Drug resistance resulting from mutations to an enzyme target is a common phenomenon that limits the lifetime of many otherwise successful drugs. Often, these mutations are discovered after clinical exposure, forcing the discovery process to progress through additional cycles. In contrast, it would be useful to be able to predict resistance mutations early in the drug design process and to incorporate that knowledge to arrive at stronger preclinical drug candidates. Toward this goal, we have applied an ensemble-based protein design algorithm to predict potential resistance mutations toward novel antifolates that target dihydrofolate reductase from Staphylococcus aureus. Positive design criteria force the maintenance of catalytic activity; negative design decreases the affinity of a lead compound.  We then created and evaluated four predicted enzymes with double mutations. Three of these double mutants maintain activity and exhibit lower affinity for the lead; a crystal structure of the top-ranked mutation reveals the basis for the loss of potency. In a refined strategy, enzymes with single nucleotide polymorphisms (SNPs) were predicted to be resistant to the lead. Of the four mutant enzymes evaluated, all four maintain activity at levels equal to the wild-type enzyme and two exhibit 16-68-fold losses in potency.

Toward the prediction of resistance mutations

243rd National Meeting (2012)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.