The Loom

E. coli Evolution Follow-up

I’m back at last from the west coast leg of the Microcosm tour.

Portland had a cloudy, melancholy charm, and at Powell’s I gave a reading in front of a collection of hand-made black velvet paintings from the nearby Velveteria. When the audience’s eyes drifted off of me, I couldn’t tell if they were lost in thought or distracted by Jimi Hendrix or a smoking clown.

The next day I headed for San Francisco, where I talked to Moira Gunn for her show Tech Nation (link to come). Then I had lunch with Kirsten Sanford, who will be interviewing me on tomorrow’s edition of This Week In Science. Then off to Santa Cruz, to talk to Robert Pollie at KUSP for his show Talk of the Bay (link to come). Finally I made my way over to Kepler’s in Menlo Park. I spoke there a few years ago, and since then they closed and were saved by the community. I was glad to be able to come back.

In the morning I flew to Seattle. I headed for Microsoft Research to give a talk, which I’m told will be online before long. I was a little spooked by the experience, because, in addition to the lunchtime crowd in the room, there were lots of people watching online elsewhere–in some cases in other countries. I had to resist the instinct to talk very loudly so that people over in China could hear me.

Then I made a quick appearance on KOMO, the ABC affiliate in Seattle. The anchor started talking about E. coli in hamburger and spinach, and I responded by describing the billions of E. coli in her. I saw her eyes widen a little in what I’m guessing was supressed horror, but she handled it like a pro.

Finally I went to Town Hall and waited for intrepid souls to wade through the downpours to hear me talk. It was great to see familiar faces (like this mug). I met blogger Geoff Arnold, who showed me Microcosm on Kindle, and since I couldn’t autograph his screen, he took a picture. (I think Town Hall will also be posting my talk–will update.)

Along the way, I wrote a blog post about some new advances in the research I describe in Microcosm. The response was terrific (thanks in part to a link from reddit), and the comments have been multiplying faster than E. coli on a warm day.

There were a few questions that came up that I thought I’d address in follow up.

–First off, the paper itself is finally online now.

Matt asks

how are they sure this citrate eating adaptation was a result of mutations, and not, say, an existing sequence of dna that was just locked in an intron or something, and then eventually shuffled to a coding region of the genome? Could they follow the genetic changes point by point, or are they still trying to figure that out?

Introns are segments of protein-coding genes that get edited out as the DNA sequence is read by a cell. E. coli and other bacteria don’t have introns. But E. coli does sometimes shuffle segments of DNA around its genome. Whether the mutations involved were shuffles, or changes to single nucleotides of DNA, or accidental duplications of DNA, etc.–all that remains to be seen.

–Phishrow asks

how common are experiments of this scale and duration? You hear about 20 odd year studies on human populations from time to time; but how many situations like this one do we have bubbling away?

Good question. I’m having a hard time thinking of anything that’s run anywhere near as long as Lenski’s 20 yr experiment (I’m thinking specifically of evolutionary experiments). A lot of great work has been done on bacteria and viruses over the course of a few hundred or few thousand generations. It’s not easy to keep something running for decades, though–especially to find the funding for it.

Ian asks

Carl – Are you aware of any long-running experiments like this where the initial bacterium has accumulated sufficient mutations that in the end it would be classified as a different type (genus or something higher) of bacterium from what it started out? The “Shigella” comment in the article above comes close.

Actually, the Shigella case shows just how hard it is to use conventional taxonomy to understand the evolution of bacteria. Shigella seemed so different from E. coli when it was first discovered that it was put in a different genus. It seemed different because it makes us sick by invading cells, something harmless strains don’t do. There are also lots of other differences–Shigella lacks some key enzymes E. coli has. But studies on their DNA revealed that several strains of E. coli had independently evolved into “Shigella” strains. What is clear is that in this cluster of lineages, there has been some dramatic evolutionary change. And now Lenski has seen some dramatic evolutionary change over the course of a few thousand generations.

–Heather raises some concerns

I’m not knowledgeable about bacteriology, nor am I opposed to evolution, but 2 facts in the article stand out: (1) contamination from foreign bacteria, including citrate-eaters, occurred often enough that the researchers had a procedure for it (toss the flask and start from the most recent frozen sample of the same line), and (2) the E Coli. can develop the ability to eat citrate by acquiring the plasmid DNA ring from a citrate-eater.

