An Optogenetic Method to Modulate Cell Contractility during Tissue Morphogenesis

Abstract

Morphogenesis of multicellular organisms is driven by localized cell shape changes. How, and to what extent, changes in behavior in single cells or groups of cells influence neighboring cells and large-scale tissue remodeling remains an open question. Indeed, our understanding of multicellular dynamics is limited by the lack of methods allowing the modulation of cell behavior with high spatiotemporal precision. Here, we developed an optogenetic approach to achieve local modulation of cell contractility and used it to control morphogenetic movements during Drosophila embryogenesis. We show that local inhibition of apical constriction is sufficient to cause a global arrest of mesoderm invagination. By varying the spatial pattern of inhibition during invagination, we further demonstrate that coordinated contractile behavior responds to local tissue geometrical constraints. Together, these results show the efficacy of this optogenetic approach to dissect the interplay between cell-cell interaction, force transmission, and tissue geometry during complex morphogenetic processes.

Graphical abstract

Figure 1

Light-Mediated CRY2-OCRL Plasma Membrane Recruitment Results in PI(4,5)P₂ and Actin Depletion from the Embryo Cortex (A and B) Cartoon depicting a surface view of the Drosophila blastoderm epithelium and schematic of the optogenetic module that we used. This is based on the interaction between CIBN and the PHR domain of CRY2 upon blue-light (488-nm) illumination. CIBN was tagged with enhanced GFP and with a CaaX anchor, which localizes it to the plasma membrane. CRY2 was fused to the catalytic domain of the Drosophila inositol polyphosphate 5-phosphatase OCRL and tagged with mCherry. In the absence of blue light, the 5-ptase is cytosolic (A). Upon blue-light illumination, the 5-ptase is recruited to the plasma membrane, where it dephosphorylates PI(4,5)P₂ to PI(4)P. This results in PI(4,5)P₂ depletion from the plasma membrane and cortical actin depolymerization (B). (C) Mean levels of mCherry::CRY2-OCRL recruitment to the plasma-membrane-anchored CIBN::pmGFP in response to 488-, 458-, 514-, and 561-nm light. The x coordinates represent time of exposure to light. Photo-activation was achieved using a continuous-wave laser of the indicated wavelength at a scanning speed of 1.27 × 10⁻⁶ s/pixel, for a total time of 1 s for the entire embryo at 30-s intervals. The y coordinates represent the log₂ ratio of CRY2-OCRL plasma membrane (pm) to cytosol (cyt) fluorescence intensities. Pooled data are represented as mean ± SD (n = 3 embryos for each condition). (D–G) Confocal images of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL. Before blue-light illumination, mCherry::CRY2-OCRL is cytosolic (D). After 1-s exposure to 488-nm light, mCherry::CRY2-OCRL is recruited to the plasma membrane (E), where CIBN::pmGFP localizes (F). (G) shows a merge of (E and F). Scale bars, 10 μm. (H–K) Confocal images of a representative embryo co-expressing a non-tagged version of CIBN, the PI(4,5)P₂ biosensor PH_PLCδ::GFP, and mCherry::CRY2-OCRL. PH_PLCδ::GFP is depleted from the plasma membrane upon illumination at 488 nm: compare signal (H) at the onset of photo-activation (T₀) and (J) 1 min after the beginning of photo-activation. Of note, the total signal intensities shown in (H) and (J) are approximately the same (3.663 versus 3.531, a.u.). (I and K) show the localization of mCherry::CRY2-OCRL at the onset (T₀) and at 1 min after the beginning of photo-activation, respectively. Scale bars, 10 μm. (L) Barplot showing mean PH_PLCδ::GFP levels in control and photo-activated (PA) embryos. Control embryos express mCherry::CRY2-OCRL and PH_PLCδ::GFP. PA embryos co-express a non-tagged version of CIBN (to avoid overlap with PH_PLCδ::GFP signal), PH_PLCδ::GFP, and mCherry::CRY2-OCRL. Both control and PA embryos were exposed to 488-nm illumination. Values along the y axis represent the log₂ ratio of the average PH_PLCδ::GFP plasma membrane fluorescence intensity at 1 min after the beginning of photo-activation (fluo_1min) to the average PH_PLCδ::GFP plasma membrane fluorescence intensity at the beginning of photo-activation (fluo_T0). Pooled data are represented as mean ± SD (n = 3 embryos for each condition; p = 1.0 × 10⁻³; two-sample Student’s t test). (M–P) Confocal images of a representative embryo co-expressing a non-tagged version of CRY2-OCRL, the actin-binding protein Moesin::mCherry, and CIBN::pmGFP. Moesin::mCherry localizes to the cell cortex at the beginning (T₀) of photo-activation (M) and is depleted 4 min after the beginning of photo-activation (O). (N and P) show the localization of CIBN::pmGFP at the beginning (T₀) and 4 min after the beginning of photo-activation, respectively. Scale bars, 10 μm. (Q) Barplot showing mean Moesin::mCherry levels in control and photo-activated (PA) embryos. Control embryos express CIBN::pmGFP and Moesin::mCherry. PA embryos co-express a non-tagged version of CRY2-OCRL (to avoid overlap with Moesin::mCherry signal), Moesin::mCherry, and CIBN::pmGFP. Both control and PA embryos were exposed to 488-nm light. The y axis displays the log₂ ratio of the average Moesin::mCherry plasma membrane fluorescence intensity at 4 min after the beginning of photo-activation (fluo_4min) to the average Moesin::mCherry plasma membrane fluorescence intensity at the beginning of photo-activation (fluo_T0). Pooled data are represented as mean ± SD (n = 3 embryos for each condition; p = 5.7 × 10⁻⁴; two-sample Student’s t test). See also Figure S1.

