PCP and Septins govern the polarized organization of the actin cytoskeleton during convergent extension

Summary

Convergent extension requires the coordinated action of the PCP proteins ^1,2 and the actin cytoskeleton ^3-6, but this relationship remain incompletely understood. For example, PCP signaling orient actomyosin contractions, yet actomyosin is also required for the polarized localization of PCP proteins ^7,8. Moreover, the actin-regulating Septins play key roles in actin organization ⁹ and are implicated in PCP and CE in frogs, mice, and fish ^5,6,10-12, but execute only a subset of PCP-dependent cell behaviors. Septin loss recapitulates the severe tissue-level CE defects seen after core PCP disruption, yet leaves overt cell polarity intact ⁵. Those results highlight the general fact that cell movement requires coordinated action by distinct but integrated actin populations actin, such as lamella and lamellipodia in migrating cells ¹³ or medial and junctional actin populations in cells engaged in apical constriction ^14,15. In the context of Xenopus mesoderm CE, three such actin populations are important (Figure S1), a superficial meshwork known as the “node-and-cable” system ^4,16-18, a contractile network at deep cell-cell junctions ^6,19, and mediolaterally oriented actin-rich protrusions which are present both superficially and deep ^4,19-21. Here, we exploited the amenability of the uniquely “two-dimensional” node and cable system to probe the relationship between PCP proteins, Septins, and the polarization of this actin network. We find that The PCP proteins Vangl2 and Prickle2 and septins co-localize at nodes, and that the node and cable system displays a cryptic, PCP- and septin-dependent anteroposterior polarity in its organization and dynamics

Blurb:

Devitt et al. explore the interplay of PCP proteins, actin, and septins during convergent extension, revealing new complexity to the polarization of the actin cytoskeleton during gastrulation.

Results and Discussion:

PCP protein and Septin localization in the actin node and cable network during Xenopus convergent extension.

We first examined localization of the core PCP proteins Vangl2 and Prickle2 in the node and cable system. During CE in zebrafish, both proteins were shown to localize to polarized foci, but the nature of these foci is unknown ^22-24. We found that fluorescent fusions to either Vangl2 or Prickle2 in Xenopus were also present in foci, and moreover, that these foci were coincident with 1-2 actin rich nodes (Figure 1A, B). PCP foci (and thus actin-rich nodes, see Figure 2, below) were not initially polarized, but became progressively biased to the anterior over time (Figure 1D-E, H), suggesting that the PCP foci in Xenopus mesenchymal cells follow the pattern of progressive polarization observed for junction accumulation of PCP proteins in epithelial cells ^2,25. The localization of these foci was confirmed by immunostaining for endogenous Vangl2 (Figure S2). By contrast, a fusion to the core PCP protein Dishevelled was localized to posteriorly biased foci that also developed progressively (Figure 1C, G, F, H), similar to that reported in zebrafish ²², but these were not co-localized with actin-rich nodes.

Figure 1: PCP protein and septin localization in actin rich nodes.

A-C. Still images of FP fusions to indicated PCP proteins (green) and Lifeact (magenta) in polarized cells, insets show magnified view of nodes. D-F. Still images of indicated PCP proteins (green) and Lifeact (magenta) in pre-polarized cells, insets show magnified view of nodes. G, H. Quantification of the position of FP-labeled PCP foci in polarized and pre-polarized cells. I-L. Images of indicated Septin subunit s(green) and Lifeact (Magenta) at nodes; Scalebar = 10 um. M. Schematic interpretation of data in A-L. N, O. Vangl2-GFP localization in control and Sept7MO cells. P. Quantification of AP positioning of Vangl2 puncta. N values and statistics are provided in Supplemental Information. See also Figures S1-S3.

Figure 2: Progressive polarization of the actin node and cable system requires PCP and Septins.

