Pickering stabilization of a dynamic intracellular emulsion

Biomolecular condensates are cellular compartments that form by phase separation in the absence of limiting membranes. Studying the P granules of C. elegans, we find that condensate dynamics are regulated by protein clusters that adsorb to the condensate interface. Using in vitro reconstitution, live observations and theory, we demonstrate that localized assembly of P granules is controlled by MEG-3, an intrinsically disordered protein that forms low dynamic assemblies on P granules. Following classic Pickering emulsion theory, MEG-3 clusters lower surface tension and slow down coarsening. During zygote polarization, MEG-3 recruits DYRK/MBK-2 kinase to accelerate localized growth of the P granule emulsion. By tuning condensate-cytoplasm exchange, interfacial clusters regulate the structural integrity of biomolecular condensates, reminiscent of the role of lipid bilayers in membrane-bound organelles. One Sentence Summary Biomolecular condensates are stabilized by interfacial nanoscale protein clusters.

condensates that become kinetically-arrested over time, in some cases becoming glass-like (2,3). Fast dynamic condensates form emulsions that coarsen down to a single large condensate (4)(5)(6)(7). In contrast, kinetically-arrested condensates form emulsions that resist coarsening over long time scales (8) . Like condensates in vitro, condensates in cells exhibit a range of dynamic behaviors, but these do not always fit theoretical predictions (9,10). For example, some condensates resist coarsening for hours, yet maintain fast exchange dynamics and dissolve within minutes in response to changes in the cellular environment (11). Several hypotheses have been put forward to explain the lack of coarsening in dynamic cellular emulsions, including physical barriers that keep condensates away from each other (5,12,13), active mechanisms that continuously regenerate small condensates (14), and chemical reactions and protein 10 gradients that suppress Ostwald ripening (14-17) (see Supplemental Text). In this study, we investigate the mechanisms that control the dynamics of P granules, condensates in the C. elegans germline. We find that P granule dynamics are controlled by nanoscale protein clusters that adsorb to the condensate interface, a phenomenon first described by Ramsden (1904) and Pickering (1907) for inorganic emulsions (18,19). 15 P granules were the first cellular condensates proposed to form by LLPS (20). At the core of P granules is a liquid-like phase assembled by PGL proteins (20)(21)(22). During most of the C. elegans life cycle, PGL condensates form stable assemblies on the cytoplasmic phase of nuclei (23). During the oocyte-to-zygote transition, PGL condensates redistribute to the cytoplasm and undergo two rapid cycles of dissolution and condensation (Fig. 1A). The first 20 cycle occurs during oocyte maturation: most PGL condensates in the maturing oocyte dissolve and reassemble minutes later in the newly fertilized egg (zygote). The second cycle occurs when the zygote becomes polarized along its anterior-posterior axis: PGL condensates in the anterior cytoplasm dissolve while PGL condensates in the posterior grow (Fig. 1A). Factors required for P granule dissolution during oocyte maturation have not yet been identified. Factors 25 required for dissolution during polarization include MEX-5, an RNA-binding protein, MEG-3 and MEG-4, two paralogous intrinsically disordered proteins, and MBK-2, a DYRK family kinase that interacts physically and genetically with MEG-3 (21,24,25). During polarization, MEX-5 becomes enriched in the anterior cytoplasm where it promotes P granule dissolution, possibly by competing with P granule proteins for RNA (21,26). In zygotes lacking mbk-2 or meg-3 meg-30 4 activity, P granules do not dissolve in the anterior cytoplasm, despite a normal MEX-5 gradient (27). In this study, we investigate how the MEGs and MBK-2 collaborate with MEX-5 to regulate P granule dynamics.
Super resolution 3D confocal microscopy confirmed that MEG-3 forms diffraction-limited clusters (<160 nm) on the PGL-3 interface in vivo and in vitro ( Fig. 1B and fig. S1). MEG-3 clusters are resistant to dilution, high temperature, and salt treatment, in contrast to PGL-3 condensates which readily dissolve in dilute conditions and at elevated temperatures (28).
Using a single-molecule method adapted from Wu et al., 2019 (29), we measured the dynamics of MEG-3 and PGL-3 molecules in P granules in vivo ( Fig. 1C and D, and movies S1 and S2).
