Dynamics of Core Planar Polarity Protein Turnover and Stable Assembly into Discrete Membrane Subdomains

Summary The core planar polarity proteins localize asymmetrically to the adherens junctions of epithelial cells, where they have been hypothesized to assemble into intercellular complexes. Here, we show that the core proteins are preferentially distributed to discrete membrane subdomains (“puncta”), where they form asymmetric contacts between neighboring cells. Using an antibody internalization assay and fluorescence recovery after photobleaching in prepupal and pupal wings, we have investigated the turnover of two key core proteins, Flamingo and Frizzled, and find that the localization of both within puncta is highly stable. Furthermore, the transmembrane core proteins, Flamingo, Frizzled, and Strabismus, are necessary for stable localization of core proteins to junctions, whereas the cytoplasmic core proteins are required for their concentration into puncta. Thus, we define the distinct roles of specific core proteins in the formation of asymmetric contacts between cells, which is a key event in the generation of coordinated cellular asymmetry.

Data show relative levels of extracellular Fmi in mutant compared to wildtype tissue of the same wings. Stars indicate P values relative to wildtype at the same timepoint, or as indicated by bars.
Fmi is internalised faster in fz,stbm mutant tissue using both antibodies; however the effect is significantly stronger using full-length Fmi antibody. Note that in this experiment, as mutant and wildtype values are ratioed, this should provide an internal correction for fall-off of the Fab antibody fragment. However, the apparently slower rate of internalisation seen with the Fab fragments may be an artefact caused by mutant tissue having a greater fraction of internalised Fab fragment than wildtype tissue, and this internalised fraction being protected from dissociating from the tissue when in endosomes and then being recycled to the plasma membrane.
(E) Fmi antibody internalisation experiment in wildtype wings using Fmi Fab antibody fragments, stained for extracellular Fmi at various chase times. Although confocal settings are constant, confocal detector background is more prominent at later timepoints, as Fab antibody fragments fall off and the remaining population becomes increasingly difficult to detect. Nevertheless, a punctate junctional Fmi population (arrows) remains at junctions at 30 min, demonstrating that antibody cross-linking is not the cause of the failure of punctate Fmi to internalise. Scale bar 2.5 µm. (L) Antibody internalisation in wings expressing ptc-GAL4, fmi-IR. Extracellular Fmi levels are reduced to 30 % of wild type levels, but asymmetric localisation and trichome polarity is normal (not shown). There is no significant difference in internalisation compared to wildtype tissue in the same wings.

Supplemental
(H,I) 28 hr pupal wing tissue showing Stbm staining in wildtype (H) or pk pk-sple13 (I) tissue from the same wing. (H',I') The outcome of threshholding and particle analysis to identify puncta, using a constant threshhold in the wildtype and mutant areas. Arrows show prominent puncta, note puncta are overall smaller in the mutant (I).
(J,K) Distribution of puncta size in 28 hr and 20 hr pupal wings (J), or in wildtype and dgo 380 , pk pk-sple13 and dsh V26 mutant tissue (K), stained for Stbm. Puncta size is in pixels (1 pixel = 47 x 47 nm), and 20 pixels was the lower cut-off for puncta, corresponding to puncta with an average diameter of 240 nm. In (K) size categories are larger, to reduce noise due to the smaller sample size. More smaller puncta are seen at 20 hr and in the mutant conditions than in wildtype at 28 hr, at the expense of larger puncta. Distributions are significantly different using a chi-squared test for wildtype at 28 hr vs 20 hr (P***) and for wildtype vs dgo (P***), pk (P**) and dsh (P***) at 28 hr.
(L,M) Quantitation of puncta size (L) or mean fluorescence intensity (M) in 28 hr and 20 hr wildtype pupal wings, and in dgo 380 , pk pk-sple13 and dsh V26 mutant clones stained for Fmi. Stars are P values compared to 28 hr wildtype wings, or as indicated by bars. Note that, unlike the other core proteins, Fmi levels are slightly increased in dsh mutant tissue (Shimada et al., 2001;Strutt and Strutt, 2008), and thus the particle analysis failed to see a reduction in Fmi puncta size in these clones.  . fmi E45 produces full-length protein, but it is not correctly processed, and a ladder of smaller bands is seen (star). Quantitation of the cleaved form of Fmi is shown.

Supplemental
(C) Graph of recovery of Fz-EYFP fluorescence after different amounts of bleaching. Bleaching was varied by altering the number of laser passes (iterations) over the bleached region, with laser power remaining constant (this method also varies the length of bleaching time, 1 pass = 1.2 sec).
A similar percentage recovery is seen with up to 50 iterations, but recovery is reduced with more iterations, suggesting that above 50 laser passes the laser-exposure may induce cross-linking of proteins and thus inhibit fluorescence recovery.
(D) Graph showing the percentage of bleaching relative to the number of bleach iterations. A bleaching amount of between 50 and 80% is considered optimal to prevent artefacts occurring in the resulting data, as this should give a good signal-to-noise ratio without damage to the tissue.