Fig 1.
Transition from spread to rounded states and localization of F-actin and myosin in the cortex of a typical rounded cell and cortical structure.
A. Merged DIC and fluorescence image of spread Swiss 3T3 fibroblasts stably transfected with Lifeact-RFP (green) cells with clearly visible signal from stress fibers. Cell outlines traced from image. Bar = 20μm. B. Image of the same cell after trypsin-induced detachment. The white outline shows the former spread shape. Yellow arrow points to the rounded cell. C. Cartoon compares the radius of a sphere required to accommodate the cell volume from the spread state (R = 50μm) versus the rounded state (R = 13 μm) for the cell on the left. D. Distribution of spread and rounded cell areas. E. Distribution of rounded cell radii. F. DIC and confocal fluorescence images of rounded CHO cell stably expressing GFP-Lifeact (green) and RFP-MLC (red) Bar = 5 μm. G. Transmission electron microscopy image of GFP immunogold staining of rounded CHO cells with stable expression of Lifeact-GFP. Black dots which represent gold particles show the position of actin filaments. Bar = 1 μm. H. Inset shows outlined region in G at higher magnification. Arrow points to F-actin immediately underlying the plasma membrane in BLiPs: arrowhead points to the F-actin in the cortex closer toward center of the cell. Bar = 0.5 μm.
Fig 2.
A. Two layer cortex model (left) and its bead-spring representation with notations employed (right). B. Folded geometries as a function of the fold number and the excess surface ratio. C. A portion of a model cell with BLiPs at steady state where the gold line represents the cortex. D. The shape of three single folds extracted from simulations for initial excess ratio, ER = 4 with different numbers of folds, N: 20 (blue), 40(red) and 50(yellow). E. The shape of three single folds extracted from simulations for number of folds N = 40 with different initial excess ratio, ER,: 6 (blue), 4 (red), 2 (yellow). All calculated shapes were scaled to have the same unit area inside the outer perimeter. The bending energy (BE) is presented for each configuration.
Fig 3.
A. Fluorescence, DIC and merged images of rounded CHO cells stably transfected with Lifeact-GFP. B. Fluorescence, DIC and merged images taken near the equatorial plane of a rounded CHO cell with the fluorescence signal coming from the PH domain of PLC-delta fused to EGFP that marks the inner leaflet of the plasma membrane (Bar = 5 μm). C. Scanning electron microscope image of the rounded state (Bar = 5 μm). D. BLiP radii distribution for N = 7096 BLiPs.
Fig 4.
BLiPs morphology generated by seed-and-growth models.
A. A realization of the Voronoi diagram. The seeds and boundaries of Voronoi cells are shown in blue; the Voronoi polygons are shown in red. B. A realization of the “seed and growth” model. C. Magnified view of rectangle in (B). D. Distribution of BLiP sizes: Voronoi model (green bars) and “seed and growth” model (red bars) normalized using the assumption that average rounded cell radius is 8 um; the blue line shows the experimentally obtained distribution. E. For the “seed and growth” model, plots of % surface area stored in BLiPs and % of total volume stored in BLiPs as a function of the number of BLiPs. F. A cross-section of “seed and growth” model. G. Transmission electron micrograph of a section of a rounded cell (Bar = 2 μm) for comparison to F.
Fig 5.
Schematic for the phase field formulation of a cell in an aqueous medium.
ϕ1, ϕ2, ϕ3, represent volume fractions of the external aqueous medium, the nematic cortex as schematically depicted by the cross-hatched region and the cytosol, respectively, with ϕ1+ϕ2+ϕ3 = 1. The entire computational domain is denoted as Ω. The cell surface is defined by the level sets ϕ1 = ϕ2 = 0.5, while the cortex-cytosol interface is defined by ϕ2 = ϕ3 = 0.5. Note that we penalize coexistence of three phases.
Fig 6.
Application of the phase field model to 2D target images.
A. Proof of principle of the 2D phase field simulation. From left to right, convergence from circular initial data to a target cell surface morphology with 25 uniform, equally spaced, “BLiPs”. B. Convergence of a 2D phase field simulation to a 2D, TEM image of a representative cell surface morphology. The red curve in panels 1–3 is the cell surface obtained from a 2D TEM micrograph, which serves as the target of the phase field model. The black contours in each panel are the initial data (left panel) which evolves to the actual cell surface in the phase field simulation. The green contours depict the interface between the cortex and interior cytosol. In the right panel, the F-actin filament orientational distribution within the nematic cortex is superimposed, as predicted by the phase field model. In these simulations, the Flory order parameter, , is set to 1. C. Phase field predictions of the pressure distribution. D. the first invariant (trace) of the dominant stored stress, the Ericksen stress, for the converged stationary morphology shown in B, right panel. E. A blow-up of the trace of the Ericksen stress inside the dashed yellow rectangular domain in D. The color bars for C, D, and E are in units of Pascals.
Fig 7.
Phase field simulation showing convergence to a target 3D cell morphology image.
(A) The target 3D cell morphology. (B) The assumed initial data for the cell surface. (C) The converged cell surface from the phase field simulation. (D) 2D planar slices at z = 0.5, y = 0.5 and x = 0.5, respectively, of the converged 3D phase field simulation shown in C. The color bar depicts level sets of the cortical phase variable ϕ2 to delineate the cortex (red is the level set ϕ2 = 1) from the pure external aqueous medium and pure interior cytosol (both blue since ϕ2 = 0), while both diffuse interfaces are depicted by the color interpolation between these level sets of ϕ2. (E) Corresponding pressure distributions (units of Pa) in the 2D slices shown in (D). (F) Distribution of the trace of the Ericksen stress (units of Pa) in the 2D slices shown in (D). The color bar for E, F is in units of Pa. For these calculations the Flory order parameter is set to 1.
Fig 8.
The orientational distribution of F-actin filaments in the nematic cortex for the steady state cell morphology associated with the 3D target morphology of Fig 7.
Since the Flory order parameter is unknown, for illustrative purposes we impose its maximum value of 1 in this simulation, and impose lower values in the S6 and S7 Figs. (A) 3D view; (B-D) 2D planar projections of the nematic director field in the cortex in the x = 0.5, y = 0.5, z = 0.5 planes, respectively. The color bar shows the magnitude of nematic director, where |p| = 1 denotes nematic phase and |p| = 0 denotes the isotropic phase.