Local Wnt3 globally orients the body axis tissue polarity and morphogenesis

(A) A key unanswered question of animal evolution is the origin of body axis patterning mechanisms, allowing initial multicellular assemblages to position specialised cell types (red) and coordinate cell polarity (represented in this diagram by oriented cilla and elongated shape) (B) Key stages of the early morphogenesis of Clytia embryos in relation to Wnt3 mRNA localisation (red). At the early gastrula (EG) stage, ectodermal cilia have formed, PCP has developed along the future oral-aboral axis, and the pointed morphology of the animal pole morphology can first be distinguished. Most axial elongation takes place during gastrulation. See also Supplementary Figure 1. (C) Schematic representation of planar cell polarity in a single epidermal cell at the EG stage. Basal bodies are positioned on the oral side at the apical cortex (translational polarity). (D) Confocal image of the apical surface of a single epidermal cell stained for actin (cyan, phalloidin labelling) and γ-tubulin immunostaining (magenta), and a representation of the PCP vector, defined by the basal body position and its mirror image across the apical surface centroid as the initial and terminal points, respectively. (E) Local PCP coordination in Control-MO and Wnt3-MO-injected embryos at the EG stage. bar= 10 µm. Top) Confocal images as in D. Bottom) PCP vector representation of the cells in the images. Dots represent cells with no clear polarisation or mitotic cells. Apical cell contours are wavy in Wnt3-MO embryos. (F) Distribution of cell orientation in individual scans (lines) and their average (histogram) for non-injected (373 cells, 7 images), control-MO injected (304 cells from 3 images) and Wnt3-MO-injected (225 cells from 6 images). The average orientation is defined as the origin (0°) for each scan. Horizontal bar: standard deviation. G) The PCP polarisation index (degree of the polarisation of cells) by the length of the PCP vector, normalised by the cell size. See supplementary method for the definition (**** p< 0.0001, Mann–Whitney U test). (H) Experimental procedures of the Wnt3-rescue experiment. Wnt3MO is injected into unfertilised eggs, then rescue mRNA solution with fluorescently labelled 10 kDa dextran is injected into a 4-cell stage blastomere. (I) Circular colour-code representation of the PCP. (J) Wide range PCP coordination observed in Wnt3-rescue experiments. Left: Mock rescue control by water injection (N=9). Right: Rescue by Wnt3 mRNA injection (N=27). Each set consists of three illustrations. top: colour-wheel representation of the PCP. Cells derived from the Wnt3 or water-injected blastomere are at the top-right of each image, indicated with grey shading. bar= 50 µm. bottom-left: confocal thumbnail image. cyan: actin; yellow: Dextran-Alexa647(injected blastomere derived cells). bottom-right: radar plot of the cell orientation. The arrows in the radar show the circular mean of the PCP orientation within the single scan. circ s.d.: circular standard deviation. (K) Axial morphology of normal, Wnt3-MO injected and Wnt3 rescued embryos at late gastrula (LG) stage. bar=50 µm. Elongated body axis morphology was restored so that cells with injected Wnt3 mRNA were located at the induced oral (posterior). See also Fig. 3C.

Wnt3-driven PCP orientation precedes visible cell alignment.

(A, B) PCP orientation in Wnt3 rescue embryo at the mid-blastula stage (A: 8 hpf, N=10) and late-blastula stage (B: 9 hpf, B, N=10). Graphical PCP representations and radar histogram plots are as in Fig. 1. (C, D) Radar plot of PCP in distal (C) and proximal (D) areas of the region shown in (B). (E) Basal body displacements after cell divisions. Bar=10 µm. Cell membranes labelled by PH-Venus (green) and basal bodies with Poc1-mCherry (magenta). Arrows indicate basal bodies positioned on opposite sides of cleavage planes located after being engaged to the spindles. Aborally displaced basal bodies migrate toward the oral side. The average PCP defined the oral-aboral axis. (F) Distribution of the angles between the cleavage orientation axis to the OA axis defined by the PCP orientation (0-90°, n=131). Cleavages along the OA axis (0-45°) are favoured (* p<0.05 chi-square test). (G) Directional migration of the basal body toward the oral cell boundary after the cell cleavage. Each line indicates the distance of a single basal body to the oral edge of the cell after the last cleavage. Basal bodies migrate toward the oral cell boundary, suggesting PCP has been established by the time PCP becomes structurally manifest. (H) Subset of (G) where the initial distance to the oral edge is less than four micrometres after the cell division.

PCP orientation by Wnt3 is Wnt/β-catenin-independent, and its propagation across the body axis requires Stbm and Fmi.

(A) Wnt3-rescue with an additional injection of a dominant-negative form of TCF (dnTCF) into the egg, N=10. (B) Wnt3-rescue by the constitutively active form of β- catenin (CA-β-cat) instead of Wnt3 mRNA, N=7. (C) Onset of axial morphology in EG embryos. Localised Wnt3 is necessary and sufficient for the onset of the elongation toward its source. Green: actin (phalloidin), magenta: nuclei (To-Pro3) and blue: Wnt3-mRNA/Dextran injected cell lineage. (D) Roles of Wnt3 and β- catenin in the axial elongation in LG embryos. The elongation was measured by the elongation index (L/W): the primary axis length (L) divided by its perpendicular length (W), **** p< 0.0001, Mann–Whitney U test). Both local Wnt3 expression and β-catenin/TCF-dependent gastrulation were required for axial elongation. See supplementary Figure 3. (E, F) Wnt3-rescue experiments with additional injections of (E) Stbm-MO (N=23) and (F) Fmi-MO (N=6). Radar plots of the PCP orientation show the manually defined proximal area (Proximal: between the dotted line and the Wnt3-positive region) and distal area (Distal: opposite area to the Wnt3-positive region across the dotted line). PCP is correctly polarised without Stbm or Fmi close to the Wnt3-injected cell clone but not in the distant area.

