Taking tests of the predictability of evolution one step further

Recent evid­ence sug­gests that if we replayed evol­u­tion from the same start­ing point, in the same envir­on­ment, it would pro­ceed in a pre­dict­able man­ner, with cer­tain genes being repeatedly selec­ted. But what about in a more real­ist­ic world, where envir­on­ments dif­fer? Dr Car­oline Turn­er provides new insights from her exper­i­ment­al study pub­lished today in Evol­u­tion Let­ters. 

A recent post on this blog, “Test­ing the Pre­dict­ab­il­ity of Evol­u­tion”, nicely intro­duced the cap­tiv­at­ing sub­ject of the repeat­ab­il­ity of evol­u­tion. Namely, if we were able to turn back time and evolve the same organ­isms under the same con­di­tions, would evol­u­tion occur in the same way? In their recent paper in Evol­u­tion Let­ters, Härer et al. con­sidered a vari­ant of this ques­tion – if sim­il­ar, but not identic­al, organ­isms evolve under the same con­di­tions, will those dif­fer­ent organ­isms adapt in the same way? Our paper, now pub­lished in Evol­u­tion Let­ters, asks the com­ple­ment­ary ques­tion: If identic­al organ­isms evolve in sim­il­ar, but not identic­al envir­on­ments, will those organ­isms adapt to com­mon selec­tion pres­sures in the same way?

To address this ques­tion, we exper­i­ment­ally evolved bac­teria in the labor­at­ory. Because bac­teria repro­duce so quickly and have large pop­u­la­tion sizes, it is prac­tic­al to watch how they evolve in a rel­at­ively short amount of time in the labor­at­ory. Exper­i­ment­al evol­u­tion in the labor­at­ory allows us to evolve many rep­lic­ate pop­u­la­tions all foun­ded from the same ancest­or. Each rep­lic­ate pop­u­la­tion is then an inde­pend­ent repe­ti­tion of evol­u­tion. Evol­u­tion in the labor­at­ory allows us to con­trol the evol­u­tion­ary envir­on­ments to a much great­er degree than is pos­sible in nature, mak­ing sure that the rep­lic­ate pop­u­la­tions evolve in as con­sist­ent of an envir­on­ment as pos­sible. The oth­er bene­fit of work­ing with bac­teria is that we can eas­ily freeze the pop­u­la­tions and later revive them from the frozen stocks. This means that we can dir­ectly com­pare the char­ac­ter­ist­ics of evolved pop­u­la­tions and their ancestors.

In our exper­i­ment, we wanted to test the degree to which our bac­teri­um, the Gram-neg­at­ive oppor­tun­ist Burk­hol­d­er­ia ceno­cepa­cia, adap­ted in the same way to a giv­en selec­tion pres­sure, even when anoth­er aspect of their envir­on­ment var­ied. We var­ied two aspects of the envir­on­ment. First, we var­ied the amount of food (sug­ar) that was avail­able to the bac­teria. We also var­ied wheth­er selec­tion favored bac­teria which grew in liquid (plank­ton­ic growth) or on sur­faces (biofilm growth). The treat­ments are sum­mar­ized below (the second high-car­bon, biofilm treat­ment is a con­trol for the effects of pop­u­la­tion size):

We evolved six rep­lic­ate pop­u­la­tions in each envir­on­ment for 90 days. As the bac­teria rep­lic­ate in each envir­on­ment, new muta­tions arise by chance due to errors in copy­ing DNA. If the bac­teria with a par­tic­u­lar muta­tion were able to com­pete more suc­cess­fully in their envir­on­ment, then those bac­teria would be able to pro­duce more off­spring than bac­teria without the muta­tion. As a res­ult, that muta­tion would reach a high­er fre­quency in the pop­u­la­tion over time.

After 90 days of evol­u­tion, we sequenced the gen­omes of the evolved pop­u­la­tions to find the muta­tions that spread in each pop­u­la­tion. We meas­ured the genet­ic sim­il­ar­ity between pop­u­la­tions evolving in each envir­on­ment. Over­all, pop­u­la­tions that evolved under a shared selec­tion pres­sure (either sug­ar level or mode-of-growth) were more genet­ic­ally sim­il­ar to each oth­er than pop­u­la­tions that evolved in envir­on­ments which shared neither selec­tion pres­sure. For example, pop­u­la­tions evolved in a high-sug­ar envir­on­ment repeatedly evolved muta­tions in a pro­tein affect­ing iron stor­age, regard­less of wheth­er they were selec­ted for plank­ton­ic or biofilm growth. These res­ults indic­ate that evol­u­tion­ary change can be repeat­able, even when envir­on­ments are not exactly the same.

How­ever, there was one excep­tion to the pat­tern of high­er genet­ic sim­il­ar­ity between pop­u­la­tions that evolved in sim­il­ar envir­on­ments. Pop­u­la­tions evolved in low-car­bon, biofilm envir­on­ments had almost no genet­ic sim­il­ar­ity with pop­u­la­tions evolved in low-car­bon, plank­ton­ic envir­on­ments. The bac­teria adap­ted to low-car­bon con­di­tions dif­fer­ently, depend­ing on wheth­er they were in an envir­on­ment select­ing for plank­ton­ic or biofilm growth.  We think this excep­tion res­ults from a deep tradeoff between invest­ing in rap­id growth or syn­thes­iz­ing poly­mers to pro­duce a biofilm. Such a tradeoff may be masked when resources are abund­ant but it emerges when resources are scarce and bac­teria are forced to com­mit to either life­style. Des­pite this excep­tion, over­all we observed a great deal of pre­dict­ab­il­ity to evol­u­tion with many par­al­lel muta­tions occur­ring both with­in treat­ments and between treat­ments with shared selec­tion pressures.

Car­oline Turn­er is a postdoc­tor­al asso­ci­ate in the Cooper Lab at the Uni­ver­sity of Pitt­s­burgh. The study is freely avail­able to read and down­load from Evol­u­tion Let­ters.