The adult cuticular wing of Drosophila is covered by an array of distally pointing hairs that reveals the planar polarity of the wing. We report here that mutations in dachsous disrupt this regular pattern, and do so by affecting frizzled signaling. dachsous encodes a large membrane protein that contains many cadherin domains and dachsous mutations cause deformed body parts. We found that mutations in dachsous also result in a tissue polarity phenotype that at the cellular level is similar to frizzled, dishevelled and prickle, as many cells form a single hair of abnormal polarity. Although their cellular phenotype is similar to frizzled, dishevelled and prickle, dachsous mutant wings display a unique and distinctive abnormal hair polarity pattern including regions of reversed polarity. The development of this pattern requires the function of frizzled pathway genes suggesting that in a dachsous mutant the frizzled pathway is functioning - but in an abnormal way. Genetic experiments indicated that dachsous was not required for the intracellular transduction of the frizzled signal. However, we found that dachsous clones disrupted the polarity of neighboring wild-type cells suggesting the possibility that dachsous affected the intercellular signaling function of frizzled. Consistent with this hypothesis we found that frizzled clones in a dachsous mutant background displayed enhanced domineering non-autonomy, and that the anatomical direction of this domineering non-autonomy was altered in regions of dachsous wings that have abnormal hair polarity. The direction of this domineering nonautonomy was coincident with the direction of the abnormal hair polarity. We conclude that dachsous causes a tissue polarity phenotype because it alters the direction of frizzled signaling.

Cadherins comprise a group of transmembrane proteins that mediate Ca2+-dependent cell adhesion (Geiger and Ayalon, 1992). They are located in adherens junctions and desmosomes. To function in adhesion, cadherins need to associate and form complexes with the actin cytoskeleton and the catenins (Nagafuchi and Takeichi, 1988). In addition to serving as mediators of adhesion the cadherin-catenin proteins are also involved in signaling pathways. In Drosophila, the armadillo (arm) gene encodes β-catenin (Peifer and Weischaus, 1990). This gene is an essential component of the wingless (wg) (wg is a member of the Wnt family) signaling pathway in Drosophila (Peifer et al., 1991). β-catenin is also involved in Wnt signaling in the Xenopus embryo (Gumbiner, 1995). In both of these cases there is evidence that the β-catenin signaling activity is independent of its role in cadherin-mediated cell adhesion (Heasman et al., 1994; Peifer et al., 1994), and involves the activation of the lef/TCF/pangolin transcription factor (van de Wetering et al., 1997; Brunner et al., 1997). However, since both activities take place in the same cell it would not be surprising if there was a linkage between signaling and adhesion (Gumbiner, 1995; Eaton and Cohen, 1996). Indeed, overexpression of cadherin can inhibit β-catenin signaling (Heasman et al., 1994; Funayama et al., 1995, Sanson et al., 1996).

The products of several Drosophila genes have been found to encode cadherin like proteins. Genetic studies indicate that two of these genes, fat (ft) and dachsous (ds), (Clark et al., 1995; Mahoney et al., 1991) have interesting roles in imaginal development. Consistent with the hypothesis that the proteins encoded by these genes have cadherin like activity, mutations in both ds and ft produce a number of phenotypes that can be interpreted as being due to effects on cell adhesion (Clark et al, 1994; Mahoney et al., 1991). However, both of these genes encode proteins that are substantially larger than the typical vertebrate cadherins or the Drosophila E-cadherin (Tepass et al., 1996; Uemura et al., 1996), raising the possibility that their function is not typical of the cadherin family.

The adult cuticle of Drosophila is decorated with large numbers of polarized structures. In any body regions these are typically aligned in parallel, giving the region a ‘tissue polarity’ (Adler, 1992; Gubb, 1993). For example, the adult wing is covered by a distally pointing array of hairs. Several groups have been investigating the genetic basis for the morphogenesis of tissue polarity (Adler, 1992; Gubb, 1993; Strutt et al., 1997; Zheng et al., 1995). We have primarily used the wing as a model system. In the pupal wing each cell extends a microvillus like prehair from the apical surface of the cell (Wong and Adler, 1993). This prehair, which gives rise to the adult cuticular hair, is formed in the vicinity of the distal-most vertex of these polygonally shaped cells. The use of this subcellular location appears to be tightly linked to the distal polarity of the hair, as mutations identified because of altered hair polarity also result in an abnormal subcellular location for prehair initiation (Wong and Adler, 1993). The frizzled (fz) signaling/signal transduction pathway, which has both cell nonautonomous and cell autonomous functions, appears to control this process. We have recently carried out a screen for new wing tissue polarity mutations and among those recovered were mutations in both ds and ft.

As has been described previously (Clark et al., 1995) strong ds mutations result in adults with abnormally shaped legs and wings. We show here that ds mutations result in a wing tissue polarity phenotype that, based on its cellular phenotype (e.g. relatively few multiple hair cells),puts them in the fz-like phenotypic group of tissue polarity mutations (Wong and Adler, 1993; Krasnow and Adler, 1994). The ds mutant polarity pattern is however, unique among tissue polarity mutants. Most notably, in some regions hair polarity is reversed and hairs point proximally.

In genetic mosaics we found that ds displayed domineering cell nonautonomy and by constructing and examining double mutants we found that the function of fz tissue polarity pathway genes was required for the development of the unique ds polarity pattern. This requirement suggested the possibility that the ds polarity pattern resulted from abnormal activity of the fz pathway. The ability of fz overexpression to produce an inturned (in) like phenotype (Krasnow and Adler, 1994) was used as a test for the cell autonomous transduction of the fz signal. We found that ds mutations did not block the ability of fz overexpression to produce an in-like phenotype, and concluded that ds, unlike dsh was not required for the transduction of the fz signal (Krasnow et al., 1995). In contrast, we found that ds mutations caused abnormalities in fz intercellular signaling (i.e. the cell nonautonomous function of fz). We saw both an enhanced extent and an altered anatomical direction for the domineering non-autonomy of fz clones. The direction of the domineering nonautonomy was the same as the local polarity of hairs. This was seen both in regions with normal distal polarity and in regions with abnormal proximal polarity. It seems likely that the abnormal polarity seen in regions of ds wings is due to the abnormal anatomical direction of fz signaling in these regions. The data suggest that ds is involved in fz signaling and raise the possibility that fz signaling will be associated with adherens junctions.

