Epithelial cell chirality emerges through the dynamic concentric pattern of actomyosin cytoskeleton

Abstract

The chirality of tissues and organs is essential for their proper function and development. Tissue-level chirality derives from the chirality of individual cells that comprise the tissue, and cellular chirality is considered to emerge through the organization of chiral molecules within the cell. However, the principle of how molecular chirality leads to cellular chirality remains unresolved. To address this fundamental question, we experimentally studied the chiral behaviors of isolated epithelial cells derived from a carcinoma line and developed a theoretical understanding of how their behaviors arise from molecular-level chirality. We first found that the nucleus undergoes clockwise rotation, accompanied by robust cytoplasmic circulation in the same direction. During the rotation, actin and Myosin IIA assemble into the stress fibers with a vortex-like chiral orientation at the ventral side of the cell periphery, concurrently forming a concentric pattern at the dorsal side. Further analysis revealed that the intracellular rotation is driven by the concentric actomyosin filaments located dorsally, not by the ventral vortex-like chiral stress fibers. To elucidate how these concentric actomyosin filaments induce chiral rotation, we analyzed a theoretical model developed based on the theory of active chiral fluid. This model demonstrated that the observed cell-scale unidirectional rotation is driven by the molecular-scale chirality of actomyosin filaments even in the absence of cell-scale chiral orientational order. Our study thus provides novel mechanistic insights into how the molecular chirality is organized into the cellular chirality, representing an important step toward understanding left-right symmetry breaking in tissues and organs.

Keywords: cell biology; cell chirality; chiral active matter; cytoskeleton; left–right asymmetry; none; physics of living systems.

Figure 1.. Chiral nuclear rotation and cytoplasmic flow in singly isolated Caco-2 cells.

(A) Rotating nucleus probed by the rotation of nuclear texture. The endpoints of the red line segments are the positions of tracked landmarks of the nucleus. (B) The cumulative angle of nuclear rotation plotted against time and (C) average angular velocity averaged over the first 10 hr (${\displaystyle n=22}$). Here, positive angle values indicate clockwise rotation. (D) Chiral cytoskeletal structure of F-actin (phalloidin) and microtubule (immunostaining). Scale bar: 20 µm. (E) Schematic diagram of the orientation of actin stress fibers and microtubule.

Figure 1—figure supplement 1.. Effect of physical environment on the nuclear rotation.

Angular velocity of nucleus on the different coating applied to the glass substrate: collagen (n = 19), non-coated (n = 21), fibronectin (n = 22) and poly-l-lysine (n = 22). p values were calculated using Mann–Whitney U test (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 1—figure supplement 2.. Actin bundles in the peripheral region of the cell.

In the control cell (DMSO), actin bundles (phalloidin, magenta) in the peripheral region of cell, which are tilted to form chiral pattern, appear to be anchored to vinculin (green), a focal adhesion protein, at their both ends. We call them stress fibers. In the cell treated with SMIFH2, one end of each actin bundle was anchored to vinculin, while the other ends were not anchored. Consequently, the actin bundles extend in the radial direction. We call them radial fibers.

Figure 2.. Identification of molecular mechanisms responsible for the chiral rotation.

Roles of F-actin, microtubule, Arp2/3, formin-mediated actin polymerization, and Myosin II activity were investigated. Cells were treated with DMSO (0.2%, control), actin polymerization inhibitor latrunculin A (2 µM), actin depolymerization inhibitor Jasplakinolide (10 nM), microtubule inhibitor nocodazole (50 µM), Arp2/3 inhibitor CK666 (200 µM), formin inhibitor SMIFH2 (40 µM), or Myosin II inhibitor blebbistatin (1 µM). (A) Snapshot images from the live image of actin dynamics in cells expressing Lifeact-RFP. Scale bar: 20 µm. (B) The cumulative angle of nuclear rotation averaged over different cells plotted against time for different conditions: DMSO (${\displaystyle n=19}$), nocodazole (${\displaystyle n=11}$), blebbistatin (${\displaystyle n=10}$), CK666 (${\displaystyle n=13}$), and SMIFH2 (${\displaystyle n=10}$). The standard deviation is represented by shaded regions. (C) Angular velocity of cells under different conditions averaged over the first 5 hr of the time-evolution plot in (B). (D) Angular velocity of control and SMIFH2-treated cells averaged over the last 5 hr of the time-evolution plot in (B): DMSO (${\displaystyle n=13}$) and SMIFH2 (${\displaystyle n=10}$). p values were calculated using the Mann–Whitney U test (${\displaystyle {\displaystyle \ast p<0.05,\ast \ast p<0.01,\ast \ast \ast p<0.001}}$). Here, positive angle values indicate clockwise rotation.

