Non-specific adhesive forces between filaments and membraneless organelles

Abstract

Membraneless organelles are liquid-like domains that form inside living cells by phase-separation. While standard physical models of their formation assume their surroundings to be a simple liquid, the cytoplasm is an active viscoelastic environment. To investigate potential coupling of phase separation with the cytoskeleton, we quantify structural correlations of stress granules and microtubules in a human-derived epithelial cell line. We find that microtubule networks are significantly perturbed in the vicinity of stress granules, and that large stress granules conform to the local pore-structure of the microtubule network. When microtubules are depolymerized by nocodazole, tubulin enrichment is localized near the surface of stress granules. We interpret these data using a thermodynamic model of partitioning of particles to the surface and bulk of droplets. This analysis shows that proteins generically have a non-specific affinity for droplet interfaces, which becomes most apparent when they weakly partition to the bulk of droplets and have a large molecular weight. In this framework, our data is consistent with a weak (lower than kbT) affinity of tubulin sub-units for stress granule interfaces. As microtubules polymerize their affinity for interfaces increases, providing sufficient adhesion to deform droplets and/or the network. We validate this basic physical phenomena in vitro through the interaction of a simple protein-RNA condensate with tubulin and microtubules.

Condensates of disordered proteins have small-world network structures and interfaces defined by expanded conformations

“…This implies that the width of the interface is at least as large as the mean end-to-end distance of a flexible PLCD. This observation is consistent with inferences reported in a recent study by Böddeker et al, 53 of condensates being defined by “fat” interfaces.…”

“…Overall, the results in Fig. 4 show that the interface, even away from Tc, is not thin, and is consistent with inferences reported in a recent study by Böddeker et al, 49 of condensates being defined by "fat" interfaces.…”

“…6. Our results suggest that interfaces between condensates and the coexisting dilute phases should be thought of as being “fat” 49 rather than “thin”. This feature of the interface is realized by the ability of disordered proteins to be relatively more expanded, both locally and globally, when compared to the dilute and dense phases.…”

The interface of condensates of the hnRNPA1 low complexity domain promotes formation of amyloid fibrils

“…In addition to promoting conformational changes and orientation 59 , interfaces of condensates can preferentially partition components 66 and accelerate or inhibit fibrillation of client molecules 67 . Overall, interfaces can therefore represent a special location for biochemical activity within condensates and contribute to the complex interplay between condensation and fibrillation.…”

Small, fat-filled lipid droplets remain spherical as they indent a nucleus, dilute the lamina, and cause rupture

“…The particular value of γ is important because the pressure (force-per-area) that an FD sustains and exerts against a nucleus, for example, is set by γ multiplied by the FD’s curvature (radius −1 ) ( Fig.1A-ii ); furthermore, any distortion from a spherical shape necessarily increases an FD’s surface area and costs an energy set by γ. Measurements of γ are currently known only for FD’s from sonicated lysates of kidney cells (Cos7 and HEK) and fly hemocyte-like embryonic cells, with γ = 2-5 mN/m for ∼3-10 µm diam FD’s (Ben M’barek et al, 2017), (Lyu et al, 2021) ( Fig.1Aii, scale ). For comparison, protein condensates have far lower γ ∼0.01 mN/m) and can be highly elongated in cells to “ conform to the local structure ” such as the cytoskeleton (Boddeker et al, 2022). A typical cell membrane will also quickly rupture under an effective tension of ∼10 mN/m (Evans et al, 1976), which implies that pressing a soft condensate and perhaps an FD against a cell membrane is unlikely to rupture a membrane.…”