Most PGL-3 molecules exhibited short-lived trajectories in P granules with an average apparent diffusion coefficient of D c = 0.056 μm 2 /s ( Fig. 1D and fig. S2). In contrast, all MEG-3 molecules exhibited restricted long-lived trajectories with an average apparent diffusion coefficient of Dc = 0.0018 μm 2 /s ( Fig. 1C and D, and fig S2). In 3/3 cases where we captured the trajectories of three labeled MEG-3 molecules in the same P granule, their relative position remained fixed over time ( Fig. 1C and D, and movie S2). Together these observations confirm that PGL-3 molecules exist primarily in a dynamic liquid-like phase, whereas MEG-3 molecules experience 15 much slower dynamics resembling solid clusters within our experimental time scales.
Solid particulates that adsorb to liquid surfaces reduce surface tension and stabilize emulsions against coarsening (so-called "Pickering agents" (30)). To explore whether MEG-3 exhibits properties of a Pickering agent, we first tested whether MEG-3 affects the interfacial properties of PGL-3 condensates assembled in vitro. "Naked" PGL-3 droplets readily wet the 20 surface of untreated glass slides ( Fig. 1E and fig. S3A). In contrast, PGL-3 droplets coated with MEG-3 clusters exhibited lower wetting on glass slides, consistent with MEG-3 clusters adsorbing to the PGL-3 interface ( Fig. 1E and fig. S3A). To determine whether MEG-3 stabilizes the PGL-3 emulsion against coarsening, we examined the evolution of a newly assembled PGL-3 emulsion overtime. We found that over the course of 180 minutes, the PGL-3 emulsion  We conclude that MEG-3 stabilizes the PGL-3 emulsion in a concentration-dependent manner, consistent with MEG-3 acting as a classic Pickering agent in vitro.
Even in the absence of MEG-3, coarsening of the PGL-3 emulsion was slow, occurring on a minute-to-hour time scale. Previous studies have shown that PGL-3 condensates are "aging Maxwell fluids" whose viscosity strongly increases overtime, eventually adopting glasslike properties (see Supplemental Text, (2,32)). To further explore the properties of the PGL-3 phase, we examined the dynamics of PGL-3 condensates challenged with excess PGL-3 protein. We found that within less than 1 minute of assembly, PGL-3 condensates became refractory to incorporation of new PGL-3, causing excess PGL-3 to form separate condensates ( Fig. 2A). Mixing between "old" and "new" PGL-3 occurred progressively, indicating that PGL-3 molecules continue to exchange between the condensed and dilute phases (Fig. 2B). Addition of MEG-3 reduced coarsening but did not affect significantly the rate of old and new PGL-3 mixing, as expected for a surface agent that does not block exchange at the PGL-3 interface ( Fig. 2A and C). These observations are consistent with PGL-3 condensates rapidly (within seconds) approaching kinetic arrest, where the soluble-to-condensate rate slows down significantly, limiting growth and coarsening.
The slow soluble-to-condensate conversion rate of the PGL-3 emulsion is consistent 15 with the apparent stability of P granules during most of germline development and suggests that active processes must operate to dissolve P granules during oocyte maturation and polarization.
MBK-2 is required for P granule dissolution during polarization and its mammalian homolog DYRK3 has also been implicated in the dissolution of condensates in mammalian cells (27,33,34). We found that recombinant DYRK3 phosphorylates PGL-3 efficiently in vitro ( fig. S4A).

