Chiral growth of adherent filopodia

Abstract

Adherent filopodia are elongated finger-like membrane protrusions, extending from the edges of diverse cell types and participating in cell adhesion, spreading, migration, and environmental sensing. The formation and elongation of filopodia are driven by the polymerization of parallel actin filaments, comprising the filopodia cytoskeletal core. Here, we report that adherent filopodia, formed during the spreading of cultured cells on galectin-8-coated substrates, tend to change the direction of their extension in a chiral fashion, acquiring a left-bent shape. Cryoelectron tomography examination indicated that turning of the filopodia tip to the left is accompanied by the displacement of the actin core bundle to the right of the filopodia midline. Reduction of the adhesion to galectin-8 by treatment with thiodigalactoside abolished this filopodia chirality. By modulating the expression of a variety of actin-associated filopodia proteins, we identified myosin-X and formin DAAM1 as major filopodia chirality promoting factors. Formin mDia1, actin filament elongation factor VASP, and actin filament cross-linker fascin were also shown to be involved. Thus, the simple actin cytoskeleton of filopodia, together with a small number of associated proteins are sufficient to drive a complex navigation process, manifested by the development of left-right asymmetry in these cellular protrusions.

Figure 1

Filopodia turn counterclockwise in cells spreading on galectin-8. (A) Confocal microscopy images of typical HeLaJW cells, nontransfected (left), transfected with GFP-myosin-X (middle), and treated with siRNA against Arp2 (right). Cells were fixed 20 min after spreading on galectin-8-coated substrate and stained with phalloidin to visualize actin (magenta). GFP-myosin-X is shown in green. (B–D) Filopodia number (B), length (C), and fractions of bent filopodia (D) in HeLaJW cells, nontransfected (control), transfected with GFP-myosin-X (Myo10), and treated with siRNA against Arp2 (siArp2), as assessed 20 min after plating on galectin-8-coated substrate. Each dot corresponds to an individual cell. Pooled data from more than three experiments in each case are presented as box and whisker plots. p values were calculated using nonparametric Mann-Whitney test. (E) Typical images of left-bent, right-bent, wavy, or straight filopodia. (F) Pie diagrams representing percentage of the filopodia with different morphology in cells treated as indicated. Sectors of the pies, corresponding to left-bent, right-bent, wavy filopodia, and straight filopodia are denoted by red, blue, green, and gray, respectively. Data on nontransfected (control), GFP-myosin-X transfected (Myo10), and Arp2 knockdown (siArp2) cells were calculated by assessment of all filopodia from pooled experiments. The number of experiments (N), number of cells assessed, and total number of filopodia are indicated under each pie. (G) Chirality index is defined as the number of left-bent minus right-bent filopodia divided by the total number of nonstraight filopodia. Each dot corresponds to an individual cell. Pooled data from no less than three experiments are presented. The bars show the mean values and error bars the standard deviation. p values were calculated using nonparametric Mann-Whitney test.

Figure 2

Dynamics and structure of turning filopodia. (A) Time course of the formation of bent filopodia by changing the direction of extension of straight filopodia in cells transfected with GFP-myosin-X (Myo10) or Arp2 knockdown cells (siArp2) during cell spreading on substrates coated with galectin-8. Cells were imaged using interference reflection microscopy (IRM), in combination with epifluorescence microscopy for visualization of GFP-myosin-X. Scale bars, 15 μm. Time after cell seeding (min) is indicated in each image. (B) Left: kymograph showing the time course of extension and turning of a typical filopodium of cells transfected with GFP-myosin-X. Notice the bent filopodium was formed from a straight filopodium, which changed the direction of extension. Right: the velocity of the filopodia tip relative to the substrate during the growth period calculated for the filopodia shown in the left. Notice that the velocity of filopodia tip is reduced before filopodia change the direction of extension. (C) Kymograph showing the dynamics of myosin-X during filopodia turning in cells expressing GFP-myosin-X (magenta) and Td-Tomato-F-tractin (green). Note that the filopodia turning is accompanied by splitting of the myosin-X patch at the tip of filopodia into two. One of the two remains associated with the former end of the filopodia and eventually disappears while the other is formed at the new filopodia tip. Five examples of this process are shown. Scale bar, 2 μm. The interval between the frames is 2 s.

Figure 3

Cryoelectron tomography images of filopodia tips. (A–F) A slice of cryo-ET images in the x-y plane of straight filopodia (A and B) and left-bent filopodia (C–F) in cells overexpressing GFP-myosin-X spreading on galectin-8-coated substrates for 20 min. Slice thickness, 8.9 nm (see Videos S4, S5, S6, S7, S8, and S9). (G) 3D rendering the isosurface of the images in (B, C, and F) (red, actin; gray, membrane). Note that in straight filopodia (A, B, and G, left) the filament cores are located symmetrically along the mid-line of filopodia and approached the filopodia tips. In the bent filopodia (C–F and G, middle and right), the bulbous membrane extensions are located asymmetrically relatively to the bulk of the actin cores, which are shifted to the right side of the bulbs. The left halves of the bulbs either did not contain actin filaments (C and G, middle), or contained splayed filaments (D and E), or the loop formed by the continuation of the core (F and G, right).

