Profiling Unbinding Events to Identify Near-Native RNA-Ligand Poses

Abstract

Predicting the structure (or pose) of RNA-ligand complexes is an important problem in RNA structural biology and drug discovery. Although one could use molecular docking procedures to rapidly sample putative poses of RNA-ligand complexes, accurately dis-criminating the native-like poses from non-native, decoy structures remains a formidable challenge. Here, we started from the assumption that native-like RNA-ligand poses are less likely to dissociate during molecular dynamics simulations, and then we used enhanced simulations to promote ligand for diverse poses of a handful of RNA-ligand complexes. By fitting unbinding profiles obtained from the simulations to a single-exponential, we identified the characteristic decay time ({tau}) as particularly effective at resolving native poses from decoys. Remarkably, a simple regression model trained to predict the simulation-derived parameters directly from structure could also discriminate poses. As such, when molecular dynamics simulations are feasible, one could use them to simulate ligand unbinding to aid in the identification of near-native poses. On the other hand, when speed is more important, one could use a simple structure-based regression model, like the ones described in this study, to analyze and filter RNA-ligand poses.