Evolution Icon Evolution
Life Sciences Icon Life Sciences

Richard Lenski’s Long-Term Evolution Experiments with E. coli and the Origin of New Biological Information (Updated)

Casey Luskin

e. coli bacteria

Intelligent Design and the Origin of Information: A Response to Dennis Venema


In this article, Part 5, we:
• Understand why Venema’s citation of Lenski’s “Long Term Evolution Experiments” do not demonstrate that “Complex, specified information can indeed arise through natural mechanisms”
• Review Michael Behe’s 2010 paper in Quarterly Review of Biology which investigated Lenski’s research and found that “mutations were decreasing or eliminating the protein’s function
• Investigate why Lenski’s E. coli bacteria that evolved the ability to uptake citrate under oxic conditions didn’t evolve anything new and likely experienced loss-of-molecular function


Other Installments:
Part 1: Intelligent Design and the Origin of Biological Information: A Response to Dennis Venema
Part 2: Why Did One Theistic Evolutionist Part Ways with BioLogos?
Part 3: What Is a Proper Test of Intelligent Design?
Part 4: The False Dichotomy Between Intelligent Design and Natural Causes
Part 5 (This Article): Richard Lenski’s Long-Term Evolution Experiments with E. coli and the Origin of New Biological Information
Part 6: Another Bogus Claim of “Novel Function Arising Through Mutation and Selection”
Part 7: Confusing Evidence for Common Ancestry with Evidence for Random Mutation and Natural Selection
Part 8: Critically Analyzing the Argument from Human/Chimpanzee Genetic Similarity

See below for the update.

Some critics of intelligent design (ID) misunderstand ID as a denial of natural causes. For example, we have recently seen how theistic evolutionist Dennis Venema wrongly suggests that, in both a scientific and theological sense, ID denies natural causes. Venema imports this misunderstanding into his proposed methods of testing ID, suggesting that if we find natural causes doing anything, then ID is refuted.

Venema writes: “any natural mechanism that can be shown to produce information would render [Stephen Meyer’s] argument that information only arises from intelligent sources null and void.”
Dennis Venema’s argument collapses into this: ‘if Darwinian evolution can do anything, then ID is wrong.’ But this is not how we test ID, for ID readily allows that natural selection and random mutation can effect some changes in populations. The right question is not ‘Can natural selection do anything?’ but rather ‘Can natural selection do everything?’
With this in mind, let’s analyze Dr. Venema’s discussion of Richard Lenski’s Long Term Evolution Experiments (“LTEE”) with E. coli.

Where’s the Behe?
Before discussing the LTEE, it’s important to note that from the beginning of his series for BioLogos on evolution and the origin of information, Venema didn’t just purport to critique Stephen Meyer’s arguments in Signature in the Cell. Rather he referred to rebutting the entire “Intelligent Design Movement” or what he called (following Judge Jones?) the “IDM.”

But if Venema is going to critique the entire “IDM” using Richard Lenski’s “Long Term Evolution Experiments,” then Venema should discuss the most relevant literature of the “IDM” that discusses those experiments. He doesn’t do that.

In Venema’s discussion of the LTEE, there is no mention of a 2010 peer-reviewed scientific paper written by the most prominent biochemist in the “IDM,” published in a prominent biology journal, extensively critiquing Lenski’s LTEE. Venema fails to note and discuss Michael Behe’s December 2010 paper in Quarterly Review of Biology (QRB), which extensively discusses and critiques Lenski’s Long Term Evolution Experiments. Instead, Venema critiques the writings of Stephen Meyer, who hasn’t commented on Lenski’s LTEE because they weren’t relevant to his arguments in Signature in the Cell about the origin of life.

By misrepresenting Meyer’s thesis as being refuted by evidence of the power of natural selection, Venema creates a straw man. Meanwhile he ignores the substantive critiques by leading ID proponents of the very evidence he raises.