A-D. Individual cells mosaically labeled with Lifeact-RFP at times indicated at left (maximum intensity projections of cortical actin 10um into cell). a’-d’. Cyan box and inlay highlight individual nodes. Heatmap showing orientation of actin fibers quantified with OrientationJ (Legend in a’; cyan = mediolateral, red = anteroposterior). a”-d” Quantification of actin filament orientation. E. Quantification of cell length-to-width ratios over time. F. Upper panel, schematic of AP localization, arrow representing cellular anterior-posterior axis. Lower panel, quantification of Lifeact-RFP labelled nodes over time. G-K. Representative images of cells with indicated manipulations. (Maximum intensity projection of cortical actin 10um into cell). Inlay shows actin node and cable, where present, indicated by magenta box). g’, g”. Schematics of quantification. h’-k’: Quantification of actin fiber alignment in cells. h”-k” Quantification of actin cable positioning. Scale= 10um. N values and statistics are provided in Supplemental Information.

We next assessed the localization of Septins, which are known to assemble into hetero-oligomeric complexes comprised of one each of four Septin subgroups (see Figure S3A, B)²⁶. We first examined Sept2 and Sept7 subunits, which are essential for in CE ^5,6. Both were enriched in anterior actin-rich nodes (Figure 1I,J ; Figure S3D,E). To represent the two remaining Septin subgroups (Figure S3A,B), we chose Sept8 and Sept9 based on their strong expression in the dorsal mesoderm during CE ²⁷, and these, too, were strongly enriched at nodes (Figure 1K, L; Figure S3F,G). The localization of Septins reflected the diverse morphology of nodes, but as a population, their localization was highly consistent, as shown by overlapping peaks of Septin and actin intensity plots (Figure S3d”-g”; Figure 1M).

In light of their colocalization, we next probed the functional interrelationship between PCP proteins and Sept7. Interestingly, we found that actin rich, Vangl2+ nodes still form in the absence of Sept7, though they are mispositioned, and not as strongly biased to anterior regions (Figure 1N-P). Conversely, KD of Vangl2 elicited a robust and generalized loss of detectable of Sept7-GFP in cells (Figure S2D-F), which could reflect either disrupted localization/diffusion of Sept7 or a generalized loss of the protein. These interdependent phenotypes, with Vangl2 required for Sept7 localization and Sept7 also required for anterior positioning of Vangl2+ nodes, reflect the complex reciprocal relationship shown previously for PCP proteins and actomyosin during CE ^7,8.

A cryptic anteroposterior polarity in the node and cable system requires PCP proteins and Sept7.

These data raise an interesting question, because while the orientation of actin fibers within the node and cable system and their myosin-driven contraction are polarized mediolaterally ^4,16-18, no aspect of the system is yet known to be polarized in the A/P axis marked by PCP protein localization. These finding prompted us to further explore polarity within the node and cable system.

First, time-lapse analysis revealed that actin fibers were initially random, but became progressively aligned and oriented over time as cells progressively polarize and elongate ^20,28 (Figure 2A-E). Interestingly, however, this progressive mediolateral polarization of fibers was coincident with the progressive polarization of actin-rich nodes to the anterior region of cells (Figure 2F), reflecting the polarization of PCP proteins described above.

PCP disruption is known to completely abolish all aspects of planar polarity during CE, and accordingly, knockdown (KD) of Vangl2 or Prickle 2 (Pk2) or expression of dominant negative Disheveled (Xdd1), severely disrupted both mediolateral alignment of actin fibers (Figure 2G, g’ - J, j’) and the anterior positioning of actin cables (Figure 2g” - j”). By contrast, Sept7 loss strongly disrupts CE at the tissue level but leaves overt cell polarity intact ⁵, and consistent with that effect, loss of Sept 7 did not affect the mediolaterally polarized alignment of individual actin filaments (Figure 2K, k’) but did significantly disrupt the anterior enrichment of actin cables (Figure 2k”). This result was paradoxical. On the one hand, the lack of effect on actin fiber alignment was consistent with the normal overall cellular polarity in Sept7 KD cells ⁵. On the other, the effect on anterior actin cable positioning suggests the presence of a cryptic Sept7-dependent anteroposterior-oriented organization of the actin system in these cells.

