Decoding molecular recognition of inhibitors targeting HDAC2 via molecular dynamics simulations and configurational entropy estimation.

Molecular recognition by enzymes is a complicated process involving thermodynamic energies governing protein-ligand interactions. In order to aid the estimation of inhibitory activity of compounds targeting an enzyme, several computational methods can be employed to dissect this intermolecular contact. Herein, we report a structural dynamics investigation of an epigenetic enzyme HDAC2 in differentiating its binding to various inhibitors within the sub-sites of its active site. Molecular dynamics (MD) simulation was employed to elucidate the intermolecular interactions as well as the dynamics behavior of ligand binding. MD trajectories of five distinct HDAC2-inhibitor complexes reveal that compounds lacking adequate contacts with the opening rim of the active site possess high fluctuation along the cap portion, thus weakening the overall affinity. Key intermolecular interactions determining the effective binding of inhibitors include hydrogen bonds with Gly154, Asp181, and Tyr308; hydrophobic interactions between Phe155/Phe210 and the linker region; and a pi-stacking with Arg39 at the foot pocket. Decomposition of the binding free energy calculated per-residue by MM/PBSA also indicates that the interactions within the internal foot pocket, especially with residues Met35, Leu144, Gly305, and Gly306, can contribute significantly to the ligand binding. Additionally, configurational entropy of the binding was estimated and compared to the scale of the binding free energy in order to assess its contribution to the binding and to differentiate various ligand partners. It was found that the levels of entropic contribution are comparable among a set of structurally similar carbamide ligands, while it is greatly different for the set of unrelated ligands, ranging from 2.75 to 16.38 kcal/mol for the five inhibitors examined. These findings exemplify the importance of assessing molecular dynamics as well as estimating the entropic contribution in evaluating the ligand binding mechanism.