XB-ART-57693

PLoS Biol
January 1, 2021;
19
(1):
e3001060.
## Ectoderm to mesoderm transition by down-regulation of actomyosin contractility.

Kashkooli L
,
Rozema D
,
Espejo-Ramirez L
,
Lasko P
,
Fagotto F
.

Abstract

Collective migration of cohesive tissues is a fundamental process in morphogenesis and is particularly well illustrated during gastrulation by the rapid and massive internalization of the mesoderm, which contrasts with the much more modest movements of the ectoderm. In the Xenopus embryo, the differences in morphogenetic capabilities of ectoderm and mesoderm can be connected to the intrinsic motility of individual cells, very low for ectoderm, high for mesoderm. Surprisingly, we find that these seemingly deep differences can be accounted for simply by differences in Rho-kinases (Rock)-dependent actomyosin contractility. We show that Rock inhibition is sufficient to rapidly unleash motility in the ectoderm and confer it with mesoderm-like properties. In the mesoderm, this motility is dependent on two negative regulators of RhoA, the small GTPase Rnd1 and the RhoGAP Shirin/Dlc2/ArhGAP37. Both are absolutely essential for gastrulation. At the cellular and tissue level, the two regulators show overlapping yet distinct functions. They both contribute to decrease cortical tension and confer motility, but Shirin tends to increase tissue fluidity and stimulate dispersion, while Rnd1 tends to favor more compact collective migration. Thus, each is able to contribute to a specific property of the migratory behavior of the mesoderm. We propose that the "ectoderm to mesoderm transition" is a prototypic case of collective migration driven by a down-regulation of cellular tension, without the need for the complex changes traditionally associated with the epithelial-to-mesenchymal transition.

PubMed ID: 33406067

PMC ID: PMC7815211

Article link: PLoS Biol

Species referenced: Xenopus laevis

Genes referenced: arhgap1 bcr cdh3 ctnnb1 ctrl fas fn1 mlc1 mylkl odc1 pcyt1a pxn rho rho.2 rhoa rnd1 rock1 rock2 stard13 tst vcl

Morpholinos: cdh3 MO4 rnd1 MO1 rnd1 MO2 rnd1 MO3 stard13 MO1 stard13 MO2

Article Images: [+] show captions

Fig 1. Distinct properties of ectoderm and mesoderm at the cellular level. (A–E) Organization of cell-matrix adhesive structures. Dissociated Xenopus ectoderm (A, C) and mesoderm (B, D) cells expressing Vinc-Che and mYFP were plated on FN, either as single cells (A, B) or as small groups (C, D) and imaged live by spinning disc confocal microscopy. y: autofluorescence of yolk platelets. Ventral: ventral z plane close to the glass. max p: Maximal z projection. (A) Ectoderm cells do not spread on FN, but adhere to it through a characteristic adhesive ring (A, A’, filled arrowheads). They typically form blebs that are continuously pushed around the cell (dashed line with arrow). Right inserts: orthogonal view (orth) showing the cross-section of the membrane and of the vinculin ring (filled arrowheads). The dashed line underlines the bottom of the bleb. (B) Mesoderm cells spread on FN and extend multiple lamellipodia. They transiently polarize during their migration, with 1 protrusion becoming the tail (t); see also time lapse S1 Fig. They form vinculin-positive FAs (filled arrowheads), generally oriented in the direction of the protrusions (arrows). (C) Ectoderm cells form compact groups, with few protrusions in the center and numerous blebs at the periphery (dashed lines). External cells emit protrusions under the more central cells (yellow arrows). Individual cells build partial adhesive structures (filled arrowheads), which together form a supracellular ring. (D) Mesoderm cells form looser groups, each cell emitting multiple lamellipodia, most of them extending outwards (white and yellow arrows indicate peripheral and internal lamellipodia, respectively), with numerous FAs oriented radially (arrowheads). Panel D’ is an enlargement of the boxed portion of panel D. Scale bars: A, C, D 10μm; B 20μm; D’ 5μm. (E) Quantification of vinculin accumulation at FAs of isolated cells, expressed as Vinc-Che fluorescence concentrated in clusters divided by the total fluorescence along the ventral cortex. A color code is used throughout the figures, including blue for control ectoderm and red for control mesoderm. The box plots show the interquartile range (box limits), median (center line and corresponding value), and min and max values without outliers (whiskers). Statistical comparison using 2-sided Student t test. For all experiments presented in this study, P values are indicated as follows: *P <0.05, **P <0.01, ***P <0.001; NS, not significant. The same color code is also used to indicate statistical comparison between 1 condition and control ectoderm (blue) or control mesoderm (red). Other comparisons are indicated by black asterisks and brackets. The numbers in parentheses correspond to the number of cells/ number of experiments. Refer to S1 Data. (F) Single-cell motility. Frames of spinning disc confocal time-lapse movies. Cells expressed paxillin fused to YFP (Pax-YFP) and membrane Cherry. (F) Ectoderm cells are immobile, anchored by their stationary adhesive ring (arrowheads) and bleb (star). Scale bars: F 5 μm; F’ 20 μm. (F’) Mesoderm cells actively migrate, rapidly remodeling protrusions and FAs (red-yellow-green color-coded arrows and arrowheads indicate successive positions respectively of 1 extending lamellipodium and the retracting tail). White arrowheads: FAs at thin protrusions. (G) Quantification of single-cell migration. Refer to S1 Data. (H) Adhesion assay. Dissociated cells were plated on the adherent substrate, either FN or recombinant cadherin-Fc fusion protein, then subjected to rotation. Adhesion is expressed as the percentage of cells remaining adherent after rotation (see Materials and methods). The column plots show averages and standard deviation of 15 experiments, a total of approximately 5,000 cells/conditions. Statistical comparison on the % adherent cells/experiment, pairwise 2-sided Student t test. Refer to S1 Data. FA, focal adhesion; FN, fibronectin; mYFP, membrane-targeted yellow fluorescent protein; Vinc-Che, Vinculin-Cherry; YFP, yellow fluorescent protein. https://doi.org/10.1371/journal.pbio.3001060.g001 | |

