How replicable are evolutionary outcomes? For example, can we expect the same changes in gene expression in replicate populations of an organism evolving at an extremely hot temperature? Riehle and colleagues from the University of California at Irvine addressed this question by growing replicate lines of E. coli at a stressfully hot temperature over 2000 generations and comparing gene expression in these heat-adapted lines to gene expression in ancestral E. coli that had evolved at a more comfortable temperature. Their results shed some light on the question `to what extent are evolutionary outcomes repeatable?'
The strength and uniqueness of the approach Riehle's team pursued are its replication of the evolutionary process. E. coli replicate fast and easily under laboratory conditions. In order to quantify the evolutionary outcome of three E. coli lines evolving at 41.5°C, the authors isolated E. coli gene messages and reverse transcribed them into cDNA. Then, to identify which genes are differentially expressed in these heat-adapted lines, they compared the quantitative abundance of almost 2000 gene transcripts of these evolved, thermally adapted E. coli lines with the message abundance of an ancestral E. coli strain that lived at 37°C. The authors' analysis focused on genes that showed a statistical significance based on the variation in expression rather than the conventional two-fold change in expression. They found nine promising `hot temperature'genes, which changed in the same direction in all three thermally adapting lines.
Among the nine genes that showed highly replicable changes in expression were two RNA chaperones of the cspA (cold shock protein) family, two transcripts that are actually not cold inducible, unlike other cold shock proteins, but constitutively expressed at 37°C. Perhaps unsurprisingly,transcripts of the gapA gene for the glycolytic protein GAPDH, which is known to enhance cell growth at extreme temperatures, were also upregulated. In addition, the authors found that all three high-temperature lines showed high expression of transcripts of proteins that are involved in oxidative stress (qpiB), redox reactions (trxA or thioredoxin) and the pyrimidine salvage pathway (upp), as well as a lipoprotein and two proteins of unknown function. Finally, they noted that 30 additional genes changed expression in two of the three lines.
To see the forest despite all those trees, the authors looked at changes in transcript expression of functionally related gene groups in addition to single genes. The authors assume that adaptive change can arise not just through one specific change in gene expression but also through several different single changes in expression, as long as the genes are members of a functionally related group. Interestingly, they found a much higher level of replicable evolutionary changes at the level of functional groups (14%) than single genes (3.8%). Thus, repeatable adaptive changes may arise from different genes that belong to the same functional group.
Gene groups that were expected to, and indeed did, change include those that function at temperature extremes, regulate the extracytoplasmic stress response and are involved in protein folding. Somewhat more surprising are the replicable evolutionary changes that the authors observed for ribosomal proteins and the translational machinery, since we consider these to be among the most evolutionarily conserved cellular processes. In summary, the authors show that the degree to which the evolutionary process repeats itself depends on our perspective. Here, it was a functional gene group perspective versus a single gene-centered view. The outcome might be different if we could look at changes in the interactions among gene products at the system level.
