Quantifying the hydrogen-bond propensity of drugs and its relationship with Lipinski's rule of five

Abstract

We define Vn as the molecular electrostatic potential (MESP) at the nucleus of a free atom, calculated excluding the contribution of that nucleus charge. The Vn value at the nucleus of a free atom and that within the atom-in-molecule state show substantial difference (ΔVn) due to alterations in the electronic configuration of the atom resulting from molecular bonding. Thus, variations in molecular structure and bonding lead to differing ΔVn values for the same type of atoms in molecules. The Vn analysis focuses on frequently occurring ring structures in drugs and several drug molecules using density functional theory (DFT) at the M06L/6-311++G(d,p) level. The summation of ΔVn for all heteroatoms acting as hydrogen bond acceptors (HBA) yields ΣΔVn(A), while ΣΔVn(H) represents the cumulative sum of ΔVn for hydrogen atoms participating in hydrogen bond donors (HBD) and those in C–H bonds near heteroatoms. The total hydrogen bond propensity of a drug molecule (EHBP) is predicted by EHBP = 0.05 (ΣΔVn(A) − ΣΔVn(H)) − 7.68n. This equation is derived using interaction energy (ΣEint) calculations for ‘n’ configurations of twenty ring molecule⋯H2O complexes. The EHBP and ΣEint showed a strong linear relationship with R = 0.974. Further, the EHBP equation is validated using the ΣEint data of FDA-approved oral drugs. The EHBP prediction, solely based on MESP data, ranges from −2.3 to −131.5 kcal mol−1 for 193 FDA-approved oral drugs. Most orally administered drugs adhering to Lipinski's rule of five criteria typically fall within the EHBP range of −16.0 to −84.0 kcal mol−1. The study also derived a relationship between EHBP and Lipinski's rule of five (Ro5) parameters by incorporating a new parameter that accounts for hydrogen atoms participating in Ow⋯H interactions from CH groups. This research presents a robust MESP-based approach for predicting hydrogen bond propensity in drug molecules, complementing drug design and optimization efforts. The correlation with Lipinski's rule of five parameters underscores the importance of hydrogen bond interactions in drug-likeness.