Fig 2. Inhibition of Rock confers ectoderm cells with mesoderm-like properties. (A-C) Induction of cell spreading and migration by Rock inhibition. (A–D) Confocal imaging of initiation of spreading and migration for single cells (A, B) and a small group of cells (C). Rock inhibitors, Y27632 (50 μM), and H1125 (1 μM) were added at time = 0’. Note that the onset of the transition is not synchronous. Arrows: nascent protrusions; filled arrowheads: ring-like adhesion; carrowheads: FAs. Scale bars: 10 μm. (D) Shift in cell morphology. Cells were classified in morphological subtypes: round and blebbing (b), round without blebs (r), polarized (p), and spread (s). In wild-type conditions, round cells are typically immotile, while polarized and spread cells migrate. A fifth category, named polarized with bleb (pb), includes cells with irregular morphology and blebs. The diagram shows the distribution of wild-type mesoderm and ectoderm cells, as well as of ectoderm cells treated for 50 minutes with 10 μM or 50 μM Y27632 (Y10 and Y50) or 1μM H1125 (H). For b and s categories, conditions were compared to control ectoderm by 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. (E) Migration speed of Rock-inhibited cells. Quantification as in Fig 1. Comparison to ectoderm control by 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; FA, focal adhesion; HSD, honestly significant difference; mYFP, membrane-targeted YFP; Vinc-Che, Vinculin-Cherry. https://doi.org/10.1371/journal.pbio.3001060.g002 | |

Fig 3. Rnd1 and Shirin are essential for gastrulation. (A) Rnd1 and Shirin expressions are enriched in the mesoderm. RT-qPCR from dissected tissue. mRNA levels in ectoderm and mesoderm, normalized to homogenously expressed ODC. Three to 6 experiments, pairwise 1-sided Student t test. Refer to S1 Data. (B–E) Whole embryo loss-of-function phenotypes: 4-cell stage embryos were injected in the dorsal side (d) with a control (COMO), Rnd1 (RndMO), or Shirin (ShiMO) morpholinos. Embryos were fixed and imaged at stage 11. (B-D) Examples of typical control mesoderm and RndMO and ShiMO phenotypes. (E) Example of a “partial” phenotype (here ShiMO). (B–E) External views from the vegetal pole. Red arrows point to the position of the dorsal blastopore lip of a control embryo, absent in RndMO (C) and ShiMO (D) embryos. (E) In the partial phenotype, the lip is present but the blastopore has remained widely open compared to control. In many embryos, the ventral blastopore is also affected, due to the diffusion of the morpholinos to the ventral blastomeres before complete separation after the 2nd cleavage. (B’–E’) Sagittally bisected embryos. In a control embryo (B’), the extent of involution (dashed black arrow) can be seen by the position of the tip of the mesendoderm (white arrow) that has moved far away from the blastopore lip (red arrow). (C’) RndMO embryo lacking any sign of involution. The white arrowhead points to the dorsal edge of the blastocoel cavity (bl), resembling that of a pregastrula embryo. (D’) Characteristic ShiMO phenotype, with flat blastocoel floor (white arrow) and thicker non-involuted dorsal marginal zone (black arrowheads), both indicative of failed involution. (E’) Partial involution (white arrow). Yellow arrowheads: thin BCR, indicative of ectoderm epiboly in all conditions. Scale bars: 200 μm. (F, F’) Score of the penetrance of the gastrulation external and internal phenotype: N, normal embryo; P, partial inhibition, C, complete inhibition. Comparison by 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; HSD, honestly significant difference; ODC, ornithine decarboxylase; RT-qPCR, reverse-transcription quantitative PCR. https://doi.org/10.1371/journal.pbio.3001060.g003 | |