Fly culture and strains

Unless otherwise stated, flies were cultured at 25oC. Many mutant and Deficiency containing stocks were obtained from the stock centers at Indiana University and Bowling Green State University. We have been carrying out a large-scale FLP/FRT screen (Xu and Rubin, 1993) to recover hair polarity mutations, which will be presented elsewhere (P. N. Adler, J. Charlton and J. Liu, in preparation). Briefly, FRT flies were mutagenized with EMS, crossed to hs-flp; FRT flies and clones induced in the F1 progeny via heat shocking larvae. The adult F1 flies were anesthetized under CO2, and one wing was removed without killing the fly. The wing was examined under a compound microscope and flies where clones with altered hair polarity, number or morphology were saved and bred to recover the mutation. In these screens we recovered mutations in ds that prompted the experiments reported here.

Cytological procedures

Pupal wings were stained with a fluorescent phalloidin to stain the actin cytoskeleton and/or with flourescent antibodies and examined using a Molecular Dynamics confocal microscope as described previously (Wong and Adler, 1993; Park et al., 1994).

Generation of genetic mosaics

Mosaic clones were generated using the FLP/FRT system (Golic and Lindquist; Xu and Rubin, 1993). To mark ds clones we used recessive mutations in the ck gene. These mutations result in multiple, split and shortened hairs that typically lie at a greater angle to the wing blade than wild-type cells (Gubb et al., 1984; C. Turner and P. N. Adler unpublished).

In the experiments where we generated fz mutant clones in a ds mutant background we used the cell autonomously acting hair morphology mutation starburst (strb) (Park et al., 1996). Clones of strb produce hairs with a distinctive and easy to identify phenotype (although difficult to photograph). We classified fz clones for the extent of domineering nonautonomy they produced in the following way. Clones that caused surrounding wild-type cells, no more than 3 cell diameters from the border of the clone, to point no more than 45° from distal were classified as showing weak domineering nonautonomy. When the disruption was more pronounced (wild-type cells as many as 5 cells from the clone were affected and pointed as much as 90° from distal) the clones were classified as showing moderate nonautonomy. For clones showing strong nonautonomy surrounding wild-type cells as many as 10 cells from the border were affected and occasional hairs pointed as much as 180o from distal. For a clone to be classified as showing extreme nonautonomy there had to be a substantial region (>5 cells) that had well aligned reversed polarity.

Scoring of mutant wings

Wings were mounted in Euparal (Asco Labs) and examined under bright-field microscopy. As part of the analysis we often made drawings of individual wings that showed the abnormal polarity pattern of wing hairs on the dorsal surface of the wing (see Figs 3 and 7). The pattern shown is of an individual wing, however, at least five other wings of that genotype were examined to ensure that the wing drawn was typical for the genotype. We quantified the ds tissue polarity phenotype by estimating the fraction of a wing (or region of a wing) with abnormal polarity and by determining the fraction or number of multiple hair cells in specific regions of the wing as described previously (Wong and Adler, 1993; Krasnow and Adler, 1994).

ds mutations result in a tissue polarity phenotype on the wing

In a screen to identify new tissue polarity mutations we recovered a mutation (UAO71) which gave a fz-like phenotype in wing clones. Rare homozygous UAO71 adult flies showed shortened wings with misplaced and severely eroded cross veins (see Fig. 1D), a short fat body and severely distorted legs. We subsequently isolated 12 additional mutations in this gene (as judged by a failure to complement UA071) that gave a similar, albeit in most cases a less severe phenotype (Fig. 1). Most of our experiments have utilized the UAO71 allele. The mutation was mapped to cytogenetic location 21D1 on the basis of being uncovered by Df(2L)al (21B8-C1; 21C8D1) and Df(2L)ast2 (21D1-2; 22B2,3) but not by Df(2L)S3 (21D2,3; 22A1). The genetic location and phenotype suggested that UAO71 could be an allele of ds. This was confirmed by the failure of UAO71 to complement ds1 and ds33K (a presumptive null allele which is associated with In(2LR)bwv1; Clark et al., 1995). We henceforth refer to UA071 as dsUAO71. The dsUA071/ds33K flies showed the same extreme phenotype we saw in the dsUA071 homozygotes and hemizygotes. We found that ds33k /Df(2L)al and ds33k /Df(2L)ast2 flies had wing tissue polarity phenotypes that were similar to those of dsUA071 homozygotes and hemizygotes both with respect to the abnormal polarity pattern and the frequency of multiple hair cells produced (Table 1). These data indicate that dsUA071 is close to a null allele for wing tissue polarity and that the ds alleles we recovered are not unusual ds alleles. The dsUAO71/ds1 flies did not show any major wing hair polarity phenotype, but did show the body shape phenotype associated with the ds1 allele (Clark et al., 1995). The ds1 allele appears not to be severe enough to routinely cause a tissue polarity phenotype on the wing. About half of the wings of either ds1/dsUA071 or ds1/ds33k flies showed a very weak tissue polarity phenotype. This consisted of a couple of multiple hair cells and/or a small region of polarity disruption in the proximal part of the wing.

Table 1.

ds is not required for the transduction of the fz signal

ds is not required for the transduction of the fz signal
ds is not required for the transduction of the fz signal
Fig. 1.

Light micrographs of Drosophila wings. All wings are oriented with proximal to the left. (A) Oregon R, (B) dsYM31, (C) dsYAC41, (D) dsUAO71/ds33k. E-P show the dorsal surface along the anterior margin of the wing. (E) Oregon R, (F) dsYM31, (G) dsYAC41, (H) dsUAO71/ds33k, (I) fzR54/fzK21 (a strong fz genotype), (J) dsUAO71/ds33k; fzR54/fzK21, (K) inIH56 (a strong in allele), (L) dsUAO71/ds33k; inIH56, (M) mwh3 (a strong mwh allele), (N) dsUAO71/ds33k; mwh3, (O) hs-fz/+ (1-hour heat shock 24 hours awp formation), (P) dsUAO71/ds33k; hs-fz/+ (1-hour heat shock 24 hours awp). Note that in all double mutants the wing vein placement abnormalities of ds were present, but hair polarity and number resemble the non-ds mutant. Arrows point in the local direction of hair polarity.

Fig. 1.

