Mechanochemical regulation of oscillatory follicle cell dynamics in the developing Drosophila egg chamber

Abstract

During tissue elongation from stage 9 to stage 10 in Drosophila oogenesis, the egg chamber increases in length by ∼1.7-fold while increasing in volume by eightfold. During these stages, spontaneous oscillations in the contraction of cell basal surfaces develop in a subset of follicle cells. This patterned activity is required for elongation of the egg chamber; however, the mechanisms generating the spatiotemporal pattern have been unclear. Here we use a combination of quantitative modeling and experimental perturbation to show that mechanochemical interactions are sufficient to generate oscillations of myosin contractile activity in the observed spatiotemporal pattern. We propose that follicle cells in the epithelial layer contract against pressure in the expanding egg chamber. As tension in the epithelial layer increases, Rho kinase signaling activates myosin assembly and contraction. The activation process is cooperative, leading to a limit cycle in the myosin dynamics. Our model produces asynchronous oscillations in follicle cell area and myosin content, consistent with experimental observations. In addition, we test the prediction that removal of the basal lamina will increase the average oscillation period. The model demonstrates that in principle, mechanochemical interactions are sufficient to drive patterning and morphogenesis, independent of patterned gene expression.

FIGURE 1:

(A) (a–d) Egg chambers labeled with 4′,6-diamidino-2-phenylindole and myosin-mCherry (surface view) at stage 8 (a), early stage 9 (b), late stage 9 (c), and stage 10 (d). Maximum-intensity projection of the z-stacks shows the early-stage apical concentrated myosin and basal accumulation of myosin after stage 9. Scale bar, 50 μm. (B) Mechanical model. Cartoon of surface view of a Drosophila egg chamber showing the D-V and A-P axes. Cells are modeled as springs of stiffness k_c in the D-V direction and are connected in the A-P direction through angular springs of stiffness k_as and preferred angle β as shown in C. (D) Zoomed-in midsection of the egg chamber. (E) Connection to the basal lamina. Each cell is identified by the angular positions of its ends, θ. (F) Biochemical model. Molecular pathway governing the activation of myosin contraction in response to tension. F_i (blue arrow) represents contractile force from the ith cell, and F_i−1 and F_i+1 (red arrows) represent forces on the ith cell by neighboring cells.

FIGURE 2:

Behavior of single follicle cells. (A) As we apply an increasing external stretching force to a single follicle cell, we see that (B) the follicle cell length increases with increasing force. However, as the force reaches a threshold, the cell starts to oscillate. At large forces the oscillations disappear and the cell continues to stretch. (C) The amount of activated Rho increases with increasing force, and there is an oscillation in the amount activated Rho. Rho reaches a maximum value at large force. (D–F) When the external force is held constant, three behaviors are seen. At low forces (D), the system settles to a steady level of activated Rho and MLC. At intermediate forces (E), the system exhibits an oscillatory limit cycle. At high forces (F), a steady state is again reached. Therefore our model predicts a Hopf bifurcation with increasing external force.

FIGURE 3:

Follicle cell length and myosin oscillations. (A) Plot showing oscillations in cell length (blue) and myosin content (red). Increase in myosin content corresponds to decrease in cell length. (B) Oscillation period distribution for different initial conditions (ICs), showing that the range is between 5 and 7 min and is independent of ICs. (C) Phase distribution of oscillations in 120 cells, showing that the oscillations are asynchronous. The phases are uniformly distributed around 2π. (D, E) Phase diagrams of oscillatory behavior with and without basal lamina. The system generally exhibits asynchronous oscillations or steady nonoscillatory behavior. There is a small synchronous oscillation regime without basal lamina (white), although this would require a high internal pressure. The red circle indicates, in our model, the region close to the physiological situation.

FIGURE 4:

Internal egg chamber pressure and maximum contractile force have opposing effects on egg chamber radius. (A) Plot showing change in egg chamber radius as a function of internal pressure (P). Increase in pressure increases the egg chamber radius (here F_max = 40nN per unit-cell width). (B) Plot showing change in egg chamber radius as a function of maximum contractile force F_max. Increase in F_max decreases the egg chamber radius (here P = 0.4 kPa).

FIGURE 5:

Effects of the basal lamina and mosaic analysis. (A) Images of basal lamina (labeled with collagen–green fluorescent protein [GFP]) and myosin (myosin-mCherry) in control conditions. The relative positions of collagen and myosin fibers remains unchanged, suggesting that basal lamina could be mechanically coupled to basal myosin. Scale bar, 20 μm. (B) Egg chambers stained with collagen-GFP from stage 8 to stage 10 in control conditions, at the beginning of collagenase treatment (t = 0 min), and after collagenase treatment (t = 30 min). (C) Experimental measurements on follicle oscillations upon disruption of basal lamina. The distribution of oscillation periods became longer. The average egg chamber width became smaller (inset). (D) Modeling predictions of oscillation period as a function of stiffness of the basal lamina. Collagenase treatment reduces basal lamina stiffness and increases oscillation period for several values of P and F_max. The predicted egg chamber radius also becomes smaller as basal lamina stiffness is reduced, in agreement with experiments. (E) It is possible to abolish myosin contraction in some follicle cells using constitutively relaxing cells (ROCK RNAi–expressing cells); these cells (green) do not oscillate. It is then possible to examine the interaction between the wild-type cells (blue and red) and mutant cells (green). (F) Experiments and modeling show that there are no changes to oscillatory period in neighboring wild-type cells (blue) or wild-type cells directly neighboring mutant cells (green). Mutant cells, however, cease to oscillate. The oscillatory period is unchanged in neighboring versus nonneighboring wild-type cells (inset).