Fig 4. Rnd1 and Shirin are essential for mesoderm spreading and migration. (A–E) Loss-of-function cellular phenotypes. (A–C) Examples of control morpholino (COMO), RndMO, and ShiMO mesoderm cells, expressing Vinc-Che and mYFP, plated on FN. (A–C) Ventral z planes, merged channels; (A’–C’) vinculin alone; (A”–C”) maximal z projections. (A”’–C”’) Schematic diagrams summarizing the general cell morphology and adhesive structures. Protrusions are indicated by arrows, FAs by concave arrowheads, vinculin ring by filled arrowheads. Dotted lines highlight the max lateral extension of the cell mass. (A) Control spread mesoderm with large protrusions and numerous FAs. t, tail. (B) Typical RndMO cell displaying a bulging body (B”’, dotted lines) and a small ventral surface with diffuse vinculin (B”’, pink line). (C) Example of a bulky ShiMO cell with intermediate adhesive structures, including small FAs at short protrusions (arrows) and ectoderm-like partial ring encompassing most of the ventral surface (arrowheads). A bleb is visible in the max projection (C”, dashed line). Y, yolk platelets. Scale bars 10 μm. (D) RndMO and ShiMO cells show a significant shift in morphology from spread to round and blebbing cell. Comparison for either of the 2 categories with corresponding COMO (red asterisks), 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. (E) Both RndMO and ShiMO inhibit cell migration. Gray asterisks: comparison with double injection RndMO + ShiMO, which significantly enhanced the migration phenotype. One-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. (F) Quantification of vinculin accumulation. Comparison to COMO using pairwise 2-sided Student t test. Refer to S1 Data. RndMO cells have little to no detectable vinculin-rich structures. ShiMO cells show high variability (see main text). (G, H) Inhibition of cell adhesion on FN and on cadherin substrates. Five experiments, a total of 375–740 cells/conditions for FN, >1,000 cells for cadherin. Statistical comparison on the % adherent cells/experiment, pairwise 2-sided Student t test. Refer to S1 Data. ANOVA, analysis of variance; FA, focal adhesion; FN, fibronectin; HSD, honestly significant difference; mYFP, membrane-targeted YFP; Vinc-Che, Vinculin-Cherry. https://doi.org/10.1371/journal.pbio.3001060.g004 | |

Fig 5. Ectopic expression of Rnd1 or Shirin confers ectoderm with mesoderm-like morphological and migratory properties. (A–G) Effect on cell morphology and vinculin distribution. (A–E) Examples of ectoderm cells co-expressing Vinc-Che and either mYFP (A, control ectoderm), Rnd1-YFP (B, C), or Shirin-YFP (Shi-YFP, D, E). (A) Typical control ectoderm cell, with its distinctive vinculin ring (arrows) and blebs (dashed lines). (B) Rnd1-expressing cells elongate, expand their ventral surface in contact with the substrate, but form only few vinculin-positive FA-like structures (arrows). (C) High Rnd1 expression: The ventral surface is expanded, but lacks vinculin FAs. Cells are bulkier (contours highlighted by dotted lines), although blebs are absent. (D) Shirin-expressing cells spread and form prominent lamellipodia with FAs (white arrows). The yellow arrow points the retracting tail. (E) High Shirin expression: Cells emit long and disorganized protrusions in all directions, but lack detectable FAs, and the cell body tends to round up (dotted lines). Y, yolk platelets. Scale bars: 10 μm. (F) Quantification of vinculin accumulation. Consistent with the loss of the ring and the paucity of FAs, most of vinculin is homogeneously distributed on the ventral surface. Refer to S1 Data. (G) Distribution of morphological subtypes. Both Rnd1 and Shirin cause a strong shift toward spread cells. See S5A Fig for additional morphometric data. Refer to S1 Data. (H, I) Effect on cell migration and adhesion. (H) Frames from time-lapse movies. Examples of Rnd1- and Shirin-expressing ectoderm cells spreading and migrating. The cell in H” spreads extensively, ending with multiple protrusions (black arrowheads) and low motility. Scale bar: 20 μm. (I) Quantification of cell migration, as in Fig 2. Different levels of Rnd1 and Shirin expression were tested (250 and 500 pg mRNA for Rnd1, 75 and 150–300 pg for Shirin). Rnd1-expressing cells show higher migration than wild-type or Shi-expressing cells. Statistical comparisons: 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. (J, K) Quantification of cell adhesion on FN and on cadherin. Four to 5 experiments, >1,000 cells per condition. Statistical comparison on the % adherent cells/experiment, pairwise 1-sided Student t test. Refer to S1 Data. ANOVA, analysis of variance; FA, focal adhesion; FN, fibronectin; HSD, honestly significant difference; mYFP, membrane-targeted YFP; Vinc-Che, Vinculin-Cherry. https://doi.org/10.1371/journal.pbio.3001060.g005 | |