This possibility prompted us to develop another metric of actin organization that was more granular than cable formation yet independent of actin fiber alignment. We therefore segmented Lifeact-GFP labeled actin fibers from our TIRF movies using SOAX software ²⁹ and used customized scripts to quantify their persistence length (L_p)(Figure 3A). This metric reflects the bending stiffness of polymers in isolation, but in the complex environment of the cell, we reasoned that it may serves instead as a more general proxy for overall actin organization. Indeed, we found that actin fibers in the anterior region of cells displayed significantly higher L_p compared to those in the posterior (Figure 3B-D).

Figure 3: PCP- and Septin-dependent anteroposterior pattern in the node and cable system.

A. Representative TIRF image of a control cell, outlined in pink. Boxes highlighting anterior (cyan) and posterior (yellow) regions of the cell. B, C. Representative anterior and posterior cropped ROIs from A. b’-c’. tSOAX segmentation. b”-c”. representative single actin filament traces used for quantification. D. Quantification of persistence length (Lp) of actin filaments in anterior (A) or posterior (P) cellular regions of control, Sept7MO, and Vangl2MO cells. E, F. Lifeact-RFP labeling the actin cortex in control and Sept7 MO cells. Boxes indicate regions used for kymograph, below. e’, f’. Kymograph of representative actin fiber coalescence over time. G. Quantification of actin filament coalescence over time, 0 sec indicates the full coalescence to a single fiber. N values and statistics are provided in Supplemental Information. See also Figure S4.

Strikingly, the anteroposterior-polarized asymmetry in L_p of actin fibers was eliminated by loss of either Vangl2 or Sept7 (Figure 3D; Figure S4), an effect resulting specifically from reduction in actin persistence lengths in anterior regions (Figure 3D). Together, these findings reveal that the actin node and cable system displays two vectors of polarity: individual actin fibers align mediolaterally in a Vangl2-dependent manner, while higher level actin organization quantified either by cable positioning or by L_p of actin fibers is patterned anteroposteriorly in a manner that requires both Vangl2 and Sept7.

Sept7 is required for progressive coalescence of the anterior actin cables

The data so far indicate that Vangl2 acts via Sept7 to control anteroposterior actin organization in the node and cable system. To avoid the additional pleiotropic effects of core PCP loss, we focused a deeper exploration on Sept7 KD, starting with a time-lapse analysis of the evolution of the Sept7 KD phenotype over time (Figure 3E, F). In movies, we observed that while individual, smaller actin cables form in the absence of Sept7, they fail to coalesce into a single, anteriorly positioned cable. Using kymographs, we quantified this phenotype, revealing that after Sept7 KD, coalescence was severely disrupted, with actin filaments remaining well-spaced on average, even at later time points (Figure 3e’, f’, G).

Sept7 controls local dynamics of polarized actin flow

Finally, we probed the effect of Sept7 loss on actin dynamics using fluorescent speckle microscopy ³⁰, in which mosaically labeled actin monomers trace both the bulk flow of actin and the turnover of individual actin filaments ^31,32. Specifically, we imaged the overall actin network with Lifeact-RFP and identified nodes with Prickle2-GFP, as shown in Figure 4A, with an enlarged region and split channels shown in Figure 4B, C. In the same cells, we also imaged single actin speckles using dye-labelled actin monomers (Figure 4A, D).

Figure 4: Sept7 controls local dynamics of polarized actin flows.

