Evolution Icon Evolution

Viral Video Overstates the Evidence About Bacterial Evolution

Casey Luskin
Photo: Richard Lenski, via Veritasium (screen shot).

Over the past week we’ve received quite a few emails from Discovery Institute followers about a new video from the YouTube science channel Veritasium. The video is titled, “The Longest-Running Evolution Experiment.” It’s about the Long Term Evolution Experiment (LTEE) hosted at Michigan State University — a project we’ve covered many times in the past here in Evolution News (see hereherehere, and here for a few samples). In fact, our most recent treatment was on an ID the Future podcast interview released on June 14, 2021 — two days before the Veritasium video went up — with Dustin Van Hofwegen, a biology professor at Azusa Pacific University in the Los Angeles area. Van Hofwegen has devoted much of his career to investigating, replicating, and fact-checking evolutionary claims about the LTEE. 

Evolution advocates love the LTEE. In 2015 New Scientist said that it “has become a poster child for evolution, causing consternation among creationists trying to explain away its compelling evidence.” This brings us to the Veritasium video, which as of this writing has over 3 million views, using the LTEE to promote evolution. Unfortunately, this means that many people are being misinformed by overstatements in the video about the evidence for E. coli evolution.

In the video, lead LTEE researcher Richard Lenski claims that the experiment provides “one of the most direct demonstrations of Darwinian adaptations by natural selection you can imagine.” One obvious and simplistic response is that after some 75,000 generations these E. coli bacteria are still…bacteria. But that doesn’t really address the claims of the video or the LTEE. After all, evolution proceeds by small steps, and small genetic changes are supposed to add up to create new complex features, and ultimately perhaps a new species. Thus, if the bacteria evolved something genuinely new at the biochemical level, then perhaps the LTEE has supported the viability of an evolutionary model where small changes build and eventually add up to something new, even entirely novel. So we need to dig deeper. 

Relative Fitness

When evolutionary biologists analyze the details of the LTEE, they are concerned primarily with relative fitness — are these organisms becoming better than their ancestors at growing in this flask environment? Yes, these organisms are getting better at growing on glucose. In the environment of the LTEE, they are in a race to consume as much glucose as they can. It’s like a race team trying to move through the course faster than their competitors. Bacteria who get “faster” by gathering glucose more quickly will outcompete their fellows. It’s simple when viewed this way. Grow faster equals more fit. 

But the details of how they get faster matter. This is microbial genetics. When microbial geneticists have looked at the details of what is happening to these bacteria, a different interpretation emerges. Faster for the organisms in the LTEE means that they are emphasizing functions involved in glucose metabolism. Nearly all other functions become a hindrance to going faster. So what happens to these hindrances? They get inactivated. For a bacterium, it is costly to make something that you aren’t using. For a bacterium in the LTEE, costly means you don’t grow as fast. And what does this mean for these bacteria? They eventually end up in the autoclave. The ones that survive are the ones that are more streamlined. It’s like auto racing, where incremental improvements happen as the race car gets more streamlined: that is, it becomes lighter and/or has less drag. Unnecessary genes create the equivalent of drag. Getting rid of these genes entails a loss of function. There’s nothing remarkable about that.

No, Citrate Metabolism Isn’t Anything New

In the video, the narrator says that “in 2003 the bacteria started doing something remarkable.” That occurred when, as Lenski describes it, “one of the twelve lineages suddenly began to consume a second carbon source, citrate.” One day, he says, a lineage “discovered that there was this nice lemony desert, and they had begun consuming that and getting a second source of carbon and energy.” Lenski makes it sound like this is an entirely new capability: “E. coli going back to its original definition of a species is incapable of that.” He gives the impression that bacteria suddenly evolved the ability to eat an entirely new food source — a “remarkable” evolutionary adaptation that they were “incapable” of performing previously. But this is simply false. Normal E. coli can eat and metabolize citrate. The pro-evolution website Evo-Ed, at Lenski’s own university, explains:

Like many organisms, E. coli has a citric acid cycle, and so metabolizes citrate while growing on various substances. It can also grow anaerobically by fermenting citrate.