Light micrographs of Drosophila wings. All wings are oriented with proximal to the left. (A) Oregon R, (B) dsYM31, (C) dsYAC41, (D) dsUAO71/ds33k. E-P show the dorsal surface along the anterior margin of the wing. (E) Oregon R, (F) dsYM31, (G) dsYAC41, (H) dsUAO71/ds33k, (I) fzR54/fzK21 (a strong fz genotype), (J) dsUAO71/ds33k; fzR54/fzK21, (K) inIH56 (a strong in allele), (L) dsUAO71/ds33k; inIH56, (M) mwh3 (a strong mwh allele), (N) dsUAO71/ds33k; mwh3, (O) hs-fz/+ (1-hour heat shock 24 hours awp formation), (P) dsUAO71/ds33k; hs-fz/+ (1-hour heat shock 24 hours awp). Note that in all double mutants the wing vein placement abnormalities of ds were present, but hair polarity and number resemble the non-ds mutant. Arrows point in the local direction of hair polarity.

The cellular phenotype of ds places it in the fz-like group of genes

Previously we found we could distinguish between fz-like and in-like phenotypes by comparing the number of multiple hair cells as a function of the fraction of the wing showing altered polarity (Krasnow and Adler, 1994). We scored this phenotype for 3 ds genotypes of varying severity and found that they all fell in the fz-like group (Fig. 2). The phenotype of the ds mutants was however, more variable from one individual to another than is typical for other tissue polarity genes.

Fig. 2.

Plot of the number of multiple hairs (in the dorsal C region) versus the fraction of the dorsal C region where hairs show abnormal polarity. Stronger alleles are indicated with darker shaded symbols. Different symbols indicate the gene (or condition) in question. From stronger to weaker, the genotypes for the different fz like genes were: fzR52/fzD21, fzHD21/fzD21 (18°C), fzR53, dsh1, pk1, dsUAO71/ds33k, dsYAC41, dsYM31, hsfzI (1 hour heat shock at 37°C at 30 hours awp), hsfzI (1 hour heat shock at 33oC at 29 hours awp), hsfzI (1 hour heat shock at 37°C at 25 hours awp), hsfzI (1 hour heat shock at 31°C at 29 hours awp). For the in like genes the genotypes were inIH56/Df(3L)rdgC (a deficiency for in), inII53/Df(in) (at 29°C), inHC31/Df(in) at 25°C, fy1, fyJN12/fy1, hsfzI (1 hour heat shock at 37°C at 34 hours awp), hsfzI (1 hour heat shock at 33°C at 34 hours awp), hsfzI (1 hour heat shock at 31°C at 34 hours awp). The data for the fz and in alleles and the early and late fz gain-of-function points comes from Krasnow and Adler (1994).

Fig. 2.

Plot of the number of multiple hairs (in the dorsal C region) versus the fraction of the dorsal C region where hairs show abnormal polarity. Stronger alleles are indicated with darker shaded symbols. Different symbols indicate the gene (or condition) in question. From stronger to weaker, the genotypes for the different fz like genes were: fzR52/fzD21, fzHD21/fzD21 (18°C), fzR53, dsh1, pk1, dsUAO71/ds33k, dsYAC41, dsYM31, hsfzI (1 hour heat shock at 37°C at 30 hours awp), hsfzI (1 hour heat shock at 33oC at 29 hours awp), hsfzI (1 hour heat shock at 37°C at 25 hours awp), hsfzI (1 hour heat shock at 31°C at 29 hours awp). For the in like genes the genotypes were inIH56/Df(3L)rdgC (a deficiency for in), inII53/Df(in) (at 29°C), inHC31/Df(in) at 25°C, fy1, fyJN12/fy1, hsfzI (1 hour heat shock at 37°C at 34 hours awp), hsfzI (1 hour heat shock at 33°C at 34 hours awp), hsfzI (1 hour heat shock at 31°C at 34 hours awp). The data for the fz and in alleles and the early and late fz gain-of-function points comes from Krasnow and Adler (1994).

Fig. 3.

Drawings of the wing hair polarity pattern on the dorsal surface of a typical mutant wing. The filled areas are regions where a polarity cannot be defined as neighboring hairs are not well aligned.

Fig. 3.

Drawings of the wing hair polarity pattern on the dorsal surface of a typical mutant wing. The filled areas are regions where a polarity cannot be defined as neighboring hairs are not well aligned.

ds mutants display a unique wing hair polarity pattern

Mutations in tissue polarity genes do not result in a random pattern of hair polarity in most regions of the wing (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993). Rather, mutations result in gene-specific stereotypic patterns. Mutations in most tissue polarity genes (e.g. fz, dishevelled (dsh), in, fuzzy (fy), Van Gogh, starry night, fritz, multiple wing hairs (mwh) result in a similar (albeit not identical) abnormal polarity phenotype. For example, in wings from flies mutant for any of these genes, hairs in the E region of the wing point toward the posterior margin (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993; R. Krasnow and P. N. Adler, unpublished). We refer to this as the fz/in polarity pattern (Fig. 3). It appears to be the default pattern that results from the inactivation of the fz pathway. Mutations in prickle (pk) result in a distinctively different pattern. For example, in a pk wing cells in the E cell tend to point anteriorly (Fig. 3). We found that ds mutant wings also displayed stereotypic polarity patterns, although the pattern was dramatically different from any other tissue polarity mutants. Notably, in strong mutants there was a region at the anterior margin with a swirl of partially reversed polarity hairs (Figs 1, 3). In other regions in the anterior/central proximal part of the wing blade there were regions of completely reversed polarity (i.e. hairs that pointed proximally) (Figs 3, 4). Such regions of reversed polarity were separated from regions of relatively normal polarity via a region of swirling (Fig. 4). In weaker alleles these patterns were seen in a less extreme form (Figs 1, 3). As is typical for other tissue polarity genes, the polarity disruptions were more severe on the ventral than dorsal surface. In many ds wings the region of reversed polarity on the ventral surface was quite large (Fig. 4A,B). The novelty of the ds pattern argued that ds did not produce a polarity phenotype by inactivating the fz signaling/signal transduction pathway as do mutations in genes such as fz, dsh or in.

Fig. 4.

Three examples of reversed hair polarity in dsUAO71/ds33k wings. (A,B) Ventral surface of the wing, and (C) the dorsal surface. In all cases proximal is to the left and distal to the right.

Fig. 4.

Three examples of reversed hair polarity in dsUAO71/ds33k wings. (A,B) Ventral surface of the wing, and (C) the dorsal surface. In all cases proximal is to the left and distal to the right.