Figure 2

Activation of CRY2-OCRL Plasma Membrane Recruitment Causes Inhibition of Apical Constrictions and Arrest of Ventral Furrow Formation (A–C) Still frames from a confocal movie of the ventral mesoderm of a representative control embryo expressing only CIBN::pmGFP 10 min before ventral furrow formation (A), 5 min before ventral furrow (VF) formation (B), and at the onset of ventral furrow formation (C). Scale bars, 10 μm. (D–F) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL at the onset of photo-activation (T₀) (D) and at 15 min (E) and 25 min (F) after the beginning of photo-activation. Photo-activation was started before the beginning of ventral furrow formation using a continuous 488-nm wave laser at a scanning speed of 1.27 × 10⁻⁶ s/pixel, for a total time of 1 s for the entire embryo at 30-s intervals. Scale bars, 10 μm. (G–I) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL at the beginning of photo-activation (T₀) (G) and at 2.5 min (H) and 5 min (I) after the beginning of photo-activation. Photo-activation was started after the beginning of ventral furrow formation and resulted in the reversion of the invagination process with alternated patches of constricted and relaxed cells. Scale bars, 10 μm. (J and K) Quantification of cell area (purple) and a-p anisotropy (green) for (J) the control embryo shown in (A–C) and for (K) the photo-activated embryo expressing CRY2-OCRL shown in (D–F). The y coordinates represent cell area expressed in squared microns (left y axis) and a-p anisotropy (right y axis). Values along the x axis represent time in minutes. Solid and dashed lines indicate the median over all cells for cell area (solid) and a-p anisotropy (dashed). Shaded regions show the interquartile range. In (J), time 0 corresponds to the time point of ventral furrow invagination. In (K), time 0 corresponds to the beginning of photo-activation. (L and M) Comparison of a-p anisotropy (L) and cell area (M) between control embryos expressing only CIBN::pmGFP (at the time point of tissue invagination) and photo-activated embryos co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL at three different time points after the beginning of photo-activation (10 min, 20 min, and 30 min). Values along the y axis represent a-p anisotropy in (L) and cell area in (M). Statistical testing of differences in cell area was performed on the log-transformed values (left y axis). Absolute values for cell areas (in squared microns) are represented on the right y axis as a reference (M). For both cell area and a-p anisotropy, the control samples are significantly different from the photo-activated samples. Each dot represents a single embryo. The crosses show group median (horizontal line) and interquartile range (vertical line). (n = 5 embryos for each condition). ^∗p ≤ 0.05; ^∗∗p ≤ 0.01; ^∗∗∗p ≤ 0.001; n.s., not significant; pairwise two-sample Student’s t tests with pooled variance and multiple testing correction with Bonferroni’s method. See also Movies S1 and S2A and S2B.

