Three-Dimensional Characterization of Active Membrane Waves on Living Cells

We measure the temporal evolution of three-dimensional membrane topography on living fibroblasts and characterize the propagation of membrane waves using a wide-field optical profiling technique. The measured membrane profiles are compared with the numerical results calculated by the active membrane model recently proposed by Shlomovitz and Gov. After the treatments of blebbistatin and latrunculin A separately, the membrane waves disappear and the membrane surfaces are flattened, verifying that the membrane waves are driven by the interactions between myosin II and actin polymerization.

Figure 1

(a) Axial response curve of wide-field optical sectioning microscopy. A dashed rectangle indicates the working region of NIWOP. (b) Typical membrane topography obtained by NIWOP. The inset shows the bright-field reflection image of this cell and the dashed square labels the observation area. (c) Membrane-height profile along the dash-dot line in (b). (d) Kymograph of the membrane topography along the dashed line in (b). On the distance axis, the origin represents the cell edge.

Figure 2

(a) Peak-to-valley amplitudes of membrane waves versus the distances to cell edges. Solid curve is a guide to the eye. (b) Wavelengths versus the distances to cell edges. Solid curve is a fitting of $\lambda (x)={\lambda }_a[1-\exp (-x/x_c)]$. (c) Wave speeds versus the distances to cell edges. Solid curve is a guide to the eye. Each data point represents the mean in a $2\mathrm{\text{−}}\mu \mathrm{m}$ distance interval, and the horizontal values are the centers of intervals. The error bars are 95% confidence intervals determined by Student’s $t$ distribution.

Figure 3

(a) Dispersion relation of membrane waves. Each data point represents the mean in a wave-number interval of $0.2\mu {\mathrm{m}}^{-1}$. (b) Frequency response of the peak-to-valley amplitudes, normalized to the maximum values of each cell. Each data point represents the mean in an angular-frequency interval of $0.008\mathrm{rad}/\sec$. The error bars are 95% confidence intervals determined by Student’s $t$ distribution. The fitting curves are based on the active membrane wave model. For details, please see the text.

Figure 4

(a) Membrane topography of a cell (i) before and (ii) 23 min after the treatment of $75\mu \mathrm{M}$ BBI. (iii) Kymograph of the membrane topography. We treated the cell at the 9th min, washed out the reagent at the 39th min, and restarted the measurement at the 41st min. (b) Membrane topography of a cell (i) before and (ii) 25 min after the treatment of 100 nM LA. (iii) Kymograph of the membrane topography.