Evolution of Bacterial Flagellum

I posted this on the original book blog back in the summer of 2008. Since then, the book blog was hacked and lost. So let me repost it again. A current update is appended at the end.

For the past several years, I have been focused on how one might facilitate the evolution of metazoa. But because of recent scientific discoveries, I should pause and comment on an old topic – the bacterial flagellum.

Five years ago, I wrote an essay that raised many questions and expressed skepticism about Nick Matzke’s hypothesis of homology between the bacterial flagellum and the F ATPase. At the time I wrote this, the skepticism was justified. But since then, new data have come in that have served to significantly strengthen Matzke’s hypothesis and undercut my skepticism.

Since 2003, Matzke hooked up with Mark Pallen, a researcher who studies the flagellum/TTSS and who uncovered sequence evidence that indicated the flagellar gene FliH is a fusion of domains that are homologous to the b subunit and the delta subunit of the F-ATPase. This was a significant piece of support for Matzke’s hypothesis. And in 2006, Matzke and Palin published a paper in Nat Rev Microbiol.and I congratulated Nick.

And it now appears as if the structure of another flagellar gene, FliJ, further supports his hypothesis:

Namba and colleagues have now solved the structure of FliJ, another protein that interacts with FliI and FliH. And what they found was clear evidence of homology with yet another protein from the F-type ATPase–the gamma subunit!

Looking back, I would say my skepticism was justified in a trivial sense. Matzke argued that FliH was homologous to the b subunit, FliJ was homologous to the delta subunit, and FliP was homologous to the gamma subunit. At the time, the data did not exist to justify such claims and as it turns out FliH is homologous to both b and delta, while FliJ is homologous to gamma. These specific predictions did not bear out.

But in a far more important and global sense, it does indeed look like Matzke’s hypothesis is correct and that the TTSS machinery is homologous to the F-ATPase.

In The Design Matrix, I explore how the concept of IC interfaces with cooption and intelligent design and offer the following as part of my approach:

Instead, independent evidence is needed to support such a hypothesis of cooption cobbling a machine together. This does not mean we need something that amounts to a proof. Nor does it mean that an exhaustive Darwinian explanation is needed. On the contrary, the evidence we need is extremely modest and lacking in detail……First, if an irreducibly complex machine did evolve into existence through cooption, then the parts must have predated the machine. They must have been doing something else prior to being recruited into the machine. Thus, some evidence of this pre-machine activity is needed. Since we cannot travel back in time, we will have to settle for traditional evidence of common descent. Do the various parts of the machine have homologs that are in turn part of a system that is more ancient than the machine?

Multiple points of homology between the components of the F-ATPase and flagellum/TTSS would clearly qualify as “various parts of the machine” having “homologs that are in turn part of a system that is more ancient than the machine.” Thus, if I were to assign a Discontinuity Score to the flagellum, it would reside within the negative realm of the continuum.

And it becomes more interesting when we realize that the themes of cooption, modularity, and subfunctionalization have apparently played critical roles in the evolution of the bacterial flagellum. This is something to dig into at a later date.


UPDATE

As you can see, I am trying to approach these issues in an open-minded and open-ended fashion, while, given the abundant evidence that evolution is in play, setting the bar quite high for any discontinuity score. In fact, the essay I wrote above follows seamlessly from something I posted to the ARN forum back in the Spring of 2001:

Now, when I infer ID behind the bacterial flagellum, I am focused on the core aspects that are shared by all bacterial flagella (I have previously explained how core IC modular systems are under functional constraint and surveying 3.5 billion years of evolutionary tinkering is a good way to detect such constraint). This core aspect of the bacterial flagellum is thus hypothesized to be essential to basic flagellar function in all bacteria while more peripheral elements, while important/essential in the context of individual species, can be tweaked in accord with the needs of the organism. And it’s when one focuses on the core constituents of “the bacterial flagellum” that we still find a hefty IC system that is not convincingly explained by cooption ,as (for one reason) there are no apparent precursors for all components of this core (remember we are talking about what happened and not about what could have happened). However, some have proposed such things as the type III secretory system and the F-ATP synthases. I simply don’t find such explanations convincing (the former probably post-dates flagella and the latter doesn’t look like a precursor). And it is not bias behind my skepticism, as both precursor states would nicely fit into a front-loading hypothesis (in fact, if anything, I’d love to fit the flagella into a front-loading hypothesis).

