Receptor dynamics in virtual screening: A dihydrofolate reductase case study.

The implications of receptor dynamics virtual screening were explored as part of proof of concept study for an interactive docking, molecular dynamics and scoring strategy. The original study was to demonstrate whether or not the iterative process proposed would relax receptor structures that began as apo (without drug) or homology models toward known holo (with drug) coordinates. Due to species differences in the original receptor test systems, the study was refined to a set of well characterized dihydrofolate reductase (DHFR) receptors. Since both apo and holo structures were not available for each DHFR receptor selected, the goal of relaxing an apo structure toward a holo structure was modified. Various complexes of enzymes, ligand and cofactor were constructed to represent intermediates in the catalytic cycle know to adopt distinct conformations, and the new goal was to demonstrate movement toward one of these conformations during molecular dynamics simulations. Additionally, drug resistant E. coli DHFR and human DHFR structures were selected to probe the possible flexibility differences giving rise to enzymatic activity. Traditional molecular dynamics simulations demonstrated correct conformational change for one enzyme cofactor complex after 70 nanoseconds of simulation. Overall, relatively few complexes experienced any conformational change, indicating the need for longer trajectories or more advanced simulation techniques. Flexible regions of the enzyme described in literature were correctly depicted by trajectory analysis. Flexibility differences between bacterial and vertebrate DHFR showed bacterial DHFR to have more localized movements of three flexible loops and vertebrate DHFR more disperse flexibility. Evaluating the free energy of binding by Molecular Mechanics - Poisson Boltzmann Surface Area (MM-PBSA) methods was shown to be more insightful when employed before and after conformational change rather than over an entire trajectory. Ligand binding values generally independent of enzyme activity demonstrated the necessity of considering binding in the context of the catalytic cycle. Traditional molecular dynamics demonstrated the capability of moving a receptor toward a known conformation, through the resolution of conformational change was depended on the amount of trajectory available. Advanced molecular dynamics methods possible capable of expediting conformational change are proposed for future studies.