Fig 6. Effect of ectopic expression of Rnd1 and Shirin on cell adhesive structures and cortical myosin. β-catenin, used as general marker for cadherin-based cell adhesions, and pMLC were localized by immunofluorescence on cryosections of whole embryos at early gastrula stage. The fluorescence along the cell periphery, defined by the β-catenin signal, was quantified and expressed relative to the median intensity of control ectoderm. (A, B) Comparison of β-catenin and pMLC levels in the dorsal ectoderm and dorsal PCM of normal embryos. (A) Diagram of the embryo with boxes indicating the regions used for quantification. (A’) Quantification. Numbers into brackets: number of embryos/number of experiments. Statistical comparison to ectoderm using 2-sided Student t test. (B) Example of dorsal region, immunolabeled for β-catenin (magenta) and pMLC (red). Nuclei were counterstained with Hoechst. (B”’) Enlarged view of the region used for quantification. (C–G) Effect of Rnd1 and Shirin ectopic expression in the ectoderm. (C) Diagram indicating the regions of the ectoderm used for quantification. For consistency, all analyses were performed on the upper lateral region (both dorsal and ventral, indicated by dashed boxes in the diagram, because it constitutes a robust landmark where the inner ectoderm layer has a stereotyped organization. (C’) Quantification. Statistical comparison using 2 sided, pairwise Student t test. Refer to S1 Data. (D-G) Examples of control (D), YFP-Rnd1-expressing (E), and YFP-Shirin-expressing (F) ectoderm, immunolabeled for the YFP-tag (green), β-catenin (far red, colored in magenta), and pMLC (red). Top panels present general views, and the other panels show enlarged portions of the inner ectoderm layer used for quantification. Note that the strong bending of the ectoderm layer is due to the partial collapse of the blastocoel cavity during fixation. White arrows point to plasma membranes marked by β-catenin (D’, E’, F’) and to the corresponding pMLC signal (D”, E”, F”). Little to no pMLC enrichment is observed in Shirin-expressing cells (F”). Arrowheads in E’ and F’ point to concentrations of β-catenin, particularly frequent in Shirin-expressing ectoderm, and which contrast with the low membrane signal (arrow). Yellow arrow in E: Rnd1-expressing ectoderm cells that have penetrated into the mesoderm layer. Yellow arrows in F: ectoderm cells expressing particularly high levels of YFP-Shirin. PCM, prechordal mesoderm; pMLC, phosphorylated MLC. https://doi.org/10.1371/journal.pbio.3001060.g006 | |