Vague Discussions vs. Precise Discussions of Lenski’s LTEE
As an initial salvo regarding Lenski’s LTEE, Venema writes:

[T]here were many possible genetic states of higher fitness available to the original strain, and random mutation and natural selection had explored several paths, all leading to a higher amount of “specified information” — information that specifies increased reproduction and survival in the original environment. All this was by demonstrably natural mechanisms, with a complete history of the relevant mutations, the relative advantages they conferred, and the dynamics of how those variants spread through a population. The LTEE is at once a very simple experiment, and an incredibly detailed window into the inner workings of evolution.

But what exactly was the “specified information” that increased? What new function was gained? Where did natural selection and random mutation produce functional, information-rich genes and proteins? Venema doesn’t say what new functions arose, what changed, or what information was gained. His claim that natural selection produced “specified information” is vague.

By contrast, in critiquing claims that the LTEE has produced something new, Behe’s 2010 Quarterly Review of Biology paper was anything but vague:

By examining the DNA sequence of the E. coli in the neighborhood surrounding the IS [insertion sequence] elements, the investigators saw that several genes involved in central metabolism were knocked out, as well as some cell wall synthesis genes and several others. In subsequent work, Cooper et al. (2001) discovered that twelve of twelve cell lines showed adaptive IS-mediated deletions of their rbs operon, which is involved in making the sugar ribose. Thus, the adaptive mutations that were initially tracked down all involved loss-of-FCT.

Several years later, when the cultures had surpassed their 20,000th generation, Lenski’s group at Michigan State brought more advanced techniques to bear on the problem of identifying the molecular changes underlying the adaptation of the E. coli cultures. Using DNA expression profiles, they were able to reliably track down changes in the expression of 1300 genes of the bacterium, and determined that 59 genes had changed their expression levels from the ancestor, 47 of which were expressed at lower levels (Cooper et al. 2003). The authors stated that “The expression levels of many of these 59 genes are known to be regulated by specific effectors including guanosine tetraphosphate (ppGpp) and cAMP-cAMP receptor protein (CRP)” (Cooper et al. 2003:1074). They also noted that the cellular concentration of ppGpp is controlled by several genes including spoT. After sequencing, they discovered a nonsynonymous point mutation in the spoT gene. When the researchers examined ten other populations that had evolved under the same conditions for 20,000 generations, they found that seven others also had fixed nonsynonymous point mutations in spoT, but with different substitutions than the first one that had been identified, thus suggesting that the mutations were decreasing the protein’s activity.

The group then decided to concentrate on candidate genes suggested by the physiological adaptations that the cells had made over 20,000 generations. One such adaptation was a change in supercoiling density; therefore, genes affecting DNA topology were investigated (Crozat et al. 2005). Two of these genes, topA and fis, had sustained point mutations. In the case of topA, the mutation coded an amino acid substitution, whereas, with fis, a transversion had occurred at the fourth nucleotide before the starting ATG codon. The topA mutation decreased the activity of the enzyme, while the fis mutation decreased the amount of fis gene product produced.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010).)

If you weren’t following all the technical language, here’s what’s going on: For the first 20,000 generations of Lenski’s LTEE, very little happened. There were a few molecular adaptations observed, yet whenever we understood their molecular basis, they involved the knocking out of genes, or decreasing protein activity — in essence, a decrease in specificity. Behe summarizes:

The fact that multiple point mutations in each gene could serve an adaptive role — and that disruption by IS insertion was beneficial — suggests that the point mutations were decreasing or eliminating the protein’s function.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010) (emphasis added).)

Unlike Venema’s discussion, Behe’s is precise, giving multiple examples and detailed descriptions of the types of changes observed in Lenski’s LTEE. And Behe found that the types of changes taking place in the E. coli tended to decrease or eliminate protein function.

Before getting into a discussion of the citrate-using strain of E. coli, Behe closes with another specific example that involved decreasing gene activity in Lenski’s LTEE:

In an investigation of global protein profiles of the evolved E. coli, Lenski’s group discovered that the MalT protein of the maltose operon had suffered mutations in 8 out of 12 strains (Pelosi et al. 2006). Several mutations were small deletions while others were point mutations, thus suggesting that decreasing the activity of the MalT protein was adaptive in minimal glucose media.

