Computation of Absolute Binding Free Energies for Noncovalent Inhibitors with SARS-CoV-2 Main Protease.

Accurate, routine calculation of absolute binding free energies (ABFEs) for protein-ligand complexes remains a key goal of computer-aided drug design since it can enable screening and optimization of drug candidates. For development and testing of related methods, it is important to have high-quality datasets. To this end, from our own experimental studies, we have selected a set of 16 inhibitors of the SARS-CoV-2 main protease (Mpro) with structural diversity and well-distributed BFEs covering a 5 kcal/mol range. There is also minimal structural uncertainty since X-ray crystal structures have been deposited for 12 of the compounds. For methods testing, we report ABFE results from 2 µs molecular dynamics (MD) simulations using free energy perturbation (FEP) theory. The correlation of experimental and computed results is encouraging, with a Pearson's r2 of 0.58 and a Kendall τ of 0.24. The results indicate that current FEP-based ABFE calculations can be used for identification of active compounds (hits). While their accuracy for lead optimization is not yet sufficient, this activity remains addressable in separate lead series by relative BFE calculations.

Assessing Metadynamics and Docking for Absolute Binding Free Energy Calculations Using Severe Acute Respiratory Syndrome Coronavirus 2 Main Protease Inhibitors.

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).

Ensemble Geometric Deep Learning of Aqueous Solubility.

Geometric deep learning is one of the main workhorses for harnessing the power of big data to predict molecular properties such as aqueous solubility, which is key to the pharmacokinetic improvement of drug candidates. Two ensembles of graph neural network architectures were built, one based on spectral convolution and the other on spatial convolution. The pretrained models, denoted respectively as SolNet-GCN and SolNet-GAT, significantly outperformed the existing neural networks benchmarked on a validation set of 207 molecules. The SolNet-GCN model demonstrated the best performance on both the training and validation sets, with RMSE values of 0.53 and 0.72 log molar unit and Pearson r2 values of 0.95 and 0.75, respectively. Further, the ranking power of the SolNet models agreed well with a QM-based thermodynamic cycle approach at the PBE-vdW level of theory on a series of benzophenylurea derivatives and a series of benzodiazepine derivatives. Nevertheless, testing the resultant models on a set of inhibitors of the macrophage migration inhibitory factor (MIF) illustrated that the inclusion of atomic attributes to discriminate atoms with a higher tendency to form intermolecular hydrogen bonds in the crystalline state and to identify planar or nonplanar substructures can be beneficial for the prediction of aqueous solubility.

OPLS/2020 Force Field for Unsaturated Hydrocarbons, Alcohols, and Ethers.

The OPLS all-atom force field was updated and applied to modeling unsaturated hydrocarbons, alcohols, and ethers. Testing has included gas-phase conformational energetics, properties of pure liquids, and free energies of hydration. Monte Carlo statistical mechanics (MC) calculations were used to model 60 liquids. In addition, a robust, automated procedure was devised to compute the free energies of hydration with high precision via free-energy perturbation (FEP) calculations using double annihilation. Testing has included larger molecules than in the past, and parameters are reported for the first time for some less common groups including alkynes, allenes, dienes, and acetals. The average errors in comparison with experimental data for the computed properties of the pure liquids were improved with the modified force field (OPLS/2020). For liquid densities and heats of vaporization, the average unsigned errors are 0.01 g/cm3 and 0.2 kcal/mol. The average error and signed error for free energies of hydration are both 1.2 kcal/mol. As noted before, this reflects a systematic overestimate of the hydrophobicity of organic molecules when the parametrization is done to minimize the errors for properties of pure liquids. Implications for the modeling of biomolecular systems with standard force fields are considered.