Update 01-03-2010: Be sure to check the plant edition of Evolving Molecular Machines over at Lab Rat’s blog!
This is a story about an event that took place 2 billion years ago. With the benefit of hindsight and a great deal of human bias, we could argue that it was one of the pivotal moments in the evolution of life on earth. What happened? Life was well underway at the time, with proto-bacteria already populating the oceans for over hundreds of millions of years. One of the cells alive at the time, swallowed an alpha-proteobacterium. Something remarkable happened: the alpha-proteobacterium did not die but survived in the host cell. Over time, the host and symbiont became to be dependent on each other. While the symbiont proved to be an excellent supplier of energy, both host and symbiont became the ancestor of all eukaryotic life on earth, which is the lineage that gave rise to plants, animals, protists and fungi. Each and every cell in our body carries the legacy of this endosymbiotic event in the form of mitochondria, which have come to be, amongst other things, the energy suppliers of our cells.
Indian Muntjac fibroblast cells: the mitochondria are visible as a highly interconnected network in red. The cell nucleus is blue, actin is stained green. Source.
Mitochondria still have properties that remind us of their once free roaming lifestyle: they carry their own genome and are surrounded by a double membrane, isolating themselves from the rest of the cell. Most of the genes that were originally carried by the alpha-proteobacterium have been re-located to the host genome. So many proteins that are required inside the mitochondrion, are in fact synthesized in the cytoplasm and have to be imported afterwards. Nowadays this arrangement is well established, and the mitochondria and the rest of the eukaryotic cell function in perfect harmony with each other. But it is very difficult to find out how this complex arrangement came to be, billions of years ago. You could argue that the endosymbiont lost its genes, and therefore required protein import machinery to remain functional. Or maybe it was the evolution of protein import machinery that drove the loss of genes in the first place?
By looking at the state of mitochondria in a wide variety of eukaryotes, scientists are able to piece together bits of this complex episode in the history of eukaryotic evolution. Any protein that is imported into the mitochondria has to pass the two membranes that surround the mitochondrium. In eukaryotes, this proccess is mediated by the TIM23 and TOM protein complexes (TOM stands for Transporter Outer Membrane, and TIM for Transporter Inner Membrane). These complexes consist of multiple protein subunits that form the importers, not quite unlike separate mechanical parts that together form a dedicated machine. But how could these molecular import machines evolve? No bacteria have protein complexes that import proteins over their double membrane, so where did these complexes come from and how did they acquire the functionality that they have now?
The import of a mitochondrium into the cell. A protein (blue) is first transported over the outer membrane by TOM, and then transported to the matrix by TIM23. Both TOM and TIM23 consist of multiple subunits, making them a sort of 'molecular machines'. Derived from here.
In a PNAS paper published in 2009, Abigail Clements and colleagues provide evidence for an interesting scenario. They focussed their analysis on those subunits that are shared by by all eukaryotes that are alive today. Two of those subunits are Tim44 and Tim14, both part of TIM23. The group was able to find two distinct proteins with sequence similarity to Tim44 and Tim14 in alpha-proteobacteria, that they have dubbed TimA and TimB. Sure enough, these proteins localized to the inner membrane of the bacteria, just like they localize to the inner membrane in eukaryotes. However, unlike Tim44 and Tim14, TimA and TimB don’t interact with each other! The team shows that TimA is associated with an ATP-dependent protease, but they found no protein interactors for TimB.
This is were the story gets interesting: the researchers set out to show that the bacterial TimB can be recruited to an eukaryotic TIM23 transporter lacking Tim14. By analyzing the 3D structure of Tim14, they identified a crucial asparigine residue that is at the basis of interaction with other subunits. This asparagine is lacking in the bacterial TimB, but the team made an artificial TimB protein that does contain this asparagine. They used this modified TimB to restore a growth defect in yeast cells that lacked a Tim14 subunit. With this experiment, the researchers say that by this single mutation, they converted the bacterial TimB to function in the TIM23 complex.
This leaves us with an interesting picture of the evolution of molecular machines: different parts and pieces are recruited from ‘salvaged parts’, generating new functionality and complexity. In this particular scenario, a protein present in the endoysymbiont got incorporated into a new complex. There’s also evidence that some subunits are derived from proteins derived from the host cell (see picture below). In a complex interplay between host and endosymbiont, eukaryotes seem to have salvaged all the raw resources that were at their disposal.
