Dissecting the molecular bridges that mediate the function of Frizzled in planar cell polarity

Abstract

Many epithelia have a common planar cell polarity (PCP), as exemplified by the consistent orientation of hairs on mammalian skin and insect cuticle. One conserved system of PCP depends on Starry night (Stan, also called Flamingo), an atypical cadherin that forms homodimeric bridges between adjacent cells. Stan acts together with other transmembrane proteins, most notably Frizzled (Fz) and Van Gogh (Vang, also called Strabismus). Here, using an in vivo assay for function, we show that the quintessential core of the Stan system is an asymmetric intercellular bridge between Stan in one cell and Stan acting together with Fz in its neighbour: such bridges are necessary and sufficient to polarise hairs in both cells, even in the absence of Vang. By contrast, Vang cannot polarise cells in the absence of Fz; instead, it appears to help Stan in each cell form effective bridges with Stan plus Fz in its neighbours. Finally, we show that cells containing Stan but lacking both Fz and Vang can be polarised to make hairs that point away from abutting cells that express Fz. We deduce that each cell has a mechanism to estimate and compare the numbers of asymmetric bridges, made between Stan and Stan plus Fz, that link it with its neighbouring cells. We propose that cells normally use this mechanism to read the local slope of tissue-wide gradients of Fz activity, so that all cells come to point in the same direction.

Fig. 1.

A summary of the experiments. The experiments are shown as clones (ellipses) that affect (or not) the polarity of the wild-type (wt) or mutant surround. Anterior of the fly is shown towards the top of the figure; all hairs made by the ventral abdominal epidermis normally point posteriorly, towards the bottom. Large red arrows indicate a change in polarity spreading up to several cells from the clone, and large red arrowheads indicate a change that is limited largely or only to the abutting cells. Small red arrows and arrowheads indicate that the polarity effects are weak. Arrows in grey indicate an effect (surmised) that is co-oriented with the extant polarity and therefore cryptic. The pale green and pale pink colours represent endogenous Vang and Fz. Vang and Fz together (wt) give pale yellow; absence of both gives white. Dark colours indicate overexpression; Stan is shown as grey. Complete experimental genotypes are listed in the Materials and methods according to this scheme: e.g. 1A is a Vang^– clone in a wild-type background. Polarising effects depicted are based on assessments described in the Materials and methods.

Fig. 2.

Examples of clones and polarity effects in the pleura. (A-H) Genotypes of the clones and the backgrounds are indicated, coded by column and row as in Fig. 1 and in the Materials and methods. The clones are marked by pawn and shavenoid, mutations that together remove hairs cell-autonomously, producing a naked patch; the direction of each hair made by adjacent receiving cells is indicated by a coloured dot (red, outwards; blue, inwards; yellow, parallel). The clones either affect (A-C,E,H) or not (D,F,G) the orientation of hairs adjacent to the clone. The adjacent hairs tend to be oriented outwards (A,C,E), inwards (B,H) or more or less randomly (D,F,G). This figure is intended to provide representative examples of the different polarising effects observed, and not to provide quantitative data, which are presented in Fig. 3 and Table 1. All pictures are at the same magnification; hair sizes vary according to the segment as well as to the position within each segment. Anterior is towards the top.

Fig. 3.

Quantification of the polarising effects of Stan^Fz and Stan^V signals on receiving cells. Boxplots of key experimental genotypes, showing the percentage of hairs facing outwards from the clones (see Materials and methods; Table 1). (A) Experiments performed in ds^–/+ flies to test the requirement for Vang in the response of fz^– receiving cells to Stan^Fz signal. Vang^– UAS.fz clones repolarise surrounding cells to point away strongly in fz^– flies (4E; box 2; green) but only weakly in Vang^– fz^– flies (5C; box 1; purple). The weak repolarising activity of UAS.fz clones in Vang^– fz^– flies is similar, irrespective of the ds genotype (compare the first box in A with the third box in B), but significantly different from the negative control (fourth box in B) and from the strong repolarising activity in fz^– flies (second box in A, first box in B). (B) Experiments performed in ds^– flies. UAS.fz clones polarise surrounding cells strongly to make hairs that point away in fz^– flies (4C; first box; pink), but only weakly in Vang^– and Vang^– fz^– flies (3C, 5C; second and third boxes; yellow). UAS.Vang clones have no detectable effect in Vang^– fz^– and fz^– flies (5A, 4A; fifth and sixth boxes; grey; the fourth box is a negative control, grey, 5D), whereas they polarise strongly to point inwards in Vang^– flies (3A; seventh box; blue).

