Distinct mechanisms of planar polarization by the core and Fat-Dachsous planar polarity pathways in the Drosophila wing

Abstract

Planar polarity describes the coordinated polarization of cells within a tissue plane, and in animals can be determined by the "core" or Fat-Dachsous pathways. Current models for planar polarity establishment involve two components: tissue-level "global" cues that determine the overall axis of polarity and cell-level feedback-mediated cellular polarity amplification. Here, we investigate the contributions of global cues versus cellular feedback amplification in the core and Fat-Dachsous pathways during Drosophila pupal wing development. We present evidence that these pathways generate planar polarity via distinct mechanisms. Core pathway function is consistent with strong feedback capable of self-organizing cell polarity, which can then be aligned with the tissue axis via weak or transient global cues. Conversely, generation of cell polarity by the Ft-Ds pathway depends on strong global cues in the form of graded patterns of gene expression, which can then be amplified by weak feedback mechanisms.

Keywords: CP: Developmental biology; Dachsous; Fat; Frizzled; PCP; planar cell polarity; planar polarity.

Graphical abstract

Figure 1

Organization of the core and Ft-Ds planar polarity pathways in the pupal wing (A′) The core proteins form asymmetric intercellular protein complexes at apicolateral cell junctions, consisting of a homodimer of the cadherin Flamingo (Fmi, red, also known as Starry Night) associated with the sevenpass transmembrane protein Frizzled (Fz, green) and the cytoplasmic proteins Dishevelled (Dsh, blue) and Diego (Dgo, magenta) on distal junctions, and the fourpass transmembrane protein Strabismus (Stbm, orange, also known as Van Gogh) and the cytoplasmic protein Prickle (Pk, cyan) on proximal junctions. (A″) Core protein complexes form local clusters of complexes of the same orientation (puncta) through the action of local feedback interactions. These might be positive interactions stabilizing complexes of the same orientation (green arrows) or negative interactions destabilizing complexes of opposite orientation (red symbols). (A‴) Core protein complexes are segregated to opposite cell edges, where they are concentrated in puncta (blue arrow, inset). This segregation may be promoted by non-local (cell-scale) inhibitory interactions (red). Trichomes (grey) emerge from distal cell edges due to core protein localization. (A⁗) In the wild-type pupal wing blade (dark grey), the overall direction of core polarity and trichome orientation is distal (black arrows). This is believed to result from weak global cue(s) (red arrow). Early in pupal wing development global cues specify a radial polarity pattern which from ∼16 h onward is re-oriented onto the proximodistal axis (Aigouy et al., 2010). (B′) Ft (teal) and Ds (purple) form trans-heterodimers between neighboring cells. (B″) Ft and Ds also cluster in cell junctions, which may result from feedback interactions (green and red symbols). (B‴) Ft-Ds complexes are localized to opposite cell edges, where they are concentrated in puncta (blue arrow, inset). In the mid-posterior half of the wild-type pupal wing, Ft is at medial (anterior) cell edges (upward in cartoon), whereas Ds is at lateral (posterior) cell edges. Trichomes emerge from distal cell edges independently of Ft-Ds localization. (B⁗) Ft-Ds polarity is radially oriented toward the wing margin (red arrows) with Ft medial (toward centre of wing) and Ds lateral (toward margin). This pattern thought to result from opposing gradients/boundaries of Ds (purple) and Fj (yellow) expression, with Fj high at the wing margin and Ds high medially.