PCP behaviour in Stbm/Fmi-mosaic/ Wnt3-rescue experiments supports the two-step model for global PCP orientation

(A-D) Wnt3-rescued early gastrula embryos, including Stbm(–) or Fmi(–) mosaic patches made by blastomere transplantation. (A, B) blastomeres from Wnt3-MO/Stbm-MO double-injected embryos were transplanted as the doner into Wnt3-rescued host embryos at the 64-cell stage. They happened to be incorporated in close/proximal (A: 658 cells, N=4) or far/distal (B: 757 cells, N=4) positions with respect to the Wnt3 source, respectively. (C, D) The same experiment was conducted with Wnt3-MO/Fmi-MO-injected blastomeres as the donor, incorporated in proximal (C: 984 cells, N=3) and distal (D: 649 cells, N=3) positions, respectively. (E) Schematic drawing of the experimental procedure of the transplantation. (F) Wnt3-rescued early gastrula embryos with Stbm(–) mosaic patch generated by additional injection of Stbm-MO at the 16-cell stage (N=1). (G) Schematic drawing of the experimental procedures for (F). The flow of the PCP orientation is indicated by grey arrows in the bottom right drawings in (A-D) and right in (G). (H) Graphical summary of Stbm-MO and Fmi-MO mosaic experiments. The graphical representation and thumbnail confocal images indicate Wnt3-positive lineages and Stbm-MO or Fmi-MO-injected lineages in yellow and blue. All vector representations are as in Figure 1.

Mechanical strains can reorient PCP.

(A-E) PCP orientation induced by transplantation from a Wnt3-MO donor into the Wnt3-MO background of a Wnt3-rescue host. A) Incomplete incorporation of host and donor cells into a smooth epidermis (N=3/8) causing long-range PCP orientation, as if the donor acts as an aboral cue, in addition to the Wnt3 oral cue. (A) PCP representation by colour-coded bars. See also the legends for Figure.1 (B) Confocal image of the apical cortex of the epidermis showing the rosette structure caused by incomplete incorporation of the donor cells. Both host and donor cells are elongated around the graft boundary. (C) Graphical summary of the PCP orientation: Wnt3-mRNA lineage in Wnt3-rescue host in yellow, Wnt3-MO graft in blue. (D) 3D reconstruction of the rosette structure from the confocal images by contouring cortical actin signals. (E) Schematic representation of the transplant experiment. (F) 3D representation of the late gastrula morphologies induced by mRNA injection of dominant negative form (RhoABC-DN) and constitutively active forms (RhoABC-CA2) of the PCP effector RhoABC, which caused reduced elongation and extra protrusions, respectively. The overall oral-aboral axis remained distinguishable based on morphology and gastrulation. (G-I) PCP is coordinated consistently with respect to the induced protrusions caused by RhoABC-CA2. (G) Colour-wheel representation of PCP in the RhoABC-CA2 injected early gastrula embryos; the presumed oral pole is to the top-right. (H) 3D reconstruction of the tissue morphology for the same specimen, with PCP flow represented by arrows. (I) Confocal image of the apical cortex of the epidermis at the induced protrusion, corresponding rectangles in G and H. Cortical actin fibres are circumferentially organised around the protrusion. Bars are 50 µm except for 10 µm in B and I. See Figure 1 for the circular colour-code representation of PCP orientation. cyan: actin; magenta: γ- tubulin; blue: donor lineage marker Dextran for B and I

Models for body axis symmetry-breaking during Clytia embryogenesis and PCP-driven axis innovation scenarios in metazoan evolution

(A) Two-step PCP axis orientation by local Wnt3. Local gradients of Wnt3 at the oral end of the Clytia embryo orient epidermal cell polarity near the Wnt3-positive area. This process does not require Wnt/β-catenin signalling. Local PCP orientation propagates to the aboral side through the action of core PCP proteins between neighbouring cells. Wnt3 expression is likely maintained by Wnt3/β-catenin feedback regulation. (B) Two possible metazoan body axis evolution scenarios from a choanoblastaea, a multicellular and non-polarised metazoan common ancestor. In the “Wnt/β-catenin-first” scenario (bottom), the choanoblastaea first acquires a mechanism to activate the Wnt/β-catenin pathway locally. Stably locally maintained Wnt ligand expression is then recruited by preexisting but non-directed core Fz/PCP interactions to orient PCP and morphogenesis. In the “PCP-first scenario” (top), the self-organising capacity of core PCP protein interactions creates tissue polarity without any specific cue to specify the direction. The advantage of coordinated polarity of cilia movement for filter-feeding, swimming or morphology could have favoured directed symmetry breaking using reliable cues. These could have initially been provided by environmental or mechanical constraints before the recruitment of localised Wnt binding to Fz to assure PCP orientation, and then the emergence of downstream beta-catenin signalling.