You can read the paper for the details, but the short response is that the researchers repeatedly checked for contamination and established that it was indeed E. coli that was eating citrate. Also, they set up the entire experiment to make it impossible for E. coli to pick up plasmids.

–Ken Finley writes

This is total horse crap. There’s nothing in the Bible to suggest that evolution exists. You’re just arbitrarily making up excuses.

If the bacteria changed, it was clearly because God willed it. He does that sometimes, you know.

Just because God helped the bacteria survived, you can’t just simply say it’s because we come from monkeys. That’s stupid and arrogant.

You’ll go to hell for your blasphemy

Sometimes I have a hard time figuring out when these sorts of posts are serious or jokes.

–Nate writes

This is a very interesting study, but I would like to point out to some people that seem to have misunderstood what happened. The bacteria did not develop a way to eat citrate, they mutated to a point where they were able to get it across their membranes. They already had the capability to digest it. Most likely a few bacteria had a few mutations which damaged their membranes and allowed citrate to get through. I would like to know what all the tradeoffs were in these bacteria as well. Losing several capabilities while gaining one doesn’t seem like a step forward to me, but in this situation it was advantageous to these bacteria because of the abundance of citrate.

He then follows up...

…it’s proof that an organism cannot gain a capability through mutations without losing several others. If, hypothetically, the same bacteria gained a dozen more capabilities, this research would tend to show that the bacteria would end up losing 3-4 times that many capabilities. If a bateria did lose that many, it would most likely no longer be viable. It does prove microevolution occurs, which we already knew, but cannot be made to prove anything past that. That type or extrapolation is foolish and ignorant.

Nate, you may want to read the paper. You’re sounding a lot more certain about the nature of the mutations than the scientists themselves are. And you’re almost certainly wrong when you suggest that the citrate-eaters just have damaged membranes. You’d have to do some serious damage to its membranes to let citrate leak in–so much damage that lots of stuff would leak in and out. That’s clearly not the case here, because the bacteria are healthy. Bacteria use special channels to draw in these molecules.

As for tradeoffs, it’s true that these citrate-eaters may not do as well eating glucose as their ancestors. I have no idea where you get the “3-4 times that many capabilities” phrase. But so what? Evolution is not about “steps forward,” like some unhindered march of progress. It’s about change, and tradeoffs are an essential part of that change.

And if you’re going to call this “microevolution,” you’ll need to define your terms here for us. Here we have seen the evolution of a capability, the lack of which was significant enough that it marked E. coli as a species. What’s micro about that?

Andrew asks,

Forgive the question, as I’m not a scientist (just a interested dabbler), but I thought that evolution was, in general, a slow process that could not be observed so quickly? Is the situation different for bacteria? Is evolution something that can be observed in a matter of years for them?

Bacteria can divide several times a day, which makes them very fast breeders. And since they’re so small, hundreds of millions can fit in a small flask. Even though mutations are incredibly rare, these sorts of numbers make it inevitable that they will arise in bacterial colonies. And natural selection combined with these numbers means that over a few years you can see clear change occurring. There’s no compelling reason to think that these processes don’t happen in bigger, slower-breeding species (like us), but it’s harder to see the changes because they take decades, centuries, or millennia. It’s also worth bearing in mind that Lenski’s experiment is a tiny embodiment of evolution in bacteria. Just remember that bacteria pack the soils, the oceans, the sea floor, and our bodies. They’ve been evolving for billions of years. And they can pool their evolutionary potential by trading genes. That makes the evolution of a new trait in Lenski’s lab all the more important.

Comments

  1. #1 Matt
    June 5, 2008

    Thanks for addressing those comments, Carl! I really can’t wait to hear about the follow-up work. Please keep us informed.

  2. #2 factician
    June 5, 2008

    For another scientist who studies long term cultures (of a different sort) check out Steve Finkel: http://college.usc.edu/faculty/faculty1003247.html

    He started starving bacteria in Roberto Kolter’s lab in Harvard (about 15 years ago) and is still starving those same cultures to this day. The difference between his experiments and Lenski’s, is that Lenski adds back fresh nutrients. Finkel lets E. coli starve for years, and watches the interesting communities that develop (and he does indeed find cannibals – cool!).