Pupal wing phenotype of ds

We observed the actin cytoskeleton in dsUAO71/ds33K pupal wings via phalloidin staining. As is the case in wild-type wings, prehair initiation began first in cells located distally and progressed proximally in a patchy pattern (Wong and Adler, 1993). In wing regions that did not show substantial polarity alterations prehairs were initiated in the vicinity of the distal vertex of the wing cells (this is what is seen in wild type; Wong and Adler, 1993). In most regions with abnormal polarity in ds wings, prehairs initiated at an alternative location along the cell periphery. In the regions with reversed polarity, prehairs initiated in the vicinity of the proximal-most vertex (Fig. 5A). In regions where hair polarity was close to random and in regions at the center of swirls, prehairs were found to initiate in the central regions of the cells (Fig. 5A). This pattern is most similar to that seen previously for pk (Wong and Adler, 1993). Because of the well known interaction between catenins and cadherins we examined the distribution of the Armadillo (β-catenin) protein in wild-type and dsUAO71/ds33K pupal wings. No differences were seen, suggesting that ds does not play a key role in governing the subcellular localization of Armadillo.

Fig. 5.

(A) A micrograph of a dsUAO71/ds33k wing stained with rhodamine phalloidin and examined in the confocal microscope. The asterisk indicates a cell where a prehair is found in the center of the apical surface of the cell. Cells with prehairs that point proximally and are located at the proximal edge of the cell are obvious in the upper left side of the panel. An arrowhead points to one such cell. Scale bar is 5 μm. (B,C) dsUAO71ckUAH21 clones. The clone boundaries are outlined. The arrows point to regions of wild-type cells with abnormal polarity. In all micrographs proximal is to the right and distal is to the left.

Fig. 5.

(A) A micrograph of a dsUAO71/ds33k wing stained with rhodamine phalloidin and examined in the confocal microscope. The asterisk indicates a cell where a prehair is found in the center of the apical surface of the cell. Cells with prehairs that point proximally and are located at the proximal edge of the cell are obvious in the upper left side of the panel. An arrowhead points to one such cell. Scale bar is 5 μm. (B,C) dsUAO71ckUAH21 clones. The clone boundaries are outlined. The arrows point to regions of wild-type cells with abnormal polarity. In all micrographs proximal is to the right and distal is to the left.

fz pathway function is needed for the ds polarity pattern

We attempted to construct double mutants between ds and fzR54/fzK21, dsh1, inIH56, and mwh3. The strong ds genotype (dsUAO71/ds33K) used results in low viability (only about half the expected number of adults emerged and most of these quickly got stuck in the food and died) and a severely distorted body shape. We did not recover any dsh; dsUAO71/ds33K adults. The double mutants between ds and fz, in, and mwh all showed the general body shape (e.g. wing, leg) and wing vein abnormalities of ds, but they all had a fz/in-like wing tissue polarity pattern (see Fig. 1). Further, in the ds; in and ds; mwh double mutants a large number of cells formed more than 1 hair as is seen in in and mwh (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993). The epistasis was not perfect in all aspects, however. For example, the number of hairs per cell in a ds; mwh double mutant was lower than in a mwh single mutant (2.68 vs 3.3). We concluded from these experiments that the function of the fz pathway genes was required for the unusual ds polarity pattern, which suggested that the fz signaling/signal transduction pathway was functioning, but in an abnormal way in ds mutants. Further evidence for an interaction between fz pathway genes and ds is that a weak ds mutant genotype (e.g. ds1/ds33k) acted as an enhancer of a hypomorphic fz allele (fzR53) as it increased the fraction of the wing that showed abnormal polarity.

We also generated flies that were mutant for ds and which carried a hs-fz transgene (Krasnow and Adler, 1994). A single 1-hour heat shock at 24 hours after white prepupae (awp) formation suppressed the distinctive ds polarity phenotype (e.g. regions of reversed polarity) (Fig. 1O,P). The phenotype obtained resembled that seen in hs-fz/+ control flies (Krasnow and Adler, 1994). Other aspects of the ds mutant phenotype (e.g. appendage shape abnormalities, wing vein placement abnormalities) were not suppressed. We suspect that the fz overexpression phenotype results from the ligand independent activation of the fz signal transduction pathway. That fz overexpression suppressed the ds polarity phenotype is consistent with fz being downstream of ds in wing tissue polarity, although other data suggest the relationship between ds and fz is more complex.

ds displays domineering cell nonautonomy

Since the screen we used to isolate the ds alleles relied on seeing a mutant phenotype in clones it was clear that neighboring wild-type cells would not completely rescue the mutant phenotype of a ds clone. As a test of whether or not wild-type cells could rescue juxtaposed ds cells we induced ds and ck twin spots. We frequently saw cells with a ck phenotype next to cells showing altered polarity, indicating that juxtaposed wild-type cells need not rescue ds cells. Indeed, in some cases the ck cells had altered polarity suggesting that ds clones display the domineering cell nonautonomy characteristic of fz and pk (Gubb and Garcia-Bellido, 1982; Vinson and Adler, 1987). To determine if this was the case we generated ds ck clones. The clones larger than 16 cells appeared more oval and less extended than typical wing clones and they also often appeared to form a bulge on the wing surface. These phenotypes could be a manifestation of altered cell adhesion. We found a majority of clones (63/117) resulted in altered polarity of neighboring wild-type cells (Fig. 5B,C). This domineering nonautonomy was size dependent, and was the majority result only when the clones were relatively large (>100 cells). Both the frequency and the magnitude of the region of wild-type cells with altered polarity was smaller than we had seen in our earlier studies on fz (Vinson and Adler, 1987; Jones et al., 1996). As we saw with fz nonautonomy, the affected wild-type hairs tended to point toward rather than away from the clone (Adler et al., 1997). In the case of fz there is a striking directionality to the domineering nonautonomy as neighboring wild-type cells distal (and anterior and posterior) but not proximal to the clone show altered polarity (Vinson and Adler, 1987). To determine if ds clones showed any directional bias to their non-autonomy we examined this for the 63 clones that displayed nonautonomy. We found 36 showed distal nonautonomy, 32 anterior nonautonomy, 30 posterior nonautonomy, and only 7 proximal nonautonomy. Thus, while ds nonautonomy is less biased in direction than fz, it is not random. That ds acts nonautonomously argues that if ds mutations produce a tissue polarity phenotype by altering the activity or function of the fz pathway it needs to affect the cell nonautonomous part of the pathway (i.e fz signaling).