Looking at Table 3 of Behe’s QRB paper, not a single example of an adaptive mutation in Lenski’s LTEE entailed a gain of a new molecular function. In fact, over the course of his entire paper, Behe goes further and explains that most of our known examples of molecular adaptations in bacteria entail “loss-of-function” mutations. Somehow, Venema doesn’t discuss any of these findings.

E coli. Could Uptake and Metabolize Citrate Before Lenski’s LTEE
Later, when referring to a different stage of the LTEE, Venema claims that a new function did arise in Lenski’s E. coli bacteria during the experiments: the ability of E. coli to metabolize citrate. Venema claims that “One of the defining features of E. Coli is that it is unable to use citrate as a food source,” but after a series of mutations “bacteria that use citrate dominate the population.” According to Venema, these experiments show “Complex, specified information can indeed arise through natural mechanisms.”

Yet Venema leaves out important details, creating an inaccurate impression. As we’ll discuss below, normal E. coli already have machinery to uptake and metabolize citrate, so the general fact that Lenski’s bacteria showed this ability is really quite unremarkable.

Unfortunately, Venema’s readers on the BioLogos will never hear that. They also won’t learn that Michael Behe has written extensively about Lenski’s research, showing that the machinery for E. coli to uptake and metabolize citrate already existed in these bacteria. This isn’t an entirely new biochemical pathway. Venema fails to note that normal E. coli already have the ability to uptake and metabolize citrate. They just can’t normally uptake it under oxic conditions; Lenski’s bacteria evolved the ability to uptake it under oxic conditions used in the experiment. Then the E. coli used their normal metabolic pathways to use citrate as a food source. Behe made this point while commenting on these claims soon after they were first published in 2008:

Now, wild E. coli already has a number of enzymes that normally use citrate and can digest it (it’s not some exotic chemical the bacterium has never seen before). However, the wild bacterium lacks an enzyme called a “citrate permease” which can transport citrate from outside the cell through the cell’s membrane into its interior. So all the bacterium needed to do to use citrate was to find a way to get it into the cell. The rest of the machinery for its metabolism was already there. As Lenski put it, “The only known barrier to aerobic growth on citrate is its inability to transport citrate under oxic conditions.” (1)

(Michael Behe, Amazon Blog, “Multiple Mutations Needed for E. Coli” (June 6, 2008).)

Likewise, Behe’s recent 2010 paper in Quarterly Review of Biology provided an extensive critique of claims that Lenski’s LTEE showed the evolution of a new pathway that could metabolize citrate. Venema doesn’t cite or mention Behe’s QRB paper, but it too explains that E. coli already had the ability to metabolize citrate. Behe explains:

Recently, Lenski’s group reported the isolation of a mutant E. coli that had evolved a Cit+ phenotype. That is, the strain could grow under aerobic conditions in a culture of citrate (Blount et al. 2008). Wild E. coli cannot grow under such conditions, as it lacks a citrate permease to import the metabolite under oxic conditions. (It should be noted that, once inside the cell, however, E. coli has the enzymatic capacity to metabolize citrate.) The phenotype, whose underlying molecular changes have not yet been reported, conferred an enormous growth advantage because the culture media contained excess citrate but only limited glucose, which the ancestral bacteria metabolized.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010).)

Thus, Behe explains that the precise genetic mechanisms that allowed E. coli to uptake citrate under oxic conditions are not known. But Behe goes further and points out that the citrate-metabolizing E. coli strains really aren’t anything new, and that previous investigations suggest that the ability of the E. coli to uptake citrate under oxic conditions might result from molecular loss-of-function:

As Blount et al. (2008) discussed, several other laboratories had, in the past, also identified mutant E. coli strains with such a phenotype. In one such case, the underlying mutation was not identified (Hall 1982); however, in another case, high-level constitutive expression on a multicopy plasmid of a citrate transporter gene, citT, which normally transports citrate in the absence of oxygen, was responsible for eliciting the phenotype (Pos et al. 1998). If the phenotype of the Lenski Cit+ strain is caused by the loss of the activity of a normal genetic regulatory element, such as a repressor binding site or other FCT, it will, of course, be a loss-of-FCT mutation, despite its highly adaptive effects in the presence of citrate. If the phenotype is due to one or more mutations that result in, for example, the addition of a novel genetic regulatory element, gene-duplication with sequence divergence, or the gain of a new binding site, then it will be a noteworthy gain-of-FCT mutation.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010).)

