Theory and Modeling of Noncovalent Binding

University Of California, San Diego, La Jolla CA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Specific, noncovalent binding is central to many biological functions, such as signaling and immunity; and most drugs are small molecules that bind a specific protein and modulate its function. This project aims fr a deeper and more quantitative understanding of noncovalent binding, coupled with increasingly predictive and insightful modeling technologies that will speed the discover of new medications. One goal is to develop best practices for computing not only bindin free energies, but also numerically precise binding enthalpies, using molecular dynamics simulations with an explicit treatment of water. We will then apply these practices to seek mechanistic explanations of often puzzling experimental calorimetric data, such as observations that ligand preorganization may strengthen binding not entropically, as expected, but enthalpically; and the common, but still poorly understood, phenomenon of entropy-enthalpy compensation. We will also work on a new method to test the force fields used in molecular simulations. Today, force fields are tested primarily by computing small molecule hydration free energies and the physical properties of pure liquids, and comparing these results with experiment. Thus, although simulations are widely used to model binding, experimental binding data plays little role in the testig and optimization of simulation force fields. Here, we aim to establish host-guest bindng data as a new benchmark for testing force fields, by assembling a panel of experimentally characterized, computationally tractable, host-guest systems, along with computational scripts which automate the calculation of their binding free energies and enthalpies and the comparison of these results with experiment. This validation engine will be used to test widely used force fields, and will also be shared freely with ther groups. Finally, we will integrate free energy simulation methods into two collaborative drug??discovery projects. One project centers on a human enzyme called soluble epoxide hydrolase, whose three-dimensional structure is known and which is therapeutically relevant to cardiovascular and inflammatory disease. In our preliminary studies, conventional docking approaches have provided little or no useful guidance; we now aim to apply more sophisticated fre energy simulation methods to this challenging case. The second project, which is at a earlier stage, concerns a bacterial virulence factor called Cif. This, too, is an epoide hydrolase, and its inhibition is expected to be helpful to patients with Pseudomonas pneumonia. We plan to use docking and free energy methods to help discover potent Cif inhibitors for use as chemical probes and, potentially, as first steps toward a ne medication.