ds is not required for the intracellular transduction of the fz signal

Loss-of-function mutations in the cell autonomously acting in- like genes results in many cells forming more than one hair (Wong and Adler, 1993; Adler et al., 1994). Thus, the in-like genes function as negative regulators of prehair initiation. Based on epistatic interactions, we previously suggested that Fz and Dsh acted upstream of In and Fy and inhibited the activity of these proteins in the vicinity of the distal vertex leading to prehair initiation at this location (Wong and Adler, 1993). The overexpression of fz a few hours prior to prehair initiation results in many cells forming more than one hair resembling a weak or moderate allele of inturned or fuzzy (Krasnow and Adler, 1994). We interpreted the induction of multiple hair cells by late fz overexpression as being due to the over activation of the fz signal transduction pathway leading to excess inhibition of In and Fy (Krasnow and Adler, 1994). Consistent with this hypothesis a similar phenotype is produced by the late overexpression of dsh, and the weak overexpression of fz acts as an enhancer of weak in and fy alleles (R. Krasnow and P. N. Adler, unpublished). We previously found that the ability of fz overexpression to produce this gain-of-function phenotype is blocked in a dsh mutant (but not a pk mutant). This argues that dsh, but not pk, is downstream of fz and required for fz signal transduction (Krasnow et al, 1995). When we overexpressed fz in a ds mutant wing we obtained an in-like phenotype (Table 1), leading us to conclude that ds is not required for fz signal transduction. It is notable that dsUA071/ds33K; hs-fz/+ flies that were not heat shocked showed a typical ds polarity pattern. In previous experiments we found that a hs-fz transgene provided enough function at 25°C (without any heat shocks) to provide substantial rescue of a null fz genotype (Krasnow and Adler, 1994). That we did not see any substantial rescue of the ds polarity pattern in these flies argues that ds does not produce a mutant polarity pattern by interfering with transcription of the fz gene.

ds alters the direction of fz intercellular signaling

We previously found that fz clones (for most alleles including null alleles) displayed distal domineering cell nonautonomy (Vinson and Adler, 1987). That is, wild-type cells distal (and in part anterior and posterior) but not proximal to the clone showed abnormal polarity. We have used this property of fz clones as an assay for fz intercellular signaling (i.e the cell nonautonomous function of fz). To determine if ds was required for fz signaling we generated fz strb clones in dsUA071/ds33K wings. The interpretation of this experiment is more challenging than for an equivalent experiment in a wild-type wing. In a wild-type wing carrying a fz clone all hairs point distally, except those influenced by the fz clone. This is not the case in a ds wing, which has large regions with abnormal polarity. We first examined fz clones found in regions of ds wings where polarity is typically normal (i.e. distal). In such regions we scored 24 fz strb clones all of which showed distally directed domineering cell nonautonomy (Fig. 6A,B). Indeed, the domineering nonautonomy appeared much enhanced over that seen for fz clones in a wild-type wing (compare with Fig. 1 in Jones et al., 1996). In the ds background we saw substantial regions of hairs with completely reversed polarity distal to the clone. Indeed, in some cases (e.g. Fig. 6A) hairs on all sides of the clone pointed directly toward the clone boundary. We confirmed that ds resulted in enhanced domineering cell nonautonomy of fz clones by comparing the extent of nonautonomy (see methods for details) in fz clones induced in either wild-type or ds wings (Table 2). Thus, ds can be considered to be a negative regulator of fz domineering nonautonomy.

Table 2.

fz nonautonomy is enhanced in ds mutant wings

fz nonautonomy is enhanced in ds mutant wings
fz nonautonomy is enhanced in ds mutant wings
Fig. 6.

fz strb clones in a dsUAO71/ds33k wing. The boundary of the clones is outlined. The arrows point in the direction of the hairs in that location. D is distal and P proximal. Note the extreme nonautonomy shown by all three clones. (A,B) Clones in regions that normally show distal polarity in ds wings. The nonautonomy here is distal, downstream and attractive domineering nonautonomy. (C) Hair polarity is typically proximal in this region on the ventral surface of a ds wing. This clone shows what appears to be proximal, downstream and attractive domineering nonautonomy. Due to this wing region not being flat we needed to assemble this panel using individual images from several different focal planes.

Fig. 6.

fz strb clones in a dsUAO71/ds33k wing. The boundary of the clones is outlined. The arrows point in the direction of the hairs in that location. D is distal and P proximal. Note the extreme nonautonomy shown by all three clones. (A,B) Clones in regions that normally show distal polarity in ds wings. The nonautonomy here is distal, downstream and attractive domineering nonautonomy. (C) Hair polarity is typically proximal in this region on the ventral surface of a ds wing. This clone shows what appears to be proximal, downstream and attractive domineering nonautonomy. Due to this wing region not being flat we needed to assemble this panel using individual images from several different focal planes.

Fig. 7.

Drawings of a wild-type wing and a region of a ds wing that shows proximal polarity. Distal is to the left and proximal to the right. The filled arrows represent hairs produced by cells that are wild type for fz; the open rectangles, cells mutant for fz. In the upper panels are wings with no fz clones, and the lower panels show the effects of fz clones in an otherwise wild-type or ds wing. Note that in the ds wing the anatomical direction of fz nonautonomy is now proximal and not distal as in wild type.

Fig. 7.

Drawings of a wild-type wing and a region of a ds wing that shows proximal polarity. Distal is to the left and proximal to the right. The filled arrows represent hairs produced by cells that are wild type for fz; the open rectangles, cells mutant for fz. In the upper panels are wings with no fz clones, and the lower panels show the effects of fz clones in an otherwise wild-type or ds wing. Note that in the ds wing the anatomical direction of fz nonautonomy is now proximal and not distal as in wild type.