Figure 2

Puncta are sites of aligned asymmetric complexes in both the core and Ft-Ds pathways (A) Diagram of adult wing indicating region analyzed below vein 4 (blue box). (B–D) Images of ds-EGFP, ft-EGFP, and fz-EGFP twin-clones next to tissue with untagged protein, revealing asymmetric cellular localizations on clone boundaries. Arrows point to high EGFP levels: Ds-EGFP on posterior junctions (B), Ft-EGFP anterior (C), Fz-EGFP distal (D). Arrowheads point to low EGFP junctions. Asterisks indicate weak EGFP puncta on anterior ((B) Ds-EGFP) and posterior ((C) Ft-EGFP) junctions. All images from 28 h pupal wings below vein 4. Distal right and posterior down here and subsequently. Scale bar, 5 μm. (B′–B‴) Selection of puncta for analysis. Ds immunolabeling (B′) was used to select puncta (B″, see STAR Methods) on clone boundaries between ds-EGFP and untagged (ds⁺) tissue which were classified according to boundary orientation. Here, puncta are false-colored depending on boundary orientation—posterior (purple), lateral (yellow), and anterior (green) (B‴). (E–G) Normalized mean intensity of EGFP on cell junctions at clone boundaries for fz-EGFP (E), ds-EGFP (F), and ft-EGFP (G). (H–J) Normalized mean intensity of EGFP in puncta and non-puncta. (K–M) Mean asymmetry of EGFP in junctions, puncta, and non-puncta. All means were compared with repeated measure ANOVA analysis ((E–G) and (K–M) are ANOVA with Tukey’s multiple comparisons test and (H–J) are ANOVA with Sidak’s multiple comparisons test to compare between selected columns), p values as indicated. Error bars are standard deviation. n = 7 wings per genotype. See also Figure S1.

Figure 3

Ds is stable on both anterior and posterior cell edges (A) Image of ft-EGFP/ds-mApple twin-clone in 28 h pupal wing. Posterior Ds-mApple (magenta) and anterior Ft-EGFP (green) co-localize in puncta (arrows). Weaker punctate posterior Ft-EGFP and anterior Ds-mApple also co-localize (arrowheads). In zoomed image (right) separation of Ft-EGFP and Ds-mApple can be seen, consistent with them being localized on opposite sides of cell junctions. (B and C) FRAP analysis of Ds-EGFP on whole anterior and posterior junctions. (B) Recovery of Ds-EGFP intensity after photobleaching. (C) Mean intensity of the stable amount of Ds-EGFP (A.U.), calculated by multiplying the stable fraction by the total fluorescence (unpaired t test p = 0.03, anterior n = 7 wings, posterior n = 8 wings). Note, ratio of Ds-EGFP on posterior versus anterior junctions in live imaging is lower than in fixed samples (Figure 2F) due to differences in sample preparation. (D and E) FRAP analysis of punctate Ds-EGFP. (D) Recovery of Ds-EGFP intensity. (E) Mean intensity of the stable amount of Ds-EGFP in puncta (unpaired t test p = 0.12, anterior n = 4 wings, posterior n = 9 wings). (F and G) FRAP analysis of Fz-EGFP on whole proximal and distal junctions. (F) Recovery of Fz-EGFP intensity after photobleaching. (G) Mean intensity of the stable amount of Fz-EGFP (unpaired t test, p = 0.0001, proximal n = 7 wings, distal n = 7 wings). (H and I) FRAP analysis of Fz-EGFP puncta on proximal and distal junctions. Puncta were identified by mCherry-Diego localization. (H) Recovery of Fz-EGFP intensity. (I) Mean intensity of the stable amount of Fz-EGFP in puncta (unpaired t test, p = 0.0003, proximal n = 7 wings, distal n = 7 wings). Table S1 contains the summarized data and 95% confidence intervals for all the FRAP experiments. See also Table S1.

