Quantum Mechanical and Molecular Mechanical Calculations on Substituted Boron Clusters and Their Interactions with Proteins

Abstract

Medicinal chemistry entails not only synthesis but also compound design. The role of rational drug design has been boosted in recent decades through the use of computer‐ aided drug design, either ligand‐ or structure‐based [1]. The former area makes heavy use of statistics and chemi‐informatics to set up dependencies between the physicochemical properties of the compounds and their biological activities. Although some properties, such as the lipophilicity (expressed as the n‐octanol/water partition coefficient, logP), can also be evaluated computationally by quantum mechanical (QM) or molecular mechanical (MM) methods, ligand‐based drug design is not the focus of this chapter. On the contrary, we discuss here structure‐based drug design, which uses three‐ dimensional (3D) structures of protein–ligand complexes to estimate affinities. The geometries are most often determined experimentally by X‐ray crystallography or nuclear magnetic resonance (NMR). In computer‐aided structure‐based drug design, the ligand’s binding pose within the protein is predicted by docking, a task that has practically been mastered for the broad organic chemistry space [2]. Scoring is thereafter used to assess which of the poses represent the native complex and to rank the compounds by affinity. In contrast, this task has, until recently, been deemed unsolved [3]. We have approached the low reliability of scoring by developing a QM‐based scoring function [4]. Its fundamental principle is QM treatment of protein–ligand noncovalent interactions and solvation [5]. We showed that such an approach can be advantageously used to unequivocally identify the ligand native pose [6], reproduce binding affinity in a series of ligands to various protein targets [5], and describe nonclassical noncovalent interactions, such as halogen bonds [7] or even covalently binding inhibitors [8]. Based on this extensive experience of ours with organic ligands and a decade‐long experience with calculations of boron clusters bound to proteins [9], we affirm that QM scoring is a general solution to the affinity prediction of boron cluster/protein binding. Quantum Mechanical and Molecular Mechanical Calculations on Substituted Boron Clusters and Their Interactions with Proteins Jindřich Fanfrlík, Adam Pecina, Jan Řezáč, Pavel Hobza, and Martin Lepšík*