We next examined fz strb clones in regions of ds wings where hair polarity is typically proximal (Figs 6C, 7). In such regions of ds wings we found 9 fz strb clones. For all of these clones, hairs both proximal and distal to the clone pointed toward the clone (Figs 6C, 7). Hairs proximal to the clone showed distal polarity while hairs distal to these clones showed proximal polarity. We assume that hair polarity in these regions would have been proximal in the absence of the clone. Thus, the fz clone resulted in hairs proximal to the clone changing their polarity from proximal to distal (Fig. 7). Hence, these fz clones displayed proximal domineering nonautonomy. This argues that the normal distal direction of fz signaling (as defined by the direction of nonautonomy) was reversed in these regions of ds wings and was now proximal. We conclude that one of the consequences of ds mutations is that in regions of altered hair polarity (e.g. proximal polarity) the direction of fz signaling is similarly altered (i.e. we see proximal domineering nonautonomy).

ds and cell adhesion

Cadherins are known to be mediators of Ca2+ mediated cell adhesion. The unusual structure of the Ds and Ft proteins raised the possibility that these proteins, which contain multiple cadherin motifs, might not function in cell adhesion. Finding that ds clones on the wing are biased towards an oval shape (minimizing contact with wild-type cells), and that the clones often bulged from the wing surface, supports the previous suggestions that ds functions in cell adhesion (Clark et al., 1995). In addition, we found that ds pupal wings were much more fragile than wild-type pupal wings, resulting in them frequently falling apart during manipulation. This also seems likely to be due to an adhesion defect. At least some of the tissue polarity phenotypes associated with ds however, resemble those seen in fz pathway mutants and seem unlikely to be due to an effect on cell adhesion. For example, the use of an altered subcellular location for prehair initiation seems unlikely to be due to altered cell adhesion. When we examined double mutants between fz pathway genes and ds we found they had the altered body shape of ds, but a tissue polarity phenotype that resembled the fz pathway mutation. That the body shape defects of ds are not suppressed in a fly that is doubly mutant for fz and ds indicates that the abnormal fz signaling seen in ds wings is not the cause of the body shape abnormalities. These observations do not rule out the possibility that the body shape defects cause the abnormalities in fz signaling, although we do not think this likely.

ds and tissue polarity on the wing

It is not terribly surprising that ds function is essential for the development of normal wing tissue polarity. Held and colleagues (1986) found that ds mutations disrupt the polarity of bristles on the legs and that they cause leg joint abnormalities that are similar to those produced by fz and sple. The overall morphological abnormalities caused by ds mutations are greater on the leg, but it appears that the effects on the wing and leg may be parallel.

ds mutations do not inactivate the fz pathway

Previous work has shown that the function of the fz pathway is essential for the formation of a wing with normal distally pointing hairs (Wong and Adler, 1993). This pathway is thought to be composed of both an intercellular signaling system and an intracellular signal transduction system (Vinson and Adler, 1987; Wong and Adler, 1993; Park et al., 1994). The cell autonomous function of this pathway leads to the restriction of prehair initiation to the vicinity of the distal vertex (Wong and Adler, 1993). Previous studies have argued that mutations in genes such as dsh, in, fy and RhoA produced a tissue polarity phenotype by interfering with the intracellular transduction of the fz signal (Wong and Adler, 1993; Krasnow et al., 1995; Strutt et al., 1997). We have found that ds is not required for the transduction of the fz signal and consistent with that conclusion it does not cause a polarity pattern that is typical of a lack of or reduction in fz pathway function.

ds causes a tissue polarity phenotype by altering the direction of fz signaling

We found that the correspondence between the direction of hair polarity (distal), the subcellular location for prehair initiation (in the vicinity of the distal vertex) (Wong and Adler, 1993) and the direction of fz domineering nonautonomy (distal) (Vinson and Adler, 1987) seen in wild-type wings was maintained in regions of ds wings with reversed polarity. In such wing regions proximally pointed hairs were formed in the vicinity of the proximal vertex, and fz clones displayed proximal domineering nonautonomy. This leads us to conclude that the cause of the abnormal hair polarity in ds wings is the abnormal direction of fz signaling and that ds cells in these regions are responding normally to an abnormal signal. This is consistent with the conclusion noted above that fz signal transduction does not require ds function. The hypothesis that ds causes a tissue polarity phenotype because it alters the direction of fz signaling also explains the need for fz pathway genes for the ds mutant polarity pattern, ds mutants having a fz-like cellular phenotype and the domineering nonautonomy of ds. It is also consistent with the ability of fz overexpression to suppress the ds polarity phenotype, since fz overexpression can ‘suppress or override the wild-type polarity pattern’ that arises from normal fz signaling.

In other experiments we have also found that in prickle mutant wings the abnormal direction of hair polarity predicts and is coincident with the abnormal direction of fz signaling (R.E. Krasnow and P.N. Adler, unpublished). We note that there is no gross abnormality in wing or body shape in pk mutants as there is in ds mutants, thus an abnormal direction of fz signaling does not rule out having a normal body shape. We think it is significant that these two tissue polarity mutants that do not produce a fz/in polarity pattern (see Fig. 3) both appear to cause a tissue polarity phenotype by altering the direction of fz signaling. We suggest that mutations that produce a fz/in like mutant polarity pattern will do so by inactivating the fz signaling/signal transduction pathway, and that mutations that produce different abnormal polarity patterns will do so by altering the direction of fz signaling.

Why does ds cause altered fz signaling?

Two types of models have been proposed to account for the role of Fz in tissue polarity and to explain the distal domineering nonautonomy of fz clones (Adler et al., 1997). The cell-by-cell signaling model suggests that the binding of ligand at one side of a cell leads to the Fz receptor becoming unevenly activated across the cell (Park et al., 1994). This leads to both prehair initiation and the relaying of the signal being localized at the distal edge of cells, which leads to hair polarity being coincident with the direction of signaling. This model does not specify how the overall polarity is established, but it seems likely that it is imposed by special ‘boundary cells’ located proximally. In this model the domineering distal nonautonomy of fz clones was ascribed to a failure of cells distal to the clone to receive the signal (Vinson and Adler, 1987). The secondary signal model suggests that the Fz receptor is activated in a gradient fashion along the proximal/distal axis of the wing by a long range gradient of a morphogen ligand (Zheng et al. 1995). Fz activation leads to the proportional production of a secondary signal, which acts more locally to polarize cells. In this model the nonautonomy of fz clones is due to a failure of clone cells to produce the secondary signal. A group of proximal cells presumably form the source for the gradient of Fz ligand. Either of these models can easily accommodate the observation that there are regions in a ds wing where the direction of fz signaling is reversed. For example, by hypothesizing that ds mutations result in a change in the fate of cells that then serve either as ectopic sources of the gradient morphogen or locations that initiate cell-by-cell signaling. Such a model might also be able to explain the altered wing shape and wing vein pattern as a consequence of the population of cells with altered cell fate. This hypothesis does not however, explain the need for ds function for tissue polarity development in all regions of the wing as was shown by our analysis of ds clones. Nor does it explain the enhanced domineering nonautonomy of fz that we also observed in ds wings. We therefore prefer hypotheses where ds mutations produce their phenotypic effects by altering the function of the fz pathway in all regions of the wing. The effects of ds mutations on fz signaling are subtle and do not appear to be due to either a simple quantitative loss or gain of function. Perhaps ds mutations alter the efficiency, stability or propagation of a polarity signal in a way that leads to the system becoming destabilized. The regions of reversed polarity might be caused by minor perturbations in fz signaling being amplified and propagated in a ds mutant wing due to the decreased stability of the system.