20
Addition of DYRK3 to pre-assembled PGL-3 condensates (2.5 μM) led to their complete dissolution over 60 minutes ( fig. S4B to E). At 5 μM PGL-3, condensates could be maintained in the presence of DYRK3, but coarsened rapidly such that by 60 minutes only a few, large PGL-3 droplets (>5 μm) remained ( Fig. 2D to E). Addition of MEG-3 did not affect the ability of DYRK3 to phosphorylate PGL-3 ( fig. S4F), yet changed coarsening dynamics. In the presence of MEG-  To examine the impact of MEG-3 on P granule dynamics in vivo, we used quantitative live-cell imaging to measure the number and size of PGL-3 condensates in wild-type and meg-3 meg-4 mutants. We began by imaging eggs in utero as they progress through oocyte maturation and fertilization (Fig. 3A). We found that the volume of PGL-3 in condensates decreased 10-fold during oocyte maturation and remained low in newly fertilized zygotes as they complete the meiotic divisions (Fig. 3B to D and fig. S5A). Total PGL-3 levels did not change during this period ( fig. S5B) consistent with a transient increase in PGL-3 solubility. We observed the same decrease in PGL-3 condensate volume in wild-type and meg-3 meg-4 oocytes, indicating that the increase in PGL-3 solubility during the oocyte-to-zygote transition is not dependent on meg-3 meg-4. The size distribution of PGL-3 condensates post-dissolution, however, was different in the two genotypes. The PGL-3 emulsion coarsened rapidly in meg-3 meg-4 zygotes with fewer larger condensates dominating the emulsion (Fig. 3D). In contrast, wild-type zygotes maintained many small PGL-3 condensates, consistent with MEG-3 stabilizing the PGL-3 emulsion against coarsening (Fig. 3D). Zygotes in late meiosis can survive outside of the uterus allowing for the acquisition of high-resolution images ex utero. These images confirmed that wild-type zygotes contain dozens of <1 μm condensates not observed in meg-3 meg-4 zygotes ( Fig. 3A and E).
These observations suggest that MEG-3 functions as a Pickering agent for the PGL-3 emulsion. To examine the physical plausibility of this hypothesis, we modeled in silico the 15 kinetics of an idealized PGL-3 emulsion in the presence or absence of a Pickering agent that lowers surface tension. To account for the intrinsic tendency of PGL-3 towards a slow solubleto-condensate conversion rate, we modeled PGL-3 dynamics under a "conversion-limited" scheme, where the soluble-to-condensate conversion rate is > 10-fold slower than the diffusionlimited adsorption/desorption rate of PGL-3, and therefore rate limiting for condensate  zygote, fuse and initiate the first mitotic division (Fig. 4A). During pronuclear formation and fusion, we observed new PGL-3 condensates appearing throughout the cytoplasm in both wildtype and meg-3 meg-4 zygotes ( Fig. 4B and C, and fig. S6A). The total volume of PGL-3 in condensates also increased ( fig. S6B), without a change in total PGL-3 ( fig. S6C), consistent with a return to low PGL-3 solubility in both genotypes. During this period, total condensate volume in the anterior half of the zygote decreased steadily eventually reaching zero, while total condensate volume in the posterior increased (Fig. 4D). Remarkably, in meg-3 meg-4 zygotes, we also observed a decrease and increase in total condensate volume in the anterior and posterior, respectively, but the amplitude of the change was greatly reduced (Fig. 4E). By mitosis, in wild-type, all PGL-3 condensates were restricted to the posterior cytoplasm; in contrast, in meg-3 meg-4 zygotes, PGL-3 condensates remained stable throughout the cytoplasm through the first cell division, suggesting that in meg-3 meg-4 mutants, the PGL-3 emulsion is no longer responding efficiently to the MEX-5-driven solubility gradient (Fig. 4A to E).
To examine condensate dynamics directly, we tracked individual PGL-3 condensates in live zygotes from pronuclear formation to pronuclear meeting ( Fig. 4F to I, and movies S3 and S4). These observations confirmed that, in meg-3 meg-4 zygotes, condensates experience very slow growth/dissolution rates during polarization, with a slight bias for decay in the anterior (k A,avg -0.19 nm/s, k P,avg = 0.06 nm/s, Fig. 4G and I). Growth/decay rates were more than 5-fold faster in wild-type zygotes with a clear bias for dissolution in the anterior and condensation in the posterior (k A,avg -1.18 nm/s, k P,avg = 0.50 nm/s, Fig. 4F and H). These findings suggest that, unlike in oocytes where PGL-3 dissolution occurs independently of meg-3 and meg-4, during polarization meg-3 and meg-4 are required to accelerate PGL-3 condensate dynamics.
