Experimental evolution of gene duplicates in a bacterial plasmid model.

The fate of gene duplicates subjected to diversifying selection was tested experimentally in a bacterial system. The wild-type TEM-1 beta-lactamase gene confers resistance to ampicillin but not to cefotaxime. Point mutations confer cefotaxime resistance, but they compromise ampicillin resistance. Thus, selection for both drug resistances in a bacterium with two copies of beta-lactamase should favor the divergence of one copy to improve cefotaxime resistance while maintaining the other copy to preserve ampicillin resistance. This selection was performed on a bacterium with identical sequences of beta-lactamase on two separate, compatible plasmids. As expected, one plasmid evolved increased cefotaxime resistance when appropriately strong cefotaxime selection was applied. However, the cefotaxime-resistant plasmid maintained sufficient ampicillin resistance to tolerate the concentration of ampicillin used, and the other plasmid was lost. Hosts carrying both the cefotaxime-resistant and wild-type plasmids were then subjected to various higher concentrations of both drugs to find conditions that would ensure the maintenance of both plasmids. In a striking contradiction to our model, no such conditions were found. The fitness cost of carrying both plasmids increased dramatically as antibiotic levels were raised, and either the wild-type plasmid was lost or the cells did not grow. This study highlights the importance of the cost of duplicate genes and the quantitative nature of the tradeoff in the evolution of gene duplication through functional divergence.