What has been the sole change? In 2001, “the latter doesn’t look like a precursor” while in 2008 it is shaping up like one. BTW, the type III secretory system still looks like it was derived from the flagellum.

Advertisements

23 responses to “Evolution of Bacterial Flagellum

  1. I’m wondering if there would still be the problem of functional intermediate states between the F-ATP synthase and the bacterial flagellum, and what these would be.

  2. Oh, all right, you’re forcing me to have another imaginary dialogue between you and Behe.

  3. Before I take that step, let me ask: We have homology between the F-ATPase and the bacterial flagellum, for two of the flagellum’s proteins (three of the F-ATPase’s). Aren’t there about 20 proteins in the flagellum’s core? The other 18 can be accounted for how?

    Is it possible that we have common design, instead of descent? Could a designer have some reason for using common proteins in different molecular machines?

  4. Hi Bilbo,

    I’m not trying to ignore you (or anyone else here). Just busy during the week. Hopefully, I can reply this weekend.

  5. I’ll give that breath-holding thing another try.

  6. can’t…hold…breath…much…longer….

  7. Pingback: Mike Gene on Evolution of the Bacterial Flagellum - Telic Thoughts

  8. Bilbo,

    You’re not alone. I do not understand why homology exerts a materialist solution vs common design either.

    I’ll have to continue to look more in depth at Matzke’s work again. Last time I looked the “possible” events were just that, “possible.”

    At this point I ask, where are the experiments? None of this is observed, it is just stated that homologous structures naturally lead to evolution without agency. But why? Why is homology a strict domain of non-telic processes?

  9. datcg,

    I’m even willing to accept the view that homology proves, or at least strongly supports, evolution. I’m just having trouble with our finding so little homology. If a flagellum is made up of 20 different proteins, and the flagellum evolved from precursor systems, where are those precursor systems, along with their homologous proteins? Have they all vanished, except for the F-ATP synthase, which only has two of the flagellar proteins? No niches, with surviving precursors?

    Now perhaps there is good reason to reject common design as an explanation for homology in this case. I would just like someone to explain why.

  10. OK, I check it out. I saw a large number of homologies between flagellar proteins and other…whatever they are. Now I’m wondering why you focused your attention on only two of them, and based your Discontinuity score on those two.

  11. Hi Bilbo,

    For a long time, the focal point about the flagellum revolved around its similarity to the type III secretory system (see arguments between Miller and Behe, for example). While the TTSS still looks like it was derived from the flagellum, Matzke and Pallen have found decent evidence for homology with the ATP synthases, which more plausibly predate the flagellum. And as I explained in the blog entry, that satisfied my very modest evidential requirement as laid out in the book.

    You also asked about common design. Yes, this remains a possibility considering a) human-like designers often borrow (as JJS explained on TT some time ago) and b) there are only a 1000 or so protein folds to go around. But to alter the discontinuity score, I’d need some type of positive argument for this explanation.

  12. Yes, but there seem to be a number of homologies listed, besides ATP synthases and the TTSS (I’m not sure what all those letters stand for.) Did those homologies influence your decision, or are you still basing it only on the two flagellar proteins?

  13. Hello Mike,

    What does it mean to “assign a Discontinuity Score to the flagellum within the negative realm of the continuum”? Does it mean “it is not designed” or “it is probably not designed” or “it is unlikely that it is designed” or “it is highly unlikely that it is designed” or something like that? Or the very opposite?

  14. Hi Bilbo,

    It was a tipping point. Recall what I wrote in 2001:

    when one focuses on the core constituents of “the bacterial flagellum” that we still find a hefty IC system that is not convincingly explained by cooption ,as (for one reason) there are no apparent precursors for all components of this core (remember we are talking about what happened and not about what could have happened). However, some have proposed such things as the type III secretory system and the F-ATP synthases. I simply don’t find such explanations convincing (the former probably post-dates flagella and the latter doesn’t look like a precursor).

