The tissue polarity gene fuzzy (fy) has two roles in the devel-opment of Drosophila wing hairs. One is to specify the correct orientation of the hair by limiting the site of prehair initiation to the distal vertex of the wing cell. The other is to control wing cell hair number by maintaining the integrity of the cytoskeletal components that direct hair development. The requirement for fy in these processes is temperature dependent, as the amorphic fy phenotype is cold sensitive. Analysis of mosaic wings has shown that the fy gene product functions cell autonomously. We have cloned the fy transcript, which encodes a novel four-pass transmembrane protein that shares significant homology with proteins encoded by vertebrate cDNAs. The fourth putative transmembrane domain does not appear to play a significant role in tissue polarity as it is deleted in a weak fy hypomorph. Expression of the fy transcript is develop-mentally regulated and peaks sharply at the time of wing cell pre-hair initiation.

The regular orientation of pattern elements on an epithelium is a common feature of animal morphology. One striking example is the organisation into parallel arrays of the bristles and hairs on the cuticle of adult insects. To produce such regular arrays, the cells of the cuticular epithelium must be polarised within the plane of the epithelium and this planar polarity must be aligned with respect to the body axes. Mutations in the Drosophila tissue polarity genes result in novel patterns of bristle and hair orientation (Gubb and Garcia-Bellido, 1982; Adler, 1992; Gubb, 1993). These patterns indicate that epithe-lial cells in tissue polarity mutants still acquire a planar polarity, although the bristle and hair alignments differ from the wild-type pattern. The study of tissue polarity genes should, therefore, provide some understanding of the mechanism by which polarised epithelial cells are aligned with the body axes. In this report, we present the phenotypic and molecular characterisation of the tissue polarity gene fuzzy (fy).

All tissue polarity mutations affect the development of wing hairs. In the wild-type wing, each cell secretes a single distally pointing hair, the initiation of which is preceded by the accu-mulation of F-actin at the distal vertex of the hexagonal pupal wing cell. In the pupal wing of tissue polarity mutants, the sub-cellular distribution of F-actin is altered and, as a direct con-sequence, the final orientation of cell hairs is changed (Wong and Adler, 1993). On the basis of differences in development of their wing hairs, the tissue polarity genes have been placed in three classes. In the pupal wing of group I mutants (frizzled (fz), dishevelled (dsh) and prickle-spiny legs (pk-sple)), the F-actin bundle is frequently displaced towards the apical centre of the cell and a single hair with aberrant polarity results. In group II mutants (inturned (in) and fy) up to three actin bundles form at the cell periphery, so that many cells secrete extra hairs as well as displaying mutant hair polarity. Group III consists of a single gene, multiple wing hairs (mwh). The pupal wing cells of mwh show a similar F-actin distribution to the group II mutants, but have additional late-forming actin bundles that produce small secondary hairs. The extra hairs formed in fy and in mutants suggest a role for these genes in maintaining the integrity of cytoskeletal components required for wing hair development, as well as in hair polarity. A similar function has been proposed for RacI, a member of the rho family of small GTPases, as the expression of a dominant negative form (RacN17) in the wing phenocopies the multiple hair phenotype of the group II genes (Eaton et al., 1995, 1996).

The wing hair phenotypes of combinations of the tissue polarity mutations have suggested epistatic relationships between the three classes of genes (Wong and Adler, 1993). These interpretations are based upon the multiple hair phenotypes. The possible epistatic relationships between these genes based on the polarity phenotypes are less clear cut, although both mwh and in are close to being epistatic to fy and pk (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993; Coulson, 1994). Flies mutant for fy or in, in combina-tion with the group I mutations, have a wing hair phenotype that more closely resembles the fy or in single mutant. In contrast, the double mutant combination of fy or in with mwh more closely resembles mwh. These observations have led to the proposal of a model in which fy and in function down-stream of the group I genes and upstream of mwh in a regu-latory pathway that controls the site of hair initiation (Wong and Adler, 1993).

