Embryo-scale tissue mechanics during Drosophila gastrulation movements

Abstract

Morphogenesis of an organism requires the development of its parts to be coordinated in time and space. While past studies concentrated on defined cell populations, a synthetic view of the coordination of these events in a whole organism is needed for a full understanding. Drosophila gastrulation begins with the embryo forming a ventral furrow, which is eventually internalized. It is not understood how the rest of the embryo participates in this process. Here we use multiview selective plane illumination microscopy coupled with infrared laser manipulation and mutant analysis to dissect embryo-scale cell interactions during early gastrulation. Lateral cells have a denser medial-apical actomyosin network and shift ventrally as a compact cohort, whereas dorsal cells become stretched. We show that the behaviour of these cells affects furrow internalization. A computational model predicts different mechanical properties associated with tissue behaviour: lateral cells are stiff, whereas dorsal cells are soft. Experimental analysis confirms these properties in vivo.

Figure 1. Embryo-scale views of the early stages of gastrulation.

(a,b) Extractions of images from a 3D time-lapse reconstruction of a Drosophila embryo during early development. Cell outlines were imaged using Gap43::mCherry. (a) Embryo cross-section views from the 3D time-lapse reconstruction at five time points during mesoderm internalization. Dorsal, ventral, lateral left and lateral right cell groups are indicated as D, V, LL and RL, respectively. t=0 min corresponds to 20% of ventral cell apical constriction (Supplementary Fig. 3). Yellow arrowhead: initial indentation of the ventral furrow. (b) Cylindrical projection of the central region of an embryo. Dashed rectangles indicate the four areas used for quantitative analysis. The projection covers the azimuth range of 450° so that the dorsal side is shown twice (marked by an asterisk). The red line marks the position at which the kymograph was recorded. (c) Average and s.d. of ventral cell surface areas over time. (d) Rate of cell surface change over time. (e) Furrow depth over time. (f) Speed of furrow ingression over time. The insets in c–f show the results for five embryos. Scale bars, 100 μm.

Figure 2. Differential distribution of the apical–medial actomyosin meshwork in the different cell populations.

(a) F-actin, labelled with MoesinABD::GFP. (b) Myo-II, labelled with sqh::GFP in the ventral, lateral and dorsal regions at a stage just before gastrulation (stage 5b) and during gastrulation (stage 6). Scale bars, 50 μm.

Figure 3. Differential behaviour of cell populations.

(a) Kymograph taken at the centre of the embryo (Fig. 1b, red line). Arrowheads mark the edges of three cell populations. Blue: region with apical constrictions; orange: ventrally displacing lateral cells; and green: dorsal spreading cells. Black arrowhead: furrow midline where the two lateral tissues meet. Yellow rectangles: edge of the future mesoderm, determined as described in Supplementary Fig. 5b. The ‘*' stresses the fact that the same dorsal tissue is represented twice. (b) Average and s.d. of left lateral (LL) and right lateral (RL) cell displacement along the DV axis over time. Positive values indicate displacement towards the ventral side. (c) Average and s.d. of major axis orientation for the LL, RL and dorsal (D) cells. The diagram below the three panels illustrates the measurements that were made: the aspect ratio of each cell (left; colour in the panels indicates cell eccentricity, with red for the highest aspect ratios) and their orientation relative to the AP axis of the embryo (right). (d) Rate of change in apical cell surface area of ventral cells (black), depth of the furrow (blue) and displacement of the two lateral cell populations (red). The solid red lines show the average displacement along the DV axis of the RL and the LL cells in time. The dashed red line shows the average displacement of RL and LL. All quantifications of cell behaviour are based on measurements the of segmented 2D surface view of the embryo.

Figure 4. Relationship between ectoderm movement and mesoderm internalization.

(a) Kymograph of a gastrulating sna twi embryo. Arrowheads: positions corresponding to dorsal edges of lateral region in wild-type (WT) embryos. The ‘*' stresses the fact that the same dorsal tissue is represented twice. (b) Lateral cell displacement in WT and sna twi embryos; >1,200 cells (four embryos) measured per genotype. Bars represent the standard deviation (s.d.). (c) Cylindrical projection of embryo with single cauterization (red arrowheads). The ‘*' stresses the fact that the same dorsal tissue is represented twice. (d) Kymograph taken along the white dashed line in c. Red and green lines mark two individual cell trajectories. Narrow section of the red line: time in which lateral cells in WT embryos are stationary; wide line: the period when WT lateral cells move. Black arrowheads: ventral midline position. (e) Furrow depth acell displacement for the embryo in d. (f) Displacement of lateral cells between 0<t<15 min plotted against the final angular position of the midline. (g) Final displacements of right (empty circles) and left (filled circles) lateral cell sheets and their sums (crosses). (h) Bilateral cauterization (red arrowheads). The ‘*' stresses the fact that the same dorsal tissue is represented twice. (i) Kymograph taken along the white dashed line in h. Green and red symbols mark the same points and times as in j. Dashed lines as in d and j, Cross-section views of the embryo along the white lines at times t₁, t₂ and t₃ marked in i. Red arrowheads as in h. Yellow arrowhead: invagination formed at t₂. Black dashed line: outline of the invaginated cell mass. Scale bars, 100 μm.

Figure 5. Lateral and dorsal cauterizations.

(a) Cylindrical projection of an embryo with bilateral cauterizations (red arrowheads). The ‘*' stresses the fact that the same dorsal tissue is represented twice. (b) Kymograph taken along the white dashed line in a. The red dashed line follows a lateral cell trajectory. (c) Cauterizations on the dorsal side of the embryo (embryo b in Supplementary Fig. 11). The ‘*' stresses the fact that the same dorsal tissue is represented twice. (d) Kymograph taken along the white dashed line in c. The green dashed lines follow two lateral cell trajectories. Black arrowhead: ventral midline. Pink arrowheads: detachment of the epithelium from the site of fixation. (e) A single fixation at the dorsal midline (embryo a in Supplementary Fig. 11). The ‘*' stresses the fact that the same dorsal tissue is represented twice. (f) Kymograph taken along the white dashed line in e. Black arrowhead: ventral midline. (g) Diagram summarizing the immobilization experiments. Green and red symbols indicate fixations that either allowed or impaired lateral cell displacement and mesoderm invagination. Further cases are summarized in Supplementary Fig. 11. Scale bars, 100 μm.

Figure 6. Computational model.

(a) Phase diagram showing shapes of embryo cross-sections generated by the model with a fixed value of cortical tensions in ventral cells and varying cortical tensions in dorsal and lateral cells. Dashed line connects states with identical cortical tension in lateral and in dorsal cells, and the red circle marks the point where the relative values of lateral and dorsal tension correspond to those measured in vivo (c). We draw attention to the fact that contrary to the situation in real embryos, the dorsal cells in the model shorten along their apical–basal axis. As the model requires conservation of volume of the cells, and it does not take the third dimension of the cell into account, this is a necessary consequence of cell widening. As we have shown in this work, the cells in fact compensate for widening by shortening in this third, AP dimension. In our model, we accept in this 2D representation the loss of height as a proxy for reduction of the AP length. (b) Isolines for values for dorsal dilation, ectoderm displacement and furrow depth achieved with the various combinations of dorsal and lateral tensions shown in a. (c,d) Experimental measurements of the initial speed of recoil of the actin meshwork in ventral, lateral and dorsal regions after laser dissection at a stage before furrow formation (c) and at the onset of furrow formation (d). (e) Shapes and furrow depth generated by the model under different cauterization conditions. Inset shows an enlarged view of the asymmetric case with a displaced furrow.