Fig 7. Differential subcellular distribution of Rnd1 and Shirin. (A, B) General distribution of Rnd1 and Shirin in mesoderm cells. (A, B) Live confocal microscope images of groups of mesoderm cells co-expressing cadherin-dTomato (Cad-Tom) and either Rnd1-YFP or GAP-deficient mutant ShirinR488A-YFP (mShi-YFP). Both Rnd1 and mShi localized to the cell cortex (concave arrowheads). On the ventral side, mShi was concentrated at protrusions (B’, arrows, see D–H), while Rnd1 was always homogenously distributed (example in A’). Rnd1, but not mShi, is concentrated at cell–cell contacts (filled arrowheads). Y, yolk platelets. (C) Quantification of Rnd1 and Shirin at cell–cell contacts, expressed as ratio of the signal intensity at cell–cell contacts divided by twice the signal along free cell edges. Rnd1 is enriched more than 2-fold at contacts, similar to cadherin. mShi is distributed homogenously along the cell periphery. Comparison Rnd1/mShi to cadherin (red) or mShi to Rnd1 (black) using 2-sided Student t test. Refer to S1 Data. (D–H) Shirin localization at the ventral surface. (D, E) Ventral surface of mesoderm cells co-expressing either mShi (D) or wtShi (E) together with Vinc-Che. (D, E) General view; D’–D”’, E’–E”’) enlargements of protrusions. mShi extensively colocalizes with vinculin at FAs (white arrowheads). (E) wtShi clusters are present throughout the ventral side of protrusions (arrow). Vinculin-positive FAs are largely confined to the periphery, only partially overlapping with wtShi clusters (orange arrowheads for vinculin, white concave arrowheads for wtShi). (F, G) Ectopic wtShi in ectoderm cells. (F) Detail of a protrusion of a fully spread cell. Similar to mesoderm, the center of the protrusion is occupied by clusters of wtShi and devoid of FAs (arrow). Small FAs are located at the periphery, close to Shirin clusters (orange and white concave arrowheads), but rarely colocalizing (white filled arrowhead). (G) Incompletely spread wtShi-expressing ectoderm cells. The left cell has lost its vinculin ring, and a wtShi-enriched protrusion is forming (arrow). The right cell still shows a weak ring lined in the inside by wtShi clusters (orange and white concave arrowheads). Scale bars: 10 μm. (H) Quantification of Shirin and Vinculin co-localization, expressed by the general Pearson’s coefficient, as well as by Mander’s coefficients, which indicates the portion of Shirin that overlap with Vinculin (M1) and the converse portion of Vinculin that overlap with Shirin (M2). Statistical comparison using 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; FA, focal adhesion; HSD, honestly significant difference; Vinc-Che, Vinculin-Cherry; wtShi, wild-type Shirin. https://doi.org/10.1371/journal.pbio.3001060.g007 | |

Fig 8. Ectopic expression of Rnd1 or Shirin modulates ectoderm cortical tension and adhesiveness. (A) Diagram of an asymmetrical cell doublet, representing the balance between cortical tensions at free edges CtA, CtB, and contact tension TAB. The orange layer symbolizes the actomyosin cortex. The curved cell–cell interface reflects unequal CtA and CtB tensions. (B–I) Examples of homotypic and heterotypic doublets, imaged by live confocal microscopy. Doublets were made by combining dissociated control ectoderm expressing mYFP (ctrl) and either Rnd1 or Shirin-expressing cells markerd with mCherry. Wild-type and Rnd1-expressing cells often displayed blebs (dashed lines). Curved interfaces indicative of tensile differences were observed for all combinations, including for homotypic doublets (e.g., panel F), but were most systematically found for heterotypic ctrl-Shirin doublets (H, I). Scale bar: 20 μm. (J–L) Relative tension measurements based on the geometry at cell vertices (see S1 Appendix). (J) Relative cortical tension between Rnd1 or Shirin-expressing cells and control ectoderm cells calculated from the ratio CtA/CtB of heterotypic doublets. The ratio for control homotypic doublets is provided for comparison. See S1 Appendix for complete measurements. Vertices flanked by a bleb (D and I) were omitted from calculations. (K) Relative strength of contact tension TAB at homotypic contacts, compared to control ectoderm-ectoderm T, the median of which was set arbitrarily at 1. See S1 Appendix for more details. (L) Relative adhesiveness α, calculated for homotypic doublets. Numbers in brackets: vertices/experiments. Statistical comparison using 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; HSD, honestly significant difference. https://doi.org/10.1371/journal.pbio.3001060.g008 | |

Fig 9. Rho/Rock regulation affects collective migration of ectoderm and mesoderm tissue explants. Tissue explants were laid on FN, and their spreading was imaged for 170 minutes. (A–D) Control ectoderm, ectoderm treated with 50-μM Y27632, and ectoderm expressing Rnd1 or Shirin. Numbers in brackets are number of explants and number of experiments. Scale bar: 100 μm. (E, F) Control mesoderm and mesoderm treated with Y26862. Red arrowheads in E indicate areas of large-scale retractions (compare 85 and 170 minutes). Scale bar: 100 μm. (G, H) Quantification of explant spreading. After segmentation, the area was calculated for the time course and normalized to the first time point. Traces show average time course curves with SD for the various experimental conditions. (G’, H’) Corresponding relative spreading after 60 minutes and 170 minutes, chosen to represent an intermediate and advanced stage of the spreading process. Statistical analysis, 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. (I–M) Delaunay triangulation of nuclei and quantification of cell dispersion. (I–L) Representative maps of triangulated nuclei after 170 minutes of imaging. X and Y labels mark the coordinates in μm, and the color-coded scale bar indicates the area of the triangles in μm2. (M) Quantification of the relative change in triangle size over time calculated by dividing the average triangle area at 170 minutes by that at 30 minutes. The 30-minute time point was chosen as it corresponds to the stage when explant had adhered to the substrate and started to spread. (N) Quantification of intercalation calculated by dividing number of nuclei at the ventral surface at 170 minutes by the number at 30 minutes. Statistical comparisons: 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; FN, fibronectin; HSD, honestly significant difference; Rock, Rho-kinases; SD, standard deviation. https://doi.org/10.1371/journal.pbio.3001060.g009 | |

