There is growing interest in uncovering the rules that govern the sequence-determinants of phase separation of intrinsically disordered proteins (IDPs). This has direct relevance for the reversible formation / dissolution of membraneless organelles and / or biomolecular condensates.<br/><br/>Here, we report the development of a theoretical / computational framework that can explain and predict the phase behavior of a subset of IDPs with prion-like domains. We show that, to first order, IDPs with prion-like domains can be modeled as flexible polymers chain comprised of multiple interacting stickers within short linear motifs that are connected by spacers. However, the spacers engender multivalency of stickers within a chain and they enable cooperative transitions such as condensation, which in turn allows for the networking of stickers via physical crosslinks that is known as percolation or gelation. The smallest stickers are individual amino acids, specifically aromatic (tyrosine and phenylalanine) and / or basic residues (arginine and lysine), but experiments show that a simple cation-pi description does not fully explain the sticker-driven specificity for phase separation. Specifically, substitutions of tyrosine (Y) with phenylalanine (F) and/or arginine (R) with lysine (K) residues drastically weaken the driving forces for phase separation.<br/><br/>We have conducted atomistic simulations and found that the differences in stickers originate in a hierarchy of interactions accessible to specific stickers (tyrosine and arginine) across distinct length scales and orientations. Based on these findings, we have devised a simple mean-field polymer theoretic description that allows us to predict the saturation concentrations based on the numbers and types of stickers, where stickers are distinguished from spacers by their intrinsic site valencies. We have incorporated the principles and parameters derived from experiments and atomistic simulations into a novel 3-dimensional lattice-model framework, LASSI (LAttice simulation module with Sticker and Spacer Interactions), which includes interactions with multiple spatial hierarchies and allows us efficient high-throughput computation. These simulations confirm the robustness of the mean-field predictions and pave the way forward for predicting the driving forces for phase separation, the overall phase behavior, and the structures of networked IDPs within condensates, directly from the amino acid sequences of IDPs.

Stickers and spacers framework for phase transitions of multivalent intrinsically disordered proteins

256th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.