Thus, previous research suggests that the adaptation which allowed these E. coli to uptake citrate under oxic conditions might be caused “by the loss of the activity of a normal genetic regulatory element.” Here’s what is likely going on here:

  • Under normal conditions, E. coli can metabolize citrate; after all metabolizing citrate is an important step in the Krebs cycle, a pathway used by virtually all living organisms when creating energy.
  • But under oxic conditions, E. coli lack the ability to transport citrate through the cell membrane into the cell. E. coli can do this under reducing conditions, but under oxic conditions E. coli can’t normally uptake citrate.
  • If Lenski’s citrate-using E. coli are like previous E. coli which were discovered uptaking citrate under oxic conditions, then it seems likely that the bacteria underwent a mutation that knocked out the regulation mechanism of a citrate-transport gene, causing over-expression, allowing the E. coli to uptake citrate under oxic conditions.

In other words, the machinery for both transporting and metabolizing citrate was already present in these bacteria. But a series of knockout mutations broke the regulation of pre-existing citrate transport mechanisms, causing over-expression of a citrate transport gene, allowing citrate to be transported under both oxic and anaerobic conditions. If this is the case, then clearly this example of Darwinian “evolution” entails the loss of a molecular function, not the gain of a new one. And there was no wholesale acquisition of the ability to metabolize or, as Venema put it, “use” citrate.

In fact, as Behe notes, we don’t really yet understand the precise molecular mechanisms that caused these E. coli to be able to uptake citrate under oxic conditions. So as far as we can tell, these changes entailed the origin of no new functional genes or proteins but might have resulted from a broken regulatory mechanism. We have not seen that natural selection and random mutation can produce functional, information-rich genes and proteins, and Venema is wrong to suggest otherwise.

Contra Venema, this example hardly shows the Darwinian evolution of a “new function,” especially since E. coli already had the ability to uptake and metabolize citrate. Venema claims that CSI has arisen, but if we don’t even know what mechanisms were involved in this change, how does he know that it is new CSI?

What do Lenski’s LTEE Really Tell Us?
In his QRB paper, Behe goes on to explain that to date, the known adaptations that have occurred in Lenski’s LTEE are either modification-of-function or loss-of-function changes:

The results of future work aside, so far, during the course of the longest, most open-ended, and most extensive laboratory investigation of bacterial evolution, a number of adaptive mutations have been identified that endow the bacterial strain with greater fitness compared to that of the ancestral strain in the particular growth medium. The goal of Lenski’s research was not to analyze adaptive mutations in terms of gain or loss of function, as is the focus here, but rather to address other longstanding evolutionary questions. Nonetheless, all of the mutations identified to date can readily be classified as either modification-of-function or loss-of-FCT.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010).)

Behe’s paper further suggests that when there are several kinds of potential adaptive mutations that might occur, loss or modification of function adaptations will be far more common than gain-of-function adaptations. He concludes:

Even if there were several possible pathways by which to construct a gain-of-FCT mutation, or several possible kinds of adaptive gain-of-FCT features, the rate of appearance of an adaptive mutation that would arise from the diminishment or elimination of the activity of a protein is expected to be 100-1000 times the rate of appearance of an adaptive mutation that requires specific changes to a gene.

(Michael J. Behe, “Experimental Evolution, Loss-of-Function Mutations and ‘The First Rule of Adaptive Evolution’,” Quarterly Review of Biology, Vol. 85(4) (December, 2010).)

The sort of loss-of-function examples seen in the LTEE will never show that natural selection can increase high CSI. To understand why, imagine the following hypothetical situation.