Fig 10. ROCK inhibition and Rho regulators modulate tissue stiffness, viscosity, and surface tension. Micropipette aspiration was used to measure physical properties of tissue explants. Explants were aspirated into the pipette at constant pressure, then pressure was reset to 0 to let the explant retract. (A, B) Examples of aspiration and retraction of control ectoderm and mesoderm explants. Aspiration pressure was 250 Pa. Pressure was released after 240 seconds. Scale bars: 100 μm. (A’, B’) Corresponding aspiration and release profiles. The blue double arrows indicate the extent of deformation of the tissue during the first 20 seconds, defined as the fast “passive” phase. The 2 slow, linear phases of aspiration and release, highlighted in red, were used to calculate viscosity and TST. Scale bars: 100 μm. (C) Examples of aspiration of control ectoderm, ectoderm treated with Y26862, or expressing Rnd1 or Shirin, and control mesoderm. Pressure was 250 Pa. Images display the frame corresponding to the deformation 220 seconds after the initiation of aspiration. The colored overlays indicate the distances of deformation during the first fast phase (20 seconds, blue) and during the subsequent slow phase (220 seconds, red). Scale bar: 100 μm. (D–I) Calculated parameters: (D, G) Length of deformation 20 seconds after initiation of aspiration, encompassing the initial passive phase. (E, H) Tissue viscosity calculated from the rates of aspiration and retraction (see Materials and methods). (F, I) TST. Numbers in brackets are number of explants and number of experiments. Statistical comparisons: 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; HSD, honestly significant difference; Rock, Rho-kinases; TST, tissue surface tension. https://doi.org/10.1371/journal.pbio.3001060.g010 | |

S1 Fig. (Related to Fig 1) (A) Mode of mesoderm locomotion. Consecutive frames from time lapse of mYFP labeled mesoderm cells migrating on FN. The behavior of the central cell is highlighted: The cell emits 1 or multiple protrusions (red arrows). One of the protrusions becomes a tail (yellow arrowhead) as the cell stretches toward another direction, and eventually retracts (red arrowheads). | |

S1 Fig. (B) Quantification of accumulation of Vinculin-Cherry in FAs: linearity between fluorescence levels in FA and total intensity (related to Fig 1A–1D). Because Vinculin-Cherry expression levels vary from cell to cell, quantification was performed for individual cells by measuring fluorescence in bright clusters (corresponding to FAs) and in the total ventral cell surface (pink on the diagram). The plot shows the average intensity of the ventral surface versus the average intensity in FAs for control mesoderm cells in 1 experiment, each dot corresponding to a single cell. It shows that accumulation at FAs is proportional to total expression levels over a wide range. Refer to S1 Data. Linearity was similarly verified for each experiment. FA, focal adhesion; FN, fibronectin; mYFP, membrane-targeted yellow fluorescent protein. https://doi.org/10.1371/journal.pbio.3001060.s003 | |