Figure 3

Ventral Furrow Formation Is Not Inhibited in Embryos Co-expressing CRY2-OCRL and CIBN::pmGFP in the Absence of 488-nm Illumination (A) Comparison of median a-p anisotropy of constricting cells in the absence of blue-light illumination between embryos co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL in a WT background (Dark); embryos co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL in a sktl (Sktl^+/−) and in a zip (Zip^+/−) heterozygous mutant background; and embryos expressing CIBN::pmGFP only in a WT background (WT). Values along the y axis represent a-p anisotropy. Each dot represents a single embryo. The crosses show group median and interquartile ranges (n ≥ 3 embryos for each condition). n.s., not significant; pairwise two-sample Student’s t tests with pooled variance and multiple testing correction with Bonferroni’s method. (B–D) Confocal images of the ventral mesoderm of representative embryos co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL in a WT background (B) and in a sktl (C) and a zip (D) heterozygous mutant background. The images were taken with 488-nm light after ventral cells started to constrict. CIBN::pmGFP is shown. Scale bars, 10 μm. (E–G’’’) Still frames from a time-lapse confocal movie of the ventral mesoderm of a representative embryo heterozygous for a loss-of-function sktl allele (sktl^+/−) and co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL 21 min before (E), 10 min before (F), and at completion of (G) ventral furrow (VF) invagination. The embryo was imaged with 561-nm light only. (G’) shows CIBN::pmGFP, (G’’) shows mCherry::CRY2-OCRL, and (G’’’) shows a merge of the two channels at completion of ventral furrow invagination. Scale bars, 10 μm. (H–J’’’) Still frames from a time-lapse confocal movie of the ventral mesoderm of a representative embryo heterozygous for a loss-of-function zip allele (zip^+/−) and co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL 20 min before (H), 10 min before (I), and at completion of (J) ventral furrow invagination. The embryo was imaged with 561-nm light only. (J’) shows CIBN::pmGFP, (J’’) shows mCherry::CRY2-OCRL, and (J’’’) shows a merge of the two channels at completion of ventral furrow invagination. Scale bars, 10 μm. See also Figure S2 and Movies S3A and S3B.

