Revealing the Ligand Binding Landscape with Advanced Molecular Simulation Methods

项目总结

Project Summary Human bodies are living systems that are constantly in flux. Pharmaceutical drugs take action within this non-equilibrium context: after a drug is ingested it is absorbed, distributed to tissues, bound (both on-target and off-target), released, metabolized and eliminated. Each of these processes occurs with a rate, and the efficacy of a drug is a largely function of these rates. In contrast, the dominant paradigm in drug discovery has been the optimization of affinity, which alone is insufficient to determine the rates of binding (kon) and unbinding (koff). Though the binding affinity is the ratio of the koff and kon, and longer residence times can lead to higher binding affinity, these are not well-correlated, as kon values can vary from diffusion-limited (109 M-1 s-1) down to <104 M-1 s-1 for protein targets with slow degrees of freedom, such G-protein-coupled receptors. Prediction of affinity is easier than kinetics as it is a state function, which depends only on the endpoints of the binding path. Binding kinetics are dependent on the molecular details encoded in the ligand binding transition state – the highest point in free energy along the binding pathway. Molecular dynamics simulation can be used to study these transition states in atomic detail, but only recently – empowered by advances in hardware and new algorithms for simulation – has it become capable of simulating unbiased ligand binding and release events, which can be coupled to long timescale protein motions. As such, little is known about the ligand binding transition state for a given protein target, and how it changes from ligand to ligand. Empowered by the WExplore enhanced sampling method (developed by the PI), the Dickson laboratory will use molecular dynamics simulation to reveal the landscape of protein-ligand conformations. WExplore can generate extremely rare ligand release pathway ensembles (events occurring only once in ~1000 seconds) without the use of biasing forces, which is a dramatic improvement upon current technology. Importantly, this will enable analysis of the ligand binding transition states for a series of ligands on two protein drug targets (soluble epoxide hydrolase (sEH), and Translocator protein 18kDA (TSPO)). This will mark the first study of the robustness of ligand binding transition states, which is a key quantity for kinetics-based drug design. Further, this work will build a method to encode properties of the transition state into screening tools that can, for the first time, screen ligands according to kinetics in a high-throughput manner. These methods will then be applied to identify new long residence time inhibitors for both sEH and TSPO, two systems where residence time has been shown to be important for drug efficacy.