Liquid-Liquid Phase Separation in Crowded Environments

Abstract

Biomolecular condensates play a key role in organizing cellular fluids such as the cytoplasm and nucleoplasm. Most of these non-membranous organelles show liquid-like properties both in cells and when studied in vitro through liquid-liquid phase separation (LLPS) of purified proteins. In general, LLPS of proteins is known to be sensitive to variations in pH, temperature and ionic strength, but the role of crowding remains underappreciated. Several decades of research have shown that macromolecular crowding can have profound effects on protein interactions, folding and aggregation, and it must, by extension, also impact LLPS. However, the precise role of crowding in LLPS is far from trivial, as most condensate components have a disordered nature and exhibit multiple weak attractive interactions. Here, we discuss which factors determine the scope of LLPS in crowded environments, and we review the evidence for the impact of macromolecular crowding on phase boundaries, partitioning behavior and condensate properties. Based on a comparison of both in vivo and in vitro LLPS studies, we propose that phase separation in cells does not solely rely on attractive interactions, but shows important similarities to segregative phase separation.

Keywords: crowding; intrinsically disordered proteins; liquid–liquid phase separation; membraneless organelles.

Figure 1

Schematic overview of a eukaryotic cell containing membranous and membraneless organelles.

Figure 2

Excluded volume effect of crowding. (A) A small particle (red) experiences a smaller excluded volume than a large particle (black sphere). The blue area represents the free space for the particle; for a large particle the effective free volume is limited. (B) Many biochemical processes display a maximum rate at some optimum crowding, as crowding enhances complexation (blue curve), but it also reduces diffusivity (red curve) (C) The effect of depletion forces, bringing molecules together reduces the total excluded volume. (Figure 2A is adapted from Minton 2001 [23], Figure 2B is adapted from Ellis 2001 [3], and Figure 2C is adapted from Richter 2008 [24]).

Figure 3

Liquid–liquid phase separation of proteins and polymers. Illustration of the difference between (A) associative phase separation which relies on attration between two macromolecules and (B) segregative phase separation which relies on repulsive interactions. (C) The three model systems from Section 4 which contains poly-U–spermine; FUS; NPM1 condensates. Poly-U–spermine condensates rely solely on charge interactions between RNA (negative) and spermine (positive). Increasing complexity, FUS has cation–π interactions between arginine (RGG motif, cation) and tyrosine ([G/S]Y[G/S] domain, π) residues within the disordered domains. RNA could play a role but is not included in the crowded studies discussed in Section 4. Finally, NPM1 has an oligomerization domain and a nuclear-binding domain in addition to charge interactions.

Figure 4

Cellular condensates might arise from both associative and segregative phase separation. (A) Phase separation results in a crowded protein-dense phase that is situated in the crowded cytosol (or nucleoplasm). (B) Three effects on phase separation caused by macromolecules: (I) LLPS is induced through depletion forces, increasing the attractive forces within the condensate. (II) LLPS is promoted through co-condensation of macromolecules through attractive interactions between the macromolecules and the proteins. (III) LLPS is promoted through segregation, as repulsive interactions between macromolecules and proteins lower the solubility of proteins.