Free energy perturbation calculations on models of active sites: Applications to adenosine deaminase inhibitors

Abstract

AbstractFree energy perturbation calculations were performed to determine the free energy of binding associated with the presence of perhaps an unusual hydroxyl group in the transition state analog of nebularine, an inhibitor of the enzyme adenosine deaminase. The presence of a single hydroxyl group in this inhibitor has been found to contribute −9.8 kcal/mol to the free energy of binding, with a 108‐fold increase in the binding affinity by the enzyme. In this work, we calculate the difference in solvation free energy for the 1,6‐dihydropurine complex versus that of the 6‐hydroxyl‐1,6‐dihydropurine complex to determine if this marked increase in binding affinity is attributed to an unusually hydrophobic hydroxyl group. The calculated ΔG associated for the solvation free energy is −11.8 kcal/mol. This large change in the solvation free energy suggests that this hydroxyl is instead unusually hydrophilic and that the difference in free energy of interaction for the two inhibitors to the enzyme must be at least ca. 20 kcal/mol. Although the crystal structure for adenosine deaminase is currently not known, we attempt to mimic the nature of the active site by constructing models which simulate the enzyme‐inhibitor complex. We present a first attempt at determining the change in free energy of binding for a system in which structural data for the enzyme is incomplete. To do this, we construct what we believe is a minimal model of the binding between adenosine deaminase and an inhibitor. The active site is simulated as a single charged carboxyl group which can form a hydrogen bond with the hydroxyl group of the analog. Two different carboxyl anion models are used. In the first model, the association is modeled between an acetic acid anion and the modified inhibitor. The second model consists of a hydrophobic amino acid pocket with an interior Glu residue in the active site. From these models we calculate the change in free energy of association and the overall change in free energy of binding. We calculate the free energies of interaction both in the absence and presence of water. We conclude from this that the presence of a single suitably placed‐CO−2 group probably cannot explain the binding effect of the‐OH group and that additional interactions will be found in the adenosine deaminase active site.