A. TIRF image of a single cell labeled with Lifeact-RFP, Pk2-GFP, and actin-647 speckles. B-D. Zoomed, split channels of the boxed region in A. Orange and purple boxes indicate regions quantified in Panel I, below). E. Heatmap and vector diagram of PIV data for the region shown in A. Arrows indicate the vector and direction of flow; legend shows heatmap indicates positive divergence of flow (blue) or negative divergent (i.e. convergence, yellow). F. Zoomed view of the boxed region of the heatmap in E. G. Line plot of Pk2-GFP pixel intensities in multiple cells. H. Line plot showing vector divergence along the same line from which Pk2 intensity was derived, above. Note maximal vector convergence is coincident with the peak of Pk2 intensity (i.e., in nodes). I. Quantification of actin speckle lifetimes in regions indicated by boxes in panels B and D, above. N values and statistics are provided in Supplemental Information.

We then used Particle Imaging Velocimetry (PIV)³³ analysis to understand the bulk flow of actin. Tracing actin speckles revealed a consistent pattern in which actin flows medially along anterior actin cables into Pk2+ nodes (Figure 4E, F; arrows indicate the vector and magnitude of particle flow). To quantify flow, we measured the divergence of speckle trajectories over time which we represented with a heat map where yellow represents vector divergence, and blue represents negative divergence (i.e convergence of two vectors). We consistently observed a dark blue region at the anterior of cells (Figure 4E, F), correlating with the Pk2+ node, indicating that actin speckles converge at the node. To quantify this effect across samples, we overlaid line plots of Pk2-GFP (Figure 4G) intensity with line plots of vector divergence (Figure 4H), and these bulk data reveal a peak of Pk2-GFP intensity at nodes coincident with the inverted peak of negative vector divergence (Figure 4G, H).

Noting that actin frequently flows to regions of enhanced depolymerization and that actin depolymerizing factors are required for normal CE ^3,4,34, we examined several such factors. We found strong enrichment of Cfl1, Twf1, and Cap in actin nodes (Figure S3H-J), suggesting that the node is a region of high actin turnover and that PCP proteins may organize actin flows through the recruitment of such factors.

Finally, we quantified actin filament turnover by tracking individual actin speckles over time, independently measuring lifetimes in nodes and in nearby cables (Figure 4B, D; orange and purple boxes, respectively). We found that the lifetime of actin speckles in nodes was significantly shorter than the lifetimes in the cable (Figure 4I, left), suggesting faster turnover at nodes and increased actin filament stability in anterior cables. Because Sept7 is required for cable formation anteriorly (Figure 3E-G), we asked how loss of Sept7 impacts dynamics of actin speckles in nodes and cables, and we observed a very specific effect: Sept7 KD significantly reduced the longer actin speckle lifetimes observed in cables (Figure 4I, right), but did not alter the shorter lifetime of actin speckles in nodes (Figure 4I). These data suggest a role for Sept7 in stabilizing actin in the cable region.

Conclusions:

Here, we exploited the actin node and cable system of Xenopus mesoderm cells engaged in CE for an in-depth analysis of the relationship between PCP proteins, Septins, and actin, providing two important insights. First, while the superficial actin node and cable system is essential for CE in Xenopus ¹⁶,¹⁷, and its dynamics are controlled by PCP ¹⁸, how the overt planar polarization of the system relates to PCP proteins has not been obvious. For example, both myosin-driven contraction of the system ^16-18 and the orientation of actin fibers within it ⁴ are polarized mediolaterally, yet PCP proteins are well-known to mark the anteroposterior axis of cells engaged in CE ^22-24,³⁵,³⁶. Our findings here are important then for revealing a previously cryptic anteroposterior-directed planar polarity of the actin node and cable system, displayed both by the anterior coalescence of actin cables (Figure 2) and by the longer persistence length of actin fibers anteriorly (Figure 3). Crucially, PCP proteins and Sept7 are essential for both.

