Session S38: Biomolecular Condensates I - Interfaces

Cellular surfaces as regulators of biomolecular condensation

All cells depend on subcellular compartments to organize diverse biochemical processes. Membraneless biomolecular condensates compartmentalize cells by enriching certain proteins and nucleic acids within assemblies that can display properties of liquid-like droplets. A key requirement for condensate function is the dynamic regulation of size, which impacts molecular exchange, reaction efficiency, and environmental responsiveness. However, condensates reconstituted from minimal components in a test tube often vastly exceed the sizes of native condensates in cells. The biophysical mechanisms that control condensate size are incompletely understood. Many condensates associate with biological “surfaces,” including (a) two-dimensional lipid bilayer membranes and (b) one-dimensional polymers of long noncoding RNA. In this talk, I will share two stories of how membrane and polymer surfaces control condensate size. In the first story, I found that attachment to membranes controls the sizes of condensates composed of the protein Whi3. Specifically, Whi3 tends to form large droplets in vitro, but Whi3 condensates in cells appear as small puncta. I discovered that Whi3 condensates are attached to endomembranes in cells. Recruiting Whi3 to membranes in vitro restricts condensate growth substantially, resulting in punctate condensates that resemble native assemblies. The slow growth was due to the slower diffusion of membrane-bound molecules compared to solution. In the second story, I found that a long noncoding RNA called NEAT1 controls the sizes of nuclear paraspeckles. NEAT1 contains binding sites for multiple proteins, including FUS and NONO. While FUS tends to form large droplets with NEAT1 in vitro, paraspeckles appear as small puncta with regular sizes of ~360 nm. I discovered that FUS-NEAT1 condensates are smaller in the presence of NONO, suggesting that NONO opposes FUS condensation by competing for binding to NEAT1. These data indicate that NEAT1 controls paraspeckle size by tuning the recruitment of competing proteins. In both stories, biological surfaces control condensate sizes without active, energy-consuming processes. Given that many condensates associate with surfaces throughout the cell, these findings reveal broadly-applicable mechanisms of size regulation.   

Polymer collapse facilitates protein phase-separation

Nuclear proteins such as transcription factors demix into liquid-like droplets in-vitro, and in-vivo phase-separate at somewhat lower concentrations onto DNA & RNA polymers. Separately, long polymers can undergo extended to collapsed transitions as solvent conditions change, and spatial rearrangements of chromosomes in three-dimensions regulate transcription. Here we use approaches from statistical physics to explore a model where proteins with a propensity to phase-separate in 3D modulate the collapse transition of a long polymer. Our analysis reveals that the surface of a long polymer is best viewed as a compressible 'scaffold' onto which a bulk fluid can phase-separate. Demixing of the bulk fluid coincides with the polymer collapse transition, and the presence of a scaffold dramatically widens the regime in which these demixing transitions occur. Polymers lacking the collapse transition or fluidity do not see a similar degree of enhancement. A simple extension of our model allows us to understand how cells may use these scaffolded phases to integrate information for sensory tasks. We draw parallels to a recent model we proposed for clusters of cytoplasmic signaling molecules that phase-separate exclusively at the surface of the plasma membrane, itself near a 2D liquid-liquid critical point. In both long polymers & critical membranes, phase-transitions in a lower-dimensional surface mediate surface-localized demixing of a bulk fluid.   

Molecular Transport across Phase Boundaries

Cells can achieve compartmentalization of biochemical processes via organelles by the selective admission of biomolecules. Organelles are enclosed by a membrane or, in the case of biomolecular condensates, by the condensate-bulk interface. 

While transport across membranes has been studied for decades, how biomolecular condensates regulate transport across their interface is less clear. This severely hampers our understanding of dynamical condensate function. A main challenge in capturing interfacial effects is the need to accurately measure several physical properties, such as diffusion and partition coefficients to meaningfully constrain our physical models. Here, we use a combination of quantitative (live) imaging techniques and theory to quantify the transport of molecules across the condensate-bulk interface in PGL-3 and FUS model condensates and compare to several theoretical models. We find that, under many conditions, the droplet interface can be explained within the experimental error bounds by taking into account the partition coefficient and different diffusivities inside and outside of the condensate. I will comment on further interfacial transport effects and discuss consequences for how cells can regulate transport into condensates given these constraints. It will be exciting to generalize our findings to more complicated (in vivo) systems and I will discuss work in progress in this direction.   

