Solvent effects on catechol's binding affinity: investigating the role of the intra-molecular hydrogen bond through a computational multi-level approach

The subtle interplay between the inter-molecular interactions established by catechol with the surrounding solvent and the intra-molecular hydrogen bond (HB) characterizing its conformational dynamics is investigated through a multi-level computational approach. First, quantum mechanical (QM) calculations are employed to accurately characterize both large portions of catechol's potential energy surface and the interaction energy with neighboring solvent molecules. The acquired information is thereafter exploited to develop a QM derived force-field (QMD-FF), in turn employed in molecular dynamics (MD) simulations based on classical mechanics. The reliability of the QMD-FF is further validated though the comparison with the outcomes of ab initio molecular dynamics, also purposely carried out in this work. In agreement with recent experimental findings, the MD results reveal remarkable differences in the conformational behavior between isolated and solvated catechol, as well as among the investigated solvents, namely water, acetonitrile or cyclohexane. The rather strong intramolecular HB, settled between the vicinal phenolic groups and maintained in the gas phase, looses stability when catechol is solvated in polar solvents, and is definitively lost in protic solvents as water. In fact, the internal energy increase associated with the rotation of one hydroxyl group and the breaking of the internal HB, is well compensated by the intermolecular HB network available when both phenolic hydrogens point toward the surrounding solvent. In such a case, catechol is stabilized in a chelating conformation, which in turn could be very effective in water removal and surface anchoring. Beside unraveling the role of the different contributors that govern catechol's conformational dynamics, the QMD-FF assembled in this work could be in future employed to model larger catechol's containing molecules, resorting to its accuracy to reliably model both internal flexibility and solvent effects, while exploiting MD computational benefits to include more complex players as for instance surfaces, ions or biomolecules.