Second, when considered as a whole, our data suggest a preliminary model for the emergence of both mediolateral and anteroposterior polarity in the node and cable system. PCP proteins (Figure 2) and Twf1⁴ are each essential for mediolateral alignment of actin fibers, arguing they act in concert and independently of Sept7 to govern this aspect of polarity. At the same time, PCP proteins are anteriorly biased (Figure 1), where they help to establish the higher-level, anteroposteriorly polarized actin organization described above. Specifically, the order observed anteriorly after loss PCP or Septins is similar to that normally observed posteriorly (Figure 3). In addition, speckle microscopy revealed that Sept7 is required to stabilize actin filaments in anterior cables. Altogether, these data suggest that PCP may recruit Sept7, where it imparts stability to actin and increases overall order, which in turn facilitates cable formation specifically in anterior regions of cells.

Finally, these data on polarization of an actin network may shed light on our emerging understanding that very local, sub-cellular mechanical heterogeneities play key roles in cell behavior ^37-39. Both experiments and theory now argue that such heterogeneities are essential for normal CE ^40-42. It is tempting to speculate that the highly local patterning of the actin cytoskeleton described here may provide a basis for local patterning of cell mechanics.

STAR Methods

Lead contact

Further information and reagent requests should be directed to John B. Wallingford (Wallingford@austin.utexas.edu).

Material availability

All reagents available from lead contact with completed MTA upon request.

Data and code availability

All data reported in this paper will be shared by the lead contact upon request.

This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and subject details

Xenopus embryo manipulations:

Xenopus laevis females were super-ovulated by injection of hCG (human chorionic gonadotropin). The following day, eggs were squeezed from the females. In vitro fertilization was performed by homogenizing a small part of a harvested testis in 1X Marc’s Modified Ringer’s (MMR) and mixing with collected eggs. Embryos were dejellied in 3% cysteine (pH 7.9) at the two-cell stage. Fertilized embryos were rinsed and reared in 1/3X Marc’s Modified Ringer’s (MMR) solution. For microinjections, embryos were placed in a 2% ficoll in 1/3X MMR solution and injected using forceps and an Oxford universal micromanipulator. Morpholino oligonucleotide or CRISPR/Cas9 was injected into two dorsal blastomeres to target the dorsal marginal zone (DMZ). Embryos were microinjected with mRNA at the 4-cell stage for uniform labeling and at the 64- to 128- cell stage for mosaic labeling. Embryos were staged according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1994). All animal work has been approved by the IACUC of UT, Austin, protocol no. AUP-2021-00167.

Method details

Plasmids, mRNA, protein, and MOs for microinjections:

Xenopus gene sequences were obtained from Xenbase (www.xenbase.org) and open reading frames (ORF) of genes were amplified from the Xenopus cDNA library by polymerase chain reaction (PCR), and then are inserted into a pCS10R vector fused with eGFP. The following constructs were cloned into pCS vector: Sept8-GFP, Sept9-GFP. These constructs were linearized and the capped mRNAs were synthesized using mMESSAGE mMACHINE SP6 transcription kit (ThermoFisher Scientific, Waltham, MA, USA). Concentration for GFP localization was titrated to the lowest concentration where we could still detect GFP signal, but no cellular phenotype was observed. The amount of injected mRNAs per blastomere are as follows: Lifeact-GFP or −RFP [50-100pg ⁴³], Pk2-GFP [35pg] ⁴⁴, Vangl2-GFP [35pg] ⁴⁴, Dvl2-GFP [5-8pg] ⁴⁵, Cap1-GFP[25pg] ⁴, Septin2-GFP [50pg] ^5,6, Septin7-GFP [35pg] ^5,6, Septin8-GFP [40pg], Septin9GFP [40pg] and Xdd1 [500pg] ⁴⁵.

Morpholinos:

Morpholinos used were previously designed to target exon-intron splicing junction (Gene Tools, Philomath, OR, USA). The MO sequences and the working concentrations include: Septin7 morpholino^5,6, 5’ TGCTGTAGAGTCAGTGCCTCGCCTT 3’, 40 ng per injection at 4-cell stage; Vangl2 morpholino⁴⁶, 5’ CGTTGGCGGATTTGGGTCCCCCCGA 3’, 20ng per injection at 4-cell stage; Pk2 morpholino⁴⁴, 5’ GAACCCAAACAAACACTTACCTGTT 3’, 20ng per injection at 4-cell stage.

