Constructing and analyzing the fitness landscape of an experimental evolutionary process

Abstract

Iterative saturation mutagenesis (ISM) is a promising approach to more efficient directed evolution, especially for enhancing the enantioselectivity and/or thermostability of enzymes. This was demonstrated previously for an epoxide hydrolase (EH), after five sets of mutations led to a stepwise increase in enantioselectivity. This study utilizes these results to illuminate the nature of ISM, and identify the reasons for its operational efficacy. By applying a deconvolution strategy to the five sets of mutations and measuring the enantioselectivity factors (E) of the EH variants, DeltaDeltaG( not equal) values become accessible. With these values, the construction of the complete fitness-pathway landscape is possible. The free energy profiles of the 5!=120 evolutionary pathways leading from the wild-type to the best mutant show that 55 trajectories are energetically favored, one of which is the originally observed route. This particular pathway was analyzed in terms of epistatic effects operating between the sets of mutations at all evolutionary stages. The degree of synergism increases as the stepwise evolutionary process proceeds. When encountering a local minimum in a disfavored pathway, that is, in the case of a dead end, choosing another set of mutations at a previous stage puts the evolutionary process back on an energetically favored trajectory. The type of analysis presented here might be useful when evaluating other mutagenesis methods and strategies in directed evolution.