Evaluating the use of absolute binding free energy in the fragment optimisation process

Binding free energy calculations that make use of alchemical pathways are becoming increasingly feasible thanks to advances in hardware and algorithms. Although relative binding free energy (RBFE) calculations are starting to find widespread use, absolute binding free energy (ABFE) calculations are still being explored mainly in academic settings due to the high computational requirements and still uncertain predictive value. However, in some drug design scenarios, RBFE calculations are not applicable and ABFE calculations could provide an alternative. Computationally cheaper end-point calculations in implicit solvent, such as molecular mechanics Poisson–Boltzmann surface area (MMPBSA) calculations, could too be used if one is primarily interested in a relative ranking of affinities. Here, we compare MMPBSA calculations to previously performed absolute alchemical free energy calculations in their ability to correlate with experimental binding free energies for three sets of bromodomain–inhibitor pairs. Different MMPBSA approaches have been considered, including a standard single-trajectory protocol, a protocol that includes a binding entropy estimate, and protocols that take into account the ligand hydration shell. Despite the improvements observed with the latter two MMPBSA approaches, ABFE calculations were found to be overall superior in obtaining correlation with experimental affinities for the test cases considered. A difference in weighted average Pearson () and Spearman () correlations of 0.25 and 0.31 was observed when using a standard single-trajectory MMPBSA setup ( = 0.64 and = 0.66 for ABFE; = 0.39 and = 0.35 for MMPBSA). The best performing MMPBSA protocols returned weighted average Pearson and Spearman correlations that were about 0.1 inferior to ABFE calculations: = 0.55 and = 0.56 when including an entropy estimate, and = 0.53 and = 0.55 when including explicit water molecules. Overall, the study suggests that ABFE calculations are indeed the more accurate approach, yet there is also value in MMPBSA calculations considering the lower compute requirements, and if agreement to experimental affinities in absolute terms is not of interest. Moreover, for the specific protein–ligand systems considered in this study, we find that including an explicit ligand hydration shell or a binding entropy estimate in the MMPBSA calculations resulted in significant performance improvements at a negligible computational cost.

Binding free energy calculations predict the potency of compounds to protein binding sites in a physically rigorous manner and see broad application in prioritizing the synthesis of novel drug candidates. Relative binding free energy (RBFE) calculations have emerged as an industry-standard approach to achieve highly accurate rank-order predictions of the potency of related compounds; however, this approach requires that the ligands share a common scaffold and a common binding mode, restricting the methods' domain of applicability. This is a critical limitation since complex modifications to the ligands, especially core hopping, are very common in drug design. Absolute binding free energy (ABFE) calculations are an alternate method that can be used for ligands that are not congeneric. However, ABFE suffers from a known problem of long convergence times due to the need to sample additional degrees of freedom within each system, such as sampling rearrangements necessary to open and close the binding site. Here, we report on an alternative method for RBFE, called Separated Topologies (SepTop), which overcomes the issues in both of the aforementioned methods by enabling large scaffold changes between ligands with a convergence time comparable to traditional RBFE. Instead of only mutating atoms that vary between two ligands, this approach performs two absolute free energy calculations at the same time in opposite directions, one for each ligand. Defining the two ligands independently allows the comparison of the binding of diverse ligands without the artificial constraints of identical poses or a suitable atom-atom mapping. This approach also avoids the need to sample the unbound state of the protein, making it more efficient than absolute binding free energy calculations. Here, we introduce an implementation of SepTop. We developed a general and efficient protocol for running SepTop, and we demonstrated the method on four diverse, pharmaceutically relevant systems. We report the performance of the method, as well as our practical insights into the strengths, weaknesses, and challenges of applying this method in an industrial drug design setting. We find that the accuracy of the approach is sufficiently high to rank order ligands with an accuracy comparable to traditional RBFE calculations while maintaining the additional flexibility of SepTop.

