Enhanced condensate fluidity in modified patchy particle models

Abstract

Biomolecular condensates are formed via liquid–liquid phase separation of proteins, often together with nucleic acids, typically driven by interactions between low-affinity binding sites. The computational study of such condensates that accounts for both the droplet-scale fluid behavior and the internal structure of the condensate requires coarse-grained models. Recently, patchy particle models, representing proteins as spheres with a repulsive core and directional attractive patches, have emerged as a powerful tool. However, these simulations typically exhibit slow dynamics and often show glass-like rather than fluid condensates as required for the description of many biomolecular condensates, establishing a need for model variants with accelerated dynamics, both to speed up simulations and to better represent fluid biomolecular condensates. Here, we study several modified patchy particle models suited to simulate the formation and dynamics of biomolecular condensates. These variants incorporate flexible patches and weak isotropic attractions between cores. They significantly accelerate the system dynamics while preserving key equilibrium characteristics of the classical patchy particle model, including the phase behavior and the local structure of the condensate. These modifications thereby enable the simulation of larger, more complex systems previously inaccessible due to prohibitive relaxation times and provide a versatile tool for studying condensate dynamics.