As Dustin Van Hofwegen put it in his recent IDTF interview, “When microbiologists look at an experiment like that, we know that E. coli does have the ability to grow on citrate. It’s used in various metabolic cycles. They have the ability to use it in the citric acid cycle. If they get it into the cell, it’s used in their metabolic processes.” Thus, no new metabolic pathways evolved in this experiment. By omitting these facts the video misinforms viewers. 

No New Genetic Information

The LTEE required 33,000 generations and many years for the bacteria to acquire the supposedly new trait. In the video Lenski says that one of his lab’s researchers wanted to explore “why did it take so long to evolve this and why has only one population evolved that ability?” The implication is that this is a complex trait that required many slow mutations to arise. Lenski says it was a “difficult” trait to evolve because it required both a “rare mutation” and also a “series of events” where multiple mutations were needed before any advantage was conferred. Van Hofwegen realized there was something fishy about these claims. As he explained to IDTF:

The only difference is that in the conditions of the [LTEE] experiment, they didn’t have a transporter. They [E. coli bacteria] didn’t have the ability to bring that citrate outside of their cells into the cells and actually use it for energy. And so when I looked at that experiment as a microbiologist I thought, all they have to do is turn that thing on. That’s really easy for bacteria to do. Why did it take them 33,000 generations to do that?

Van Hofwegen draws a comparison to a light switch. Normal E. coli have the metabolic pathways to live off citrate, and they have the ability to transport it into their cells. But under the conditions of the experiment that “light switch” was turned off. The bacteria didn’t need to evolve a new metabolic pathway or a new transport feature to eat citrate. All they needed to do was turn on their transporter under the oxic conditions of the LTEE experiment. The organisms used the “light switch” to express their citrate transporter. So how did they do it? 

A 2016 peer-reviewed study in the Journal of Bacteriology, “Rapid Evolution of Citrate Utilization by Escherichia coli by Direct Selection Requires citT and dctA,” co-authored by Van Hofwegen and biologists Scott Minnich and Carolyn Hovde, has the answer. In their research they witnessed the same trait, the ability to use this “lemony dessert,” arise in under 100 generations and 14 days. This result was repeatable 46 times. They found that the trait is not very genetically complicated — again, akin to flipping a switch — and that there is more to the story than is being been told. Indeed, their paper shows that no new genetic information arose during the evolution of this trait. 

Three Primary Mutations

Although normal E. coli can eat citrate, they cannot uptake and metabolize it under oxic conditions. In the LTEE, bacteria evolved the ability to uptake citrate under oxic conditions — what is called the “Cit+ phenotype.” But did anything “new” evolve? At the genetic level, Minnich and his co-authors’ research says the answer is no. To understand why, consider the three primary mutations required to produce the Cit+ phenotype: 

  • A mutation allowed the E. coli to express an antiporter protein, CitT, under oxic conditions. CitT permits one molecule of citrate to be imported into the cell in exchange for one of three less “valuable” molecules with less carbon: succinate, fumarate, or malate. However, the gene for this antiporter protein already existed previously. No new gene evolved. CitT is usually switched off in E. coli when oxygen is present, but this mutation allowed it to be turned on. What caused it to become turned on? Biochemically speaking, a switch that normally represses expression of the gene that produces CitT under oxic conditions was broken via a mutation that captured a new promoter, so the citrate-uptake pathway got turned on under oxic conditions. This isn’t the evolution of a new molecular feature; it’s the breaking of a molecular feature — a repressor switch. 
  • There was a duplication mutation of the gene for the CitT antiporter protein, allowing the bacteria to produce more of that protein. This allowed the bacteria to uptake more citrate under oxic conditions. This too does not involve the evolution of anything new — it only involves making more of something already present. 
  • Another gene duplication mutation occurred for the gene that produces the protein DctA, a succinate importer. This allowed some of the succinate that had been lost in exchange for citrate to be recovered and transported back into the cell. Again, this is just making more of something already present. Nothing new arose. 

