Functional Conformational Dynamics in Complex Protein Systems

Abstract

Proteins are remarkable nano-scale molecular machinery; they are involved in almost all life processes by performing assorted functions in the presence of thermal fluctuations from their environments. The coupling between solvent fluctuations and a protein is of particular importance to the protein’s large amplitude motions on μs-toms timescales, which largely overlap with the timescales of protein-protein interactions and enzymatic turnovers. Understanding the functional roles of conformational dynamics in complex protein systems demands quantitative measurements and rigorous interpretations of protein conformational distributions. Single-molecule Forster resonance energy transfer (smFRET) has emerged as a powerful technique to reveal conformational distributions by removing the ensemble averaging. It is shown in this work that a set of orthogonal conformational modes—that afford a linear combination to describe conformational dynamics—are contained in experimentally measured conformational distributions. To contextualize conformational modes and their potential functional roles, two specific examples, gramicidin S synthetase A (GrsA) and human insulin-degrading enzyme (IDE), are studied. Application of high-resolution smFRET, time-dependent bulk FRET spectroscopy, kinetics profiling and structural modeling shows that the adenylation-thioesterification cascade is gated by the conformational transitions in the adenylation and peptide carrier protein domains in GrsA. Such conformational transitions occur via the newly identified on-pathway intermediate conformations. In the case of IDE, integration of the multiple smFRET mapping and a newly developed statistical sampling framework enables the structural computation and the extraction of predominant conformational modes in a data-driven manner from multiple smFRET distance distributions. Compared to IDE monomer, the protein-protein interaction in IDE dimer profoundly alters the conformational modes and confines the motion amplitudes. These motions could allow IDE to populate in the hitherto undetermined open conformations to complete the catalytic cycle. That these modes are similar to those from the elastic network modeling for both IDE monomer and dimer suggests that large-amplitude motions could be directly encoded in three-dimension protein architectures in complex protein systems. It is anticipated that the multi-platform biophysical experimentation and statistical framework presented here could be generalized to visualize and probe conformational modes, or more generally, dynamic molecular structures in other complex protein systems of interest.