Hierarchical Ensembles of Intrinsically Disordered Proteins at Atomic Resolution in Molecular Dynamics Simulations.

Gerhard Hummer,Lukas S. Stelzl,Lisa M. Pietrek

doi:10.1021/acs.jctc.9b00809

Abstract

Intrinsically disordered proteins (IDPs) constitute a large fraction of the human proteome and are critical in the regulation of cellular processes. A detailed understanding of the conformational dynamics of IDPs could help to elucidate their roles in health and disease. However, the inherent flexibility of IDPs makes structural studies and their interpretation challenging. Molecular dynamics (MD) simulations could address this challenge in principle, but inaccuracies in the simulation models and the need for long simulations have stymied progress. To overcome these limitations, we adopt a hierarchical approach that builds on the "flexible-meccano" model reported by Bernadó et al. (J. Am. Chem. Soc. 2005, 127, 17968-17969). First, we exhaustively sample small IDP fragments in all-atom simulations to capture their local structures. Then, we assemble the fragments into full-length IDPs to explore the stereochemically possible global structures of IDPs. The resulting ensembles of three-dimensional structures of full-length IDPs are highly diverse, much more so than in standard MD simulation. For the paradigmatic IDP α-synuclein, our ensemble captures both the local structure, as probed by nuclear magnetic resonance spectroscopy, and its overall dimension, as obtained from small-angle X-ray scattering in solution. By generating representative and meaningful starting ensembles, we can begin to exploit the massive parallelism afforded by current and future high-performance computing resources for atomic-resolution characterization of IDPs.

Highlights

Proteins with intrinsically disordered regions constitute a large fraction of the human proteome.[1]
As expected from its derivation, the chain-growth algorithm generates ensembles in which each structure appears with a Boltzmann weight, eq 2, which we confirmed by comparing chains with up to 26 residues to chains grown by brute-force generation of self-avoiding random walks (Figure S2)
We generated highly diverse structures, much more diverse than what one would sample in a typical Molecular dynamics (MD) simulation and as diverse as a similar-sized flexible meccano ensemble (Figure S6A)

Summary

Introduction

Proteins with intrinsically disordered regions constitute a large fraction of the human proteome.[1]. Some intrinsically disordered proteins (IDPs) transiently sample structures and some fold upon binding to their partners, while others remain unfolded even in an ultrahigh-affinity complex.[2] Disordered regions and IDPs play essential roles in cell-signaling,[3] where their flexibility may be vital. Of IDPs in biomolecular condensates formed by liquid−liquid phase separation may be a general organizing principle in cell biology.[4] Dysregulation of liquid−liquid phase separation and aggregation of IDPs may be pathological mechanisms in many neurological diseases. The paradigm for IDPs is arguably defined by α-synuclein (aS).[5] aS was suggested to adopt α-helical conformations throughout the whole sequence in solution in its monomeric state but recent work suggests that it is better described as a disordered random coil.[6]

Methods

Results

Conclusion