  3. #3 clear as mud
    June 5, 2008

    1) re: Ken Finley’s comments… You’ve just seen Poe’s Law in action, I think.
    2) Thanks for the response to Nate. I didn’t see the subsequent comments until now, but you addressed the issues raised. I shouldn’t have sounded so course, and I *did* miss the glucose-deficiency wording in your post. But your followups (including the “3-4 times” thing and the leaky membrane concept) were perfect.

    And in response to Nate’s remaining question to me, I got a PhD in molecular microbiology and after a number of years as a post-doc, left the sciences and am now working in public health.

    Oh, and that leads to my last point – question for Carl… Now that I’m out of the sciences, and now that I’m becoming interested in reading papers again, I suddenly sadly aware that not being in a science dept means – no access to papers… including PNAS. Waaaah (again)! Can’t read it :(

    So, how on Earth DOES a non-scientist get to see such papers?

  4. #4 Luna_the_cat
    June 5, 2008

    @clear_as_mud — make friends with people who have access to the papers!

    And/or ask nicely online. Amazing what you can turn up that way.

  5. #5 JJ
    June 5, 2008

    Hope you had a good time in Santa Cruz! I love this town!

  6. #6 Dave Finn
    June 8, 2008

    While the experiment you describe says something about neo-Darwinian evolution I am not clear that it says anything much about evolution in general. One of the most serious logical errors in evolutionary theory concerns the concept of a direction to evolution. The logical error being simply that of considering it to be a single subject. There are several ways in which the statement that there is no direction to evolution can be taken as true. There is also one crucial sense in which it is false. Environments contain multiple species and whenever the environment changes the species within it are likely to be affected differently. Under these conditions the unsuccessful species are in the same environment as a species with a genome better adapted to the new environment than is their own. At that time, and for those species there is a direction to evolution. When there is such a direction the best survival strategy is for the failing organism to adjust its own genome toward that of the successful species by, for example, taking genes from the plasmid of a thriving bacterium. Any organism that survives because it has a genome that implements this strategy will pass on the strategy and it will become ubiquitous. The statistics and distribution of genes makes it clear that this is a major, and almost certainly the predominant, mechanism in the creation of novel species.

  7. #7 Keith Robison
    June 9, 2008

    Not exactly evolutionary biology, but relevant to it, is the 100+ year experiment in seed germination; caches of seeds buried in the ground are unearthed every 20 years and germination attempted; only one species sprouted last time, but that’s still amazing!

    Here’s a press release on it that Google dug up:
    http://newsroom.msu.edu/site/indexer/672/content.htm

  8. #8 Ashwan
    June 10, 2008

    @clear_as_mud: I could download this particular paper from PNAS because apparently India is “sponsored” by PNAS. In any case, I think PNAS makes all their content free after 6 months.

    You can also find most of Lenski’s papers on his own website: https://www.msu.edu/~lenski/

  9. #9 clear as mud
    June 11, 2008

    @Ashwan: Thanks! Yes, I remembered that some groups like PNAS (and ASM?) only keep the newer papers for pay. But PNAS is sponsoring India? Do they play ads every 15 minutes? (joke)

  10. #10 marcel
    June 11, 2008

    I had to resist the instinct to talk very loudly so that people over in China could hear me.

    I thought this was to help “people over in China” understand you as an English speaker.

  11. #11 Larry Fafarman
    June 13, 2008

    Here is my scenario of what happened:

    A mutation occurred at around 20,000 generations and the citrate-eating ability appeared when one of the bacteria bearing that mutation had a different mutation at around 31,500 generations. IMO the first mutation was very unusual or rare because (1) it apparently took about nine years to occur (44,000 generations in 20 years is about 2200 generations per year) and (2) it apparently appeared in only one of twelve lines of bacteria, even though all twelve lines were descended from a single individual. I think that the second mutation is a fairly common one because it was often expressed again in populations started by the unfrozen preserved populations of 20,000 generations or later, and the reason why this second mutation took so long to be expressed the first time — about 11,500 generations (from the 20,000th to the 31,500th) or 5 years — was that bacteria with the preliminary first mutation were scarce because the preliminary first mutation conferred no advantage in survival. After the preliminary first mutation occurs, appearance of the citrate-eating ability would be just a matter of time if the second mutation were a common one.

