Abstract
Computational capabilities are rapidly increasing, primarily because of the availability of GPU-based architectures. This creates unprecedented simulative possibilities for the systematic and robust computation of thermodynamic observables, including the free energy of a drug binding to a target. In contrast to calculations of relative binding free energy, which are nowadays widely exploited for drug discovery, we here push the boundary of computing the binding free energy and the potential of mean force. We introduce a novel protocol that leverages enhanced sampling, machine learning, and ad hoc algorithms to limit human intervention, computing time, and free parameters in free energy calculations. We first validate the method on a host–guest system, and then we apply the protocol to glycogen synthase kinase 3 beta, a protein kinase of pharmacological interest. Overall, we obtain a good correlation with experimental values in relative and absolute terms. While we focus on protein–ligand binding, the strategy is of broad applicability to any complex event that can be described with a path collective variable. We systematically discuss key details that influence the final result. The parameters and simulation settings are available at PLUMED-NEST to allow full reproducibility.
Highlights
Inchemistry, free energy is still the most relevant and challenging physiochemical parameter to predict computationally
The advances in free energy perturbation (FEP) have enabled the frequent application of FEP in drug discovery to estimate the relative binding free energy (RBFE).[18−20] When FEP simulations are applied to predict RBFEs, the ligand is alchemically transformed into another one through intermediate steps
The procedure is detailed in the following paragraphs, and it can be summarized in the main steps listed below: i) Generation of a molecular dynamics (MD) trajectory describing the rare event under investigation, for example, the association/ dissociation of protein−ligand complexes; ii) Identification of a preliminary minimum free energy path by a machine-learning path-finding algorithm[42] and optimization of the distance between consecutive frames (i.e., root mean square displacement (RMSD)) by the equidistant waypoints algorithm; iii) Reconstruction of the potential of mean force (PMF) by WT-MetaD on path CVs (PCVs); iv) Estimation of the standard binding free energy by processing the PMF plus the standard volume correction via a NanoShaper-based[40] technique, purposely developed for the present study
Summary
In (bio)chemistry, free energy is still the most relevant and challenging physiochemical parameter to predict computationally. Because free energy is a state function, the choice of the intermediate states is arbitrary, making the approach very flexible.[21,22] Recent progress in computer hardware and software has made it feasible to apply FEP (or other alchemical) methodologies to absolute binding free energy (ABFE) predictions,[1,23−26] creating the possibility of directly comparing the binding affinities across chemically different molecules that bind the same target or targets of the same family.[27,28] attractive, the routine application of FEP approaches to ABFE calculations is still limited because they do not fully consider how key phenomena (e.g., induced fit and desolvation) contribute to the binding affinity.[29,30] Their broad application to drug discovery is limited by the higher computational cost of ABFE relative to RBFE studies. FEP provides minimal details about binding intermediates, transient pockets, and molecular mechanisms because these calculations rely on unphysical paths (i.e., the alchemical transformations)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
