained inviolate. Even after some 2 billion years of evolution, they’re still doing the same fundamental job. The bacterial additions had to be integrated around this core network. This explains why the archaea-like genes of yeast have a central importance out of all proportion to their small number.

That is not to say that the addition of the bacterium wasn’t a pivotal event. Nick Lane has written about this subject extensively in his beautiful book, Life Ascending, and cautions against over-interpreting the new study. “The impression might be created that the host cell was somehow ‘in charge’ all along and the mitochondria had a relatively minor, certainly less important, involvement in the evolution of the eukaryotic cell,” he says. “That would be totally wrong.”

“The host cell was utterly transformed by the mitochondria,” he says. “The fact that a few archaeal genes seem to be more important than bacterial genes should not distract from the fact that the eukaryotic cell is not an archaeon. It has transformed utterly and has thousands of new genes (even in yeast), none of which would have been possible without the mitochondria.”

Bill Martin, who also supports the ring of life model, thinks that it’s important that bacteria have “made a greater quantitative contribution to yeast (and eukaryote) genomes than archaea”. To him, this runs counter to the classical tree of life model, where archaea are the sisters to eukaryotes. It can only be explained by a “symbiotic origin of eukaryotes”, where archaea and bacteria both contributed to the origin of these complex cells.

Tom Cavalier-Smith, who still champions the tree of life model, sees things differently. To him, the archaea-like genes of yeast were simply those present in the last common ancestor of eukaryotes and archaea; they reflect the fact that the two groups are sisters. According to this view, the first eukaryotes had already evolved many of their complex features before they swallowed the bacterium that would eventually become mitochondria. The bacteria-like genes in yeast and other modern eukaryotes either come from these domesticated bacteria, or they’re slow-changing remnants of extremely ancient bacterial ancestors.

Cavalier-Smith also says that the new study doesn’t account for the fact that genes evolve at very different rates. Some of the most important genes that shaped the evolution of the early eukaryotes have changed so much that they can’t be ascribed to either archaea or bacteria. Those genes tell an important story, but they’re largely ignored by Cotton and McInerney.

And if there’s one thing that everyone agrees on, it’s that yeast genes are only part of the picture. For a start, yeast can survive without mitochondria. In eukaryotes that can’t, such as animals, Lane thinks that knocking out the bacterial genes would be far more costly. Martin agrees – he thinks that the “importance” of the archaeal hand-me-downs might vary from one species to another. “Time and more analyses will tell,” he says.

This is not a debate that’s going to be settled any time soon. Indeed, Woese emailed me simply to say that he didn’t want to get involved. Perhaps this is inevitable. The origin of eukaryotes – whether through the branching of a tree or the fusion of a ring – was a critical event that took place billions of years ago. Its singular importance makes it both endlessly fascinating and perhaps endlessly difficult to resolve.

Reference: PNAS http://dx.doi.org/10.1073/pnas.1000265107
If the citation link isn’t working, read why here

Twitter.jpg Facebook.jpg Feed.jpg Book.jpg

15 thoughts on “Tree or ring: the origin of complex cells

  1. It seems, at least from my reading of your first paragraph, that you are suggesting that the endosymbiotic origin of the mitochondria is fiercely debated, that an organism closely related to the modern alpha-proteobacteria gave rise to the mitochondria is no longer debated at all. As your article goes on to describe quite well, what is currently debated has a lot to do with the ancestral host cell that presumably lies at the base of the eukaryotes.

    What I do appreciate is that you have provided some sampling of the diversity of opinions on this subject. At a conference last year on the Tree of Life in Halifax, NS, Canada several of the authors mentioned (McInnerney and Martin) gave some excellent talks summarizing some of their work in this area. As well as Jim Lake, who you provided a link to his Ring of Life paper, Martin Embley has also written some recent papers, all quite excellent, concerning the Eocyte Hypothesis which ties into this. I think the evidence is mounting towards eukaryotes branching from within the Archea.

    Kudos on a great article Ed. It isn’t often that something in the blogosphere appears in an area that I work in.

  2. Ah yes, I wasn’t casting aspersions on the endosymbiotic theory in the first paragraph. T’was quite a difficult intro to write – it needed enough of a hook for a beginner while still staying true to the actual debate. I’m hoping people new to the field won’t be confused, but I might tweak it later. Thanks for the feedback.

    This is probably one of the hardest pieces I’ve written. The topic is vast, it’s very technical and I’m trying to explain this to readers who may not understand the terms “eukaryote” or “archaea”, let alone the relationships between them. Hopefully, it’s done the job reasonably.

