Aqueous Desolvation and Molecular Recognition: Experimental and Computational Studies of a Novel Host-Guest System Based on Cucurbituril
Molecular recognition is arguably the most elementary physical process essential for life that arises at the molecular scale. Molecular recognition drives events across virtually all length scales, from the folding of proteins and binding of ligands, to the organization of membranes and the function of muscles. Understanding such events at the molecular level is massively complicated by the unique medium in which life occurs: water. In contrast to recognition in non-aqueous solvents, which are driven largely by attractive interactions between binding partners, binding reactions in water are driven in large measure by the properties of the medium itself. Aqueous binding involves the loss of solute-solvent interactions (desolvation) and the concomitant formation of solute-solute interactions. Despite decades of research, aqueous binding remains poorly understood, a deficit that profoundly limits our ability to design effective pharmaceuticals and new enzymes. Particularly problematic is understanding the energetic consequences of aqueous desolvation, an area the Toone and Beratan groups have considered for many years.
In this dissertation, we embark on a quest to shed new light on aqueous desolvation from two perspectives. In one component of this research, we improve current computational tools to study aqueous desolvation, employing quantum mechanics (QM), molecular dynamics (MD) and Monte Carlo (MC) simulations to better understand the behavior of water near molecular surfaces. In the other, we use a synthetic host, cucurbituril (CB), in conjunction with a de novo series of ligands to study the structure and thermodynamics of aqueous desolvation in the context of ligand binding with atomic precision, a feat hitherto impossible. A simple and rigid macrocycle, CB alleviates the drawbacks of protein systems for the study of aqueous ligand binding, that arise from conformational heterogeneity and prohibitive computational costs to model.
We first constructed a novel host-guest system that facilitates internalization of the trimethylammonium (methonium) group from bulk water to the hydrophobic cavity of CB with precise (atomic-scale) control over the position of the ligand with respect to the cavity. The process of internalization was investigated energetically using isothermal titration microcalorimetry and structurally by nuclear magnetic resonance (NMR) spectroscopy. We show that the transfer of methonium from bulk water to the CB cavity is accompanied by an unfavorable desolvation enthalpy of just 0.49±0.27 kcal*mol-1, a value significantly less endothermic than those values suggested from previous gas-phase model studies. Our results offer a rationale for the wide distribution of methonium in biology and demonstrate important limitations to computational estimates of binding affinities based on simple solvent-accessible surface area approaches.
To better understand our experimental results, we developed a two-dimensional lattice model of water based on random cluster structures that successfully reproduces the temperature-density anomaly of water with minimum computational cost. Using reported well-characterized ligands of CB, we probed water structure within the CB cavity and identified an energetically perturbed cluster of water. We offer both experimental and computational evidence that this unstable water cluster provides a significant portion of the driving force for encapsulation of hydrophobic guests.
The studies reported herein shed important light on the thermodynamic and structural nature of aqueous desolvation, and bring our previous understanding of the hydrophobic effect based on ordered water and buried surface area into question. Our approach provides new tools to quantify the thermodynamics of functional group desolvation in the context of ligand binding, which will be of tremendous value for future research on ligand/drug design.
natural bond orbital
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.
Rights for Collection: Duke Dissertations