S2 Fig. Localization of MLC and Rock (related to Fig 2). (A–C) Differential MLC accumulation at the cell cortex. Ectoderm and mesoderm cells expressing MLC-Cherry (MLC-Che) and mYFP. (A) Ectoderm cells show strong accumulation around the cell body (arrows) and part of the blebs (arrowhead). (B) Mesoderm cells show irregular cortical MLC, mostly at the concave regions near or between protrusion. (C) Quantification of cortical MLC, expressed as the ratio of cortical /cytoplasmic fluorescence intensities. Blebs and protrusions were excluded from the measurements. Statistical comparison using 2-sided Student t test. Refer to S1 Data. Scale bars: A’ 5 μm, B’ 10 μm, B” 5 μm. (D–K) Subcellular localization of Rock1-YFP and Rock2-YFP in ectoderm and mesoderm cells. Selected single planes from live confocal microscopy, either near the glass (ventral) or about 5–10 μm above (medial). Concave white arrowheads point at examples of Rock1/2 accumulation. (D, E, H, I) Localization relative to the cell cortex and to Vinculin-Cherry labeled cell-matrix adhesive structures (red arrowheads). (F, G, J, K) Localization relative to cell–cell contacts, marked by cadherin-dTomato (red arrows). (D, E) In the ectoderm, Rock1 and 2 have both a cortical localization. Levels are low on the ventral side inside the adhesive ring, but stronger outside of the ring, particularly for Rock2. (F, G) Levels are very low at cell–cell contacts. (H, I) In the ventral face of mesoderm cells, Rock1 tend to be enriched in the central part, Rock2 at the periphery of the protrusions. Both are low at FAs. They both accumulate at the cortex along cell free edges (medial planes). (J, K) Levels are low at cell–cell contacts. Y: autofluorescence of yolk platelets, abundant in mesoderm cells. FA, focal adhesion; MLC, myosin light chain; mYFP, membrane-targeted YFP; Rock, Rho-kinases. https://doi.org/10.1371/journal.pbio.3001060.s004 | |

S3 Fig. (Related to Fig 2) (A, B) Area expansion for single cells after treatment with Rock inhibitors Y27632 (50 μM) and H1125 (1 μM). Average and SD of 107 cells (A) and 34 cells (B). (C) Changes in vinculin distribution. Images from a time-lapse movie of a small group of 3 cells expressing Vinculin-Cherry, treated at time = 0 with Y27632. Filled arrowheads: ring-like adhesion; concave arrowheads: FAs. Scale bars: 10 μm. (D, E) Opposite effects of Rock and MLCK inhibition on cell adhesion. Ectoderm and mesoderm adhesion to FN or cadherin was measured after treatment with Rock inhibitors Y27632 (Y, 50 μM), H1125 (H, 1 μM), or the MLCK inhibitor ML7. Five experiments, a total of 1,000–2,000 cells/conditions. Statistical comparison to control ectoderm or mesoderm, comparing the % adherent cells/experiment, pairwise 2-sided Student t test. Refer to S1 Data. FA, focal adhesion; FN, fibronectin; SD, standard deviation. https://doi.org/10.1371/journal.pbio.3001060.s005 | |

S4 Fig. (Related to Fig 4) (A–D) Rescue of Rnd1MO and ShiMO spreading and migration phenotypes. Four-cell stage embryos were injected in the dorsal side with COMO, RndMO, RndMO + YFP-Rnd1 mRNA (rescue), ShiMO, or ShiMO + YFP-Shirin mRNA (rescue). Dissociated mesoderm cells were plated on FN and time-lapse movies were recorded. The fourth condition represents RndMO or ShiMO cells treated with 50μM Y27632 Rock inhibitor (Y). Statistical comparions: 1-way ANOVA followed by Tukey HSD post hoc test. Red asterisks: Comparison to COMO. | |

S4 Fig.continued (E) Migration speed for different cell morphology categories. Analysis of data from Fig 4I. Red asterisks: comparison to COMO. One-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; FN, fibronectin; HSD, honestly significant difference. https://doi.org/10.1371/journal.pbio.3001060.s006 | |

S5 Fig. (Related to Fig 5) (A) Morphometry of Rnd1 and Shirin induced spreading of ectoderm cells. The diagrams illustrate typical cell shapes. Corresponding images can be found in main Fig 5A–5E. These shapes were analyzed based on the following parameters: (A’) Area of the ventral contact surface (red in the schemes in A). (A”) Circularity of the ventral surface, which depends both on the roundness and regularity/convolution of the shape. (A”’) Ratio between the ventral area and the maximal cell area, calculated from maximal z projections. Refer to S1 Data. Blebs were excluded from measurements. Rnd1- and Shirin-expressing cells were here subdivided in 2 categorizes, low and high expression, based on the YFP fluorescence intensity. Note that these 2 categories overlap but are not equivalent Rnd1 expression levels had no significant impact on any parameter. Shirin expression had no effect on contact surface area, but high levels stimulated formation of convoluted protrusions (lower circularity) but decreased ventral/max area, reflecting the fact that many of them rounded up (fourth cell shape in panel A; see main Fig 5E). | |

S% Fig continued (B) Distribution of speed for ectoderm cells expressing Rnd1 or Shirin, compared to wild-type ectoderm and mesoderm. Brackets: range of high speed, comparable to mesoderm, achieved mainly by Rnd1-expressing cells. Refer to S1 Data. https://doi.org/10.1371/journal.pbio.3001060.s007 | |