We determined the effectiveness of absolute binding free energy (ABFE) calculations to refine the selection of active compounds in virtual compound screening, a setting where the more commonly used relative binding free energy approach is not readily applicable. To do this, we conducted baseline docking calculations of structurally diverse compounds in the DUD-E database for three targets, BACE1, CDK2 and thrombin, followed by ABFE calculations for compounds with high docking scores. The docking calculations alone achieved solid enrichment of active compounds over decoys. Encouragingly, the ABFE calculations then improved on this baseline. Analysis of the results emphasizes the importance of establishing high quality ligand poses as starting points for ABFE calculations, a nontrivial goal when processing a library of diverse compounds without informative co-crystal structures. Overall, our results suggest that ABFE calculations can play a valuable role in the drug discovery process.

Carbohydrates are key biological mediators of molecular recognition and signaling processes. In this case study, we explore the ability of absolute binding free energy (ABFE) calculations to predict the affinities of a set of five related carbohydrate ligands for the lectin protein, concanavalin A, ranging from 27-atom monosaccharides to a 120-atom complex-type N-linked glycan core pentasaccharide. ABFE calculations quantitatively rank and estimate the affinity of the ligands in relation to microcalorimetry, with a mean signed error in the binding free energy of -0.63 ± 0.04 kcal/mol. Consequently, the diminished binding efficiencies of the larger carbohydrate ligands are closely reproduced: the ligand efficiency values from isothermal titration calorimetry for the glycan core pentasaccharide and its constituent trisaccharide and monosaccharide compounds are respectively -0.14, -0.22, and -0.41 kcal/mol per heavy atom. ABFE calculations predict these ligand efficiencies to be -0.14 ± 0.02, -0.24 ± 0.03, and -0.46 ± 0.06 kcal/mol per heavy atom, respectively. Consequently, the ABFE method correctly identifies the high affinity of the key anchoring mannose residue and the negligible contribution to binding of both β-GlcNAc arms of the pentasaccharide. While challenges remain in sampling the conformation and interactions of these polar, flexible, and weakly bound ligands, we nevertheless find that the ABFE method performs well for this lectin system. The approach shows promise as a quantitative tool for predicting and deconvoluting carbohydrate-protein interactions, with potential application to design of therapeutics, vaccines, and diagnostics.

Absolute binding free energy calculations with explicit solvent molecular simulations can provide estimates of protein-ligand affinities, and thus reduce the time and costs needed to find new drug candidates. However, these calculations can be complex to implement and perform. Here, we introduce the software BAT.py, a Python tool that invokes the AMBER simulation package to automate the calculation of binding free energies for a protein with a series of ligands. The software supports the attach-pull-release (APR) and double decoupling (DD) binding free energy methods, as well as the simultaneous decoupling-recoupling (SDR) method, a variant of double decoupling that avoids numerical artifacts associated with charged ligands. We report encouraging initial test applications of this software both to re-rank docked poses and to estimate overall binding free energies. We also show that it is practical to carry out these calculations cheaply by using graphical processing units in common machines that can be built for this purpose. The combination of automation and low cost positions this procedure to be applied in a relatively high-throughput mode and thus stands to enable new applications in early-stage drug discovery.

Absolute binding free energy (ΔG) calculations can rank structurally diverse compounds, which could be useful for early-stage drug discovery. Unfortunately, for flexible systems, it can be challenging to sample the receptor conformations necessary to obtain converged ΔG calculations. Here, we address this challenge by leveraging extensive molecular dynamics simulations of apo SARS-CoV-2 main protease (MPro) that were conducted on the Folding@Home distributed computing system. A Markov state model (MSM) was built to compute the equilibrium probability of each snapshot. Representative snapshots were selected from clusters defined based on occupancy fingerprints of the catalytic site. The binding potential of mean force (BPMF), the binding free energy between a ligand and rigid receptor configuration, was computed between the representative snapshots and 130 drug leads from the COVID Moonshot, an open-source drug discovery project. ΔGs were computed using an exponential average of BPMFs based on implicit ligand theory (ILT). ΔG calculations recapitulated experimental values with a Pearson R of 0.55 and a mean-adjusted root-mean-square error of 1.6 kcal/mol. Accuracy and computational costs were found to be intermediate between docking and previous free energy calculations with a fully flexible receptor. Moreover, in 88% of systems, the calculated ΔG of the native binding pose (RMSD from crystallographic <3 Å) was within 1 kT of the top-ranked pose.