The fz gene encodes a seven-pass transmembrane protein that functions both non-cell autonomously, to transmit tissue polarity information, and cell autonomously, to direct the site prehair initiation (Vinson et al., 1989; Krasnow and Adler, 1994). The cell-autonomous function of Fz is proposed to act through the product of the dsh gene (Krasnow et al., 1995), which encodes an evolutionary conserved protein that is also part of the wingless (wg) signaling pathway (Theisen et al., 1994; Klingensmith et al., 1994; Noordermeer et al., 1994). Both Fz and an homologous Drosophila protein, Dfz2, have been shown to mediate wg signaling in a cell culture assay (Bhanot et al., 1996). On the basis of its in vivo expression pattern Fz2 has been proposed to be a receptor for wg, although an analogous Wnt ligand for Fz has not been identified. Therefore, the wg and tissue polarity signaling pathways act through Fz-like receptors and converge on dsh. Although the downstream effectors of wg signaling in Drosophila have been well characterised (see, for example, Klingensmith and Nusse, 1994), little is known about the genes proposed to act down-stream of fz and dsh in the tissue polarity pathway. To date, only the in gene has been molecularly characterised and found to encode a novel protein containing two putative membrane-spanning domains (Park et al., 1996). Here we show that fy, a second downstream component of this pathway, also encodes a novel transmembrane protein.

Phenotypic analysis

Flies were cultured at 25°C on yeasted cornmeal agar unless otherwise stated. Second chromosomes used for mapping and for phenotypic analysis were fy2bw, fy3cn bw sp, T(Y;2) fy4cn bw sp, cn bw sp, fy5b pr, fy6b pr cn, Df(2L)N22-14 and Df(2L)N22-5. Wings for phenotypic analysis were dissected from flies that had been stored at −70°C or in a 1:1 solution of 70% ethanol:glycerol at room temperature and were mounted in aquamount (BDH Ltd) with isopropanol. A representative hair polarity phenotype for each genotype was drawn after examining at least six mounted wings. Mitotic clones of fy2 were generated by crossing fy2bw males to f36a; f+30B M(2)z/CyO virgin females and irradiating F1 larvae (dose 1000 R; 300 R/minute, 100 kV, 15 mA, 2 mm Al filter) at between 72 and 96 hours after egg laying (AEL). Wings from irradiated f36a; f+30B M(2)z/fy2bw males were mounted as described above and f36a; fy2M+ clones identified by the f wing hair phenotype.

Nucleotide sequence analysis of cDNA and genomic clones

EcoRI restriction fragments from the genomic phage clone W5-p9 (Neumann-Silberberg and Schuepbach, 1993) that flank the site of the fy3 deletion were used to screen 5×105 clones of an Oregon R imaginal disc plasmid cDNA library by the protocols described in Brown and Kafatos (1988). The longest fy cDNA recovered (fcJ) was subcloned into pBluescript SK+ (Stratagene Ltd.) and deletion constructs from 5′ to 3′ of the sense strand made using exonuclease III. Single-stranded templates were produced by using the R408 helper phage (Stratagene Ltd) and were sequenced from the M13-20 primer following the Sequenase protocol (US Biochemical Corp.). Sequence data from these clones were assembled into a contig using the GCG (Genetics Computer Group, Inc.) GELASSEMBLE program. The fcJ cDNA was subcloned into pBluescript KS+ and the antisense strand sequenced from oligonu-cleotide primers complementary to the sense strand by the same protocol. The same primers were used to characterise subclones of a second fy cDNA (FcH) and the 3.2 kb BamHI fragment from phage W5-p9 containing the corresponding region of genomic DNA. The best open reading frame within the composite cDNA sequence was identified by the Positional Base Preference method in the Standard Staden Programs for Nucleotide Interpretation. The encoded polypeptide was compared to protein database entries using the Blastp and Fasta programs and to DNA databases using the tBlastn program (Altschul et al., 1990; Pearson and Lipman, 1988). Predic-tion of helical transmembrane domains was performed by the PHDhtm program of the PredictProtein server (Rost et al., 1995) and refined by alignment with the peptide sequence encoded by a hom-ologous D. virilis locus (S. C., unpublished results).

Rescue of fy2 phenotype

The 3.2 kb BamHI restriction fragment spanning the fy transcription unit (see Fig. 4) was subcloned from phage W5-p9 into the pWhite Rabbit transformation vector (Dunin-Borkowski and Brown, 1995). A solution of 1 μg/μl pWhiteRabbitfyB3.2 and 0.25μg/μl pHsπΔ2-3 DNA in Spradling buffer was microinjected into y w embryos following standard protocols (Spradling, 1986). G0 flies were crossed back to y w and w+ G1 progeny crossed to a w; Sco/In(2LR)O, Cy; In(3LR)TM2, Ubx130/In(3LR)TM6B, Hu stock to balance the inserts. To test for rescue of the fy2 phenotype, w; In(2LR)O Cy/+; fyB1/In(3LR)TM2 Ubx130males were crossed to y w; fy2 homozygous females and the fy2/In(2LR)O Cy; fyB1/+ progeny crossed to each other. Wings from w; Cy+ and w+; Cy+ progeny were mounted and hair polarity patterns compared.

