Combined Theoretical and Computational Approach for Calculating Sequence-specific Phase Diagrams of Thermoresponsive Intrinsically Disordered Homopolypeptides

Abstract

Peptide-based homopolymers comprise of multiple repeats of short peptide motifs. These intrinsically disordered homopolypeptides undergo different types of responsive phase transitions. Of particular interest are homopolypeptides that undergo thermoresponsive phase transitions whereby dense condensates are formed in response to increases above or decreases below sequence-specific critical solution temperatures. We recently developed a sequence design algorithm that combines heuristics derived from simplified all-atom simulations with Monte Carlo sampling and a sequence generating genetic algorithm [1]. This design approach can be deployed for rapid generation of hundreds to thousands of novel sequences that are predicted to have the desired thermoresponsive phase behavior. However, the most rigorous way of knowing if the designed sequence has the desired thermoresponsive features is to have access to the full temperature-dependent phase diagram. Here, we deploy a novel approach that combines computation with theory to enable the calculation of desired phase diagrams. To achieve this, we adapt, refine, and deploy the Gaussian cluster theory of Raos and Allegra [2] that was developed to model coil-to-globule transitions and phase transitions of homopolymers as a function of solution temperature. We extract temperature-dependent two- and three-body interaction coefficients and the predicted theta temperatures from all-atom simulations of interactions among the repeating units that make up homopolypeptides. Using these interaction coefficients as inputs, we use the Gaussian cluster theory to calculate sequence-specific phase diagrams for homopolypeptides. We showcase the joint computational and theoretical approach by applying it to calculate phase diagrams for homopolypeptides that have different types of thermoresponsive phase behaviors.