Figure 4

Two-Photon Activation Allows Local Modulation of Cell Contractility (A) Mean levels of mCherry::CRY2-OCRL plasma membrane recruitment in response to 488-nm light (one-photon illumination) and to 900-, 950-, and 1,000-nm light; two-photon illumination, achieved using a femtosecond (140-fs) pulsed laser at a repetition rate of 80 MHz. A single z stack was illuminated 5 μm from the apical surface, for a total scanning time of 1 s. In between two consecutive scans, there was an interval of 30 s and the imaging of mCherry::CRY2-OCRL at 561 nm. The protocol for 488-nm illumination is the same as described in Figure 1C. The values on the x axis correspond to the time of two-photon or 488-nm illumination. The y coordinates represent the log₂ ratio of mCherry::CRY2-OCRL plasma membrane (pm) to cytosol (cyt) fluorescence intensities. Pooled data are represented as mean ± SD (n = 3 embryos for each condition). (B) Mean levels of mCherry::CRY2-OCRL plasma membrane recruitment in response to 488- and 950-nm light. For λ = 488 nm, a single z stack was illuminated for 1 s. For λ = 950 nm, five consecutive z stacks separated by 1 μm were illuminated for ∼500 ms, for a total scan time of 2.5 s. In between two consecutive pulses, there was an interval of 30 s and the imaging of mCherry::CRY2-OCRL at 561 nm. The values on the x axis correspond to the time of two-photon or 488-nm illumination. The y axis displays the log₂ ratio of mCherry::CRY2-OCRL plasma membrane to cytosol fluorescence intensities. Pooled data are represented as mean ± SD (n = 3 embryos for each condition). (C) Mean levels of mCherry::CRY2-OCRL plasma membrane recruitment in response to excitation with a 950-nm laser at three different powers (3.0 mW, 1.5 mW, and 0.7 mW). Photo-activation was achieved by illuminating five consecutive z stacks separated by 1 μm for ∼500 ms each (total scan time = 2.5 s). In between two consecutive scans, there was an interval of 30 s and the imaging of mCherry::CRY2-OCRL at 561 nm. The values on the x axis correspond to the time of two-photon illumination. The y coordinates represent the log₂ ratio of mCherry::CRY2-OCRL plasma membrane to cytosol fluorescence intensity. Pooled data are represented as mean ± SD (n = 3 embryos for each condition). (D) Confocal image of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL 10 min after photo-activation, achieved as described in (C). mCherry::CRY2-OCRL is shown. The red box indicates the photo-activated area, which corresponds to (D’). In (D’), CIBN::pmGFP is shown. (E) Mean levels of mCherry::CRY2-OCRL plasma membrane recruitment with respect to the photo-activated area. The x coordinates represent cell IDs, with 0 indicating cells that are completely included in the photo-activation box (yellow cells in D and D’); −1 and 1 indicating cells that are partially included in the photo-activation box (green and blue cells in D and D’, respectively); −2, −3, −10 and 2, 3, 10 indicating, respectively, cells that are two, three, and ten cells away from cell 0 on either side of the photo-activation area. The y coordinates represent the log₂ ratio of mCherry::CRY2-OCRL plasma membrane to cytosol fluorescence intensities. Pooled data are represented as mean ± SD (n = 3 embryos for each condition). (F–I) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (F) and 10 min (G) and 20 min (H) after the beginning of two-photon illumination. The red box in (F) indicates the photo-activated area. Photo-activation was started before the beginning of ventral furrow formation and was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s. (I) Confocal image of the same embryo 30 min after the beginning of photo-activation. CIBN::pmGFP is shown. Scale bars, 10 μm. See also Figures S3, Figure S4, and Movie S4.

Figure 5

Local Inhibition of Apical Constriction in a Subgroup of Ventral Cells Causes Arrest of Ventral Furrow Formation and Coordinated Contractions (A–C) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (A) and at 9 min (B) and 20 min (C) after photo-activation. The red box in (A) indicates the photo-activated area. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 3.0 mW). (D) Confocal image of the same embryo 25 min after the first pulse of light. CIBN::pmGFP is shown. Cells in non-activated areas (white boxes) are hyperconstricted (n = 5 embryos). (E–G) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (E) and at 9 min (F) and 25 min (G) after photo-activation. The red box in (E) indicates the photo-activated area. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 1.5 mW). (H) Confocal image of the same embryo 31 min after the first pulse of light. CIBN::pmGFP is shown. Differently from (D), some of the cells contained within the photo-activated area are constricted, while some other cells are not. Notably, neighboring non-activated cells (white boxes) display a similar pattern of contractility to photo-activated cells (n = 3 embryos). (I–K) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (I) and at 9 min (J) and 15 min (K) after photo-activation. The red box in (I) indicates the photo-activated area. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 0.7 mW). (L) Confocal image of the same embryo 20 min after the first pulse of light. CIBN::pmGFP is shown. Both photo-activated cells and neighboring non-activated cells constricted and ventral furrow formed (n = 3 embryos). (M–O) Quantification of cell area (purple) and a-p anisotropy (green) for the photo-activated cells in (A–C, red box) (M), for the photo-activated cells in (E–G, red box) (N), and for the photo-activated cells in (I–K, red box) (O). The y coordinates represent cell area expressed in squared microns (left y axis) and a-p anisotropy (right y axis). The x axis represents time in minutes. Solid and dashed lines indicate the median over all cells for cell area (solid) and a-p anisotropy (dashed). Shaded regions show the interquartile range. (P) Mean levels of mCherry::CRY2-OCRL recruitment to the plasma membrane in response to different laser powers (3.0 mW, 1.5 mW, and 0.7 mW). The y axis shows the log₂ ratio of the membrane to cytoplasmic fluorescence intensity of mCherry::CRY2-OCRL in photo-activated cells. The x axis displays time in min, with 0 corresponding to the beginning of photo-activation. Pooled data are represented as mean ± SD (n ≥ 3 embryos for each condition). Scale bars, 10 μm. See also Movies S5A, S5B, and S5C.