    But, as I noted in the OP, seven years later there is an apparent precursor – “In 2001, “the latter doesn’t look like a precursor” while in 2008 it is shaping up like one.”

  15. Hi Da Vinci,

    What does it mean to “assign a Discontinuity Score to the flagellum within the negative realm of the continuum”? Does it mean “it is not designed” or “it is probably not designed” or “it is unlikely that it is designed” or “it is highly unlikely that it is designed” or something like that? Or the very opposite?

    It simply means it looks like it evolved, which is not the same as “not designed” (I do not subscribe to the Traditional Template). The Discontinuity score is simply the score keeper’s best guess with regard to the evidence that pertains to the evolution of a particular biological feature.

    The scoring is explained in the book:

    First, in the spirit of the Explanatory Continuum, we should not treat each criterion as a binary choice. Instead, each criterion should exist along a continuum with each feature numerically scored. Consider Figure 10-1. Imagine we are trying to score an analogy between a biological feature and something that is known to be designed. It shows a range from -5 to 5. A score of 0 would represent a thoroughly ambiguous situation where we cannot decide if the two systems are analogous or not analogous. The same score might also occur if there is a lack of information that can be used to make the distinction. If the analogy between the biological and designed feature is deemed to be modest in strength, we give it a score of 1. The stronger the analogy, the higher the score, where 5 would represent the strongest possible analogy that we could realistically imagine. If the biological feature and designed object are deemed to be different in kind, but only to a modest degree, we give it a score of -1. The more the designed object and biological feature are deemed to be different in kind, the lower the score. The same basic scoring strategy can be applied with any of the Design Matrix criteria.

  16. Hi Mike,

    If I understand it, the other homologies in the list you directed me to are not from precursor systems; just the two flagellar proteins that are homologous with the three F-ATP synthases. Please correct me if I’m mistaken about this.

    Anyway, the question seems to boil down to where to place the bar for determining Discontinuity. As you explained, “ I am trying to approach these issues in an open-minded and open-ended fashion, while, given the abundant evidence that evolution is in play, setting the bar quite high for any discontinuity score.

    First, I can see why one would place the bar high, after the initial seeding of life, given your hypothesis of front-loaded evolution. Thus gaps in metazoa would not be given much weight for Discontinuity.

    But second, it’s not clear to me (nor to you, I bet) that the bacterial flagellum came after the initial seeding.

    In that case, and third, I think the bar should be lower when considering gaps between the flagellum and other ancient features of bacteria.

    Therefore, and fourth, two homologous flagellar proteins with a (supposed) precursor F-ATP synthase should not be enough to score Discontinuity in the negative column. Certainly let’s give Matzke credit and lower the score. But negative? C’mon.

  17. I probably won’t get back until after the weekend, so I thought I would add this:

    If we lower the bar for scoring Discontinuity, and then look at the homologous proteins, we should:

    1) Lower our Discontinuity score.

    2) Wonder why only two homologous proteins. Where are homologies for the other 18+ proteins?

    3) Wonder why the designer would front-load the design for the flagellum, instead of just designing directly, allowing some bacteria to have motility right from the start.

    4) Wonder why the TTSS, a simpler system than the flagellum, devolves from the flagellum, instead of the flagellum evolving from the TTSS. Wouldn’t we expect the latter, if the flagellum was front-loaded?

    5) Wonder if the homology can be accounted for by common design. After all, on the front-loaded scenario, these three F-ATP synthase proteins were directly designed so that they could be used in the flagellum later on. So the designer wanted to use these proteins in the flagellum in the first place. That would support a common design argument.

  18. If we lower the bar for scoring Discontinuity, and then look at the homologous proteins, we should:
    1) Lower our Discontinuity score.
    2) Wonder why only two homologous proteins. Where are homologies for the other 18+ proteins?

    See below.

    3) Wonder why the designer would front-load the design for the flagellum, instead of just designing directly, allowing some bacteria to have motility right from the start.