It is possible that fz signaling takes place at the adherens junction and that ds mutations alter the structure or composition of the junction in a way that alters fz signaling. For example, ds could promote the assembly of a Fz receptor complex at the junction. An ‘incomplete’ complex formed in a ds mutant might be unstable leading to aberrant signaling. Alternatively, the effect of ds on fz signaling could be more indirect as is discussed below.

fz, Wnts and cadherins

It is now clear that the fz family of receptors recognize Wnt’s as ligands (Bhanot et al., 1996). wingless, which encodes the prototypical Wnt in Drosophila, has important signaling functions in Drosophila development (Ingham,1996; Eaton and Cohen, 1996; Peifer, 1995). The arm gene, which encodes a β-catenin is essential for the transduction of the wg signal. In this case the role of arm is not linked to it’s role in cell adhesion, but rather to it’s role in signal transduction (Heasman et al., 1994; Peifer et al., 1994; Brunner et al., 1997; van de Wetering et al., 1997). Since no mutants in Wnt encoding genes have yet been found to have a fz-like wing tissue polarity phenotype the identity of the fz ligand for wing tissue polarity is unclear (but see Theisen et al., 1994). Thus far, only dishevelled has been shown, in a convincing way, to function in both the wg segment polarity and fz tissue polarity pathways (Krasnow et al., 1995; Klingensmith et al., 1994; Theisen et al., 1994). Indeed, experiments argue that arm does not function in tissue polarity (Strutt et al., 1997; Perrimon, 1996; Orsulic and Peifer, 1996; P. N. Adler, unpublished). Our data shows that an unusual cadherin like protein (Ds) is essential for the function of the fz tissue polarity pathway. Perhaps there is a novel Arm-like protein that functions in the fz pathway and which binds the Ds protein. Mutations in ds might affect the function of this pathway indirectly by increasing the free concentration of the putative Arm-like protein. A precedent for this comes from observations that the overexpression of E-cadherin (or portions of this protein) can have effects on wg signal transduction (Heasman et al., 1994; Funayama et al., 1995, Sanson et al., 1996).

We thank Birgit Kerber and Simon Collier for their comments on the manuscript. This work was supported by the NIH (GM37136).

