Active transport enables protein condensation in cells

Abstract

Multiple factors drive biomolecular condensate formation. In plants, condensation of the transcription factors AUXIN RESPONSE FACTOR 7 (ARF7) and ARF19 attenuates response to the plant hormone auxin. Here, we report that actin-mediated movement of cytoplasmic ARF condensates enhances condensation. Coarse-grained molecular simulations of active polymers reveal that applied forces drive the associations of macromolecules to enhance phase separation while giving rise to dense phases that preferentially accumulate motile molecules. Our study highlights how molecular motility can drive phase separation, with implications for motile condensates while offering insights into cellular mechanisms that can regulate condensate dynamics.

Fig. 1.. ARF condensates form in the cytoplasm and undergo active motion.

(A) Schematic of the phase diagram for systems undergoing thermoresponsive phase separation. The green curve represents the phase boundary. For a given temperature, concentrations 2, 3, and 4 lie within the two-phase regime. (B) Representative image showing the fluorescence intensity of ARF19 in the nucleus versus the cytoplasm. (C) Quantification of the dilute phase fluorescence intensities for ARF19 in the nucleus versus the cytoplasm. a.u., arbitrary units. (D) Time-lapse images of an ARF condensate (white) moving along the LifeAct-tagged actin filaments (red) in the cytoplasm during transient expression in tobacco. (E) Analysis of local MSDs over short-time intervals following the approach of Arcizet et al. (17) Here, the trajectories R(t) is defined as R(t) = 〈[r(t) − r₀]²〉^0.5, where r(t) is the position of the condensate at time t and r₀ is the position at t = 0. The trajectories, shown here for ARF7 condensates, are colored on the basis of the exponent for the MSDs evaluated over short-time windows.

Fig. 2.. Condensate motility is aided by the actin network.

(A) Joint probability densities for angles and speeds of the tracked pARF19:ARF19-mVenus condensates from seedlings incubated for 2 hours in a mock (DMSO) or 10 μM LatB treatment. Data collected using TrackMate (58) are shown for five distinct condensates captured from immature trichoblasts over the course of 5 min at a frame rate of 2.3 frames/s. (B) Marginal probability densities for areas of condensates were computed for condensates corresponding to different speed intervals. These were compared to the distributions obtained upon LatB treatment (dashed curve). For high-speed condensates, we observe a pronounced shift toward condensates of large areas. (C) KL divergence computed for each distribution as the posterior and the distribution obtained from LatB treatment as the prior. (D) Distance R(t) traversed by ARF19 condensates at time t after a 2-hour treatment with mock (DMSO) or 10 μM LatB. Local scaling of MSD with lag time is used to color the trajectories based on the exponents. Values of α > 1 indicate active motion. (E) Area of condensates after a 10-min treatment with mock (DMSO) or LatB. (F) Angle and speed distributions for YFP-ARF19 condensates in young root epidermal cells of wild-type and myo3KO seedlings carrying pUBQ10:YFP-ARF19. (G) Confocal images of young root epidermal cells of wild-type and myo3KO seedlings carrying pUBQ10:YFP-ARF19. (H) Mean YFP-ARF19 condensate areas (±SD) in immature trichoblasts of wild type (Col-0), myo2KO, or myo3KO carrying pUBQ10:YFP-ARF19.

Fig. 3.. Simulations show how motility enhances the driving forces for condensation.

(A) Langevin dynamics simulations were deployed to study the effects of forces applied along the z-direction to each polymer in a simulation box comprising 650 distinct bead-spring polymers. We apply F_act at each time step and study the phase behavior of polymers using multiple chain simulations. (B) Equilibrium snapshots from the simulations of two-phase systems for different strengths of the applied force F_act at the end of the simulation. (C) Radial density profile of the polymer beads from the center of the condensate. COM, center of mass. (D) Coexistence curves, derived from the radial density profiles, are used to delineate the two-phase regime bracketed by the coexisting dilute phase (left arm) and dense phase (right arm) at different values of F_act. (E) Variation of ρ_sat as a function of F_act shows the log-linear relationship between these quantities. (F) Velocity of molecules in the dense phase plotted as a function of the magnitude of the active force F_act. The units of velocity are $\mathrm{\sigma }/\mathrm{\tau }\prime$, where $\mathrm{\tau }\prime =200\mathrm{\tau }$ and $\mathrm{\tau }$ is the natural unit of time. (G) Radial density profile of the polymer beads in the direction perpendicular to applied force (z = 0 plane) from the center of the condensate. (H) Radial density profile in the direction parallel to applied force from the center of the condensate (x = 0 plane and y = 0 plane). (I) Interfacial widths in the direction parallel and perpendicular to the applied force. The error bars denote the standard error of the mean over the course of the simulation trajectory.

Fig. 4.. Passive molecules are left behind in the dilute phase as F_act increases in magnitude.

(A) Phase boundaries for an active-passive mixture as a function of active force for different fractions of passive chains f_passive. (B) ρ_sat as a function of the active force F_act for different fractions of the passive polymers f_passive. (C) Number of active chains in the dense phase normalized by total chains in the system as a function of the active force (F_act) for different fractions of the passive polymers f_passive. (D) Fraction of the passive chains in the dense phase as a function of the active force (F_act) for different fractions of the passive polymers f_passive. The error bars are the standard error of the mean over the course of the simulation trajectory.