Actin Speckle Microscopy:

For fluorescent actin speckle microscopy, labeled actin monomers (Alexa 647 conjugated Actin, Catalog number A12374, ThermoFisher, Waltham, MA, USA; initial experiments were performed with Alexa 568, yielding the same results) were reconstituted in a 2mg/ml stock solution. A 0.01ug/ul working solution was prepared for each 10nl injection, and 0.1 ng was injected into dorsal blastomeres of 4-cell stage embryos. This concentration was determined empirically by titrating the amount injected until no cellular phenotype was observed and labeled actin monomers were sparse enough to detect individual “speckles” ^31,32.

Live imaging of Keller explants:

mRNAs were injected at the dorsal side of 4 cell stage embryos, and the dorsal marginal zone (DMZ) tissues were dissected out at stage 10.5 using forceps, hairloops, and eyebrow hair knives. Each explant was mounted to a fibronectin-coated dish in Steinberg’s solution, and cultured at 16°C for half a day before imaging with a Zeiss LSM700 confocal microscope (Jena, Germany), or Nikon N-STORM combined TIRF/STORM microscope (Minato City, Tokyo, Japan).

Inhibitor treatments:

Inhibitors were solubilized in DMSO to a stock concentration of 10mM for Latrunculin B and 100mM for (+) Blebbistatin. Explants with labeled actin and Vangl2 were treated with 1uM LatrunculinB (diluted 1:10,000 in Steinberg’s solution)(CAS Number: 76343-94-7), a cell permeable actin polymerization inhibitor, or 100uM (+) Blebbistatin (diluted 1:1,000 in Steinberg’s solution)(CAS Number 856925-75-2), a non-muscle myosin inhibitor. Samples were treated with DMSO (vehicle control) or specific inhibitors and imaged immediately after application.

Image analysis:

Images were processed with the Fiji distribution of ImageJ and Photoshop (Adobe) software suites, and figures were assembled in Illustrator (Adobe) and Affinity Designer (***).

Cell length and width measurements were manually taken in Fiji along the long axis and through the centroid of the cell.

AP positioning of node and PCP proteins measurements were taken by manually drawing a vertical line though the AP axis of a cell through the node and cable region. Intensity plots were measured using the multiplot feature in Fiji. The pixels of highest fluorescent intensity were identified for proteins of interest and the corresponding position within the cell was noted. To compare across many samples, this position was scaled between 0 and 1 with 0 representing the most anterior position within a cell and 1 representing the most posterior position.

Similarly, Anterior cable intensity measurements were taken by drawing a vertical line through the midline of the cell with a line width matching the average width of a single cell (150 pixels). Using the multiplot feature in Fiji, the average actin intensity averaged across the anterior posterior axis of a cell was quantified and averaged over several samples.

Actin cable orientations were taken from TIRF images of single cell cortical actin network using the OrientationJ plugin in Fiji (http://bigwww.epfl.ch/demo/orientation/).

Actin cable bundling measurements were taken using kymographs and manual measurements between two points in Fiji. Kymographs were taken across the anteroposterior axis of individual cells near the actin node. We then measured the distance between adjacent actin filaments over time over specified time intervals.

Divergence measurements of actin speckles were taken using the PIVlab software in Matlab. Specific xy coordinates within the cell were identified to link divergence measurements obtained in PIVlab/Matlab with intensity values demarcating the node taken with Fiji.

Fluorescent actin speckle lifetime measurements were taken by tracking individual speckles using MosaicSuite (http://mosaic.mpi-cbg.de) object tracking plugin in Fiji. Custom scripts were written in Matlab to count the lifetime of each speckle. We defined the node by localization of anterior PCP proteins (Vangl2-GFP or PK2 GPF) and actin (Lifeact-RFP). We defined the cable region by actin fiber morphology and localization directly adjacent to the node. Each anterior and posterior actin speckle ROI is 100px x 100px.

