American Scientist

To the Editors:

I found Zachary D. Blount’s article “Replaying Evolution” in the May–June issue one of the most interesting I’ve read in a long time. I hate to ruin a good experiment by suggesting a butterfly effect, but I have a question: How did Richard Lenski ensure that the initial population of Escherichia coli was identical? From my lab days long ago, I would have taken the 12 samples from a single colony grown in a petri dish assuming that all the cells in that colony were identical. Did Lenski do something similar, and because DNA sequencing is now so relatively inexpensive, have any studies been done on the original sample to ensure all the cells in it were identical?

Frank M. Archer
Delta, British Columbia, Canada

Dr. Blount responds:

The question is a good one. Dr. Lenski was, of course, limited by the technology of the time when he started—back when genome sequencing was not an option. He did nearly what Mr. Archer would have done. He picked individual isolated colonies grown from the plating of a clonal culture to start the populations, so that any mutation that might produce a “butterfly effect” would necessarily have arisen independently in the replicate populations.

However, one nuance that I didn’t mention in the article is that six of the populations were founded from an ancestral strain named REL606, and six from one named REL607. REL607 is a mutant of REL606 with a single base-pair mutation that restores function to the araA gene and thereby enables it to grow on arabinose, or gum sugar. This “Ara” marker is neutral with regard to fitness under the conditions of the Long-Term Evolution Experiment (LTEE). The marker is useful because Ara⁺ colonies are white (or pink) on tetrazolium arabinose indicator medium, while Ara^– colonies are a deep red. As the LTEE cultures are always transferred with alternation of Ara^– and Ara⁺ populations, the marker allowed for detection of any inadvertent cross-contamination that might occur. (We always plate cells on tetrazolium arabinose agar medium when we periodically freeze samples. If such cross-contamination is detected, we can restart the affected population from a previously frozen sample. Dr. Lenski has discussed some of the mistakes that occasionally happen in a blog post: http://bit.ly/2s9eDqU.)

We’ve since discovered, by genome sequencing, that a secondary mutation occurred in the recD gene when the REL607 ancestor was isolated. That mutation is also selectively neutral in the LTEE environment, based on many competitions that have been performed. And although the protein encoded by recD is involved in DNA repair, that particular mutation has had no discernible effect on the rate or spectrum of mutations seen in the LTEE, again based on the extensive genome sequencing performed in recent years (see the 2016 Nature paper on this topic by Olivier Tenaillon and colleagues), nor has it had any effect on the fitness trajectories of the populations. So there seems to have not been any “butterfly effect” associated with the two mutations that initially distinguished the two sets of six LTEE populations. (If there had been, then it’s interesting that the LTEE would have essentially been like two sets of 6 replays going on at the same time, rather than one set of 12 replays.)

Of course, it’s possible that those two mutations might yet have some systematic effect down the road, although so many other mutations and phenotypic changes have occurred in the 12 populations over the course of the LTEE that it would probably be hard to determine and show definitively that the two starting mutations were responsible. In any case, we haven’t seen any indication so far that they have mattered.