The role of membrane stiffness and actin turnover on the force exerted by DRG lamellipodia

Abstract

We used optical tweezers to analyze the effect of jasplakinolide and cyclodextrin on the force exerted by lamellipodia from developing growth cones (GCs) of isolated dorsal root ganglia (DRG) neurons. We found that 25 nM of jasplakinolide, which is known to inhibit actin filament turnover, reduced both the maximal exerted force and maximal velocity during lamellipodia leading-edge protrusion. By using atomic force microscopy, we verified that cyclodextrin, which is known to remove cholesterol from membranes, decreased the membrane stiffness of DRG neurons. Lamellipodia treated with 2.5 mM of cyclodextrin exerted a larger force, and their leading edge could advance with a higher velocity. Neither jasplakinolide nor cyclodextrin affected force or velocity during lamellipodia retraction. The amplitude and frequency of elementary jumps underlying force generation were reduced by jasplakinolide but not by cyclodextrin. The action of both drugs at the used concentration was fully reversible. These results support the notion that membrane stiffness provides a selective pressure that shapes force generation, and confirm the pivotal role of actin turnover during protrusion.

Figure 1

Push and retraction by a lamellipodium. (a) Low-resolution image of a bead trapped in front of a lamellipodium emerging from the soma of a DRG neuron in control conditions. (b and c) High-resolution images during a push. At t₁ the bead is in the optical trap (b), and when the lamellipodium grows, at t₂, it pushes the bead (c). The cross indicates the center of the optical trap. (d) The three components F_x, F_y, and F_z of the force exerted when the lamellipodium pushes the bead. (e–h) As in a–d but in the presence of cyclodextrin. (i–l) As in a–c but during retraction and in the presence of jasplakinolide. (j) At t₁ the bead is in the optical trap. (k) At t₂, when the lamellipodium retracts, the bead is pulled away from trap. (l) The lamellipodium retracts and displaces the bead both laterally and vertically. In panels a–h the trap stiffness is k_x,y = 0.015, k_z = 0.005, and in panels i–l it is k_x,y = 0.07, k_z = 0.03.

Figure 2

Maximal force and velocity. (a) Histograms of the maximal protruding velocity during vertical pushes v_max, in control conditions and in the presence of jasplakinolide or cyclodextrin. (b) As in panel a but for maximal force exerted in vertical pushes F_max. (c) The relation between v_max and F_max during vertical pushes. (d–f) Superimposed Fv relationships from five individual vertical pushes in the presence of jasplakinolide (d), in control conditions (e), and in the presence of cyclodextrin (f).

Figure 3

Reversibility of the effect of jasplakinolide and cyclodextrin after washout. (a–c) Lamellipodia emerging from DRG GCs moving in cyclic waves of protrusions (b) and retractions (c); the dotted line represents the leading edge of lamellipodia in panel a. (d) Maximal protrusion/maximal retraction of lamellipodium versus time. The single dotted line represents the time of cyclodextrin addition, and the double dotted lines indicate the time of washout. (e) Histograms of the wave period in control conditions and in the presence of cyclodextrin and after washout. (f) Histograms of maximal protruding velocity in control conditions and in the presence of cyclodextrin and after washout. (g–i) As in d–f but in the presence of jasplakinolide.

Figure 4

Fv relationships during pushes and retractions. (a–d) Average Fv relationship, <Fv>_0.2, normalized to F_max for (a) vertical pushes, (b) lateral pushes, (c) vertical retractions, and (d) lateral retractions. The numbers of individual Fv relationships that were averaged in control conditions and in the presence of jasplakinolide and cyclodextrin were equal to (a) 23, 14, and 15, respectively; (b) 20, 14, and 14, respectively; (c) 23, 18, and 15, respectively; and (d) 14, 16, and 14, respectively.

Figure 5

Increase of noise during pushes in control conditions and in the presence of cyclodextrin but not in the presence of jasplakinolide. (a) The longitudinal components of the bead displacement during a lateral push in the presence of cyclodextrin show a clear increase in noise. (b) Relation between force and variance for lateral pushes in control conditions in the presence of cyclodextrin and in the presence of jasplakinolide. Data for control and jasplakinolide were taken from our previous work (26).

Figure 6

Elementary events underlying force generation in control conditions and in the presence of cyclodextrin are less pronounced in the presence of jasplakinolide. (a–c) Magnification of the z component during push in control conditions (a), in the presence of cyclodextrin (b), and in the presence of jasplakinolide (c). Original traces were filtered by the nonlinear diffusion algorithm (26), resulting in a smooth component and jumps. Jumps were detected infrequently during a push in the presence of jasplakinolide, but very often during a push in control conditions and in the presence of cyclodextrin. (d–f) Density of forward j⁺ and backward j⁻ jumps during pushes in control conditions (d), in the presence of cyclodextrin (e), and in the presence of jasplakinolide (f). (d) The fitting was performed with the values of 148 and 146 events/s for the jump frequency of positive and negative jumps, A₊ and A₋, respectively, and 5 and 4.8 nm for the mean size of positive and negative jumps, j^+∗ and j^−∗, respectively. (e) In the presence of cyclodextrin, the values of A₊, A₋, j^+∗ and j^−∗ were 226 and 224 events/s and 4.6 and 4.3 nm, respectively. (f) In the presence of jasplakinolide, the values of A₊, A₋, j^+∗ and j^−∗ were 48 and 44 events/s and 2.4 and 2.3 nm, respectively.

Figure 7

Effect of jasplakinolide and cyclodextrin on membrane stiffness. (a) Low-resolution image of an AFM cantilever in front of a GC emerging from the soma of a DRG neuron. (b and c) Force-displacement curves, displaying the AFM cantilever deflection as a function of its vertical z position when the cantilever was moved toward the soma (b) or GCs (c). (d–g) Bars indicate the value of Young's modulus, E, obtained from the best fit of the force displacement curves with the Hertz model (method 1 using Eq. 1) before and after the same neuron was treated with cyclodextrin, when the cantilever was moved toward the soma (d) or the GCs (e). Data are the mean ± SE. (f and g) As in d and e, but bars indicate the value of E in control conditions and in the presence of jasplakinolide. (h) Force-displacement curves fitted with the standard Hertz model by considering only indentations of <50 nm (method 2 using Eq. 1) and fitted with the corrected Hertz model (method 3 using Eq. 2). (i and j) Bars indicate the value of E for the same GCs shown in e obtained by using method 2 (i) or method 3 (j). Membrane stiffness obtained by the three methods decreased after addition of 2.5 mM cyclodextrin.