BioCentury
ARTICLE | Distillery Techniques

Other

November 2, 2016 9:47 PM UTC

A computational model for predicting how mutations alter activity could aid the design of therapies that overcome drug resistance. For any protein-compound pair, the model utilizes molecular dynamic simulations to build a library of the non-static conformational states for each variant of the protein containing one or more point mutations, then performs in silico docking between each conformational state and the compound to identify variants that adopt high binding affinity conformations. When applied to the β-lactamase CD248 endosialin (TEM1) and 14 variants, the model identified several TEM1 variants predicted to have high activity against the β-lactam antibiotic cefotaxime, and results were confirmed by expressing each variant in E. coli and measuring its catalytic activity against cefotaxime. Next steps includes using results from the model to design non-covalent inhibitors of TEM1...