  3. It was quite well done, I was very pleasantly surprised. This area of molecular evolution tends not to get much press in the science blogosphere so it was refreshing to have something on the topic.

  4. I think many of those quoted would be disappointed if any of the positions espoused above turned out to be right. Life is at its best when it defies expectations.

    Congratulations on an excellent exposition.

  5. Very interesting stuff. But I’m left wondering: what about the nucleus!? This is a big, big difference between Eukaryotes and both archaea and bacteria. Reading up on this really quickly on Wikipedia, it appears that some families of single-celled Eukaryotes don’t have mitochondria at all, which would seem to indicate that mitochondria arose after the nucleus, which would seem to indicate that they weren’t involved in the divergence of Eukaryotes from the other forms of life (though obviously the absorption of mitochondria had dramatic effects on the later radiation of Eukaryotes).

  6. An enormous question and one that’s almost LESS clear than the origin of eukaryotes. Firstly, those mitochondria-less eukaryotes: they’re not really relevant to the origin of the nucleus but they’re a key part to the duelling eukaryote-origin debate. They’ve been put forward as evidence that the ancestral eukaryotic cell was well on the way to being complex before it swallowed the bacterium that would eventually become the mitochondria (supporting the tree of life model). However, it’s since been shown that these species have lost their mitochondria over time or changed them into something else – they still have a bacteria-archaea mish-mash in their genomes. It seems that all known eukaryotes either have or had mitochondria, which fits with the ring of life model where the bacterium-swallowing event was the key moment that led to the genesis of eukaryotes. The nucleus is more complicated. I’ll get to that later…

  7. What does this “ring” represent?
    Basis of all lineage is a tree, in the case of conjunktion of
    different cells into some merger, the tree has to be
    enhanced with some conjunktion, which may lead to a
    “woven” structure.
    I do not see any reason in that ring, exept maybe repersenting:
    “We do not know”


  8. Nice post Ed! This research field, while extremely interesting, is really hard to capture well, but you’ve done a great job!

    I would like to provide a bit more evidence for the ‘nephew’-model that regards archae and eukaryotes as closely related though. Recently it was discovered that some Planctomycetes have membrane coat proteins and are capable endocytosis. These findings make the case that some bacteria evolved very ‘complex’ or eukaryote-like features before the endysymbiosis with an alpha-proteobacterium happened, really convincing. I’ve written two blogposts on these findings here and here.

    These findings are also not wholly uncontroversial, as nothing in this field really is (as you experiences firsthand). It does make it all the more exciting for us standing at the sidelines of course!

  9. @Jason: Ed already touched on this, but this is something that the lab I am a part of has as a major research focus, mostly working on evolutionarily interesting species and groups of protists from the point of view of the evolution of mitochondria. While there are many ‘amitochondriate’ species out there, as Ed says, genomic data know shows that they still have many genes of mitochondrial origin in their nuclear genomes. If you recall mitochondrial genomes are reduced, with most of their genes having been lost or transfered to the host nucleus over time. So in things without mitochondria, to still have many of those genes, is strong evidence that they once had mitochondria but have secondarily lost them as a consequence of shifts in lifestyle from aerobic to anaerobic organisms.

    In fact, ‘classical’ mitochondria are just one type of organelle in a broader group of related ones that we call Mitochondria-Related or Mitochondria-Like Organelles (MRO/MLO). This includes anaerobic mitochondria that use a molecule other than oxygen as the terminal electron acceptor as well as Hydrogenosomes and Mitosomes which are highly reduced forms. Hydrogenosomes use an alternative method of energy generation from the electron transport chain and produce hydrogen gas as a by-product and typically lack a genome altogether. Mitosomes are even more highly reduced and the only conserved and uniform function so far seems to be the production of iron-sulfur clusters.

    Loss of mitochondria and alteration to MLO’s other than typical mitochondria seems to have happened frequently and independently across the tree of eukaryote species but so far everywhere we look we do find evidence that that all eukaryotes once had mitochondria of some sort at the very least. If there were any species of true ‘amitochondriate’ eukaryotes that diversified prior to the acquisition of the alpha-proteobacterial endosymbiont those lineages have died out. More likely I think the acquisition and retention of the endosymbiont was the defining moment that really made ‘eukaryotes’ as a group and diversification happened after that.

