Simulation of Biomolecular Condensates with Martini3-IDP

Abstract

Inside our cells, some proteins remain flexible and dynamic rather than folding into rigid shapes. These intrinsically disordered proteins (IDPs) can assemble into liquid-like biomolecular condensates, like oil droplets in water solution, to organize cellular biochemistry without encapsulating lipid membranes while continuously exchanging components with their environment. Studying these ever-changing assemblies is challenging, but computer simulations act as a “computational microscope.” I first recalibrated the popular simulation tool Martini, and created Martini3-IDP, which correctly reproduces the flexibility and expanded shapes of IDPs, all while remaining compatible with the existing Martini toolkit. Using this tool, I uncovered three key insights. First, condensates containing both structured and disordered domains, as in real proteins, have a very different internal architecture and slower internal motions than simplified disordered-only models. Second, the condensate scaffold proteins can reshape the conformation of client proteins that enter the droplet, revealing design rules for how condensates might alter protein function. Third, I built a realistic computational model of cellular P-body, incorporating six key proteins and RNA mimics. This multicomponent condensate revealed a highly uneven interior with mobile solvent pockets and component-specific behavior. Together, this work provides a powerful simulation tool and new molecular insights into how cellular droplet based organization works, from internal architecture to client reprogramming.