Figure 2—figure supplement 1.. Microtubule inhibition by nocodazole.

Typical snapshots of microtubule of a cell 0.5 and 3.5 hr after the addition of nocodazole (50 µM).The scale bar is 20 µm.

Figure 2—figure supplement 2.. Effect of formin depletion on the nuclear rotation.

(A) RNA-sequencing (RNA-seq) analysis showing gene expression levels (transcripts per million; TPM) of major mammalian formin family members (DIAPHs and DAAMs) in Caco-2 cells. Circles represent individual samples (circles of the same color indicate replicates), and bars indicate mean values (n = 3). Western blot showing protein levels of DIAPH2 (B) and DAAM1 (C) in Caco-2 cell treated with siRNAs.GAPDH (bottom row) was used as an internal control. (D) Angular velocity of nucleus of Caco-2 cells that were treated with siRNA for DIAPH2 or DAAM1. The rotation of the nucleus was tracked for 5 hr following the start of the live imaging. Negative control (N.C.) (n = 10), DIAPH2 (n = 18), and DAAM1 (n = 17). p values were calculated using Mann–Whitney U test (*p < 0.05, **p < 0.01, ***p < 0.001). Snapshot images from the live image of actin dynamics in cells expressing Lifeact-RFP in DIAPH2 (E, E’) and DAAM1 (F, F’) depleted cells. Scale bar: 20 µm.

Figure 2—figure supplement 3.. Effect of Myosin II depletion on the nuclear rotation.

(A) Angular velocity of nucleus of Caco-2 cells that were treated with siRNA for Myosin II A and/or B heavy chains. The rotation of the nucleus was tracked for 5 hr following the start of the live imaging. Negative control (N.C.) (n = 10), Myosin II A (n = 19), Myosin II B (n = 15), and Myosin IIA and B (n = 20). p values were calculated using Mann–Whitney U test (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Western blot showing protein levels of Myosin IIA (top row) and IIB (middle row) heavy chains in Caco-2 cells treated with siRNAs. GAPDH (bottom row) was used as an internal control. Quantification of fold change relative to the negative control (N.C.), normalized to GAPDH protein level, is shown in bar graphs (bottom panel). Images of Myosin IIA (C) or Myosin IIB (D) (immunostaining, green) and actin filaments (phalloidin, magenta) in cells treated with siRNAs.

Figure 2—figure supplement 4.. Effect of vinculin depletion on the nuclear rotation.

(A) Angular velocity of nucleus of Caco-2 cells that were treated with siRNA for vinculin. The rotation of the nucleus was tracked for 5 hr following the start of the live imaging. Negative control (N.C.) (n = 10) and vinculin (n = 17). p values were calculated using Mann–Whitney U test (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Western blot showing protein levels of vinculin in Caco-2 cell treated with siRNAs. GAPDH (bottom row) was used as an internal control. Quantification of fold change relative to the negative control (N.C.), normalized to GAPDH protein level, is shown in bar graphs (bottom panel). (C) Images of actin filaments (phalloidin, magenta) and vinculin (immunostaining, green) in cells treated with siRNAs. Scale bar: 20 µm.

Figure 3.. Organization of F-actin and Myosin IIA.

(A) Control cells treated with DMSO show a chiral tilted pattern of F-actin and Myosin II visualized by phalloidin and immunofluorescence with an antibody against Myosin IIA, respectively. (B) SMIFH2 (40 μM) treated cells show a concentric pattern of F-actin and Myosin II. (C) The chiral tilted pattern of F-actin and Myosin II is suppressed in cells treated with blebbistatin (1 μM). The bottom panel shows vertical cross-sections. Scale bars: 20 μm (horizontal) and 5 μm (vertical).