  10. “Loss of mitochondria and alteration to MLO’s other than typical mitochondria seems to have happened frequently and independently across the tree of eukaryote species”
    Indeed a few months ago an anaerobic eumetazoan (a loriciferan) was discovered in the Mediterranean: http://www.biomedcentral.com/1741-7007/8/30
    So, even animals (albeit very little ones) can do without mithocondria
    Anyway, the origin of complex cells story is so fascinating. Endosymbiosis between single-celled organisms leading to the “invention” of the nucleus, mithocondria, plastids etc. seems so improbable (I think it’s one of the favourite arguments of the creationists, as an example of the presumed “irreducible complexity”), yet today we can see in action examples of very strange symbiogeneses between two or more organisms which give rise to completely different organisms. Think of lichens (especially those, like the wonderful Lobaria pulmonaria, which are the union of THREE kinds of organisms), or the truly astonishing (and, I think, quite relevant to Ed’s story) Mixotricha paradoxa, which has FOUR bacterial symbionts, some of which function as cilia, and other as mithocondria

  11. Funny, this exact topic briefly surfaced while chatting with supervisor today 😉

    I second Dan’s comment about how pleasantly surprising your post is for this topic! The typical analysis goes “OMG RING OF LIFE = SOOOO COOL” and kinda ends there. You explored more theories of eukaryogenesis than even Koonin in his recent review on the subject. Then again, Koonin seems to have a thing against Tom. Still, not a valid reason to leave out the most sensible and natural explanation of origin of eukaryotes.

    However, as my supervisor pointed out, you don’t get much press for reporting that something evolved the normal way. Rings and ancient extinct chimaeras from hell tend to be a bit more eye-catching…

    Too many people get carried away with molecular phylogenies and forget they’re just ONE component of life in general. Trees do not speak Truth or anything like that. They should be treated like measurements, and understood that many, MANY things can skew the hell out of them. And time doesn’t change some of those inherent biases and artefacts — eg., a very recent massively-multi-gene tree of life screwed up the eukaryotic phylogeny as badly as the Woese tree, and in a very similar way. Because they used maximum parsimony (cladists, lol). Parsimony likes to put microsporidia (fungi) and diplomonads and such as basal eukaryotes. We like to think likelihood and Bayesian methods are superior, but who knows what kind of consistent artefacts we’ll later learn they have…!

    So how do we reconstruct the tree then, if phylogenies are only suggestions? Well, one way would be to actually consider as much data as possible, from many, many fields. I’d have more faith in an analysis spanning multiple disciplines than some giant tree with 5000 concatenated genes or whatever (which invariably coincides with shitty taxon sampling too, grrr). Tom Cavalier-Smith is amazing at that, but sadly he appears to be the lone voice. As such, he pretty much has to push extreme hypotheses just to get people to look up from their molecular trees and pay attention.

    In order to resolve this issue, we must look at data and theories from cell biology, biochemistry, paleontology, ecology, theoretical biology, etc. Molecular biology alone will not suffice.

    Oh, and the Three Domain dogma (including the “ring of life”) seriously needs to be let go for a while. Not thrown out, just let go. There appears to be vicious opposition to any alternative to eubacterial holophyly. “Trees don’t show it”. Well, yeah, trees also showed amitochondriate early eukaryotes in the 1990’s. The matter is not actually settled. Science would progress so much more smoothly if people paid attention to the actual support of accepted ideas. Evolution, of course, is solid. Eubacterial monophyly, perhaps not so much.

    One of these days I’ll finally write up my own post on the subject. After I read one more of Tom’s megapapers… (hard work)

    @Jason You might be interested in reading about the Archezoa Hypothesis, and it’s eventual demise 😉

    @Walter RE anoxic loriciferan — I’d be careful with the conclusion that they lack mitochondria, unless you mean the canonical type (non-MLO). It may be something like the partially reduced mitochondria (MLO) of Blastocystis or that anoxic ciliate (Dan Gaston’s been involved in a paper on that subject). Whether that counts as “without mitochondria” becomes iffy. The only eukaryotes that can be argued to be truly MOSTLY amitochondriate are microsporidia — ATP gets *IMPORTED* into the mitosome (mitochondrial relic) in that case! (also, the authors had no strong reasons to conclude those organelles were ‘hydrogenosomes’ – they WERE careful enough to call them ‘hydrogenosome-like’ organelles, to give them credit. But the media kinda got carried away with the whole hydrogenosome idea…)

    And Myxotricha is indeed awesome. Its relatives, fellow Parabasalians as well as Oxymonads, are also completely full of endo- and episymbionts. They have hydrogenosomes, and free hydrogen is a prized commodity, so plenty of methanogens cuddle up with the shriveled mitochondrial relics.