Persistence length measurements of actin filaments:

The open-source biopolymer tracking software TSOAX v0.1.8 ²⁹ (https://www.lehigh.edu/~div206/tsoax/) was used to track actin fibers from TIRF movies automatically. The ridge threshold, stretch-factor, and gaussian std (pixels) parameters were set to 0.01, 1.5, and 1 respectively. These values were chosen as they resulted in accurate filament labeling. The remaining parameters were unchanged from default values. TIRF movies were processed prior to TSOAX analysis with the application of a gaussian blur of sigma radius = 1 to enhance the accuracy of fiber tracking. TSOAX output containing the spatial and temporal coordinates that define labeled fibers and track groupings was further analyzed to quantify the persistent length of each fiber using customized Python and MATLAB scripts. Briefly, we treated each actin filament as a non-equilibrated semi-flexible polymer. Its persistent length $P$ can be obtained via an equation derived from the worm-like chain model (WLC)⁴⁷:

$$\langle R^2\rangle =2sP\ell \left(1-\frac{sP}{\ell }\left(1-e^{\frac{\ell }{sP}}\right)\right)$$

where $\ell$ is the length of an arbitrary segment of the filament, $R$ is the end-to-end distance of the segment, and $s$ is set to 1.5 for non-equilibrated fibers. The measure of the mean square of $R$ is a function of $\ell$ for each fiber, and the weighted least squares is used to measure.

Quantification and statistical analysis:

Statistical analyses were carried out using Microsoft Excel, Matlab (Natick, MA, USA), and GraphPad Prism (Graphpad, San Diego, CA, USA) software. Each experiment was conducted on multiple days and included biological replicates. Due to non-normality observed and low n-number in some observations, we used non-parametric tests for statistical comparisons throughout the manuscript. We used Mann Whitley U tests, Kruskal-Wallis tests, or Kolmogorov–Smirnov tests for statistical comparisons of two groups, multiple groups, or distributions, respectively. MEAN and SD are plotted along with individual data points throughout the manuscript.

N values and Statistics for each figure are as follows:

Figure 1:

Panel G: Polarized: Kruskal-Wallis test; Vangl2 vs Pk2 p>0.999, Vangl2 vs Dvl2 p<0.0001, Pk2 vs Dvl2 p<0.0001; Vangl2 GFP n= 74 cells from 10 embryos; Pk2GFP n= 45 cells from 3 embryos; Dvl2-GFP n= 75 cells from 9 embryos)

Panel H: Pre-polarized: Kruskal-Wallis, p>0.999 n.s., Vangl2 GFP n= 31 cells from 7 embryos; Pk2GFP n= 42 cells from 9 embryos; Dvl2-GFP n= 84 cells from 3 embryos).

Panel P: Mann-Whitney test, control vs Sept7 MO, p<0.0001****, n=34 control cells from 3 embryos; n= 60 Sept7MO cells from 6 embryos).

Figure 2:

Panels A-F: Hr1 n= 103 cells, Hr2 n=117 cells, Hr3 n=98 cells, Hr4 n=73 cells).

Panels G-K: Vangl 2 MO: Kolmogorov-Smirnov test, actin filament orientation p<0.0001; n= 38 control cells from 12 embryos; n=46 morphant cells from 13 embryos; Pk2MO, KS test, actin filament orientation p<0.0001; n= 23 control cells from 4 embryos; n= 23 morphant cells from 4 embryos). D) dominant negative (DN) disheveled (Xdd1) (control vs Xdd1, Kolmogorov-Smirnov test, actin filament orientation p<0.0001; n=23 control cells from 13 embryos; n= 38 mutant cells from 8 embryos) and E) Sept7 morphant cell, inlay of actin node and cable, magenta box highlighting region of inlay (control vs Sept7MO, Kolmogorov-Smirnov test, actin filament orientation p=0.5596; n= 26 control cells from 10 embryos; n= 31 morphant cells from 10 embryos).