Consider an imaginary order of insects, the Evolutionoptera. Let’s say there are 1 million species of Evolutionoptera, but ecologists find that the extinction rate among Evolutionoptera is 1000 species per millennium. The speciation rate (the rate at which new species arise) during the same period is 1 new species per 1000 years. At these rates, every thousand years 1000 species of Evolutionoptera will die off, while one new species will develop–a net loss of 999 species. If these processes continue, in 1,000,001 years there will be no species of Evolutionoptera left on earth.

If Behe is correct, then Darwinian evolution at the molecular level faces a similar problem. If, all other things being equal, a loss or modification of function adaptation is generally 100-1000 times more likely than gain of function adaptations, then eventually an evolving population might run out of molecular functions to lose or modify. Neo-Darwinian evolution cannot forever rely on examples of loss or modification-of-function mutations to explain molecular evolution. At some point, there must be a gain of function.

Vaguely Appealing to Vast Probablistic Resources Won’t Work
Venema closes his post on the LTEE by saying: “what the IDM claims is impossible, these ‘tiny and lowly’ organisms have simply been doing — and it only took 15 years in a single lab in Michigan. Imagine what could happen over 3,500,000,000 years over millions of square miles of the earth’s surface.”

But vague appeals to vast eons of time and huge population sizes are unconvincing. You just have to do the math. As David Abel reminds us:

Mere possibility is not an adequate basis for asserting scientific plausibility. A precisely defined universal bound is needed beyond which the assertion of plausibility, particularly in life-origin models, can be considered operationally falsified. But can something so seemingly relative and subjective as plausibility ever be quantified? Amazingly, the answer is, “Yes.” … One chance in 10200 is theoretically possible, but given maximum cosmic probabilistic resources, such a possibility is hardly plausible. With funding resources rapidly drying up, science needs a foundational principle by which to falsify a myriad of theoretical possibilities that are not worthy of serious scientific consideration and modeling.

(David L. Abel, “The Universal Plausibility Metric (UPM) & Principle (UPP),” Theoretical Biology and Medical Modelling, Vol. 6:27 (Dec. 3, 2009).)

In the case of E. coli and citrate, the bacteria already had the ability to uptake and metabolize citrate, and simply found a way to transport it under different conditions. It’s likely this occurred by overexpressing pre-existing transport mechanisms. Does this imply that anything and everything “could happen over 3,500,000,000 years over millions of square miles of the earth’s surface”? Well, ID proponents aren’t interested in making vague and ambiguous appeals to vast amounts of probabilistic resources. They want to test these questions, and follow the evidence where it leads.

As discussed here, ID proponents have asked just how long it takes to evolve traits that require multi-mutation features. A multi-mutation feature requires multiple mutations to be present before there is any advantage given to the organism. Doug Axe’s research makes assumptions very generously favoring Darwinian evolution. He assumed the existence of a huge population of asexually reproducing bacteria that could replicate quickly — perhaps nearly 3 times per day — over the course of billions of years. Yet even here, complex adaptations requiring up to six mutations with neutral intermediates can become fixed. Beyond that, things become implausible.

If only slightly maladaptive intermediate mutations are required for a complex adaptation, only a couple (at most two) mutations could be fixed. If highly maladaptive mutations are required, the trait will never appear. Axe discusses the implications of his work:

[T]he most significant implication comes not from how the two cases contrast but rather how they cohere — both showing severe limitations to complex adaptation. To appreciate this, consider the tremendous number of cells needed to achieve adaptations of such limited complexity. As a basis for calculation, we have assumed a bacterial population that maintained an effective size of 109 individuals through 103 generations each year for billions of years. This amounts to well over a billion trillion opportunities (in the form of individuals whose lines were not destined to expire imminently) for evolutionary experimentation. Yet what these enormous resources are expected to have accomplished, in terms of combined base changes, can be counted on the fingers.

(Douglas D. Axe, “The Limits of Complex Adaptation: An Analysis Based on a Simple Model of Structured Bacterial Populations,” BIO-Complexity, Vol. 2010(4):1-10.)

If Axe is correct then we cannot always assume, as Venema seems to do, that sufficient probabilistic resources exist to produce complex features we see in life.

