Machine Learning Guided Ligand-Protein Simulation Approach Elucidates the Binding Mechanism of Abscisic Acid

Abstract

Rising temperatures and climate change are threatening agricultural productivity and have motivated efforts to elucidate the molecular mechanisms of drought resistance in plants. Abscisic acid (ABA) is a key plant hormone that imparts drought resistance in plants. Crystal structures of receptors (PYLs) involved in ABA perception have shed light on the interactions ABA makes in the binding pocket of the receptor but these structures provide limited knowledge about the dynamic mechanism of ABA binding. Mechanistic understanding of ABA recognition by plants could provide new avenues for development of drought resistant agrochemicals. In this work, we have performed extensive atomistic molecular dynamics (MD) simulations using Markov state models (MSMs) based adaptive sampling protocols to characterize ABA binding pathway for two ABA receptors in A. Thaliana, AtPYL5 and AtPYL10 receptors belonging to different sub-classes of ABA receptors to identify the major bottlenecks in their binding pathways. Our results not only explain the binding mechanism of ABA molecule but also provide atomistic information about the intermediate states along the binding pathways. We have also identified a new non-productive pose, which we call the inverted state, where ABA gets trapped in the binding pocket while forming an unproductive complex. We validate our results with previously published NMR and Hydrogen/Deuterium exchange experiments to show that receptors retain their flexibility even after ABA binding. Our findings shed light on the role of specific receptor residues in ABA binding, explore role of water in binding and characterize major barriers to ABA binding. Our results demonstrate the efficacy of MD simulations and MSM framework in elucidating the mechanism of ABA signaling. Knowledge of ABA binding mechanism will help in improving drought resistance in plants by informing genetic manipulations and agrochemical discovery efforts.