Figure 4

Reduction of cell adhesion to galectin-8 by thiodigalactoside (TDG) abolished filopodia chirality. (A and B) Confocal microscopy images of HeLaJW cells expressing GFP-myosin-X (Myo10) (A) and depleted of Arp2 (siArp2) (B) spreading on galectin-8-coated substrate for 20 min in the absence or in the presence of 10 mM galectin-8 ligand TDG. Cells were stained with phalloidin to visualize actin (magenta). GFP-myosin-X is shown in green. (C and D) Pie diagrams representing percentage of the filopodia with different morphology in cells treated as indicated. Color coding is the same as in Fig. 1. The number of experiments (N), cells, and filopodia assessed are indicated under each pie. Note that incubation with TDG significantly decreased the fraction of left-bent filopodia (C), and increased the fraction of right-bent filopodia (C and D). (E–G) Filopodia number (E), length (F), and fractions of bent filopodia (G) in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments for each condition are presented as box and whisker plots. p values were calculated using nonparametric Mann-Whitney test. (H) Filopodia chirality index in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments are presented. The bars show the mean values and error bars the standard deviation. p values were calculated using nonparametric Mann-Whitney test. Note that treatment with TDG reduced filopodia chirality.

Figure 5

Myosin-X is required for filopodia chirality in Arp2 knockdown cells. (A) Confocal microscopy images of Arp2-depleted cells (siArp2), Arp2, and myosin-X-depleted cells (siArp2+siMyo10), and Arp2-depleted cells expressing GFP-myosin-X (siArp2+GFP-Myo10) fixed 20 min after spreading on galectin-8-coated substrate and stained with phalloidin to visualize actin (magenta). GFP-myosin-X is shown in green. (B) Pie diagrams representing percentage of the filopodia with different morphology in cells treated as indicated. Color coding is the same as in Fig. 1. The numbers of experiments (N), cells and filopodia assessed are indicated under each pie. Note that myosin-X knockdown significantly reduced the fraction of left-bent filopodia, while its overexpression increased that fraction. (C) Filopodia number, length, and fractions of bent filopodia in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments for each condition are presented as box and whisker plots. p values were calculated using nonparametric Mann-Whitney test. (D) Filopodia chirality index in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments are presented. The bars show the mean values and error bars the standard deviation. p values were calculated using nonparametric Mann-Whitney test. Note that knockdown of myosin-X reduced filopodia chirality index.

Figure 6

The effects of knockdowns of mDia1, mDia2, FMNL2, and VASP on filopodia chirality. (A) Pie diagrams representing percentage of the filopodia with different morphology in cells treated as indicated. Color coding is the same as in Fig. 1. The numbers of experiments (N), cells, and filopodia assessed are indicated under each pie. In the left column, the data for nontransfected (control) cells and cells with single knockdown of mDia1, mDia2 FMNL2, or VASP are presented. The middle column shows the data for GFP-myosin-X-expressing cells (Myo10), without or with the knockdown of mDia1, mDia2 FMNL2, or VASP, respectively. In the right column, the data for Arp2-depleted cells without or with additional knockdown of mDia1, mDia2 FMNL2, or VASP are shown. (B) Filopodia chirality index in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments are presented. The bars show the mean values and error bars the standard deviation. p values were calculated using nonparametric Mann-Whitney test. Note that depletion of mDia1 and VASP reduced chirality in Arp2 knockdown cells, while depletion of mDia2 and FMNL2 slightly increased the filopodia chirality index.

Figure 7

The effects of knockdown of fascin and DAAM1 in filopodia formation and filopodia chirality. (A) Pie diagrams representing percentage of the filopodia with different morphology in cells treated as indicated. Color coding is the same as in Fig. 1. The numbers of experiments (N), cells, and filopodia assessed are indicated under each pie. In the left column, the data for nontransfected (control) cells and cells with single knockdown of DAAM1 or fascin are presented. The middle column shows the data for GFP-myosin-X-expressing cells (Myo10), without or with the knockdown of DAAM1 or fascin, respectively. In the right column, the data for Arp2-depleted cells without or with additional knockdown of DAAM1 or fascin are shown. Note that knockdown of DAAM1 increased the fraction of right-bent filopodia, while knockdown of fascin reduced the fraction of straight filopodia and increased the fraction of right-bent and wavy filopodia. (B) Filopodia chirality index in cells treated as indicated. Each dot corresponds to an individual cell. Pooled data from two experiments are presented. The bars show the mean values and error bars the standard deviation. p values were calculated using nonparametric Mann-Whitney test. Note that knockdown of DAAM1 reduced chirality index more strongly than knockdown of fascin.