Characterisation of fy mutant alleles

The fy3 deletion was characterised by amplifying a DNA fragment spanning the deleted EcoRI restriction site from fy3 homozygote and progenitor genomic DNA using the oligonucleotide primers 5′-GCTACACGGACTGTCTGCTG-3′ and 5′-GCACTCATGGCATGT-GCATG-3′ following the PCR protocol described in Collier et al. (1992). PCR products were purified using the Prep-a-Gene kit (Bio-Rad), heat-denatured and sequenced from the same primers following the standard Sequenase protocols (US Biochemical Corp.). To identify the fy2 and fy5 mutations, the 3.1 kb BamHI/XhoI fragment spanning the fy locus (see Fig. 4) was amplified from fy2 homo-zygote genomic DNA using the flanking oligonucleotide primers 5′-CTGACCACATACTTCTTGATTC-3′ and 5′-CACACATCT-TACAGTTGGATC-3′. The PCR product was digested with BamHI and XhoI, cloned into pBluescript KS+ and characterised by single-stranded sequencing from fy-specific oligonucleotide primers as described above. Clones from three independent PCR reactions were sequenced to eliminate the possibility that base changes had arisen from nucleotide misincorporation by Taq polymerase. The fy2 mutation also introduces a novel BglII restriction site into the fy coding region, the presence of which has been confirmed by com-parative Southern blot analysis.

Northern and in situ analyses of fy expression

For northern analysis, developmentally staged wild-type (Canton-S) animals were homogenised in a guanidinium isothiocyanate solution and total RNA isolated according to the method of Chomczynski and Sacchi (1987). Poly(A)+ RNA was prepared from the total RNA using oligo(dT)-cellulose by the method of Aviv and Leder (1972). Approx-imately 0.5 μg of poly(A)+ RNA from each developmental stage was separated on an agarose-formaldehyde gel, blotted to a nylon membrane and hybridised with radiolabelled FcJ cDNA and Rp49 probes following standard procedures. For whole-mount in situ hybridisation, sense and antisense digoxigenin-labelled RNA probes were prepared using the FcH fy cDNA as a template. Hybridisation of the riboprobes to pupal wings was performed as described in Sturtevant et al. (1993) but with hybridisation and washing steps increased to 55°C.

The fy gene is not easily mutated

The original fy allele (fy1) has been lost. A mutant with a similar phenotype was recovered in an EMS screen and desig-nated fy2 due to its’ similar map position (Grell, 1969), although a test of allelism was not possible. An X-ray muta-genesis screen has been undertaken in this laboratory to create new fy alleles. Of 35,000 chromosomes screened, however, only two (fy3 and fy4) showed a fy phenotype in combination with the fy2 allele (Clare Henchcliffe, unpublished results). Polytene chromosome analysis has shown that the fy3 chro-mosome is cytologically normal and the fy4 mutation is asso-ciated with a T(Y:2) translocation that breaks in the interval 29B4-C2. We have recently acquired two EMS-induced fy alleles fy5 (previously fyJN11) and fy6 (previously fyJN12) isolated in Paul Adler’s laboratory (Univer-sity of Virginia, USA).

The fy2 and fy3 alleles behave as amorphs as they do not give appreciably stronger phenotypes in com-bination with a deletion of the region (Df(2L)N22-14) than when homozygous. The fy5 and fy6 alleles are weak hypomorphs displaying only mild deviations from wild-type wing hair polarity and a low incidence of wing cells displaying extra hairs. Flies of the genotypes fy2/fy5 and fy2/fy6 have significantly stronger phenotypes than fy5/Df(2L)N22-14 and fy6/Df(2L)N22-14, respec-tively, implying that the fy2 allele displays some antimor-phic character in combina-tion with hypomorphic fy alleles. Since stocks homozygous for the fy2 or fy3 alleles can be maintained, the fy gene appears not to be required for viability or fertility. In contrast, the fy4 translocation is lethal in combination with a deletion of the region (Df(2L)N22-14), although this lethality maps to an adjacent locus (Glynnis Johnson and S. C., unpublished results).

