Liquid–Liquid Phase Separation in the Presence of Macromolecular Crowding and State-dependent Kinetics

Abstract

Biomolecular condensates formed via liquid–liquid phase separation (LLPS) are increasingly being shown to play major roles in cellular self-organization dynamics in health and disease. It is well established that macromolecular crowding has a profound impact on protein interactions, particularly those that lead to LLPS. Although synthetic crowding agents are used during in vitro LLPS experiments, they are considerably different from the highly crowded nucleo-/cytoplasm and the effects of in vivo crowding remain poorly understood. In this work, we applied computational modeling to investigate the effects of macromolecular crowding on LLPS. To include biologically relevant LLPS dynamics, we extended the conventional Cahn–Hilliard model for phase separation by coupling it to experimentally derived macromolecular crowding dynamics and state-dependent reaction kinetics. Through extensive field-theoretic computer simulations, we show that the inclusion of macromolecular crowding results in late-stage coarsening and the stabilization of relatively smaller condensates. At a high crowding concentration, there is an accelerated growth and late-stage arrest of droplet formation, effectively resulting in anomalous labyrinthine morphologies akin to protein gelation observed in experiments. These results not only elucidate the crowder effects observed in experiments, but also highlight the importance of including state-dependent kinetics in LLPS models, and may help in designing further experiments to probe the intricate roles played by LLPS in self-organization dynamics of cells.

WNK1, a molecular crowding sensor, links phase separation to cellular physiological stress

“…This is consistent with other proteins that undergo PS and form membrane‐less compartments 5 . However, PS triggered by molecular crowding caused by hyperosmotic stress appears to be a common mechanism by which multimeric proteomes exert biological functions 3 …”

“…PS can be triggered by various cellular stresses, including temperature, pH, and hyperosmolarity 2 . Meanwhile, PS can produce membrane‐free biomolecular condensates, in which specific biomolecules are concentrated and serve biological functions 3 . For instance, the phosphorylation of the Alzheimer's disease‐related protein tau by cyclin‐dependent kinase 2 is accelerated in condensates, 4 and biomolecular condensates are implicated in the regulation of postsynaptic density at neuronal synapses in synaptic signaling and plasticity 3 .…”

“…Meanwhile, PS can produce membrane‐free biomolecular condensates, in which specific biomolecules are concentrated and serve biological functions 3 . For instance, the phosphorylation of the Alzheimer's disease‐related protein tau by cyclin‐dependent kinase 2 is accelerated in condensates, 4 and biomolecular condensates are implicated in the regulation of postsynaptic density at neuronal synapses in synaptic signaling and plasticity 3 . Numerous studies have demonstrated the significance of biomolecular condensates in physiology and pathophysiology 2,3 .…”

“…3 For instance, the phosphorylation of the Alzheimer's disease-related protein tau by cyclindependent kinase 2 is accelerated in condensates, 4 and biomolecular condensates are implicated in the regulation of postsynaptic density at neuronal synapses in synaptic signaling and plasticity. 3 Numerous studies have demonstrated the significance of biomolecular condensates in physiology and pathophysiology. 2,3 In addition to triggering PS, hyperosmotic stress can also activate WNK1.…”

Atomistic Simulation of Lysozyme in Solutions Crowded by Tetraethylene Glycol: Force Field Dependence

“…Computational approaches are a valuable supplement to experimental measurements for the investigation of the structure, dynamic, association, and charge regulation [ 14 , 20 , 23 , 40 , 46 , 47 , 48 , 49 , 50 , 51 ] as well as the liquid-liquid phase separation of biomolecules in a crowded environment [ 52 , 53 , 54 , 55 , 56 , 57 ]. Despite the expensive computational cost in modeling crowded environments, model accuracy is also an issue.…”