Figure 4

Ft-Ds puncta are mixed with complexes in both orientations (A and B) Schematic of EGFP/mApple twin-clone experiments to examine puncta composition. Cells express a protein tagged with EGFP (green) or mApple (magenta), and puncta are examined at clone boundaries between cells (grey indicates unlabeled trans-binding partners). (A) A situation where individual puncta are strongly polarized, but some puncta are polarized in the opposite orientation. (B) Mixed puncta with complexes primarily in one orientation but with a proportion in the opposite orientation. (C and D) Images of twin-clone tissue in a 28 h pupal wing. (C) ds-EGFP/ds-mApple twin-clones. Arrowheads indicate punctate anterior Ds-EGFP co-localizing with posterior Ds-mApple in neighboring junctions. (D) EGFP-stbm/mApple-stbm twin-clones. Distal EGFP-Stbm does not form distinct puncta. Scale bars, 5 μm. (E and F) Scatter plots of EGFP versus mApple intensity in puncta on twin-clone boundaries. Puncta from multiple wings pooled. (E) Puncta between ds-EGFP/ds-mApple tissue divided into those on the anterior (magenta) or posterior (green) boundary relative to EGFP (n = 6 wings). (F) Puncta between EGFP-stbm/mApple-stbm tissue divided into those on proximal (green) or distal (magenta) boundaries (n = 5 wings). In both cases, puncta groups are either EGFP > mApple or mApple > EGFP, consistent with the model in (B) where puncta on a boundary of a particular orientation show the same overall polarity. (G and H) Diagrams illustrating deduced distributions of stable complexes in puncta. (G) Ft-Ds puncta have stable Ds on both posterior and anterior sides of junctions, consistent with puncta being mixed and stable Ft-Ds heterodimers being present in both orientations. (H) Core pathway puncta only have stable Fz on distal cell edges, consistent with stable core pathway complexes being present in only one orientation. See also Figure S2.

Figure 5

The core pathway but not the Ft-Ds pathway de novo self-organizes coordinated cellular polarity (A, E, and I) Images of 28 h pupal wings showing effects of protein expression for the indicated periods, immunolabeled to show normalized protein intensity (greyscale) and cell polarity marked by yellow lines. (A) Fz-EYFP expression in an fz background, immunolabeled for Fmi. (E) Ds expression in a ds background, immunolabeled for Ft. (I) Ds expression in a ds background with uniform Fj expression, immunolabeled for Ft. Scale bars, 5 μm. (B, F, and J) Circular weighted histograms, representing polarity angles of individual cells for different lengths of protein expression. Data pooled from multiple wings (n = number of wings). (B) Fz-EYFP expression in an fz background. (F) Ds expression in a ds background. (J) Ds expression in a ds background with uniform Fj expression. (C, G, and K) Average cell polarity (direct average polarity magnitude) versus length of expression. (C) Fmi polarity upon Fz-EYFP expression in an fz background. (G) Ft polarity upon Ds expression in a ds background. (K) Ft polarity upon Ds expression in a ds background with uniform Fj expression. Note, slight increase as Ft changes from a diffuse membrane localization to a punctate distribution between 0 and 4 h, with no increase thereafter. We attribute the small increase to a random punctate distribution having a less uniform membrane localization than a diffuse distribution. (D, H, and L) Locally coordinated polarity magnitude (neighbor vector average polarity) versus length of expression. (D) Fmi polarity upon Fz-EYFP expression in an fz background. (H) Ft polarity upon Ds expression in a ds background. (L) Ft polarity upon Ds expression in a ds background with uniform Fj expression. In (L), again note no increase beyond 4 h. All means were compared with ANOVA analysis (Tukey’s multiple comparisons test to compare all columns) and p values indicated. Error bars are standard deviation. See also Figure S3.