    Flagellar motility may not be all that important given that most clades of bacteria lack flagella. Cyanobacteria, arguably among the most important life forms on this planet, do not have flagella.

    4) Wonder why the TTSS, a simpler system than the flagellum, devolves from the flagellum, instead of the flagellum evolving from the TTSS. Wouldn’t we expect the latter, if the flagellum was front-loaded?

    Good point. And it may turn out that the TTSS is older. Right now I would just notice that the TTSS seems to be mostly about establishing symbiotic relationships with eukaryotic cells, so its possible the TTSS would have no function at such an early time.

    5) Wonder if the homology can be accounted for by common design. After all, on the front-loaded scenario, these three F-ATP synthase proteins were directly designed so that they could be used in the flagellum later on. So the designer wanted to use these proteins in the flagellum in the first place. That would support a common design argument.

    Interesting point. Like I said, I do not rule out common design; I just need some positive indicator it is in play. In other words, someone needs to develop what would be considered hallmarks of common design.

  19. First, I can see why one would place the bar high, after the initial seeding of life, given your hypothesis of front-loaded evolution. Thus gaps in metazoa would not be given much weight for Discontinuity.
    But second, it’s not clear to me (nor to you, I bet) that the bacterial flagellum came after the initial seeding.
    In that case, and third, I think the bar should be lower when considering gaps between the flagellum and other ancient features of bacteria.

    Yes, this was the argument I was standing on in 2001. I never bought into the IC arguments as they pertained to blood clotting or the immune system, but the flagellum was different in that it could be plausibly included in a seeding event.

    Therefore, and fourth, two homologous flagellar proteins with a (supposed) precursor F-ATP synthase should not be enough to score Discontinuity in the negative column. Certainly let’s give Matzke credit and lower the score. But negative? C’mon.

    You keep arguing that there are only two homologous flagellar proteins. But this does not appear to be the case.

    There are three with the ATP synthase alone. FliI is homologous to alpha/beta, FliH is homologous to both b and delta, while FliJ is homologous to gamma. And I suspect Matzke is going to be right about FliQ and the c subunit. That’s a very significant chunk of the export system.

    Then there is the motor. MotA and MotB are homologous to iron transporters, ExbB and ExbD. FliG is homologous to the magnesium transporter, MgtE.

    What’s mostly left is the driveshaft/hook/filament, which accounts for about half of the flagellar proteins. They are all homologous with each other. So instead of having 10 unaccounted proteins, you really have one (that was rapidly unpacked by serial gene duplication and subfunctionalization).

    What data do you need to move the discontinuity score into the negative realm?

    Also, one final thought about the relationship with seeding and the flagellum. One problem is that a prime candidate for seeding among the original consortium would be cyanobacteria. Not only are there ancient fossils of such bacteria, they play a key role in terraforming and supporting the whole biosphere. And guess what? So far, not one species of sequenced cyanobacteria has been determined to have a flagellum or any flagellar genes. The flagellum is simply not needed for dispersal.

  20. Bilbo: 2) Wonder why only two homologous proteins. Where are homologies for the other 18+ proteins?

    I think you got it wrong. Only 2 proteins don’t have homologs, others have. You should check Matzke-Pallen paper again.

  21. Hi Mike, I tried sending a reply a few hours ago, but I don’t think it went through. Let me try again.

    I asked you about the list of homologues that was on Matzke’s list, and how many might be from precursor systems. Perhaps you didn’t understand my question, but now that you’ve indicated that most of the flagellar proteins appear to have presursor homologies, my major objection vanishes into smoke. Your scoring Discontinuity in the negative column makes much more sense now. I see how this probably puts you at odds with Behe and Dembski. And even though I suspect they may be right, I would like to see you continue on your hypothesis and see where it leads.

  22. Hi Mike, I’m trying to reply for the third time. Now that you have made it clear that most of the homologous proteins on Matzke’s list refer to possible precursors, I admit your point, that the score for Discontinuity for the Bacterial Flagellum should be in the negative column. At least, on the assumption that rm + ns can account for most of evolution.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s