S6 Fig. (Related to Fig 6) Effect of Rnd1 and Shirin ectopic expression. (A) Loosening of ectoderm tissue upon expression of Shirin. Immunostained section of a YFP-Shirin expressing embryo showing a loosely organized ectoderm, characterized by the presence of large intercellular spaces (asterisks) and heterogenous β-catenin signal, weak signal along membranes except for strong local concentrations (arrowheads). Scale bar, 10 μm. (B–E) Cell orientation. (B–D) The main axis of deep ectoderm cells (double arrows) tend to orient roughly perpendicular to the inner surface of the tissue (dashed line). Rnd1-expressing cells show variable orientation. Shirin-expressing cells align parallel to the surface. Scale bars, 10 μm. (E) Quantification of the angle between the cell axis and the tissue interface. Numbers into brackets correspond to number of cells/embryos/experiments. Refer to S1 Data. (F–I) Analysis of β-catenin (green) and pMLC (red) in ectoderm explants. (F) Diagram, section of a control ectoderm explant (scale bar, 50 μm) and quantification. Statistical comparison using 1-way ANOVA followed by Tukey HSD post hoc test. (G–I) Examples of ectoderm explants. (G) β-catenin and pMLC signal along cell edges is highest in control (arrows). H,I) Explants expressing Rnd1 or Shirin. β-catenin tends to accumulate at cell vertices (concave arrowheads). pMLC levels are lower except for some cells (I, asterisks) that have rounded up and display high pMLC throughout the cell. Little to no β-catenin is seen between the round cells. Scale bars, 20 μm. (H) Effect of Rnd1 and Shirin expression on phosphorylation of MYPT. Dissected ectoderm tissues were analyzed by western blot. GAPDH was used as loading control, and the pMYPT signal was expressed as relative ratio, normalized to ectoderm control set to 1.0. Three independent experiments, statistical analysis using 1 sample, 2-sided t test. Refer to S1 Data. ANOVA, analysis of variance; HSD, honestly significant difference; MYPT, myosin light chain phosphatase; pMLC, phosphorylated MLC. https://doi.org/10.1371/journal.pbio.3001060.s008 | |

S7 Fig. (Related to Fig 9) (A) Example of ectoderm explant showing late partial spreading, which is only observed beyond the 120 minutes. (B) Examples of traces for single explants, illustrating the irregular expansion of mesoderm explants interrupted by retractions. In contrast, expansion of Y26862-treated mesoderm is smooth. (C) Quantification of average triangle size at the initiation of spreading (30 minutes) and the end of the time lapse (170 minutes). Refer to S1 Data. https://doi.org/10.1371/journal.pbio.3001060.s009 | |

S8 Fig. (Related to Fig 9) Knockdown of Rnd1 or Shirin affect collective properties of mesoderm tissue explants. Analysis of spreading, dispersion, and intercalation of mesodermal explants under various conditions was performed as for experiments presented in Fig 9. (A–D) Control mesoderm, mesoderm treated with Y26862, and mesoderm from embryos injected with Rnd1 MO or Shirin MO. Red arrowheads in A and D indicate areas of large-scale retractions (compare 85 and 170 minutes). Scale bar: 100 μm. (E) Average time course curves with SD for the various experimental conditions. (E’) Corresponding relative spreading after 60 minutes and 170 minutes. https://doi.org/10.1371/journal.pbio.3001060.s010 | |

S8 Fig. (Related to Fig 9) Knockdown of Rnd1 or Shirin affect collective properties of mesoderm tissue explants. (continued) Analysis of spreading, dispersion, and intercalation of mesodermal explants under various conditions was performed as for experiments presented in Fig 9. (F–K) Delaunay triangulation of nuclei in order to measure cell dispersion. (F–I) Representative plots of triangulated nuclei after 170 minutes of imaging. X and Y labels mark the coordinates in μm, and the color-coded scale bar indicates the area of the triangles in μm2. (J) Quantification of average triangle size at the initiation of spreading (30 minutes) and the end of the time lapse. (K) Quantification of the relative change in triangle size over time calculated by dividing the average triangle area at 170 minutes by that at 30 minutes. (L) Quantification of intercalation calculated by dividing number of nuclei at the ventral surface at 170 minutes by the number at 30 minutes. Statistical comparisons: 1-way ANOVA followed by Tukey HSD post hoc test. Refer to S1 Data. ANOVA, analysis of variance; HSD, honestly significant difference; SD, standard deviation. https://doi.org/10.1371/journal.pbio.3001060.s010 |