In addition to meg-3 and meg-4, P granule dissolution during polarization requires the DYRK kinase MBK-2 (35,36). Genetic epistasis experiments have shown that MBK-2's dissolution activity is dependent on meg-3 and meg-4 (27) (Supplementary text). Consistent with these observations, using an epitope tagged allele of endogenous MBK-2, we found that MBK-2 is recruited to PGL-3 condensates during polarization and this recruitment is enhanced by meg-3 and meg-4 ( Fig. 4J and K). MEG-3 and MBK-2 levels varied more than five-fold among PGL-3 condensates ( fig. S7A to D). Condensate growth rates during polarization were also heterogeneous in a manner that did not correlate with initial condensate size ( Fig. 4F  Since their description by Ramsden and Pickering in the early 1900s (18, 19), Pickering 30 agents have been used widely to stabilize emulsions in the pharmaceutical, energy and food industries (22-24). Unlike surfactants (amphiphilic molecules that insert at interfaces), Pickering agents are nanoscale solid particulates that adsorb to interfaces upon partial wetting by both phases. Adsorption is energetically favored and balances the drive to reduce interfacial area, stabilizing the emulsion against Ostwald ripening and coalescence (31). Many types of solid particulates have been shown to function as Pickering agents, from silica to denatured proteins (37-39). The first described intrinsically disordered protein, casein, functions as a natural Pickering agent in homogenized milk (40). To our knowledge, the intrinsically disordered protein MEG-3 is the first characterized example of an intracellular Pickering agent and we speculate that other self-assembling biopolymers will exhibit similar properties. In somatic cells, PGL droplets are covered by EPG-2 clusters that may function like MEG-3 to regulate the size and dynamics of PGL droplets in preparation for autophagy (41). Artificial protein-RNA assemblies that adsorb to the surface of stress granules have been reported to influence their size and coalescence (42). mRNAs that accumulate on the surface of protein condensates could also serve as stabilizing agents (43,44). Given the rich diversity of biopolymers in cells, it is tempting to speculate that biopolymers acting as Pickering agents will prove a general organizing principle for biomolecular condensates. By interacting with enzymes like the DYRK kinase MBK-2, biological Pickering agents also regulate interfacial exchange to control the influx and efflux of molecules in and out of condensates in response to environmental changes. In this 15 respect, surface-adsorbed biopolymers fulfill a boundary function for biomolecular condensates, reminiscent of the role of lipid bilayers and associated machineries (e.g. channels) in membrane-bound organelles. In fact, it has been speculated that membranes first arose from liposomes that functioned as Pickering stabilizers for aqueous emulsions (45).

Acknowledgments:
We thank the Johns Hopkins Integrated Imaging center (S10OD023548) for microscopy support. We thank the Lavis lab for HaloTag Ligands JF 549 and JF 646 , the Griffin lab for MEG-3::Halo strain, We thank the Waugh Lav for TEV protease (pRK793, Addgene), Tony Hyman, the Baltimore Worm club and the Seydoux lab for many helpful discussions. G and H) Histograms plotting the size distribution of PGL condensates assembled as in (F). Each data point indicates the fraction of total PGL-3 condensate volume represented by condensates binned by radius from 80 images (as in F) collected in 4 independent experiments. Lines fit the data to a log normal distribution.

Fig. 2. MEG-3 stabilizes PGL-3 condensates against kinase accelerated coarsening.
A) Photomicrographs of a PGL-3 488 emulsion (magenta) 60 min post initial assembly and at   3::mCherry (white). Photomicrographs were captured in live adult hermaphrodites (in utero) or after dissection out of the uterus (ex utero). Representative photomicrographs represent ~20% of oocyte volume and ~80% of zygote volume. The anterior (left) bias for PGL condensates in meg-3 meg-4 zygotes correlates with anterior displacement of the oocyte nucleus (and associated P granules) that occurs immediately prior to fertilization. Scale bars are 10 μm.
B-E) Histograms of PGL condensate volumes measured from images captured as in A representing 100% of oocyte and zygote volumes. Circles indicate the volume of individual PGL-3 in condensates binned by condensate radius in wild-type (green) and meg-3 meg-4 (black) oocytes and zygotes. Volumes are higher in E compared to D due to higher detection sensitivity ex utero.