Figure 6

Flat gradients of Ds and Fj result in loss of anteroposterior polarity and reduced sorting of Ft-Ds in puncta (A and B) Images of 28 h pupal wings with fbxl7-EGFP/mCherry-fbxl7 twin-clones in a wild-type background (A) or with uniform Ds and Fj expression (B). (A) Higher levels of Fbxl7-EGFP (arrows) and mCherry-Fbxl7 (arrowheads) can be seen on anterior cell junctions on clone boundaries. (B) Arrows indicate puncta on anterior and posterior junctions where Fbxl7-EGFP and mCherry-Fbxl7 co-localize. Scale bar, 5 μm. (C and D) Scatter plots of intensity of EGFP and mCherry in puncta on twin-clone boundaries (red dots) and control puncta in heterozygous tissue (black dots), in a wild-type background (C) or with uniform Ds and Fj (D). Data pooled from six (C) or five (D) wings. (E) Quantitation of mean degree of heterogeneity of Fbxl7 polarity in boundary and control puncta, in wild-type (n = 6) or uniform Ds and Fj (n = 5) wings. “0” indicates completely mixed and “1” completely unmixed (all Fbxl7 on one side of junctions). Errors bars are standard deviation. An ANOVA with Sidak’s multiple comparisons test was used to compare pairs of samples for significance. (F) Ratio of Fbxl7-EGFP/mCherry-Fbxl7 levels in puncta on twin-clone boundaries on the anterior of Fbxl7-EGFP clones in a wild-type background (n = 6 wings) showing anteroposterior polarization or with uniform Ds and Fj expression (n = 3 wings) showing no overall polarity. Error bars are standard deviation. Samples were compared using an unpaired t test. (G and H) Scatter plots of intensity of EGFP versus mCherry in individual puncta on boundaries, considering only those on the anterior (green) or posterior edges (magenta) of EGFP clones, in a wild-type background (G) or with uniform Ds and Fj (H).

Figure 7

Fbxl7 is not required for Ft-Ds planar polarity but does stabilize Ft-Ds in puncta (A and B) APF pupal wings (28 h) carrying loss-of-function fbxl7 clones below vein 4, marked by loss of β-gal immunolabeling (magenta), immunolabeled for Ds (A) or Ft (B). (A′) Ds immunolabeling with polarity nematics overlaid in blue reveals anteroposterior planar polarity inside and outside clone. Scale bar, 10 μm. (C) Mean intensity of Ds or Ft immunolabeling, shown as a ratio of intensity in fbxl7 tissue versus wild-type tissue in the same wing. Error bars are standard deviation, n = 11 wings. One-sample t tests were used to determine if the ratios differed from 1. (D) Average cell polarity magnitude for wild-type tissue and fbxl7 clone tissue in 28 h pupal wings immunolabeled for Ds. All clones were below vein 4. Values from the same wing are linked by black bars, n = 10. Paired t tests were used to compare values in the same wing. Polarity magnitude is relatively low as most clones analyzed lay near the wing margin. (E and F) Mean intensity of the stable amount of Ds-EGFP (E) or Ft-EGFP (F) determined using FRAP in puncta and non-puncta regions of wing discs in control tissue (purple or teal) or fbxl7 mutant tissue (green). Error bars are standard deviation. Samples compared using ANOVA with Sidak’s multi-comparison test. Number of wings analyzed n = 6 (Ds-EGFP) or n = 5 (Ft-EGFP) for control regions, n = 7 (Ds-EGFP) or n = 6 (Ft-EGFP) for fbxl7 mutant tissue. (G) Mean intensity of the stable amount of Ds-EGFP determined using FRAP analysis in puncta and non-puncta regions of wing discs in control tissue (purple) or fbxl7 overexpression tissue (green) at 29°C. Error bars are standard deviation. Samples compared using ANOVA with Sidak’s multi-comparison test. Number of wings analyzed n = 3 for control regions, n = 5 for the UAS-fbxl7 overexpression tissue. (H) Mean intensity of the stable amount of Ds-EGFP determined using FRAP analysis in puncta in wing discs. UAS-fbxl7 overexpression is temporally induced in an fbxl7 mutant background using Act-GAL4,tub-GAL80^ts by shifting animals from 18°C to 29°C for the time indicated. Control tissue without overexpression is shown in the left column at 18°C, and right column at 29°C. Error bars are standard deviation. Last two columns compared using ANOVA with Sidak’s multi-comparison test. Number of wings analyzed n = 2 for 29°C control regions (grey bar), n = 3 for the UAS-fbxl7 24 h overexpression tissue (maroon). See also Figure S4 and Table S1.