Actin-myosin network reorganization breaks symmetry at the cell rear to spontaneously initiate polarized cell motility

Abstract

We have analyzed the spontaneous symmetry breaking and initiation of actin-based motility in keratocytes (fish epithelial cells). In stationary keratocytes, the actin network flow was inwards and radially symmetric. Immediately before motility initiation, the actin network flow increased at the prospective cell rear and reoriented in the perinuclear region, aligning with the prospective axis of movement. Changes in actin network flow at the cell front were detectable only after cell polarization. Inhibition of myosin II or Rho kinase disrupted actin network organization and flow in the perinuclear region and decreased the motility initiation frequency, whereas increasing myosin II activity with calyculin A increased the motility initiation frequency. Local stimulation of myosin activity in stationary cells by the local application of calyculin A induced directed motility initiation away from the site of stimulation. Together, these results indicate that large-scale actin-myosin network reorganization and contractility at the cell rear initiate spontaneous symmetry breaking and polarized motility of keratocytes.

Figure 1.

F-actin network movement in stationary and motile keratocytes. (A and B) Keratocytes were fixed and stained with phalloidin to visualize F-actin. (A) F-actin in stationary keratocytes was denser near the cell body than the periphery. (B) Motile keratocytes had a criss-cross pattern of F-actin staining in the lamellipodia and F-actin bundles at the cell rear. (C and E) AF546-phalloidin FSM images of F-actin networks in live stationary (C) and motile (E) keratocytes. (D and F) F-actin flow field relative to the substratum extracted by multiframe correlation tracking of speckle movement. Vectors indicate network displacements between consecutive frames. Color encodes flow speed. (D) The F-actin network in stationary keratocytes flowed centripetally inwards. (F) The F-actin network in motile keratocytes had a small retrograde flow in the lamellipodium and large inwards movement at the rear sides. (G) During symmetry breaking and motility initiation of keratocytes, the cell must transition from being stationary with radially symmetric centripetal actin flow to being polarized with decreased actin flow speed at the cell front. C and E are available as Videos 1 and 2 (available at http://www.jcb.org/cgi/content/full/jcb.200706012/DC1).

Figure 2.

The prospective rear edge moves before the front edge during motility initiation. (A) Phase-contrast image sequence of two representative keratocytes breaking symmetry and initiating motility. (B, left) Cell outline superimposed on the first and last image in the image sequence for the same cells as in A. (B, right) Time sequence of the cell outline during motility initiation, with color denoting time elapsed (in seconds). (C) Protrusion/retraction map of the cell edge (in polar coordinates) over time. Blue boxes are regions of continuous rear retraction. Red boxes are regions of continuous front protrusion. The rear of the cell exhibited three phases during motility initiation: I, slow rear retraction; II, fast rear retraction; and III, persistent movement. A is available as Videos 3 and 4 (available at http://www.jcb.org/cgi/content/full/jcb.200706012/DC1).

Figure 3.

Initial events in symmetry breaking include an increase in F-actin network flow speed at the prospective cell rear and a reorientation of perinuclear F-actin network flow. (A) Representation of an F-actin flow field displacement vector by its radial (v_r) and tangential (v_θ) component. The centripetal deviation (φ) was positive for counter-clockwise rotation and negative for clockwise rotation of the original flow vector relative to v_r. (B and C) F-actin flow field of a stationary cell (B) and a cell initiating motility (C) with the prospective direction of cell movement to the right. (left) F-actin network flow field; flow vectors are colored according to speed. (right) Scalar maps of radial velocity and centripetal deviation of F-actin flow vectors. White line is the delineation of the perinuclear and peripheral regions. (C, left) Phase I: the radial velocity increased slightly at the prospective cell rear. The magnitude of the centripetal deviation increased in the perinuclear region, corresponding to a reorientation of the F-actin flow along the prospective direction of motion. (right) Phase II: large increase in radial velocity at the prospective cell rear. The increased magnitude of centripetal deviation remained. (D and E, top) Time courses of radial velocity in the front (blue), rear (red), left (brown), and right (green) perinuclear (bold) and peripheral (standard) regions. Data in D originate from the cell in C and showed large transients before sustained polarization. (bottom) Time courses of centripetal deviation in the left (brown) and right (green) perinuclear (bold) and peripheral (standard) regions. Phases I and II are indicated by horizontal bars. B and C are available as Videos 5 and 6 (available at http://www.jcb.org/cgi/content/full/jcb.200706012/DC1).

Figure 4.

Decrease in F-actin network flow at the cell front occurs during maturation of the polarized form. (A) F-actin network flow vectors relative to the substratum in the front central lamellipodium (boxed area) overlaid on an AF546-phalloidin FSM image. (B) Development of cell speed versus F-actin flow speed at the front relative to the substratum and parallel to the direction of movement over phase III (maturation) for five cells. During maturation, cells increased in speed, whereas the F-actin retrograde flow speed decreased in the lamellipodia.

Figure 5.

