Today we hear creationists pleading for science classes to “teach the controversy” over evolution.  By this they mean that science teachers should be free to discuss whether or not evolution is true or supported by the evidence, to teach the “strengths and weaknesses”.  This is unfortunate.  There is simply no such controversy within the scientific community, and wasting class time debating settled science prevents students from studying real and interesting controversies in evolutionary biology.  One such controversy is over the role of historical contingency in evolution.  On one side of this debate are people like the late Stephen Jay Gould who hold that contingency plays a large role in evolution, making the process largely unpredictable.  Gould held that if we could go back in time and replay the “tape of life” from some point in the past, we would arrive at a world wildly different from ours today.  The other side counters that natural selection strongly limits the adaptive possibilities of organisms, that there are many roads to only a few evolutionary destinations.  In contrast to Gould, this side holds that replaying evolution would yield a world highly similar to ours. A new paper in PNAS addresses this debate and suggests that historical contingency is an important feature of evolution.

The paper is out of the lab of Richard Lenski at Michigan State University.  Lenski has been running the long-term evolution experiment (LTEE) since 1988.  For 20 years, Lenksi’s group has been continuously growing 12 separate populations of E. coli and monitoring their evolution.  These bacteria are grown in a minimal media containing limiting quantities of glucose as well as an excess of citrate, which E. coli cannot metabolize in aerobic growth.  Lenksi’s group noticed that, after about 33,000 generations, one of the populations had evolved the ability to utilize citrate (referred to as the Cit+ phenotype).

33,000 generations is a long time.  Lenski points out that over that time, each population would have experienced billions of mutations, more than enough for each population to have sampled every possible point mutation several times.  The fact that it took so long to evolve the Cit+ phenotype in a single population suggested two possibilities.  The first is that Cit+ phenotype requires an extremely rare mutation, like a chromosomal inversion.  The other possibility is that the Cit+ phenotype arose through a series of two or more mutations, meaning that its evolution was contingent on the particular history of the population in which it arose.

To determine which of these possibilities was correct, Lenski’s group went back to E. coli samples taken over the course of the LTEE and attempted to replay the evolution of Cit+ from various generations.  If the Cit+ function required a single, extremely rare mutation, then it should evolve at the same slow rate in all of the replays.  On the other hand, if evolution of Cit+ is contingent on a specific potentiating mutation, then it should only evolve in replays from later generations.  In their replays, the group found that Cit+ ability only evolved in samples older than 20,000 generations, and most often from samples older than 30,000 generations.  Cit+ never evolved from younger samples.

However, this result could also be explained if the population had evolved hypermutability, or an increased mutation rate.  The paper notes that 4 of the other populations had become hypermutable over the course of the LTEE.  If older generations had an increased mutation rate, they would experience a single, rare mutation with greater frequency than earlier generations, so the rare mutation hypothesis could not be discarded.  To test this, Lenski’s grew generational samples and watched for the evolution of a marker trait, the ability to metabolize arabinose.  They found that the mutation rate from Ara- to Ara+ was identical in samples from early and late generations.  Therefore, the emergence of Cit+ in later generations was not due to an increased mutation rate.

These results support the hypothesis that evolution of Cit+ is contingent on the particular mutational history of the population.  In other words, this particular population experienced a mutation or set of mutations by, at the earliest, 20,000 generations that potentiated later mutation to Cit+.  Other populations had different evolutionary histories that did not produce this potentiating genetic background.  Therefore, they did not evolve the capacity to utilize the excess citrate in their environment.

So Gould was right, historical contingency can be a key factor in evolution.  What is really cool about Lenski’s work is that they were actually able to rewind the “tape of life” and replay evolution from various points in history.  Their results show that replaying the evolution of E. coli from earlier generations produced a different result:  the bacteria never evolved to utilize citrate.  The implication of this work is that the living world we see today is likely a contingent result of a unique evolutionary history.

Of course, creationists don’t appreciate the results of this work.  They only see it as a threat to their pseudoscientific beliefs, so they’ve been working hard to spin the paper to support their naive religious viewpoint.  I’ll discuss their objections in the next post.

Historical contingency and the evolution of a key innovation in an experimental population of Escherichia coli. Proc Natl Acad Sci U S A. 2008 Jun 10;105(23):7899-906.

DOI: 10.1073/pnas.0803151105

This entry was posted on Sunday, June 15th, 2008 at 10:45 am and is filed under Science. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.