Accurate estimation of protein-ligand (PL) binding free energies is a crucial task in medicinal chemistry and a critical measure of PL interaction modeling effectiveness. However, traditional computational methods are often computationally expensive and prone to errors. Recently, deep learning (DL)-based approaches for predicting PL interactions have gained enormous attention, but their accuracy and generalizability are hindered by data scarcity. In this study, we propose LumiNet, a versatile PL interaction modeling framework that bridges the gap between physics-based models and black-box algorithms. LumiNet utilizes a subgraph transformer to extract multiscale information from molecular graphs and employs geometric neural networks to integrate PL information, mapping atomic pair structures into key physical parameters of non-bonded interactions in classical force fields, thereby enhancing accurate absolute binding free energy (ABFE) calculations. LumiNet is designed to be highly interpretable, offering detailed insights into atomic interactions within protein-ligand complexes, pinpointing relatively important atom pairs or groups. Our semi-supervised learning strategy enables LumiNet to adapt to new targets with fewer data points than other data-driven methods, making it more relevant for real-world drug discovery. Benchmarks show that LumiNet outperforms the current state-of-the-art model by 18.5% on the PDE10A dataset, and rivals the FEP+ method in some tests with a speed improvement of several orders of magnitude. We applied LumiNet in the scaffold hopping process, which accurately guided the discovery of the optimal ligands. Furthermore, we provide a web service for the research community to test LumiNet. The visualization of predicted inter-molecular energy contributions is expected to provide practical value in drug discovery projects.

Absolute binding free energy calculations for molecules binding to the GPCR GCGR at the protein/bilayer/water interface.

Recent advances in relative protein-ligand binding free energy calculations have shown the value of alchemical methods in drug discovery. Accurately assessing absolute binding free energies remains a challenging endeavour, mostly limited to small model cases. We demonstrate accurate absolute binding free energy estimates for 128 pharmaceutically relevant ligands across 7 proteins using a highly parallelizable non-equilibrium method. These calculations also provide detailed physical insight into the structural determinants of binding, identifying subtle rotamer rearrangements between protein apo and holo states that are crucial for binding. The challenge behind absolute binding free energy calculations stems in large part from the need to explicitly account for the protein’s apo state. In this work we present several approaches to obtain apo state ensembles, including a novel rigorous method to generate protein-ligand ensembles for the ligand in its decoupled state. Altogether, we present an effective open-source protocol for prospective application in drug discovery.

Absolute binding free energy (ABFE) calculations can be an important part of the drug discovery process by identifying molecules that have the potential to be strong binders for a biomolecular target. Recent work has used free energy perturbation (FEP) theory for these calculations, focusing on a set of 16 inhibitors of the severe acute respiratory syndrome coronavirus 2 main protease (Mpro). Herein, the same data set is evaluated by metadynamics (MetaD), four different docking programs, and molecular mechanics with generalized Born and surface area solvation. MetaD yields a Kendall τ distance of 0.28 and Pearson r2 of 0.49, which reflect somewhat less accuracy than that from the ABFE FEP results. Notably, it is demonstrated that an ensemble docking protocol by which each ligand is docked into the 13 crystal structures in this data set provides improved performance, particularly when docking is carried out with Glide XP (Kendall τ distance = 0.20, Pearson r2 = 0.71), Glide SP (Kendall τ distance = 0.19, Pearson r2 = 0.66), or AutoDock 4 (Kendall τ distance = 0.21, Pearson r2 = 0.55). The best results are obtained with "superconsensus" docking by averaging the 52 results for each compound using the 4 docking protocols and all 13 crystal structures (Kendall τ distance = 0.18, Pearson r2 = 0.73).