Figure 6

The Number of Cells in which Apical Constriction Is Inhibited Influences the Capability of Neighboring Cells to Form a Furrow (A–C) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (A) and at 9 min (B) and 15 min (C) after the beginning of photo-activation. The red box in (A) indicates the photo-activated area, which partially overlaps, in the a-p direction, the presumptive mesoderm. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 3.0 mW). (D) Confocal image of the same embryo taken 17 min after the first pulse of light. CIBN::pmGFP is shown (n = 3 embryos). (E–G) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (E) and at 9 min (F) and 17 min (G) after the first pulse of local photo-activation. The red box in (E) indicates the photo-activated area, which partially overlaps, in the a-p direction, the presumptive mesoderm. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 3.0 mW). (H) Confocal image of the same embryo taken 19 min after the first pulse of light. CIBN::pmGFP is shown (n = 3 embryos). (I–K) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (I) and at 9 min (J) and 28 min (K) after the beginning of photo-activation. The red box in (I) indicates the photo-activated area. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 3.0 mW). (L) Confocal image of the same embryo taken 31 min after the first pulse of light. CIBN::pmGFP is shown (n = 3 embryos). (M–O) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (M) and at 8 min (N) and 20 min (O) after the beginning of photo-activation. The red box in (M) indicates the photo-activated area. Photo-activation was achieved with 950-nm laser light, for a total scanning time of 2.5 s every 30 s (laser power = 3.0 mW). (P) Confocal image of the same embryo taken 23 min after the beginning of photo-activation. CIBN::pmGFP is shown (n = 3 embryos). (Q) Comparison of mean photo-activation levels among embryos where photo-activation was achieved in boxes of different sizes and positions within the furrow primordium. The x coordinates represent samples. The y coordinates represent the log₂ ratio of CRY2-OCRL plasma membrane (pm) to cytosol (cyt) fluorescence intensity. Pooled data are represented as mean ± SD (n ≥ 3 embryos for each condition). Scale bars, 10 μm. See also Movies S6A and S6B and Movies S7A and S7B.

Figure 7

The Geometry of the Ventral Furrow Primordium Determines the Degree of Anisotropy of Cell Shape Changes in Individual Cells (A and B) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (A) and 9 min after the beginning of photo-activation (B). Red boxes in (A) indicate photo-activated areas, and the white box indicates a non-activated area where cells can constrict. Photo-activation was achieved by illuminating each area with 950-nm laser light, for a total scanning time of 2.5 s every 30 s. (C) Confocal image of the same embryo taken 10 min after the beginning of photo-activation. CIBN::pmGFP is shown. Anisotropy along the embryo a-p axis was calculated only for constricting cells (white box) in the non-activated area (n = 5 embryos). (D and E) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (D) and 8 min after the beginning of photo-activation (E). Red boxes in (D) indicate photo-activated areas, and the white box indicates a non-activated area where cells can constrict. Photo-activation was achieved as described in (B). (F) Confocal image of the same embryo taken 9 min after the first pulse of light. Anisotropy along the embryo a-p axis was calculated only for constricting cells (white box) in the non-activated area (n = 8 embryos). (G) Comparison of median a-p anisotropy of non-activated constricting cells between samples in dependence of the distance between the photo-activated areas. Values along the y axis represent a-p anisotropy. The degree of a-p anisotropy is higher when the boxes are placed farther apart than when they are closer together. Each dot represents a single embryo. The crosses show group median and interquartile range. p = 4.3 × 10⁻³, two-sample Welch’s t test. (H–J) Still frames from a confocal movie of the ventral mesoderm of a representative embryo co-expressing CIBN::pmGFP and mCherry::CRY2-OCRL before photo-activation (H) and at 8 min (I) and 16 min (J) after the beginning of photo-activation. The red boxes in (H) indicate photo-activated areas. Photo-activation was achieved as described in (B). (K) Confocal image of the same embryo taken 18 min after the first pulse of light (n = 3 embryos). Scale bars, 10 μm. See also Figure S5 and Movies S8A, S8B, and S9.