    Also, I am disturbed by numerous claims that the results of this study refute the ideas of Michael Behe — IMO that is not the case. Michael Behe’s response to this study is at –

    http://www.amazon.com/gp/blog/post/PLNK3U696N278Z93O

  12. #12 drew
    June 13, 2008

    Larry,

    That’s not all, it appears as though there were subsequent mutations that made the citrate uptake weaker and then even stronger than previous generations.

    And re Behe’s response:

    “One of the major points of the book was that if only one mutation is needed to confer some ability, then Darwinian evolution has little problem finding it. But if more than one is needed, the probability of getting all the right ones grows exponentially worse. “If two mutations have to occur before there is a net beneficial effect — if an intermediate state is harmful, or less fit than the starting state — then there is already a big evolutionary problem.” (4) And what if more than two are needed? The task quickly gets out of reach of random mutation.”

    Here in this study we have, an original mutation that didn’t seem to do much of anything, a second that conferred a great benefit, subsequent mutations that seem harmful and then more mutations that make the benefit even stronger than it was before. Behe specifically tells us that this is “out of reach of random mutation”. I’d call that wrong.

    And then he goes on to say “If the development of many of the features of the cell required multiple mutations during the course of evolution, then the cell is beyond Darwinian explanation.”

    Taken together with what is above he seems to be saying that this strain of E. coli—which Lenski demonstrates clearly arose from successive random mutations coupled with selection—couldn’t have possibly arisen out of successive random mutations coupled with selection. Of course that assumes that what he means by “Darwinian explanation” is random mutations coupled with selection. I’d also call that wrong.

    Where he goes completely wrong is that he seems to assume that all mutations must occur at the same time in the same replication, which anyone who understands evolutionary theory understands to probably not be the case and which Lenski et al. clearly demonstrate to not be the case.

  13. #13 melior
    June 13, 2008

    I appreciate your lucid writing style, and look forward to reading your latest book.

    I can’t resist adding a fascinating E. coli bit to ponder that I picked up reading one of Jay Ingram’s collections of strange science tales:

    These bacteria comprise roughly one fourth, by mass, of human feces.

    So you and I and all of us are evolutionary playscapes for it as well.

  14. #14 Rosie Redfield
    June 13, 2008

    This is actually a question about the research described in the Slate article. I study genetic exchange in bacteria, and found the statement that an E. coli strain had acquired hundreds of genes in only 15 years very surprising.

    I read the Slate article, and then looked for the original paper. I think it must be Manning et al. PNAS 105:4868. But I can’t find anything in either that suggests the genes were acquired in 15 years. Rather, the authors of the Manning et al paper estimate the most recent common ancestor to have lived about 20,000 years ago.

    Can anyone clear this up?

  15. #15 Larry Fafarman
    June 13, 2008

    drew said (June 13, 2008 12:26 PM) –

    Larry,

    That’s not all, it appears as though there were subsequent mutations that made the citrate uptake weaker and then even stronger than previous generations.

    But some of the citrate-eating bacteria would not have been affected by any subsequent mutations that made the citrate uptake weaker, and these unaffected citrate-eaters should have continued to thrive. Instead the citrate-eaters nearly disappeared after rising to 19% of the population.

    Taken together with what is above he seems to be saying that this strain of E. coli—which Lenski demonstrates clearly arose from successive random mutations coupled with selection—couldn’t have possibly arisen out of successive random mutations coupled with selection.

    Michael Behe never said or implied that “this strain of E. Coli . . . .couldn’t have possibly arisen out of successive random mutations coupled with selection.” And so far as I know he never argued before that this couldn’t happen — he only argued that evolution becomes increasingly difficult exponentially as the number of mutations required for a single beneficial trait increases — and I agree with him.

    melior said (June 13, 2008 12:47 PM) —

    These bacteria comprise roughly one fourth, by mass, of human feces.

    So you and I and all of us are evolutionary playscapes for it as well.

    OK, but this is a controlled experiment — the bacteria were given lots of citrate and little glucose to eat in order to give a big advantage to bacteria that develop the ability to eat citrate.

  16. #16 Owlmirror
    June 13, 2008

    the bacteria were given lots of citrate and little glucose to eat in order to have a population of bacteria that was not supposed to grow.

