An Integrative Approach to Single-Molecule FRET Spectroscopy and Molecular Dynamics Simulations for the Study of Intrinsically Disordered Proteins

Abstract

Recent technological advances in single-molecule techniques, such as single molecule fluorescence resonance energy transfer (smFRET) spectroscopy, have paved the way for the study of protein structural dynamics under biologically relevant conditions. Unlike high-resolution structure determination techniques such as x-ray crystallography, however, smFRET spectroscopy provides limited information spatially. Fortunately, the smFRET data combined with computational techniques such as molecular dynamics (MD) simulations could provide enough information to model the structural dynamics of protein at a spatiotemporal resolution. The starting point of these simulations is often based on crystal structure or other high-resolution structures of proteins. Unfortunately, such structures are not available for intrinsically disordered proteins (IDPs). Here we have employed enhanced sampling techniques in combination with all-atom MD and smFRET data to model the structural dynamics of IDPs. We have specifically employed state-of-the-art dimensionality reduction techniques in conjunction with enhanced sampling schemes to assimilate MD trajectories to make them consistent with the smFRET measurements. The novel methodology developed in this work uses the smFRET data within enhanced sampling techniques to first generate an ensemble of protein conformations and then predict FRET efficiency distributions for candidate labeled proteins. An iterative procedure is then used to efficiently design smFRET experiments based on the computational predictions and refine the conformational ensemble of protein, until a convergence is reached. The final step involves employing machine learning techniques to determine both the state distributions and their kinetics from smFRET-informed MD trajectories. We have successfully used this approach to study the structural dynamics of the C-terminal domain of membrane insertase Alb3, which is known to be intrinsically disordered. The novel approach developed here lays the groundwork for the efficient application of smFRET technique for the study of IDPs.