Figure 3.

Panel D. Kruskal-Wallis test, (Control A vs P p=0.0046**, Sept7MO A vs P p>0.999 n.s.,

Vangl2 MO A vs P p>0.999 n.s., Control A vs Sept7MO A p=0.0481*, Control A vs Sept7MO P p= 0.0037**, all other pairwise comparisons p>0.999 n.s.)

Panel G. (control vs Sept7 MO, n=16 control cells from 6 embryos; n= 19 Vangl2MO cells from 6 embryos). Scale= 10um.

Figure 4.

Panel G, H: n=7 control cells from 6 embryos.

Panel I: (Kruskal-Wallis test, control node vs control cable p=0.0093, control node vs Sept7MO node ns p>0.9999, control cable vs Sept7MO cable p<0.0001, Sept7MO node vs Sept7MO cable p>0.999, control cable vs Sept7MO nodes p=0.0033, nodes vs Sept7MO cable n.s. p=0.5498; n= 31 control node, n=29 control cable from 12 embryos; n= 13 Sept7MO node, n= 19 Sept7MO cable from 7 embryos).

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Experimental subjects
Xenopus laevis (male and female)	Xenopus1	https://xenopus1.com/
Antibodies
Rat anti-Vangl2 antibody (GST-tagged recombinant Vangl2 N-term fragment)	Millipore-Sigma	MABN750
Antibodies
Rat anti-Vangl2 antibody (GST-tagged recombinant Vangl2 N-term fragment)	Millipore-Sigma	MABN750
Chemicals, peptides, and recombinant proteins
hCG (human chorionic gonadotropin)	Merck	Code# 133754
Alexa 647 conjugated Actin	ThermoFisher	Cat# A12374
Latrunculin B	AbCam	Cat# Ab144291
Blebbistatin	Millipore Sigma	Cat# 203390
Critical commercial assays
mMessage mMachine SP6 transcription kit	ThermoFisher	Cat# AM1340
Oligonucleotides
Vangl2 morpholino	Park & Moon46	Gene-tools.com
Pk2 morpholino	Butler & Wallingford44	Gene-tools.com
Septin7 morpholino	Kim et al.5	Gene-tools.com
Recombinant DNA
Lifeact-RFP	Riedl, et al. 43	N/A
Vangl2-GFP	Butler & Wallingford44	N/A
Pk2-GFP	Butler & Wallingford44	N/A
Dvl2-GFP	Wallingford, et al.45	N/A
Septin7-GFP	Shindo & Wallingford5	N/A
Septin2-GFP	Shindo & Wallingford5	N/A
Sept8-GFP	Cloned for this paper	N/A
Sept9-GFP	Cloned for this paper	N/A
Twf1-GFP	Devitt, et al.4	N/A
Cfl2-GFP	Devitt, et al.4	N/A
CAP1-GFP	Devitt, et al.4	N/A
DN Dishevelled (Xdd1)	Wallingford, et al.45	N/A
Software and algorithms
Prism	GraphPad Software	Graphpad.com
Excel	Microsoft	Microsoft.com
Matlab	MathWorks	Mathworks.com
Python	Python	Python.org
Orientation J	Püspöki, et al.	http://bigwww.epfl.ch/demo/orientation/
MosaicSuite	Sbalzarini & Koumoutsakos	http://mosaic.mpi-cbg.de
TSOAX v0.1.8	Xu, et al.	https://www.lehigh.edu/~div206/tsoax/

Highlights:

PCP proteins localize to the actin node and cable system during Xenopus gastrulation

The node and cable system displays a cryptic anteroposterior polarity

Anteroposterior polarity of the node and cable system is PCP and Septin-dependent

Sept7 controls local dynamics of polarized actin flow