I followed an in-vitro Directed Evolution protocol using two clinically-relevant carbapenemases. We recovered several mutations commonly observed in clinical settings, but also previously unknown large-effect ones. Importantly, strong epistatic constraints largely determined evolutionary outcomes in this system: we uncover at least four distinct adaptive pathways in which the identity of the first mutation markedly affected the identity of subsequent adaptive steps (see Figure 1). These independent pathways most probably representing different solutions to the activity-stability trade-off, which we are in the process of verifying through protein purification, kinetic parameters and stability assay. Furthermore, using the 6 most prevalent mutations observed in the random in-vitro evolution step, we created all their possible 64 combinations via site directed mutagenesis. Then, the resistance profile was evaluated by minimal inhibitory concentration with the aim of defining epistatic constrains. Despite all mutations being beneficial in the ancestral background, the highest value of MIC was obtained with only the combination of two mutations. Most of the combination of the mutations conferred a beneficial or neutral effect, while a few of the combination represent a deleterious effect. Taken together, our results illustrate how strong epistasis imposes a high degree of contingency in the evolutionary pathways of two important carbapenemases, and that its evolution can be predictable based on the identity of the few first-step, beneficial mutations.
During the three-month secondment, I conducted directed mutagenesis experiments on the chromosome expressed penicillin binding protein PBP3, a protein essential for cell growth and division and are therefore critical targets for β-lactam antibiotics. Several mutations in PBP3 conferring resistance to cephalosporins were described previously. Our objective was to evaluate if there were presented epistatic constrains in the mutations controlling the system. Using CRISPR/Cas9-mediated error prone genome editing methodology (CREPE), a technique that allows the high-throughput generation of mutants, I generated a library of mutants within a targeted gene in its native genomic context. Moreover, I performed reconstruction experiments demonstrating selective effects of candidate mutations separately and in combination.
