Development of a Virtual Screening Pipeline for the Identification of Small Molecule PPIP5K2 Inhibitors

Abstract

Phosphate homeostasis is essential for survival in eucaryotes. It is mediated by inositol phosphates, which are synthesized by a sequence of kinases. One such kinase in humans is diphosphoinositol pentakisphosphate kinase 2 (PPIP5K2). However, the exact mechanism of phosphate homeostasis in general and PPIP5K2 specifically are still subject to research. To date knockouts and mutations are used for research, but an inhibitor for PPIP5K2 is heavily sought after. Bartsch and Fiedler, who carry out research concerning PPIP5K2, performed a high-throughput screening of 60 000 compounds revealing a single ATP-competitive inhibitor. With an IC50 of 15:5 M the affinity is comparably poor. The presented thesis aims at utilizing computer-aided drug design (CADD) methods to identify additional and ideally more potent PPIP5K2 ligands. We employed the binding mode of the lead compound to perform a pharmacophore-based virtual screening. A second approach deriving pharmacophore features from the PPIP5K2 apo structure was included to broaden the chemical space of potential active compounds further. We wrapped the pharmacophore screenings in a pipeline with additional CADD methods to evaluate the screening hits in more detail. Hits not fitting the binding pocket were discarded first. Then we used molecular docking to place the molecules in the binding pocket and filtered them based on matching the respective pharmacophore hypothesis and on the binding affinity score. Selected hits were simulated using molecular dynamics (MD) and the stability was evaluated by calculating the root-mean-square deviation (RMSD) and fluctuation (RMSF). We identified nine candidates from the lead-based screening and eight from the apo structure-based screening. All candidates conform to their respective pharmacophore, score sufficiently high and are stable in MD simulations. Proposed candidates were handed over to our collaboration partners at FMP and will be tested invitro. If confirmed active, these compounds will aid building a library of PPIP5K2 ligands to be used in research and for further virtual screening and lead optimization campaigns.