Protein Unfolded States are Characterized by the Duality of Sequence-Specific Conformational Preferences and Ensemble-Averaged Features of Canonical Random Coils

Abstract

A consensus regarding rigorous, quantitative descriptions of unfolded states of autonomously foldable proteins under folding conditions has remained elusive despite its fundamental importance for a variety of processes in vivo. Here, we combine time-resolved Förster Resonance Energy Transfer (FRET) using multiple non-perturbing, amino-acid-sized dye pairs, paramagnetic relaxation enhancement (PRE) experiments with multiple labels, equilibrium and time-resolved small angle X-ray scattering (SAXS), extensive all-atom simulations, and polymer theory to construct a rigorous, atomistic descriptions for unfolded states of different proteins under their respective folding conditions. Our descriptions summarize sequence-specific contact distributions while provide a framework for quantifying the amplitudes of conformational fluctuations and correlations among these fluctuations. The picture that emerges suggests that unfolded states are conformationally heterogeneous, and this heterogeneity is sequence-specific both in terms of the biases for native as well as non-native contacts. Despite significant conformational heterogeneity, quantifiable biases toward the sequence-specific folded state topologies result from conformational fluctuations in unfolded states. Importantly, we find that converging on a coherent picture for heterogeneous ensembles such as unfolded states under folding conditions requires a combination of readouts from multiple type of experiments that are then integrated with theory and simulation. These methods are also directly applicable to studies of sequence-to-conformation relationships of intrinsically disordered proteins.