Modes of mitochondrial evolution: some proteins of the protein import machinery are derived from the endosymbiont, while others can be traced back to the host. Picture taken from 3th reference.
I don’t want to end on a sour note, but I do have a few gripes with this paper: I think the story would’ve been more convincing if the researchers had shown that TimB is physically associated with the rest of the TIM23 complex. In addition, the researchers don’t mention a possible function of TimB in bacteria, which I consider a prerequisite if you want to advocate the ‘re-purposing’ of subunits and modules. I also found that the horrible word ‘pre-adaptation’ makes an appearance in this paper, but that will be the subject of another post..
Clements, A., Bursac, D., Gatsos, X., Perry, A., Civciristov, S., Celik, N., Likic, V., Poggio, S., Jacobs-Wagner, C., Strugnell, R., & Lithgow, T. (2009). The reducible complexity of a mitochondrial molecular machine Proceedings of the National Academy of Sciences, 106 (37), 15791-15795 DOI: 10.1073/pnas.0908264106
Dolezal, P. (2006). Evolution of the Molecular Machines for Protein Import into Mitochondria Science, 313 (5785), 314-318 DOI: 10.1126/science.1127895
Alcock, F., Clements, A., Webb, C., & Lithgow, T. (2010). Tinkering Inside the Organelle Science, 327 (5966), 649-650 DOI: 10.1126/science.1182129
You might also like:
OK…this is slightly scary, I literally had a lecture on this last week, and I was just getting ready to plan an essay for it! The ‘inside’ and ‘outside’ diagram is amazing; did the transport machinery come from the symbiont or the cell?
In terms of this question: “Or maybe it was the evolution of protein import machinery that drove the loss of genes in the first place?” Definitely that way around. You need the import machinery before you can loose the genes, otherwise the symbiont would just die. Gene transfer is a ratchet system, genes only go from chloroplast to nucleus (usually by chloroplasts lysing and bits of their genes getting stuck in the nucleus) and once they are in the nucleus they stay there. Protein import systems make the gene in the chloroplast mostly redundant (and it’s safer in the nucleus anyway) so it gets lost.
I am very tempted to write a sister post for this, about chloroplast import mechanisms. Would you mind?
That’s a cool coincidence! Concerning ‘outside’ vs. ‘inside’, the evolution of the core protein import machinery was probably mainly driven from the endosymbiont. The subsequent addition of subunits more likely reflect changes the cell imposed on the import complexes. I really like this view of host and endosymbiont tinkering and salvaging protein subunits along the way.
It reminds me of how the ISS gets built up – module by module, expanding on the core that was laid down in 1998. In the end, the machine that we end up might not be the best solution, but the form and nature of the initial complex determines the shape and function for the future. Amazing stuff :).
I really like your explanation of gene loss! Right now I’m wishing that my curriculum included a course on endosymbiont evolution ;).
I’d love to see your post on chloroplast import! Be sure to let me know when it’s done, so I can place a link in this article.
Great post; wish I had this last year when taking a protistology course — one of the major endosymbiosis guys teaches that, so endosymbiosis was definitely a significant item on the menu. Had to read everything directly from literature though, as no textbook or other reliable non-expert source really covers it properly, and admittedly was quite daunting at first!
Now, I do have a couple comments about the intro though.
This is a story about an event that took place 2 billion years ago.
Maybe it’s better to write 1-2 billion years? The fossil data for early eukaryotic origin is quite rather weak (Cavalier-Smith 2006 Phil Trans R Soc B; Cavalier-Smith 2006 Biol Direct); and from phylogeny it makes a lot more sense for archaeans and eukaryotes to be fairly recent clades (although all the preceding divergences (cyanobacteria, glycobacteria, posibacteria (‘gram positives’, and the ‘gracilicutes’, including the fairly late-diverging proteobacteria necessary for our eukaryogenesis), etc) could have exploded ultra rapidly in a very short period of time, but then why the relative stagnation in the eukaryote clade after its emergence?).
But even if we don’t subscribe to the eubacterial root (which is the only one that makes any modicum of sense, IMNSHO), instead going by the root being between eubacteria and archaea+eukarya, then the proto-eukaryote that swallowed the proteobacterium would’ve either been a proto-archaeon or an actual archaeon (if archaea are paraphyletic to eukaryotes, which would be awkward as multiple origins or origin and subsequent loss of the strange archaeal membrane lipids would be required). Now, I use bacteria=prokaryotes, but most microbiologists would twitch uncontrollably upon seeing that. So proto-bacteria wouldn’t be a good choice of words, perhaps.
