Size Conservation Emerges Spontaneously in Biomolecular Condensates Formed by Scaffolds and Surfactant Clients

Abstract

Biomolecular condensates are liquid-like membraneless compartments that contribute to the spatiotemporal organization of proteins, RNA, and other biomolecules inside cells. Some membraneless compartments, such as nucleoli, are dispersed as different condensates that do not grow beyond a certain size, or do not present coalescence over time. In this work, using a minimal protein model, we show that phase separation of binary mixtures of scaffolds and low-valency clients that can act as surfactants—i.e., that significantly reduce the droplet surface tension—can yield either a single drop or multiple droplets that conserve their sizes on long timescales (herein ‘multidroplet size-conserved’), depending on the scaffold to client ratio. Our simulations demonstrate that protein connectivity and condensate surface tension regulate the balance between these two scenarios. Multidroplet size-conserved behavior spontaneously arises at increasing surfactant-to-scaffold concentrations, when the interfacial penalty for creating small liquid droplets is sufficiently reduced by the surfactant proteins that are preferentially located at the interface. In contrast, low surfactant-to-scaffold concentrations enable continuous growth and fusion of droplets without restrictions. Overall, our work proposes one potential thermodynamic mechanism to help rationalize how size-conserved coexisting condensates can persist inside cells—shedding light on the roles of general biomolecular features such as protein connectivity, binding affinity, and droplet composition in this process.

Controlling cluster size in 2D phase-separating binary mixtures with specific interactions

“…We show that such size-control process, although it can lead to stationary states within observation time (Fig. 6), is not of obvious thermodynamic origin, as instead recently found for analogous processes in different systems 19,52 . On the contrary, it can be driven by a kinetic trap whose depth increases with bond strength and unevenness in concentrations.…”

“…However, re-entrant behaviour of the dense phase with respect to concentration has been observed in several phase-separating systems, suggesting that increasing concentration of a given species does not necessarily promote the formation of large assemblies. While in some cases there appears to be an electrostatic-related change in the interactions 13–15 , in others the effect seems to be fully stoichiometric 16–19 and most likely functional 20 .…”

“…Computational work on biological condensates has built upon various models, ranging from simple lattice models 21,22 , to off-lattice coarse-grained polymers 23,24 , stickers and spacers 25,26 , or patchy colloids 17–19,22,27 . In many studies, though, self-assembly is represented within the scaffold-client context, which assumes the existence of a set of proteins (scaffold) that can alone phase-separate through homotypic interactions 2,19,27,28 . While many systems correspond to this descriptions, others, especially in 2D, do not and feature mainly specific heterotypic interactions 3,5,17 .…”

Thermodynamic origins of two-component multiphase condensates of proteins

“…To do this, we use the Kirkwood-Buff expression 121 to relate the interfacial free-energy density g to the normal and tangential components of the pressure tensor, and the mean value theorem to simplify the result 122 for planar interfaces into g = (L z /2)(P norm − P tang ), where g is the interfacial free-energy density, L z is the length of the simulation box along which the interface occurs, P norm = P zz is normal to the interface and P tang = P xx = P yy is the tangential pressure, and the division by 2 accounts for the fact that there are two interfaces in our simulation set-up. 120 Although the pressure tensor has many possible denitions, from virial to mechanical expressions, the interfacial-free energy density is independent of this arbitrary choice; 123 we use the atomic virial pressure tensor in our calculation, which gives the same results as the molecular virial. 124 We compute only the interfacial free-energy density for the interface between the dense and dilute phases (i.e.…”

Fluorescent protein lifetimes report increased local densities and phases of nuclear condensates during embryonic stem cell differentiation

“…A general feature of liquid droplets exhibiting a single phase and separated from bulk liquid, much like an oil droplet in water, should exhibit homogeneous density throughout the interior of the condensate, except, perhaps, for its edges [33][34][35] . Fluorescence-based reporters of local density or crowdedness within a given region of interest (ROI) in an imaged cell could help probe this feature of condensates.…”