Exchange Dynamics of Biomolecular Condensates

Biomolecular condensates are dense assemblies of biological molecules that segregate out of the intracellular milieu via liquid-liquid phase separation. One of their key features is that they constantly exchange components with the cellular environment. It has previously been shown that this material exchange can influence condensate function via the rate of biochemical reactions occurring in condensates and the speed with which a condensate responds to environmental changes. Motivated by recent experimental evidence that condensates of a single type of biomolecule can consist of multiple species with drastically different mobilities, we theoretically derive diffusion-reaction equations to model a two-species system, where the two species represent molecules of different mobilities within the condensate. Using finite difference methods, we study the diffusion processes and characteristic timescales that mediate the dynamical exchange of the dilute and dense phases of such a two-species system.   

Non-linear partitioning of client proteins to the bulk and interface of biomolecular condensates

The ability of biomolecular condensates to sequester client proteins is thought to be central to their function. For example, partitioning enables the compartmentalization of enzyme reactions. Further, recent studies have suggested that condensates can be stabilized through the localization of client proteins to the interface. However, systematic investigations of the partitioning power of condensates are limited. Here, we quantify bulk and interfacial partitioning of tubulin to FUS-polyU droplets using light microscopy. We find that classical partitioning coefficients are insufficient to characterize these phenomena. Specifically, bulk and surface concentrations increase non-linearly with tubulin concentration.   

Lightning the Interface: the Electrochemical Features of Biomolecular Condensates

Phase transition of biomacromolecules can result in a solvent density transition between the dilute and the dense phases. This solvent density transition can lead to a new electrochemical equilibrium between the dilute and the dense phases. Here we discuss the mechanism by which the unique electrochemical environment of condensates can encode different electrochemical features. We introduce how an associative transition of biomacromolecules can result in a segregative transition of ions, forming an electric potential gradient between the dilute and the dense phases, which corresponds to an interfacial electric double layer. We find that this electric double layer serves as an active electric field that can drive spontaneous redox reactions and represents a new functional feature of condensates. This work uncovers that the functions of biomolecular condensates can be determined by their electrochemical environments, expanding our understandings on condensate function beyond the biomolecules driving or participating in phase transition.    

Nucleation pathways of multicomponent biomolecular condensates: a cautionary tale of the classical nucleation theory

Intracellular phase separation plays a crucial role in regulating important biological processes, such as transcription and DNA organization. In certain cases, condensates assemble via processes analogous to nucleation and growth in abiotic systems (e.g. first-order liquid-gas phase transition). However, a quantitative description of the nucleation landscape must also account for both the multicomponent nature of bimolecular condensates and the crowded intracellular environment.

Here, we address this question by combining theory, molecular simulation, and continuum-scale numerical methods. To illustrate the idea, we consider a 3-component system where the third component occupies only a minority volume fraction but can lower the free-energy barrier for nucleating one of the two majority components. We find that, at low mobilities and in the presence of significant thermal fluctuations, the preferred nucleation pathway can deviate significantly from predictions of Classical Nucleation Theory (CNT), in terms of both the concentration profiles and the free-energy barrier. We will discuss scenarios in which this discrepancy becomes important and the resulting biological implications.   

Merging-limited coarsening of nanoscale condensates

Droplet coarsening is a common phenomenon in which smaller droplets naturally grow into larger ones to minimize their interfacial free energy and achieve global thermodynamic equilibrium. Here, we observe that coacervate droplets of small sizes (ranging from tens to hundreds of nanometers) remain stable over hours with significantly slower coarsening rates than predicted by classic theories. Using scaling analysis, Monte Carlo simulations, and analytical theory, we demonstrate that the anomalously stable coacervates can be explained by a merging-limited coarsening (MLC) mechanism, in which merging probability becomes markedly low among coacervates of sizes smaller than a critical value, which is controlled by the internal viscosity and interfacial tension of the droplets. We find that biological condensates typically exhibit large critical sizes, making them prone to undergo slow coarsening through the MLC mechanism. Such merging-limited coarsening may represent a universal mechanism underlying condensate size control in synthetic systems and living cells.   

Capillary forces from biomolecular condensates on microtubules

The microtubule-based spindle organizes and segregates chromosomes during cell division. Thousands of microtubules nucleate rapidly and exhibit complex cooperative behavior, which is facilitated by numerous microtubule-associated proteins. A biomolecular condensate of the microtubule-associated protein TPX2 has been found to wet microtubule surfaces, in order to recruit and concentrate factors important for microtubule nucleation. Here, using atomic force microscopy, we directly measure the attractive capillary interactions on TPX2-coated microtubule surfaces. We find that the capillary interactions are significant and can therefore underly the functional role of TPX2 in microtubule nucleation and organization. The capillary interactions are influenced by the concentration of salt, which can affect the phase behavior of TPX2. We construct thermodynamically-consistent mathematical models based on a modified Voorn-Overbeek theory to explain our observations.   