Figure 4.. Expansion microscopy (ExM) imaging of F-actin and Myosin IIA.

Maximum intensity projection (MIP) images of F-actin (A, C) and Myosin IIA (B, D) in DMSO (A, B) and SMIFH2 (C, D) treated cells. The color indicates the height along the ${\displaystyle z}$-axis, where the height was measured after the samples were swollen (color bar, right). Magnified views of the white boxes are shown in the right top panels, and corresponding outlines of F-actin are shown in the right bottom panels, where the bold and dotted lines indicate stress fibers and dorsal actomyosin fibers, respectively. The vertical cross-sections (${\displaystyle xz}$) are shown in the bottom panels, where the bold and dotted lines indicate the peripheral and dorsal inner regions, respectively. Scale bars: 20 µm (horizontal) and 10 µm (vertical). (A’, C’) Composite F-actin images of the ventral (red) and dorsal (green) sides. (A’) In the DMSO-treated cell, the thickness of the ventral and dorsal sides is 2.7 and 6.5 µm, respectively. (C’) In the SMIFH2-treated cells, the thickness of the ventral and dorsal sides is 4.2 and 5.7 µm, respectively.

Figure 4—figure supplement 1.. Expansion microscopy (ExM) imaging of actin filaments and Myosin II in cells treated with blebbistatin.

Maximum intensity projection (MIP) images of F-actin (A) and Myosin IIA (B) in blebbistatin-treated cells. The color indicates the height along the z-axis, where the height was measured after the sample was swollen (colorbar, the most right). Magnified views of the white boxes are shown in the right top panels, and corresponding outlines of F-actin are shown in the right bottom panels, where the bold and dotted lines indicate thick and thin fibers, respectively. The vertical section (xz) images are shown in the bottom panels, where the bold and dotted lines indicate peripheral and dorsal inner region, respectively. Scale bars: 20 µm (horizontal), 10 µm (vertical).

Figure 5.. Dynamics of F-actin in Caco-2 cells live-imaged by lattice light-sheet microscopy (LLSM).

(A) Maximum intensity projection (MIP) image of Caco-2 treated with DMSO. (B) Snapshot images in the green rectangle in (A), obtained from the ${\displaystyle z}$-slice at ${\displaystyle z=0}$ is defined as the plane closest to the substrate. The green line is the same as the green circle in (A). Arrowheads indicate the position of filaments. (C) Snapshot images in the red rectangle in (A) obtained from the ${\displaystyle z}$-slice at ${\displaystyle z=0.5\mu \text{m}}$. The red line is the same as the red circle in (A). Arrowheads indicate the position of filaments. (D) Kymograph along the green circle in (A), obtained from a slice at ${\displaystyle z=0}$. (E) Kymograph along the yellow line in (A), obtained from the ${\displaystyle z}$-slice at ${\displaystyle z=0}$. (F) Kymograph along the red circle in (A), obtained from the ${\displaystyle z}$-slice at ${\displaystyle z=0.5\mu \text{m}}$. inset: schematic diagram of F-actin (black lines) passing through the circle. (G) Kymograph along the yellow line in (A), obtained from the ${\displaystyle z}$-slice ${\displaystyle z=0.5\mu \text{m}}$. (H) MIP image of Caco-2 treated with SMIFH2 (${\displaystyle 40\mu \text{m}}$). (I) Kymograph along the red circle in (H), obtained from the MIP image. (J) Kymograph along the yellow line in (H), obtained from the MIP image. (K) Schematic diagram of F-actin structure in control and SMIFH2 treated cells. Stress fibers (dark red) were immobile, while the dorsal actin fibers formed an ‘actomyosin ring’ (light blue), moved in centripetal and clockwise directions. Scale bar: 10 µm.

Figure 6.. Angular velocity obtained by the particle image velocimetry (PIV) analysis.

Spatial profile of angular velocity (color code) obtained from the time average of the PIV vector field in a control cell (A: DMSO) or in a cell treated with SMIFH2 (B) superimposed on a snapshot F-actin image. Scale bar: 20 µm. (C) Average angular velocity as a function of the distance from the center. (D) Average angular velocity as a function of a distance scaled by the inner radius of the actomyosin ring of individual cells. Here, positive angular velocity indicates clockwise rotation. Sample averages for two conditions are indicated by the solid lines. Error bars and shaded areas represent standard errors of the means (SEM).