Adler
,
P. N.
(
1992
).
The genetic control of tissue polarity in Drosophila
.
BioEssays
4
,
735
741
.
Adler
,
P. N.
,
Charlton
,
J.
and
Park
,
W. J.
(
1994
).
The Drosophila tissue polarity gene inturned functions prior to wing hair morphogenesis in the regulation of hair polarity and number
.
Genetics
137
,
829
836
.
Adler
,
P. N.
,
Krasnow
,
R. E.
and
Liu
,
J.
(
1997
).
Tissue polarity points from cells of higher toward cells of lower Frizzled levels
.
Curr. Biol.
7
,
940
949
.
Bhanot
,
P.
,
Brink
,
M.
,
Samos
,
C. H.
,
Hsieh
,
J. C.
,
Wang
,
Y.
,
Macke
,
J. P.
,
Andrew
,
D.
,
Nathans
,
J.
and
Nusse
,
R.
(
1996
).
Dfz2, a new member of the frizzled family in Drosophila functions as a receptor for Wingless
.
Nature
382
,
225
230
.
Brunner
,
E.
,
Peter
,
O.
,
Schweizer
,
L.
and
Basler
,
K.
(
1997
).
Pangolin encodes a Lef-1 homolog that acts downstream of Armadillo to transduce the Wingless signal in Drosophila
.
Nature
385
,
829
833
.
Clark
,
H.
,
Brentrup
,
D.
,
Schneitz
,
K.
,
Bieber
,
A.
,
Goiodman
,
C.
and
Noll
,
M.
(
1995
).
Dachsous encodes a member of the cadherin superfamily that controls imaginal disc morphogenesis in Drosophila. Genes Dev.
9
,
1530
1542
.
Eaton
S.
and
Cohen
,
S.
(
1996
).
Wnt signal transduction: More than one way to skin a (β)-cat?
Trends Cell Biol.
6
,
287
280
.
Funayama
,
N.
,
Fagotto
,
F.
,
McCrea
,
P.
and
Gumbiner
,
B. M.
(
1995
).
Embryonic axis induction by the armadillo repeat domain of beta-catenin: evidence for intracellular signaling
.
J. Cell Biol.
128
,
959
968
.
Geiger
,
B.
and
Ayalon
,
O.
(
1992
).
Cadherins
.
Annu. Rev. Cell Biol.
8
,
307
332
.
Golic
,
K. G.
and
Lindquist
,
S.
(
1989
).
The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome
.
Cell
59
,
499
509
.
Gubb
,
D.
(
1993
).
Genes controlling cellular polarity in Drosophila
.
Development
supplement,
269
277
.
Gubb
,
D.
and
Garcia-Bellido
,
A.
(
1982
).
A genetic analysis of the determination of cuticular polarity during development in Drosophila melanogaster
.
J. Embryol. Exp. Morph.
68
,
37
57
.
Gubb
,
D.
,
Shelton
,
M.
,
Roote
,
J.
,
McGill
,
S.
, and
Ashburner
,
M.
(
1984
).
The genetic analysis of a large transposing element of Drosophila melanogaster: the insertion of a w+ rst+TE into the ck locus
.
Genetics
91
,
54
64
.
Gumbiner
,
B.
(
1995
).
Signal transduction by β-catenin
.
Curr. Opin. Cell Biol.
7
,
634
640
.
Heasman
,
J.
,
Crawford
,
A.
,
Goldstone
,
K.
,
Garner-Hamrick
,
P.
,
Gumbiner
,
B.
,
McCrea
,
P.
,
Kinter
,
C.
,
Noro
,
C. Y.
, and
Wylie
,
C.
(
1994
).
Overexpression of cadherin and underexpression of β-catenin inhibit dorsal mesoderm induction in early Xenopus embryos
.
Cell
79
,
791
803
Held
,
L. I.
,
Duarte
,
C. M.
and
Derakhshanian
,
K.
(
1986
).
Extra tarsal joints and abnormal cuticular polarities in various mutants of Drosophila melanogaster
.
Roux Arch.
195
,
145
157
.
Ingham.
P.W.
(
1996
).
Has the quest for a Wnt receptor finally frizzled out
.
Trends Genet.
12
,
382
384
.
Jones
,
K. H.
,
Liu
,
J.
and
Adler
,
P. N.
(
1996
).
Molecular analysis of EMS induced frizzled mutations in Drosophila melanogaster
.
Genetics
142
,
205
215
.
Klingensmith
,
J.
,
Nusse
,
R.
and
Perrimon
,
N.
(
1994
).
The Drosophila segment polarity gene disevelled encodes a novel protein required for response to the wingless signal
.
Genes. Dev.
8
,
118
130
.
Krasnow
,
R. E.
, and
Adler
,
P. N.
(
1994
).
A single frizzled protein has a dual function in tissue polarity
.
Development
120
,
1883
1893
.
Krasnow
,
R. E.
,
Wong
,
L. L.
and
Adler
,
P. N.
(
1995
).
Dishevelled is a component of the frizzled signaling pathway in Drosophila
.
Development
121
,
4095
4102
.
Mahoney
,
P. A.
,
Weber
,
U.
,
Onofrechuk
,
P.
,
Beissmann
,
H.
,
Bryant
,
P. J.
and
Goodman
,
C.
(
1991
).
The fat tumor suppressor gene in Drosophila encodes a novel member of the cadherin gene superfamily
.
Cell
67
,
853
868
.
Nagafuchi
,
A.
and
Takeichi
,
M.
(
1988
).
Cell binding fucntion of E-cadherin is regulated by the cytoplasmic domain
.
EMBO J.
7
,
3679
3684
.
Orsulic
,
S.
and
Peifer
,
M.
(
1996
).
Cell-cell signalling: Wingless lands at last
.
Curr. Biol.
6
,
1363
1367
.
Park
,
W. J.
,
Liu
,
J.
and
Adler
,
P. N.
(
1994
).
The frizzled gene of Drosophila encodes a membrane protein with an odd number of transmembrane domains
.
Mech. Dev.
45
,
127
137
.
Park
,
W. J.
,
Liu
,
J.
and
Adler
,
P. N.
(
1996
).
The Drosophila tissue polarity gene inturned acts cell autonomously and encodes a novel protein
.
Development
122
,
961
969
.
Peifer
,
M.
(
1995
).
Cell adhesion and signal transduction: the Armadillo connection
.
Trends Cell Biol.
5
,
224
229
.
Peifer
,
M.
,
Sweeton
,
D.
,
Casey
,
M.
and
Wieschaus
,
E.
(
1994
).
Wingless signal and Zeste-white 3 kinase trigger opposing changes in the intracellular distribution of Armadillo
.
Development
120
,
369
380
.
Peifer
,
M.
and
Wieschaus
,
E.
(
1990
).
The segment polarity gene armadillo encodes a functionally modular protein that is the Drosophila homolog of human plakoglobin
.
Cell
63
,
1167
1178
.
Peifer
,
M.
,
Rauskolb
,
C.
,
Williams
,
M.
,
Riggleman
,
B.
and
Wieschaus
,
E.
(
1991
).
The segment polarity gene armadillo interacts with the wingless signaling pathway in both embryonic and adult pattern formation
.
Development
111
,
1029
1043
.
Perrimon
,
N.
(
1996
).
Serpentine proteins slither into the Wingless and Hedgehog fields
.
Cell
86
,
512
516
.
Sanson
,
B.
,
White
,
P.
, and
Vincent
,
J-P
(
1996
).
Uncoupling cadherin vbased adhesion from wingless signalling in Drosophila
.
Nature
383
,
627
629
.
Strutt
,
D.I.
,
U.
Weber
and
M.
Mlodzik
(
1997
).
The role of RhoA in tissue polarity and Frizzled signaling
.
Nature
387
,
292
295
.
Tepass
,
U.
,
Gruszynski-DeFeo
,
E.
,
Haag
,
T.A.
,
Omatyar
,
L.
,
Torok
,
T.
and
Hartenstein
,
V.
(
1996
).
Shotgun encodes Drosophila E-cadherin and is preferentially required during cell rearrangement in the neurectoderm and other morphologically active epithelia. Genes Dev.
10
,
672
685
.
Theisen
,
J.
,
Purcell
,
J.
,
Bennett
,
M.
,
Kansagara
,
D.
,
Syed
,
A.
and
Marsh
,
L.
(
1994
).
dishevelled is required during wingless signaling to establish both cell polarity and cell identity
.
Development
120
,
347
360
.
Uemura
,
T.
,
Oda
,
Hiroki
,
Kraut
,
R.
,
Hayashi
,
S.
,
Kataoka
,
Y.
and
Takeichi
,
M
(
1996
).
Zygotic Drosophila E-cadherin expression is required for processes of dynamic epithelial cell rearrangement in the Drosophila embryo
.
Genes Dev.
10
,
659
671
.
Vinson
,
C.
and
Adler
,
P. N.
(
1987
).
Directional non-cell autonomy and the transmission of polarity information by the frizzled gene of Drosophila
.
Nature
329
,
549
551
van de Wetering
,
M.
,
Cavallo
,
R.
,
Doijes
,
D.
,
von Beest
,
M.
,
von Es
,
J.
,
Loureiro
,
J.
,
Ypama
,
A.
,
Hursh
,
D.
,
Jones
,
T.
,
Bejsovec
,
A.
,
Peifer
,
M.
,
Mortin
,
M.
and
Clever
,
H.
(
1997
).
Armadillo coactivates transcription dirven by the product fo the Drosophila segment polarity gene TCF
.
Cell
88
,
789
799
.
Wong
,
L. L.
and
Adler
,
P. N.
(
1993
).
Tissue polarity genes of Drosophila regulate the subcellular location for prehair initiation in pupal wing cell
.
J. Cell Biol.
123
,
209
221
.
Xu
,
T.
and
Rubin
,
G. M.
(
1993
).
Analysis of genetic mosaics in developing and adult Drosophila tissues
.
Development
117
,
1223
1237
Zheng
,
L.
,
Zhang
,
J.
and
Carthew
,
R.
(
1995
).
frizzled regulates mirror-symmetric pattern formation in the Drosophila eye
.
Development
,
121
,
3045
3055
.