Event: QuantumBio to Present at the 2024 CCG UGM

Title: X-ray/Cryo-EM density-driven structure preparation for CADD and AlphaFold/OpenFold model feedstock: true MM & QM/MM real-space refinement in MOE/DivCon​

QuantumBio’s CEO, Dr. Lance Westerhoff, will be attending the Chemical Computing Group UGM & Conference 2024 in Montreal, Canada, from June 25 through June 28.

Abstract:

Successful structure-based drug discovery/design campaigns are dependent upon accurate protein:ligand structure determination and characterization. Gaining an understanding of the protein:ligand complex structure along with the proper protonation and explicit solvent effects is crucial for obtaining meaningful results from docking, thermodynamic calculations, active site exploration, and ultimately lead optimization.

Typically, macromolecular X-ray/Cryo-EM model refinement procedures utilize simplistic restraints localized on bond lengths, angles, and so on to ensure the reasonable geometry of the model after the refinement. A significant drawback of using conventional stereochemical restraints is that these intra-molecular restraints, no matter how accurate, do not account for critical atomic, intra- and inter-molecular interactions such as hydrogen bonds, dispersion, electrostatics, polarization, and charge transfer. This leads to errors and omissions in the models which subsequent users of these models (e.g. Computational and Medicinal Chemists) address through protonation and pKa “guesstimates” and data-independent molecular mechanics (MM) and quantum mechanics (QM) based structure minimization techniques.

Our work at QuantumBio has led to a sea change in our ability to accurately refine protein:ligand structure through the tight integration of the DivCon, linear-scaling, semiempirical, quantum mechanics (SE-QM) engine with crystallographic methods. This technology leads directly to improved models that address both the concerns of Structural Biologists (e.g. agreement with density, accurate biochemistry, etcetera) and those of Computational and Medicinal Chemists (e.g. low ligand and biomolecular strain, improved understanding of protein:ligand interactions, etcetera).

With these methods, not only can we better refine the structure, but we can also use model agreement with density to accurately determine tautomer/protomer states, flip/rotamer states, solvation characteristics, and binding modes found within the protein:ligand complex. For example, it is generally extremely difficult to experimentally determine the tautomeric state of the ligand and of the surrounding active site – even at higher resolutions. Theoretically, protonation can be established using neutron diffraction; however, experimental requirements such as reliance on very large crystals and on deuterium exchange limit the method’s suitability in SBDD. With XModeScore, we found that SE-QM-based macromolecular refinement is highly sensitive to the choice of a protonation state/tautomer form of ligands and residues. With this method, we are able to determine the correct protonation states – and by extension, rotomers, stereoisomers, flip-states, and so on – based on routine data even at lower resolutions commonly found in SBDD.

[1] Borbulevych, O. Y., Plumley, J. A., Martin, R. I., Merz K. M. & Westerhoff, L. M. (2014). Acta Cryst., D70, 1233.
[2] Borbulevych, O. Y., Martin, R. I. & Westerhoff, L. M. (2018). Acta Cryst., D74, 1063.
[3] Borbulevych, O. Y., Martin, R. I. & Westerhoff, L. M. (2021). J. Comput. Aided Mol. Des., 35, 433.
[4] Zheng, Z., Borbulevych, O. Y., Liu, H., Deng, J., Martin, R. I. & Westerhoff, L. M. (2020). J. Chem. Inf. Model., 60, 5437.