Figure 6—figure supplement 1.. Particle image velocimetry (PIV) analysis of cell treated with DMSO and SMIFH2.

PIV analysis of cell treated with DMSO and SMIFH2. Spatial distribution of azimuthal velocity (top), radial velocity (middle), and angular velocity (bottom) for cell treated with DMSO (control) (A–D) and with SMIFH2 (E–H). (I) Angular averaged azimuthal velocity plotted against the distance from the cell center. (J) Angular averaged radial velocity plotted against the distance from the cell center. Scale bar: 20 µm. Positive azimuthal and angular velocities indicate clockwise rotation.

Figure 7.. Theoretical model for the chiral cytoplasmic flow.

(A) Actomyosin generates a force dipole and a torque dipole. (B–D) Numerical simulation of Equation 1 assuming that the cell shape is axisymmetric around the cell center. (B) Actomyosin is distributed along the dorsal membrane with a concentric orientation. (C) Azimuthal velocity ${\displaystyle v_{\phi }}$ showing negative values indicating that the flow is generated in a clockwise direction. (D) Velocity in the radial ${\displaystyle \rho }$- and ${\displaystyle z}$- directions, (${\displaystyle v_{\rho }}$, ${\displaystyle v_z}$), indicated by vectors. Circulating flow is generated in the ${\displaystyle \rho }$-${\displaystyle z}$ plane. (E) Top: A concentric orientational field on a ring generates active torque in the center direction (magenta arrow). Middle: Active torque (magenta clockwise arrow) generated by a concentric orientational field on a ring. Bottom: A side view of a cell from the outside toward the cell center. The concentration of actomyosin increases in ${\displaystyle z}$ (gray color), leading to a gradient of active torque (magenta clockwise arrow) in the ${\displaystyle z}$ direction, resulting in a rotational flow clockwise (black arrows). (F) Flow profile along the dorsal side showing an inward sinistral swirling pattern. (G) Flow profile at the ventral side showing an outward dextral swirling pattern. (H) Angular velocity averaged in the ${\displaystyle z}$ direction plotted along the radial direction, showing a peak at around ${\displaystyle \rho /{\rho }_a=1}$. Here, ${\displaystyle {\rho }_a}$ is the leftmost position where ${\displaystyle S\ge 0.8}$.

Figure 7—figure supplement 1.. Cell shape used in the numerical simulation.

Cell shape used in the numerical simulation. Here, cell shape is assumed to be axisymmetric with respect to the cell center. ${\displaystyle Z_0}$ is the cell height, ${\displaystyle R_0}$ is the cell radius, and ${\displaystyle r_1}$, ${\displaystyle {\rho }_2}$ and α are the parameters that determine the cell shape.

Figure 7—figure supplement 2.. Comparison of radial and azimuthal velocities to determine the model parameters.

(A–H) Radial and azimuthal velocities shown in Figure 6—figure supplement 1. (I) Radial and azimuthal velocities in the numerical simulation. In the seplots, the radial velocity is positive in the centripetal direction.

Figure 8.. Depletion of dorsal actin and myosin by Rho Activator II stops the rotating motion of the nucleus.

(A) Actin (magenta) and Myosin II (yellow) showing localization with the dorsal marker Ezrin (cyan) in the DMSO-treated cell. (B) SMIFH2-treated cell showing an increase in dorsal actomyosin and a decrease in ventral actomyosin. (C) Rho Activator II (CN03) treated cell showing a decrease in dorsal actomyosin and an increase in ventral actomyosin. Control (DMSO) and SMIFH2-treated cells showed clockwise (CW) rotation, while CN03-treated cells did show rotation. Scale bars: 20 µm (horizontal) and 5 µm (vertical). Ratio of dorsal F-actin (D) and MyoII (E) to the ventral ones. p values were calculated using the Mann–Whitney U test. ${\displaystyle {\displaystyle \ast p<0.05}}$, n.s.: ${\displaystyle p\geqq 0.05}$.

Author response image 1.

Author response image 2.