Fig. 4.

The Stan system in PCP – a model. (A-C) The imagined disposition of Stan (blue), Fz (red) and Vang (green) proteins in sending (left) and receiving (middle and right) cells in three experiments. Stan, alone or in complex with Fz (Stan^Fz), accumulates on the apical cell membrane only when engaged in asymmetric Stan<<Stan^Fz or Stan^Fz>>Stan bridges; otherwise, it dwells transiently on the apical surface before being endocytosed, as suggested by Strutt et al. (Strutt et al., 2011). Fz complexes with Stan to make Stan^Fz that binds to Stan on the surface of abutting cells (the extracellular domain of the Stan molecule in the complex is shown in black to depict its altered state). Fz also blocks the hair-repressing activity of Stan (the latter is indicated by the blue inhibitory arrows, the thickness of the arrows reflects the number of Stan molecules not in complex with Fz, and hence able to repress hair outgrowth). Vang helps Stan form stable intercellular bridges with Stan^Fz. (A) For UAS.fz clones in Vang^– fz^– flies, the only bridges that can form are Stan^Fz>>Stan bridges between the UAS.fz sending cell and the abutting Vang^– fz^– receiving cell. Thus, higher amounts of Stan would accumulate on the left side of the Vang^– fz^– receiving cell than on the right side, repressing hair outgrowth on the left side, and biasing the cell to project its hair from the right side. No Stan<<Stan^Fz or Stan^Fz>>Stan bridges can form between this cell and the next Vang^– fz^– receiving cell to the right, limiting the polarising effect of the sending cell to only the abutting Vang^– fz^– receiving cell. (B) The situation for UAS.fz clones in fz^– flies is similar to that in A, except that Vang is now present in the fz^– receiving cell, helping to drive Stan in that cell to form stable bridges with Stan^Fz on the sending cell. Thus, higher levels of Stan accumulate on the left side of the receiving cell than in A, resulting in a stronger polarising effect. (C) fz^– UAS.Vang clones in Vang^– flies create a confrontation between fz^– sending cells and Vang^– receiving cells. Such sending cells can form only Stan<<Stan^Fz bridges with the abutting receiving cells; moreover, excess Vang in the sending cell will promote the formation of these bridges. Thus, large amounts of Stan^Fz, which lack the capacity to repress hair formation, will accumulate on the left side of Vang^– receiving cell. This first Vang^– cell engages with a second Vang^– cell on its right and could form both Stan<<Stan^Fz and Stan^Fz>>Stan bridges. Any Stan<<Stan^Fz bridges that form between the first and second Vang^– cells will locally repress hair outgrowth on the right side of first cell, biasing the hair on this cell to the left side and towards the fz^– sending cell. (D) Control of Stan PCP by morphogen gradients. A morphogen gradient directs stepwise changes in the level of Fz activity from one cell to the next. The resulting differences in Fz activity between cells determine the number of Stan molecules in each cell that are engaged in asymmetric Stan<<Stan^Fz bridges with each neighbour. This number should be highest along the interface with the neighbour with the most Fz activity, and lowest along the interface with the neighbour with least. Formation of such bridges stabilises both Stan and Stan^Fz, on the apical cell surface, protecting them from endocytosis and recycling; these accumulations may be increased by intra- and intercellular feedbacks. Stan and Stan^Fz have opposite effects on cell polarity on each side of asymmetric Stan<<Stan^Fz bridges, repressing hair formation on the Stan side, while allowing it on the Stan^Fz side – thus directing all cells to point hairs in the same direction, down the tissue-wide gradient of Fz activity.