Force Field Optimization Guided by Small Molecule Crystal Lattice Data Enables Consistent Sub-Angstrom Protein–Ligand Docking

Abstract

Accurate and rapid calculation of protein-small molecule interaction free energies is critical for computational drug discovery. Because of the large chemical space spanned by drug-like molecules, classical force fields contain thousands of parameters describing atom-pair distance and torsional preferences; each parameter is typically optimized independently on simple representative molecules. Here, we describe a new approach in which small molecule force field parameters are jointly optimized guided by the rich source of information contained within thousands of available small molecule crystal structures. We optimize parameters by requiring that the experimentally determined molecular lattice arrangements have lower energy than all alternative lattice arrangements. Thousands of independent crystal lattice-prediction simulations were run on each of 1386 small molecule crystal structures, and energy function parameters of an implicit solvent energy model were optimized, so native crystal lattice arrangements had the lowest energy. The resulting energy model was implemented in Rosetta, together with a rapid genetic algorithm docking method employing grid-based scoring and receptor flexibility. The success rate of bound structure recapitulation in cross-docking on 1112 complexes was improved by more than 10% over previously published methods, with solutions within <1 Å in over half of the cases. Our results demonstrate that small molecule crystal structures are a rich source of information for guiding molecular force field development, and the improved Rosetta energy function should increase accuracy in a wide range of small molecule structure prediction and design studies.

Structural modeling of the hERG potassium channel and associated drug interactions

“…Although GALigandDock utilizes Rosetta’s latest betanov2016 energy function (Park et al, 2021), which accounts for torsional strain energies of the ligand, in addition to using fitted van der Waals distances, solvation potentials, and partial charges, it was difficult to make conclusions between drug conformation/binding modes and interface scores in our models. Comparative ligand interface scores and probability density plots for all 2000 docked dofetilide, terfenadine, and E4031 structures can be found in Figures 10 , 11 , and 12 , respectively.…”

“…Additionally, we show that non-inactivating mutations S620T and S641T block the inward shift of F627 via differing mechanisms, preventing potential conformational changes of the SF during inactivation. Lastly, hERG channel structural models exhibiting elements associated with their respective mutations (including the WT channel model) were combined with drug models of dofetilide, terfenadine and E4031 and subjected to docking studies using the Rosetta GALigandDock protocol (Park et al, 2021). Results show that while all drugs bound to key residues Y652 and F656 across all models, corroborating existing experimental evidence, only the S641A inactivation-enhancing mutant provided alternative state-dependent binding opportunities.…”

“…Additionally, we show that non-inactivating mutations S620T and S641T block the inward shift of F627 via differing mechanisms, preventing potential conformational changes of the SF during inactivation. Lastly, we used Rosetta GALigandDock protocol (Park et al, 2021) to explore interactions of dofetilide, terfenadine and E4031 with the WT and mutant hERG channel structural models. Our results show that while dofetilide, terfenadine and E4031 bound to key residues Y652 and F656 across all models, in agreement with existing experimental evidence, only the S641A inactivation-enhancing mutant models revealed alternative state-dependent drug binding.…”

Computational Enzyme Engineering Pipelines for Optimized Production of Renewable Chemicals

“…Still, engineering or design of an enzyme–substrate complex requires almost atomic-level accuracy, which may not be reached with the molecular docking methods mentioned previously. Recently, a novel approach called GALigandDock was reported that enabled small molecule docking with sub-angstrom accuracy in the Rosetta framework by force field optimization ( Park et al, 2021 ). GALigandDock outperformed the top current docking approaches in a cross-docking benchmark set providing new directions for automated enzyme–substrate complex modeling.…”

Efficient Hit-to-Lead Searching of Kinase Inhibitor Chemical Space via Computational Fragment Merging

“…Step V: Selecting top-scoring inhibitors Finally, we re-refine and re-score the top-scoring compounds from the preceding step, by energy minimization using Rosetta [36,37] (Figure 1b, Step V). This step serves to ensure that the selected substituents on the shared hinge-binding core are indeed compatible with one another, both structurally and energetically.…”