Systematic Computational Design and Identification of Low Picomolar Inhibitors of Aurora Kinase A

Abstract

Aurora kinase A (AKA) has served as an effective molecular target for the development of cancer therapeutics. A series of potent AKA inhibitors with the (4-methoxy-pyrimidin-2-yl)-phenyl-amine (MPPA) scaffold are identified using a systematic computer-aided drug design protocol involving structure-based virtual screening, de novo design, and free energy perturbation (FEP) simulations. To enhance the accuracy of the virtual screening to find a proper molecular core and de novo design to optimize biochemical potency, we preliminarily improved the scoring function by implementing a reliable hydration energy term. The overall design strategy proves successful to the extent that some inhibitors reveal exceptionally high potency at low picomolar levels; this was achieved by substituting phenyl, chlorine, and tetrazole moieties on the MPPA scaffold. The establishment of bidentate hydrogen bonds with backbone groups in the hinge region appears to be necessary for the high biochemical potency, consistent with the literature X-ray crystallographic data. The picomolar inhibitory activity also stems from the simultaneous formation of additional hydrogen bonds with the side chains of the hinge region and P-loop residues. The FEP simulation results show that the inhibitory activity surges to the low picomolar level because the interactions in the ATP-binding site of AKA become strong by structural modifications enough to overbalance the increase in dehydration cost. Because of the exceptionally high biochemical potency, the AKA inhibitors reported in this study are anticipated to serve as a new starting point for the discovery of anticancer medicine. © 2018 American Chemical Society