A workflow for exploring ligand dissociation from a macromolecule: Efficient random acceleration molecular dynamics simulation and interaction fingerprint analysis of ligand trajectories

Abstract

The dissociation of ligands from proteins and other biomacromolecules occurs over a wide range of timescales. For most pharmaceutically relevant inhibitors, these timescales are far beyond those that are accessible by conventional molecular dynamics (MD) simulation. Consequently, to explore ligand egress mechanisms and compute dissociation rates, it is necessary to enhance the sampling of ligand unbinding. Random Acceleration MD (RAMD) is a simple method to enhance ligand egress from a macromolecular binding site that does not require the user to choose a ligand egress reaction coordinate. It thus enables the unbiased exploration of ligand egress routes. Furthermore, the RAMD procedure can be used to compute the relative residence times of ligands. When combined with a machine-learning analysis of protein-ligand interaction fingerprints (IFP), molecular features that affect ligand unbinding kinetics can be identified. Here, we describe the implementation of RAMD in GROMACS 2020, which provides significantly improved computational performance, with scaling to large molecular systems. For the automated analysis of RAMD results, we developed MD-IFP, a set of tools for the generation of IFPs along unbinding trajectories and for their use in the exploration of ligand dynamics. We demonstrate that the analysis of ligand dissociation trajectories by mapping them onto the IFP space enables the characterization of ligand dissociation routes and metastable states. The combined implementation of RAMD and MD-IFP provides a computationally efficient and freely available workflow that can be applied

The structure of Prp2 bound to RNA and ADP-BeF3−reveals structural features important for RNA unwinding by DEAH-box ATPases

“…To accelerate the dissociation of DHP from the complex, we used random-accelerated MD simulations (RAMD; Lü demann et al, 2000). The GROMACS code (version 2020.1) extended for RAMD was taken from https://github.com/ HITS-MCM/gromacs-ramd (Kokh et al, 2020). For the simulations reported in this study, we used the following RAMD settings.…”

G Protein-Coupled Receptor-Ligand Dissociation Rates and Mechanisms from τRAMD Simulations

“…The RAMD protocol for computing relative residence times was reported in Refs. 16,17 . Here, we briefly outline the main steps.…”

“…Each starting snapshot is then used to generate a series of at least 15 RAMD dissociation trajectories. We retained the default parameters of the RAMD protocol as described in detail elsewhere 46 : the ligand displacement was evaluated every 100 fs, and a ligand displacement of less than 0.025 Å led to selection of a new random force orientation; unbinding was considered to have occurred when the ligand-protein COM separation distance reached 65 Å; the trajectory coordinates were recorded every 2 ps. The length of the RAMD trajectory was limited to 24h wall-clock time by the set-up of the compute cluster used.…”

“…We recently developed the RAMD protocol for computing relative protein-ligand dissociation rates and exploring the mechanisms of protein-ligand unbinding 16,17,18 . The basis of RAMD is the Random Acceleration MD (RAMD) 19 simulation method, in which an additional randomly oriented force is applied in an adaptive manner to the ligand center of mass to facilitate the ligand unbinding process, thereby making simulation of dissociation of a drug-like compound possible on the nanosecond time-scale without a-priori knowledge of the egress route or mechanism or extensive parameter adjustment.…”

Contact Map Fingerprints of Protein-Ligand Unbinding Trajectories Reveal Mechanisms Determining Residence Times Computed from Scaled Molecular Dynamics

“…It led some of us to develop the τ -Random Acceleration Molecular Dynamics (τ -RAMD) 16 method for the evaluation of τ of a drug binding to its pharmacological target. It has been coupled with ML analysis 17 and MD-IFP (Molecular Dynamics Interaction Fingerprints) 18 for providing robust estimates of the residence time and insights into the features important for residence time using Machine Learning on the ligand's exit trajectories and applied to systems of varying levels of complexity. 19,20 Related less com-putationally intensive approaches that rely on the availability of a training set, such as COM-BINE analysis, 21 demonstrate remarkable performance by extracting protein-specific multilinear relationships between dissociation rate k of f and Lennard-Jones and Coulombic perresidue descriptors.…”