    PS: If anyone wants references to anything, let me know. Too lazy to write them out + link to them right now…

  12. The ‘Tree of Life’ diagram shows the last universal common ancestor (LUCA) at the initial branch point at the bottom, with one branch going off to the bacteria, the other to the archea and eukaryotes. While this is commonly assumed to be the case, it is exceedingly hard to show that it really was so. Essentially, all we can do is measure ‘distances’ between taxa (shorn of innumerable complexities, compare the DNA sequences from two species – the more dissimilar they are, the more distantly related they are.) This suffices to determine the three branches of the tree of life, but it does not tell us where the root is. In normal phylogenetics, we use an outgroup – some taxa which we have good a priori reason to believe are diverged prior to the last common ancestor of the group we care about. E.g. if we wanted to determine where the root of the placental mammals fell, we would include a bunch of marsupials in our analysis. To determine where LUCA falls in the tree of life, we want to compare with something which branched prior to LUCA – which clearly is not available.

    From a mathematical point of view, the root could fall anywhere on that tree. Some possibilities we can reject due to fossil evidence (e.g. LUCA was not a multicelled eukaryote.) Others would require wildly implausible changes in evolution rates (e.g. LUCA falling within the gram positive bacteria would require the gram positive bacteria to have evolved at a very slow rate and everything else at a very high rate.) However, these considerations still leave large portions of the tree as plausible locations for LUCA.

    This is all just expanding on Psi Wavefunction’s comment: we can’t take eubacterial monophyly for granted. (Or archeal monophyly for that matter. Given the evidence that the Last Eukaryotic Common Ancestor (a.k.a. Fred) had a bacterial-derived mitochondria, we can however take eukaryotic monophyly as proven.)

    I am aware (via a textbook) of work from over a decade ago which used a pair of genes hypothesized to have resulted from a gene duplication event prior to LUCA to root the tree, and this study found the traditional rooting illustrated above. I haven’t looked at this issue since then, so I’m not up to date on this.

  13. Just in case anyone’s wondering how the Tree of Life was rooted (which MW alludes to), they actually used what is called ‘Paralogue Rooting” (I think this was Iwabe et al 1989: http://www.pnas.org/content/86/23/9355.abstract); since the outgroup to all life is rocks, as some say, they used a genome duplication event that occurred BEFORE LUCA. Then, you build a phylogeny of both paralogues (both genes need to remain, eg. in the case of Iwabe et al, two subunits of ATPase) — then, and here comes the magic, each of the paralogues acts as an outgroup to the other, thereby rooting your tree for you.

    While ingenious, this method, like any other phylogenetic technique, is subject to artefacts and biases. Thing is, if your trees themselves are crappy, your paralogue rooting will be no more reliable. And we don’t actually know entirely yet just how crappy our trees are. The 1990’s eukaryotic trees with the Archezoa looked sexy at the time, and were later followed by a massive collective *facepalm* moment. Who knows if our contemporary bacterial trees may suffer a similar fate someday…

    Iwabe et al’s rooting has supported by further phylogenies AFAIK, but there could be consistent errors. That’s a problem with feeling comfy with lots of replicates — sometimes, replicates may lie consistently because of some yet-unknown methodological issue. Or maybe they got it right this time, who knows…

    Archaeal monophyly isn’t unanimously supported yet, but growing molecular evidence seems to push in that direction. Besides, it makes craploads more sense from a morphological/biochemical perspective… switching to the archaeal membranes and then back to the eubacteral/eukaryotic ones would be awkward. Postulating a separate membrane origin for Eub and Arch+Euk, with Euks later resembling Eubs, is also absurd. It makes a lot more sense for Archaeal membranes to be derived (explained more extensively in Cavalier-Smith 2006 Biol Direct, for example), especially once one lets go of the idea of them being super ancient or whatever. The name is likely a serious misnomer. Weird or super derived does not necessarily imply ancient, contrary to what intuition may suggest…

  14. Like no other blog, your articles usually grab my attention all the way through. But this one lost me. You start off talking about partnerships and then say…”But the greatest partnership of all is far more ancient.” implying you are going to tell us about this. Then you talk about different sorts of cells for a while. I got about 3/4 of the way through and gave up frustrated at having had no hints about what the partnership may be. You usually keep the storyline going so well.

  15. Just wanted to say that I appreciate the constructive criticism. No writer wants to lose their readers but lost readers never tell you about it. Except for the few occasions like this.

    So the “greatest partnership of all” bit is the union between bacteria and archaeon, which created the first eukaryote. It comes halfway through the piece when I first mention the ring of life. Looking back on it, that was probably not signposted clearly enough.

    This was a tough piece to write. Before I could go into any of that, I needed to explain what archaea were, what the standard tree of life model was, what makes eukaryotes special and so on. Spent more time on it than any other piece in recent memory. Sorry that it didn’t work out for you though. Will try harder next time.

Leave a Reply

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