Reentrant liquid condensate phase of proteins is stabilized by hydrophobic and non-ionic interactions

Abstract

Many cellular proteins have the ability to demix spontaneously from solution to form liquid condensates. These phase-separated structures form membraneless compartments in living cells and have wide-ranging roles in health and disease. Elucidating the molecular driving forces underlying liquid-liquid phase separation (LLPS) of proteins has thus become a key objective for understanding biological function and malfunction. Here we show that proteins implicated in cellular phase separation, such as FUS, TDP-43, and Annexin A11, which form condensates at low salt concentrations via homotypic multivalent interactions, also have the ability to undergo LLPS at high salt concentrations by reentering into a phase-separated regime. Through a combination of experiments and simulations, we demonstrate that phase separation in the high-salt regime is mainly driven by hydrophobic and non-ionic interactions. As such, it is mechanistically distinct from the low-salt regime, where condensates are stabilized by a broad mix of electrostatic, hydrophobic, and non-ionic forces. Our work thus expands the molecular grammar of interactions governing LLPS of cellular proteins and provides a new view on hydrophobicity and non-ionic interactions as non-specific driving forces for the condensation process, with important implications for the aberrant function, druggability, and material properties of biomolecular condensates. One Sentence SummaryProteins implicated in cellular phase separation can undergo a salt-mediated reentrant liquid-liquid phase transition.

Expansion of Intrinsically Disordered Proteins Increases the Range of Stability of Liquid–Liquid Phase Separation

“…Atomistic Molecular Dynamics (MD) simulations can characterize the conformational ensembles of single proteins and protein complexes [ 56 , 57 , 58 ], pinpoint the link between chemical modifications and sequence mutations, and the modulation of protein–protein and protein-DNA interactions [ 59 , 60 , 61 , 62 ], reveal the conformational heterogeneity of IDRs within small aggregates [ 63 ], and guide the development of chemically accurate coarse-grained models for LLPS [ 64 , 65 ]. Furthermore, the predictive and explanatory power of atomistic simulations is constantly being ramped up by the collective efforts to develop even more accurate atomistic force fields for IDRs [ 66 , 67 ].…”

Widespread occurrence of the droplet state of proteins in the human proteome

“…Native and amyloid states are stabilized by specific interactions including hydrogen bonds, ionic interactions, and van der Waals contacts typical of ordered states and enthalpic in nature (9,10). By contrast, in droplets, transient short-range aromatic cation-p and p-p, dipoledipole, electrostatic and hydrophobic interactions have been observed, providing lowspecificity, weak-affinity contacts characteristic of disordered states (11)(12)(13)(14)(15)(16). These observations have lead to a series of prediction methods (11,13,(17)(18)(19), which focused on specific side-chain interactions.…”

Targeted modulation of protein liquid-liquid phase separation by evolution of amino-acid sequence

“…We have coupled our genetic-algorithm technique to the coarsegrained model of Dignon et al 33 which is one of the best models currently available for probing the phase behaviour of protein solutions. This model has been validated against the singlemolecule experimental radii of gyration of a wide range of IDRs, 33 is residue-specific, has been shown to reproduce well the experimental phase behaviour of various proteins under different conditions, 21,37,75–77 and is computationally sufficiently inexpensive that it affords the determination of bulk LLPS properties for many sequences. Furthermore, the model accounts for key physicochemical aspects that determine the phase behaviour of proteins, such as the charge, size, relative hydrophobicity and flexibility of amino acids.…”

“…17–19 In the context of protein solutions, LLPS is principally driven by dipole–dipole, charge–charge, cation–π and π-stacking interactions, 20 whose strengths are modulated by the experimental conditions, including the temperature, pH and salt concentration. 21 In biomolecular systems, the multivalency in mixtures is thus the main physical parameter that defines the ability of a system to undergo LLPS: 6,15,22–25 biomolecules with higher valencies can establish a larger number of weak attractive interactions with other species and hence form a more stable condensate.…”