Molecular Simulations of the Binding of Small Molecule Inhibitors Targeting the Active Site of Bacterial RNase P

With the recent development of the first generation of RNA-targeting drugs, such as risdiplam, it has been increasingly of interest to understand the factors that influence ligand binding in nucleic acid systems. Nucleic acid targets pose unique challenges to existing drug discovery workflows that have traditionally focused on protein targets. In the present work we extend the capability of state-of-the-art workflows that employ alchemical free energy simulations to applications involving RNA-targeting small molecules. These workflows integrate recently developed alchemical enhanced sampling, optimized phase space overlap lambda schedules, and new methods for treating end states using reservoir replica exchange to achieve unprecedented precision in binding free energy predictions. Accuracy of these predictions is enhanced through quantum mechanical+machine learning corrections achieved efficiently through an indirect “book-ending” approach that uses non-equilibrium switching simulations integrated into the workflow. As an example application, we examine bacterial ribonuclease P (RNase P), an RNA enzyme that performs an essential role in the maturation of tRNA during translation and potential antibiotic target. We use a combination of all-atom molecular dynamics and quantum mechanical/molecular mechanical simulations to explore binding of a newly discovered small molecule inhibitor which binds to the active site of RNase P and traps the ES complex. These simulations are then used to identify key interactions between the ligand and RNase P that are hypothesized to drive ligand-dependent inhibition observed in experiment. These results are then used to guide the design of new inhibitor analogs that can be tested using alchemical simulations. The extended free energy workflow offers a powerful new tool for the design of small molecule drugs that target nucleic acids.