Coarse-Grained Simulations of Intrinsically Disordered Proteins in the Context of Liquid-Liquid Phase Separation

Abstract

Cellular assembly of highly concentrated RNA and proteins into phase separated liquid- or hydrogel-like phases can provide cellular compartmentalization in the absence of lipid membranes. The details of the interactions, and driving forces which promote phase separation into these types of assemblies are still not well understood and may require difficult and expensive experimental techniques, or impractically large atomistic simulations. To address this, we have proposed a novel framework for conducting coarse-grained simulations of the phase coexistence using knowledge-based interaction potentials and sampling techniques capable of mapping the phase coexistence behavior[1]. Using this approach, we are also able to determine intermolecular contacts and diffusion of chains in the dense phase to help determine the properties of these high-density phases, and the driving forces of their condensation. Presently, we have successfully determined the effects of mutations to highly interacting regions of commonly studied peptide sequences including TDP-43, FUS, hnRNPA2 and LAF-1. Additionally, we show that small, transient helical structures can also have an influence on the ability of the peptide to phase separate through the use of rigid constraints, or more flexible Go-like potentials. Our goal is to provide researchers with the tools to predict the effects of mutations, post-translational modifications, changes to solution conditions, and many other factors on the protein's ability to phase separate under physiological conditions. Much effort is being directed toward improved nonbonded pairwise potentials, and the design of efficient sampling methods and other theoretical models to allow for high-throughput screening of a large number of protein sequences. [1] Dignon, G. L.; Zheng, W.; Kim, Y. C.; Best, R. B.; Mittal, J.PLoS Computational Biology (under review).