Insight into binding mode of nitrile inhibitors of Plasmodium falciparum Falcipain-3, QSAR and Pharmacophore models, virtual design of new analogues with favorable pharmacokinetic profiles

We have carried out computational optimization of antiparasitic azadipeptide nitrile inhibitors (AZN) of falcipain-3 (FP3) of Plasmodium falciparum (Pf), a cysteine protease of the papain superfamily, using structurebased drug design and computer-assisted combinatorial chemistry. Three-dimensional (3D) models of complexes of inhibitor -FP3 for a training sets of published AZN analogs with experimentally determined inhibitory potencies were prepared by in situ modification of the crystal structure of PfFP3 inhibited by K11017 (Protein Data Bank entry 3BWK). We have used molecular mechanics, conformational searching and implicit solvation model to compute Gibbs free energies of inhibitor -FP3 receptor complex formation and built quantitative structure-activity relationships (QSAR) model by correlating the experimental inhibitory potencies with the computed binding affinities. The model was able to explain 97% of the FP3 inhibition data variation and was further validated with help of 3D-QSAR pharmacophore model generation (PH4). Structural information obtained from the 3D models of the AZN -FP3 complexes and the PH4 guided us in designing virtual combinatorial libraries of novel AZN analogs. Comparative analysis of the active site interactions directed us in the selection of building blocks used in the libraries. The initial virtual library was focused by means of computationally predicted oral bioavailability and subsequently in silico screened with the PH4 pharmacophore models to identify new AZN inhibitor candidates. Their inhibitory activities predicted by the QSAR model fall into the low nanomolar concentration range.