Alchemical absolute binding free energy (ABFE) calculations have substantial potential in drug discovery, but are often prohibitively computationally expensive. To unlock their potential, efficient automated ABFE workflows are required to reduce both computational cost and human intervention. We present a fully automated ABFE workflow based on the automated selection of λ windows, the ensemble-based detection of equilibration, and the adaptive allocation of sampling time based on inter-replicate statistics. We find that the automated selection of intermediate states with consistent overlap is rapid, robust, and simple to implement. Robust detection of equilibration is achieved with a paired t-test between the free energy estimates at initial and final portions of a an ensemble of runs. We determine reasonable default parameters for all algorithms and show that the full workflow produces equivalent results to a nonadaptive scheme over a variety of test systems, while often accelerating equilibration. Our complete workflow is implemented in the open-source package A3FE (https://github.com/michellab/a3fe).

Absolute Binding Free Energy Calculations Using Molecular Dynamics Simulations with Restraining Potentials

Apoptosis signal-regulating kinase 1 (ASK1), a key component of the mitogen-activated protein kinase (MAPK) cascades, has been identified as a promising therapeutic target owing to its critical role in signal transduction pathways. In this study, we proposed novel pyridin-2-yl urea inhibitors exhibiting favorable physicochemical properties. The potency of these compounds was validated through in vitro protein bioassays. The inhibition (IC50) of compound 2 was 1.55 ± 0.27 nM, which was comparable to the known clinical inhibitor, Selonsertib. To further optimize the hit compounds, two possible binding modes were initially predicted by molecular docking. Absolute binding free energy (BFE) calculations based on molecular dynamics simulations further discriminated the binding modes, presenting good tendency with bioassay results. This strategy, underpinned by BFE calculations, has the great potential to expedite the drug discovery process in the targeting of ASK1 kinase.

Alchemical absolute binding free energy calculations are of increasing interest in drug discovery. These calculations require restraints between the receptor and ligand to restrict their relative positions and, optionally, orientations. Boresch restraints are commonly used, but they must be carefully selected in order to sufficiently restrain the ligand and to avoid inherent instabilities. Applying multiple distance restraints between anchor points in the receptor and ligand provides an alternative framework without inherent instabilities which may provide convergence benefits by more strongly restricting the relative movements of the receptor and ligand. However, there is no simple method to calculate the free energy of releasing these restraints due to the coupling of the internal and external degrees of freedom of the receptor and ligand. Here, a method to rigorously calculate free energies of binding with multiple distance restraints by imposing intramolecular restraints on the anchor points is proposed. Absolute binding free energies for the human macrophage migration inhibitory factor/MIF180, system obtained using a variety of Boresch restraints and rigorous and nonrigorous implementations of multiple distance restraints are compared. It is shown that several multiple distance restraint schemes produce estimates in good agreement with Boresch restraints. In contrast, calculations without orientational restraints produce erroneously favorable free energies of binding by up to approximately 4 kcal mol-1. These approaches offer new options for the deployment of alchemical absolute binding free energy calculations.

Generalized Born (GB) implicit solvent models are widely used in molecular dynamics simulations to evaluate the interactions of biomolecular complexes. The continuum treatment of the solvent results in significant computational savings in comparison to an explicit solvent representation. It is, however, not clear how accurately the GB approach reproduces the absolute free energies of biomolecular binding. On the basis of induced dissociation by means of umbrella sampling simulations, the absolute binding free energies of small proline-rich peptide ligands and a protein receptor were calculated. Comparative simulations according to the same protocol were performed by employing an explicit solvent model and various GB-type implicit solvent models in combination with a nonpolar surface tension term. The peptide ligands differed in a key residue at the peptide-protein interface, including either a nonpolar, a neutral polar, a positively charged, or a negatively charged group. For the peptides with a neutral polar or nonpolar interface residue, very good agreement between the explicit solvent and GB implicit solvent results was found. Deviations in the main separation free energy contributions are smaller than 1 kcal/mol. In contrast, for peptides with a charged interface residue, significant deviations of 2-4 kcal/mol were observed. The results indicate that recent GB models can compete with explicit solvent representations in total binding free energy calculations as long as no charged residues are present at the binding interface.