Enhanced Sampling of Ligand Binding Coupled to RNA Conformational Dynamics

As only 1.5 % of the human genome is coded into proteins, noncoding RNAs have emerged in recent days as potential targets of therapeutic drugs. However, due to their high structural flexibility compared to conventional protein targets, it is critical to gain detailed mechanistic insights into beyond-millisecond timescale RNA conformational transitions coupled to ligand binding for the rational design of RNA-targeted drugs. Molecular dynamics (MD) simulations have been a powerful tool for studying the atomistic details of ligand-binding mechanisms but their timescales are limited to a few microseconds (μs). Enhanced sampling methods can overcome this challenge by applying a biasing force to overcome high free energy barriers. In this work, we performed different variants of the On-the-fly Probability Enhanced Sampling (OPES) algorithm in the presence of a funnel-like restraint to compute the underlying free energy landscape and kinetics of ligand binding to the theophylline RNA aptamer. Our simulations capture the coupling between the cytosine-27 (C27) nucleotide flipping transition and the ligand binding coordinate. We could reproduce the relative trends of the binding affinities and the residence times for three different ligands to the RNA aptamer using the multiple walker OPES and the OPES-flooding algorithms, respectively. Our simulations sample two distinct binding mechanisms, conformational selection and induced fit, and demonstrate that the preference for the mechanism can be determined by the binding affinity. Combining such mechanistic insights with the free energy and kinetics obtained from the enhanced sampling simulations can provide a molecular-level understanding of the chemistry and biology of RNA-ligand recognition.