Charge pattern matching as a 'fuzzy' mode of molecular recognition for the functional phase separations of intrinsically disordered proteins

Biologically functional liquid–liquid phase separation of intrinsically disordered proteins (IDPs) is driven by interactions encoded by their amino acid sequences. Little is currently known about the molecular recognition mechanisms for distributing different IDP sequences into various cellular membraneless compartments. Pertinent physics was addressed recently by applying random-phase-approximation (RPA) polymer theory to electrostatics, which is a major energetic component governing IDP phase properties. RPA accounts for charge patterns and thus has advantages over Flory–Huggins (FH) and Overbeek–Voorn mean-field theories. To make progress toward deciphering the phase behaviors of multiple IDP sequences, the RPA formulation for one IDP species plus solvent is hereby extended to treat polyampholyte solutions containing two IDP species plus solvent. The new formulation generally allows for binary coexistence of two phases, each containing a different set of volume fractions $({\varphi }_1,{\varphi }_2)$ for the two different IDP sequences. The asymmetry between the two predicted coexisting phases with regard to their ${\varphi }_1/{\varphi }_2$ ratios for the two sequences increases with increasing mismatch between their charge patterns. This finding points to a multivalent, stochastic, 'fuzzy' mode of molecular recognition that helps populate various IDP sequences differentially into separate phase compartments. An intuitive illustration of this trend is provided by FH models, whereby a hypothetical case of ternary coexistence is also explored. Augmentations of the present RPA theory with a relative permittivity ${ϵ}_r\left(\varphi \right)$ that depends on IDP volume fraction $\varphi ={\varphi }_1+{\varphi }_2$ lead to higher propensities to phase separate, in line with the case with one IDP species we studied previously. Notably, the cooperative, phase-separation-enhancing effects predicted by the prescriptions for ${ϵ}_r\left(\varphi \right)$ we deem physically plausible are much more prominent than that entailed by common effective medium approximations based on Maxwell Garnett and Bruggeman mixing formulas. Ramifications of our findings on further theoretical development for IDP phase separation are discussed. (paper)