Thus, the mutational pathway involves: (1) breaking something at the molecular level (a repressor), (2) making more of something already present (citrate importer), and (3) making more of something already present (succinate importer). Breaking features at the molecular level, or making more of some pre-existing components, have been long known to be possible under Darwinian evolution. As Minnich and his co-authors explain in their paper: “No new genetic information (novel gene function) evolved.” They also write, “The LTEE has not substantiated evolution in the broader sense by generation of new genetic information, i.e. a gene with a new function.” They conclude:

Finally, because this adaptation did not generate any new genetic information and only required expanded expressions of two existing transporters (citT and dctA), generation of E. coli Cit+ phenotypes in our estimation do not warrant consideration as a speciation event. 

For microbiologists, a key question is why this paper’s research observed the Cit+ phenotype arise so rapidly, whereas Lenski’s LTEE required a long time for the same thing to happen. A commentary in the Journal of Bacteriology that accompanied the paper explains: 

[T]he primary message of the paper by Van Hofwegen et al. is that the series of events used to explain adaptation in the short-transfer LTEE (and in speciation) might need to be revised. … It would appear that the delay in the LTEE experiments may not reflect need for a neutral potentiation step, but the difficulty of intermittent selection to act on frequent copy number variants. The bottleneck in serial dilutions is hard to cross when initial improvements are due to an unstable copy number variant that is counter-selected during the intervening rapid growth period. 

John Roth and Sophie Maisnier-Patin, “Reinterpreting long-term evolution experiments — Is delayed adaptation an example of historical contingency or a consequence of intermittent selection,” Journal of Bacteriology, Vol. 198:1009-1012 (April, 2016)

The phrase “may not reflect need for a neutral potentiation step” means that no complex sequence of neutral mutations was needed to set up the Cit+ phenotype. Essentially, this research shows that when one imposes strong selection for growth on citrate, the story isn’t one of neutral evolution evolving a complex feature, but one where each step gives a successive advantage, and no step creates anything genetically new. Under the right selection pressures, this relatively simple phenotype can arise very quickly. This research shows that Lenski’s work is not the impressive story of a complex evolutionary pathway, as many have claimed. Most importantly, this paper shows that Lenski’s work did not demonstrate the evolution of any new biochemical features. Rather, it takes pre-existing transporter proteins and over-expresses them in an unusual environment — but only by breaking a molecular switch. Biochemically, these molecules are only doing what they were already designed to do. 

Or as Van Hofwegen put it in his IDTF interview, “The organisms developed the strategy to respond to that situation by using the traits they already had.” He explains that normal bacteria have lots of pathways to eat many different kinds of food sources. It would be wasteful for all of those pathways to be active all the time, especially when most food sources aren’t present. But he notes that when bacteria are stressed, they induce new mutations via transposons in order to turn on other metabolic pathways that already exist. That’s exactly what happened here. The LTEE’s result reflects a normal system of control and regulation in biology.  

Is Evolution Boundless?

In the video, Lenski also suggests his research shows evolution is boundless, refuting an older model where an evolving population eventually reaches a limit or “asymptote” of fitness. Instead he says that bacterial evolution follows a “power law model” where there is effectively no boundary to what can evolve. As Lenski puts it, “What our experiment shows is that even in the absence of environmental change, there are so many opportunities of smaller and smaller magnitude to continue to make progress that in fact progress probably would never stop even in a constant environment.” But to make this claim, about the details of what the bacteria are actually doing genetically, Lenski needs to demonstrate added functional information, not merely observe the genome being streamlined, i.e., jettisoning otherwise useful functions for the purpose of growing best on glucose and/or citrate. Escherichia coli only have ~4,000 genes. Sooner or later, with this streamlining strategy of removing genetic information, they will hit a ceiling, or rather an asymptote.