Summarizing Venema’s Argument Regarding the LTEE
In short, Venema’s argument regarding the LTEE collapses into common misconceptions about ID, which go something like this:

  • (1) ID holds that Darwinian evolution cannot do anything.
  • (2) If Darwinian evolution can do something then it can do anything.
  • (3) Lenski’s experiments show Darwinian evolution can allow E. coli bacteria to evolve a “new function” of metabolizing citrate.
  • (4) Therefore ID is wrong, and given enough time, Darwinian evolution can do anything we “imagine.”

At each step in his argument, the facts and/or the logic is wrong:

  • Regarding (1): In fact, ID does not hold that Darwinian evolution can’t do anything. Rather it claims that natural selection can do some things, but not everything. ID proponents readily acknowledge (as Behe has) that “if only one mutation is needed to confer some ability, then Darwinian evolution has little problem finding it.” The problem comes when multiple mutations are required to produce some new structure — and as Axe’s research shows, this is where Darwinian evolution typically gets stuck.
  • Regarding (2): Darwin-defenders have a long history of over-extrapolating from the data. ID is scientifically cautious and concludes that no one single experiment can show that Darwinian evolution can do everything we ask of it. A single experiment showing the ability of Darwinian evolution to do X simply shows the ability of Darwinian evolution to do X; it doesn’t necessarily show Darwinian evolution can do Y, Z, and A, B, and C, etc. ID says we need to test hypotheses carefully and not over-extrapolate from observed data.
  • Regarding (3): In fact Lenski’s experiments did not show the Darwinian evolution of an entirely new function. E. coli bacteria already had the ability to uptake and metabolize citrate, and the experiments simply showed they evolved the ability to uptake it under oxic conditions. This very likely required the loss of a molecular function.
  • Regarding (4) ID proponents do not think it is wise or scientifically accurate to vaguely invoke vast eons of time or vast population sizes to document the alleged power of Darwinian evolution. ID cautions that Darwinian evolutionists often assume that there are sufficiently vast probabilistic resources to accomplish any task imaginable, but that assumption might not be valid. Rather than simply assuming, Doug Axe’s research finds that adaptations requiring more than six neutral mutations, or two maladaptive mutations, to provide a functional advantage would not arise in the history of the earth.

Subsequent research by Axe and Ann Gauger suggests that it would not be uncommon for Darwinian evolution to face obstacles that exhaust the probabilistic boundaries as found by Axe’s research. In 2011, they published research in BIO-Complexity that found at least seven mutations (probably many more) would be necessary to convert one protein into a supposedly closely-related protein.

While Darwinians may (or may not) claim that this was a real evolutionary pathway, it’s the type of pathway that is often claimed to have been traversed by natural selection over the course of life’s history. The fact that this simple conversion required more mutations to produce a new function than would be allowed under Axe’s mathematical models shows that there may be real obstacles to the Darwinian evolution of new proteins. Venema’s citation of Lenski’s LTEE certainly does not show otherwise.

[Update, 9/25/2012: About a year after this article was first posted, Richard Lenski’s lab has now published a paper which tries to elucidate some of the biochemical changes that allowed E. coli to metabolize citrate under oxic conditions in the lab. Ann Gauger has an excellent analysis of this paper at “Innovation or Renovation?,” where she observes:

In his paper in Quarterly Review of Biology, Dr. Michael Behe pointed out that E. coli was already capable of using citrate for anaerobic growth (when no oxygen was available). He postulated that a change in gene regulation could turn on citrate transport and permit growth on citrate under aerobic conditions.

After an enormous amount of work, having sequenced the genomes of many clones along the lineages that led to the ability to use citrate, as well as lineages that never did, and testing the phenotypes of identified mutations, Blount et al. have now reported that Behe was largely right. The key innovation was a shift in regulation of the citrate operon, caused by a rearrangement that brought it close to a new promoter.

Gauger concludes: “But does this adaptation constitute a genuine innovation? That depends on the definition of innovation you use. It certainly is an example of reusing existing information in a new context, thus producing a new niche for E coli in lab cultures. But if the definition of innovation is something genuinely new, such as a new transport molecule or a new enzyme, then no, this adaptation falls short as an innovation. And no one should be surprised.”]