Redesigned GAP to activate oncogenic GAP

Ras-positive cancer constitutes a major challenge for medical treatment. Hot spot residues Gly12, Gly13 and Gln61 constitute the majority of oncogenic mutations which are associated with detrimental clinical prognosis. Here we pre-sent a two-step mechanism of GTP hydrolysis of the wild type Ras.GAP complex using QM/MM free energy calculations with the finite-temperature string method. We found that the deprotonation of the catalytic water takes place via the Gln61 as a transient Bornsted base. We obtained reaction profiles for key oncogenic Ras mutants G12D and G12C, reproducing the experimentally observed loss of catalytic activity, and validating our reaction mechanism. Using the optimized reaction path,we devised a fast and accurate simplified QM/MM reaction path optimization procedure, to design GAP mutants that activate G12D Ras. We identified 10 GAP residues that we mutated to any other possible amino acids (except for Gly), and the activation barrier was determined for 180 single mutants. Our simplified protocol gave excellent accuracy with the full QM/MM optimized paths for all but 1 outlier on top selected GAP mutants. To further enable ultra-fast screening, we built a machine learning framework to perform a fast prediction of the barrier heights, which was tested both on the single mutation data as well as on top predicted double mutations. Our approach enables a fast and accurate screening at the level of DFT-based QM/MM reaction path optimizations to design protein sequences that help restore catalytic activity of oncogenic Ras.