(now, the root between eubacteria and eukaryotes+archaea would allow early origins for the latter, but makes very little sense from a cell biological perspective (eg. Cavalier-Smith 2006))
But this leads us to bigger problems:
Life was well underway at the time, with proto-bacteria already populating the oceans for over hundreds of millions of years. One of the cells alive at the time, swallowed an alpha-proteobacterium.
Bacteria cannot phagocytose. There is one weird case, Bdellvibrio, which manages to get inside what appears to be the space between the outer and inner membranes (http://www.nature.com/nrmicro/journal/v2/n8/fig_tab/nrmicro959_F1.html). This often used as ‘proof’ that bacteria are full of cryptic endosymbiotic events (or that “eukaryote = eubacterium eats archaeon” anyway…), but note that it doesn’t penetrate the inner membrane. The outer one is weird and porous and not quite the same thing as THE cytoplasmic membrane (inner membrane in double-membraned eubacteria, most likely homologous in all bacteria), so this is NOT a case of [involuntary] phagocytosis!
AFAIK, no archaeon can do phagocytosis either. Thus, a proto-bacterium or proto-archaeon could not have engulfed an endosymbiont. The thing that did was already fairly diverged from its bacterial brethren; probably with a fairly developed actin cytoskeleton and a membrane trafficking system (Cavalier-Smith 2008 Int J Biochem Cell Biol). Thus, it was by definition a proto-eukaryote. This may seem pedantic, but I think it’s important to note that the creature that endosymbiosed the alpha-proteobacterium was already quite weird.
It is rather strange that no primarily amitochondriate eukaryotes have been found [yet?]. This would imply that the changes associated with gaining the ability to phagocytose and the subsequent ‘enslavement’ of the mitochondrion happened mind-numbingly quickly; and/or that the mitochondrion gave SUCH an advantage to its host that they completely wiped out (by competition) all relatives without one. Probably a bit of both.
Oh, and I’d love to cite someone besides Cavalier-Smith (and read papers that are NOT 80 pages of DENSE on CRACK…), but he’s about the only one out there who can actually integrate cell biology, phylogeny, molecular genetics, genomics, paleogeology, climate, ecology, etc. into one comprehensive synthesis that actually makes some sort of coherent sense. Everyone else seems overly obsessed with endosymbiosis (eg. JA Lake, who also fails membrane topology 101 in his 2009 Nature hypothesis piece) or LGT and this bizarre attitude of hopelessness to the whole question of the early history of life.
Cavalier-Smith’s hypotheses are a brilliant mix of the insane, the rigorous, the factual and the utterly absurd into one integrated masterpiece encompassing as many fields as humanly (almost superhumanly, even) possible. It would be great if some bacteriologists followed along and tested + modified his theories as needed, rather than simply dismissing them…
[apologies for long comment! I just finished reading all 56 long pages of Cavalier-Smith 2006 Biol Direct, and bacteria are clogging my mind right now...]
Thanks for that comment! Actually, it’s not a comment.. it’s a blog post in itself! Seriously, I think you could write a great piece on it. I would definitely be interested!
I never – ever realized that bacteria don’t phagocytose! This is a serious eye-opener for me. Most readings of the endosymbiont theory I came across gloss over this. I’ll be a lot clearer about this the next time I write something about mitochondrial origins! The whole issue does raise interesting questions though (as you already mentioned!).. The problem of course is that since both phagocytosis and mitochondria are exclusive to eukaryote lineages, it is difficult to reconstruct what exactly happened in our ancestor.
There’s some work being done on dating the emergence of the actin/myosin remodeling system right? I know there’s nice work on the origins of membrane trafficking (see how I snuck in that self-reference ;) ?), what do you think of this? Could the PVC cluster be the bacterial phylum that eukaryotes are most closely related to? They do seem to share some remarkable characteristics with eukaryotes..
I definitely want to read some work by Cavalier-Smith now. I rarely come across papers that I can call insane, absurd and a masterpiece at the same time, so this should be a welcome change of pace. I guess it does take someone with a stroke of genius and the capability of integrating insights from a wide variety of fields to create some order in this mess! I’m still working my way through the recent issue of Phil Transactions R Soc though..
Again thanks for bringing this up!