Phase behavior of mixtures of intrinsic disordered proteins and rod-like colloidal particles

We report on studies of the phase behavior of aqueous dispersions composed of intrinsic disordered proteins (IDPs) [1] and submicrometer-sized rod-like colloidal particles such as fd-virus particles [2]. Here we focus on results with fd. In isolation, IDPs form condensates through spontaneous liquid-liquid phase separation in solutions with low ion concentration; in IDP-fd mixtures, the IDPs interact with the fd via hard-core and Coulomb interactions. Using fluorescence microscopy we have found that at low fd concentration, condensates form and have a core-shell structure with the fd aggregated on the surface. On the other hand, at high colloidal particle concentration, phase separation of IDPs is suppressed, with substantially fewer and smaller condensates observed. Notably, at intermediate fd concentration, the fd and the IDPs form well-mixed condensates in which the fd are highly concentrated (i.e., compared to the background liquid phase). We will describe the experiments and offer some analysis of the phase behavior of these mixtures and the microscopic structure of the condensates.

 

[1] Garabedian, M V., et al. Biochemistry 61.22 (2022): 2470-2481.

[2] Dogic, Z. and Fraden, S. Current opinion in colloid & interface science 11.1 (2006): 47-55.   

Predicting morphology of biomolecular condensates from protein interaction networks

The formation of membraneless organelles in living cells is widely regarded as a result of near-equilibrium phase separation. Various condensates can be further assembled into higher-order structures by forming thermodynamically stable interfaces between immiscible phases. Using a minimal model of a protein interaction network, we demonstrate how a "shared" protein species that partitions into both phases of a multiphase condensate can function as a tunable surfactant that modulates the interfacial properties. We use Monte Carlo simulations and free-energy calculations to identify conditions under which a low concentration of this shared species is sufficient to trigger a wetting transition. We also describe a numerical approach based on classical density functional theory to predict density profiles and surface tensions directly from the model protein interaction network. Finally, we show that the wetting phase diagrams that emerge from our calculations can be understood in terms of a simple model of selective adsorption to a fluctuating interface. Our work shows how a low-concentration protein species might function as a biological switch for regulating condensate morphologies.   

Size, shape, and fluctuations of condensates in non-equilibrium liquid-liquid phase separation

Equilibrium phase separation, in the absence of chemical reactions, leads at long times, to a condensate of system size due to the interfacial tension of smaller-sized domains. In contrast, additional long-range (Coulomb) interactions competing with interfacial tension are known to stabilize the condensate size at intermediate-length scales. Examples of such tension long-range interacting (TLR) systems are – Rayleigh instability of charged liquids, block copolymer melts, binary solvent with antagonistic salt, and biomolecular condensates. In the latter case, the chemical reactions (production and degradation of proteins, RNA molecules, etc.) involve the input of energy (activity) and are coupled to the equilibrium aspects of phase separation. In such non-equilibrium phase separation, the slow chemical kinetics of the constituents play an antagonistic role to fast molecular diffusion (Ostwald ripening) and lead to a non-equilibrium steady state. For first-order chemical kinetics, the non-equilibrium term maps to a Coulomb interaction in the effective free energy. In the mean-field limit, for infinite periodic systems, the transition between various morphologies (sphere, cylinder, lamellar, etc.) depends on the relative concentrations of the solute-rich and the solute-poor domains (chemical composition in the other equilibrium cases). An important finding of our theory is that for finite (but very large) systems with lamellar microstructure, the sample aspect ratio enters the system with free energy and the steady-state domain size. While the lamellar phase is locally stable, its restoring force to undulations is related to the curvature of the undulations and does not depend on the extra area of the layer (effectively zero tension) for needle-like and periodic systems.   

Designing the Morphology of Phase-Separated DNA Condensates

The physics and morphology of biomolecular condensates that form via liquid-liquid phase separation underpin numerous biological processes. Biomolecular condensates can also form complex multiphase morphologies, such as the nested configuration of nucleoli that facilitate the production of ribosomes. Here, we investigate how to design the morphology of multiphase condensates that form in solutions of DNA nanostars. Since the morphology of multiphase condensates is dictated by surface energies between pairs of phases, we developed coarse-grained molecular dynamics simulations to investigate how surface energies depend on the properties of nanostars (size, number of arms, sequences of sticky ends on each arm), temperature, and electrostatic screening. We also systematically studied morphologies of condensates with two types of DNA nanostars. We found that Janus-like morphologies are ubiquitous because the two condensed phases have similar surface energies. On the other hand, nested morphologies are rare because they require the two condensed phases to have drastically different surface energies, which is only possible for highly asymmetric types of DNA nanostars (different numbers of arms, sizes, and distribution of sticky ends).