Motility initiation requires Rho kinase–dependent myosin activity. (A, top) Keratocyte fixed and stained for tubulin to visualize microtubules. (bottom) Live stationary keratocyte expressing YFP-myosin regulatory light chain. (B) Motility initiation frequency of stationary keratocytes within 30 min of a temperature shift from 20 to 30°C. Successful motility initiation was defined as persistent polarized morphology and cell movement over at least four cell lengths. Shape changers were cells that had irregular morphologies. Depolymerization of microtubules with 1 μg/ml nocodazole, inhibition of PI-3 kinase with 50 μM LY294002, and inhibition of MLCK with 10 or 25 μM ML-7 had no effect on the frequency of motility initiation (P = 0.7498, P = 0.0173, and P = 0.5562, respectively). Myosin II inhibition with 40 or 100 μM blebbistatin and Rho kinase inhibition with 10 or 25 μM Y-27632 decreased the frequency of motility initiation (P < 0.0001). In contrast, myosin phosphatase inhibition with 10 or 25 nM calyculin A increased the frequency of motility initiation (P < 0.0001). When two drug concentrations were used, there was no significant difference between the results (P > 0.2); results from the two concentrations were pooled. (C) Stationary keratocytes were treated with the indicated drugs for 10–20 min. F-actin was visualized by fixing the cells and staining with phalloidin. Treatment with ML-7 and calyculin A retained the circular bands of F-actin in the perinuclear region. Treatment with blebbistatin or Y-27632 reduced F-actin in the perinuclear region, and circular bands were no longer visible (yellow brackets). (D and E) Changes in perinuclear and peripheral F-actin radial velocity and directional coherence (see Materials and methods) before and after treatment with the indicated drugs. Error bars indicate the SD of the mean over time; gray lines indicate correspondence between data points representing the same cell before and after treatment. In some cases, the SD is smaller than the size of the data point. (D) Blebbistatin treatment decreased the perinuclear and peripheral radial velocity; conversely, calyculin A treatment increased the perinuclear and peripheral radial velocity. Y-27632 or ML-7 treatment had no effect. (E) Blebbistatin and Y-27632 treatment decreased the directional coherence of the actin flow in the perinuclear and peripheral zones. The effect was greater in the perinuclear zone than the peripheral zone. ML-7 and calyculin A had no effect.

Figure 6.

Local application of calyculin A can stimulate motility initiation. (A and B) Changes in F-actin network radial velocity (A) and centripetal deviation (B) in phases I and II of spontaneous symmetry breaking in the presence of 20 nM calyculin A. (top) Scalar maps of the radial velocity and centripetal deviation. White line is the delineation of perinuclear and peripheral regions. (bottom) Time course of radial velocity and centripetal deviation. Phases I and II are indicated by horizontal bars. (A) The radial velocity at the cell rear increased during phases I and II. (B) The magnitude of the centripetal deviation increased during phases I and II, reflecting a reorientation of the F-actin flow along the prospective direction of motion. (C) Local application of calyculin A to one side of a stationary keratocyte. Red pseudocolor indicates the drug flow from the micropipette (positioned in the bottom left corner). The keratocyte polarized and initiated motility away from the source of calyculin A. C is available as Video 7 (available at http://www.jcb.org/cgi/content/full/jcb.200706012/DC1).

Figure 7.

Symmetry breaking and motility initiation are initiated at the cell rear and perinuclear region and propagate throughout the entire cell. (A) Model for symmetry breaking and motility initiation in keratocytes. (left) Schematic of F-actin network flow in the peripheral (gray) and perinuclear (purple) regions. (right) Cell cross section schematic. In a stationary keratocyte, the F-actin network flows centripetally inwards, driven by net polymerization at the periphery and myosin contraction and net depolymerization in the perinuclear region. In phase I, an increase in perinuclear contractility causes an increase in the F-actin network flow speed at the prospective cell rear and partial polarization of the perinuclear F-actin network flow in the direction of eventual cell movement. The rear edge retracts slightly and decreases in thickness. A further increase in F-actin network flow speed at the prospective cell rear, which is caused by a further increase in perinuclear contractility, leads to phase II. The rear edge retracts, and adhesive coupling to the substrate decreases at the rear. The cell body moves forward slowly. In phase III, perinuclear actin bands transform into an actin axle. F-actin flow speed at the cell front decreases relative to the substratum, whereas the adhesive coupling to the substrate and net F-actin polymerization increases. This leads to protrusion of the front edge, and the entire cell moves rapidly and persistently. (B, left) Symmetry breaking is driven by coordinated actin–myosin contraction and requires perinuclear F-actin bands. (right) Model of actin–myosin organization in stationary keratocytes. At the periphery of the lamellipodium, myosin bipolar filaments cross-link a dendritic F-actin network without contraction. Toward the cell body, the activity of large myosin II clusters contracts and reorganizes the dendritic network to form F-actin bands and bundles. This is dependent on Rho kinase.