    Fixed. The whole point is that it was not intended to “give a big advantage to bacteria that develop the ability to eat citrate”.

    You need to read up on what a “control” is in science. The evolution of Cit+ was, in this case, completely unexpected.

    This was just as serendipitous as the discovery of penicillin.

  17. #17 drew
    June 13, 2008

    “Michael Behe never said or implied that “this strain of E. Coli . . . .couldn’t have possibly arisen out of successive random mutations coupled with selection.”"

    Then I guess I didn’t understand when he said that “If the development of many of the features of the cell required multiple mutations during the course of evolution, then the cell is beyond Darwinian explanation”.

    Let’s see here, “If the development of many of the features of the cell required multiple mutations during the course of evolution” which we have here a feature which required at least 2, if we know that this trait required multiple steps, others almost certainly did as well. “Then the cell is beyond Darwinian explanation.” -This I read as “successive random mutations coupled with selection cannot explain the cell”. By including this line in an article (or blog entry) about development of a trait in a specific strain of E. coli, it leads one to conclude that he’s using the phrase to refer to the topic at hand (aka citrate eating bacteria).

    Or perhaps my assumption that “successive random mutations coupled with selection” is what he meant by “Darwinian explanation”, was wrong. If that is the case please explain to me what is meant by “Darwinian explanation”.

  18. #18 Larry Fafarman
    June 14, 2008

    Owlmirror (comment #16),

    You didn’t even quote me correctly! I actually said,

    — the bacteria were given lots of citrate and little glucose to eat in order to give a big advantage to bacteria that develop the ability to eat citrate.

    – and you misquoted me as follows:

    the bacteria were given lots of citrate and little glucose to eat in order to have a population of bacteria that was not supposed to grow.

    Why were the bacteria “not supposed to grow”?

    The whole point is that it was not intended to “give a big advantage to bacteria that develop the ability to eat citrate”.

    Giving the bacteria lots of citrate and insufficient glucose — as was done in this experiment — does give a big advantage to bacteria that develop the ability to eat citrate. I am still trying to get an answer to my question of whether evolution of citrate-eating E. coli bacteria was one of the goals of the experiment.

    You need to read up on what a “control” is in science.

    I did not use “control” in the sense of, say, a “control group.” What I meant was that the conditions of the experiment were closely controlled.

    The evolution of Cit+ was, in this case, completely unexpected.

    Zachary Blount, a co-author of the paper, said that evolution of Cit+ was not unexpected — such evolution had been observed once before. What I am trying to find out is whether trying to induce such evolution was one of the purposes of the experiment — as I said, the condition of lots of citrate and insufficient glucose tends to promote such evolution.

  19. #19 Larry Fafarman
    June 14, 2008

    drew said (comment #17),

    Let’s see here, “If the development of many of the features of the cell required multiple mutations during the course of evolution” which we have here a feature which required at least 2, if we know that this trait required multiple steps, others almost certainly did as well. “Then the cell is beyond Darwinian explanation.”

    Behe didn’t say that an evolution requiring two mutations is impossible — he just called it “a big evolutionary problem” and said that the problem gets exponentially worse when more mutations are needed. Here is what he actually said,

    I think the results fit a lot more easily into the viewpoint of The Edge of Evolution. One of the major points of the book was that if only one mutation is needed to confer some ability, then Darwinian evolution has little problem finding it. But if more than one is needed, the probability of getting all the right ones grows exponentially worse. “If two mutations have to occur before there is a net beneficial effect — if an intermediate state is harmful, or less fit than the starting state — then there is already a big evolutionary problem.” (4) And what if more than two are needed? The task quickly gets out of reach of random mutation.

  20. #20 Hoovooloo
    June 17, 2008

    Yes, Mr. Fafarman, he called it “a big evolutionary problem.” These results, however, show that in a relatively short timescale (the 1st mutation occurred about 3/4 of the way into the experiment, so the other mutations occurred in ~5 years), 2-3 mutations could occur that provided the bacteria with a new, complex trait. If this sort of trait can occur in 5 years, and evolution has billions of years to work, this would seem to refute Behe’s point, would it not? In addition, this experiment showed that the trait can propagate, likely at a reduced speed, even when a new mutation harms the ability to propagate (note “harmed”–not eliminated), thus directly refuting the second point Behe made.