The amorphic fy phenotype is cold sensitive

Mutant fy flies display an altered pattern of bristle and hair ori-entation. The microchaetae and macrochaetae of the notum and abdomen are turned towards the midline, rather than pointing posteriorly, and the bristles of the anterior wing margin stands more erect than wild type (Grell, 1969). The cuticular hairs, which usually point distally on the appendages and posteriorly on the rest of the body, are arranged into novel patterns of polarity. The wing hair patterns of fy mutant flies of the same genotype that have been cultured at the same temperature are almost identical. Diagrammatic representations of the hair polarity pattern on the dorsal wing surface of fy3/Df(2L)N22-14 mutant flies cultured at 18°C and 29°C and also the wild-type polarity of cn bw sp progenitor flies are shown in Fig. 1. The hair polarity pattern on the ventral surface of the mutant wing is similar to the dorsal surface at the margins, but shows a more pronounced deviation from wild type towards the interior of the wing blade.

Fig. 1.

Diagrammatic representation of hair polarity patterns on the dorsal wing surface of fy mutant and progenitor flies. Vectors delineate the direction of hair polarity for each region of the wing blade. Shaded areas represent regions for which a polarity could not be assigned either because of variation between individual wings, or because all wings show close to random orientation. Each diagram represents an average phenotype from six wings of flies raised contemporaneously. The erect triple row of the anterior wing margin associated with the fy phenotype is not represented. Next to each diagram is a photograph of the C region of a wing of the same genotype immediately anterior to the posterior cross vein. Phenotypes represented are; (A) cn bw sp (fy3 progenitor) homozygous flies cultured at 18°C, (B) fy3/Df(2L)N22-14 flies cultured at 29°C and (C) fy3/Df(2L)N22-14 flies cultured at 18°C.

Fig. 1.

Diagrammatic representation of hair polarity patterns on the dorsal wing surface of fy mutant and progenitor flies. Vectors delineate the direction of hair polarity for each region of the wing blade. Shaded areas represent regions for which a polarity could not be assigned either because of variation between individual wings, or because all wings show close to random orientation. Each diagram represents an average phenotype from six wings of flies raised contemporaneously. The erect triple row of the anterior wing margin associated with the fy phenotype is not represented. Next to each diagram is a photograph of the C region of a wing of the same genotype immediately anterior to the posterior cross vein. Phenotypes represented are; (A) cn bw sp (fy3 progenitor) homozygous flies cultured at 18°C, (B) fy3/Df(2L)N22-14 flies cultured at 29°C and (C) fy3/Df(2L)N22-14 flies cultured at 18°C.

Flies carrying either the fy2 or the fy3 allele display a significantly stronger phenotype when cultured at 18°C than at 29°C (Fig. 1B,C). The phenotype at 25°C is intermediate but much closer to that at 29°C. The stronger phenotype is char-acterised by an increase in the overall deviation from wild-type hair polarity and the proportion of the wing for which a specific polarity cannot be assigned. These changes are more evident towards the interior of the wing blade than at the margins (Fig. 1B,C). As both the fy2 and fy3 alleles are amorphs, this cold-sensitivity appears to derive from a temperature-dependent requirement for the fy gene product, rather than from the products of these alleles being inherently cold sensitive.

A proportion of fy wing cells secrete up to four hairs instead of the single hair produced by wild-type cells. This phenotype is not uniform across the wing blade. Cells close to the wing margin still secrete a single hair, with the exception of cells of the first two rows next to the posterior margin, between the alula and the distal tip of the L4 vein, which frequently produce two (Fig. 2A). There is also a region of wild-type hair number a few cell diameters either side of the distal region of the wing veins, particularly L3 and L5 on the dorsal surface (Fig. 2B). This possibly correlates with the fact that L3 and L5 are dorsal veins, although a reciprocal phenotype with respect to the ventral veins is less apparent on the ventral wing surface. In general, cell hair number increases with the deviation of hair polarity from wild type (e.g. Fig. 1B,C). An exception to this rule is the behavior of cells with reversed (distal-to-proximal) polarity, which usually retain wild-type hair number (Fig. 2C).

Previous studies have found that fy flies do not have the ectopic inverted joints in the tarsi of the leg that are associated with some tissue polarity mutants (Held et al., 1986; Coulson, 1994). However, double mutant combinations of the fy2 allele with either pk or mwh show ectopic tarsal joints at a low frequency, which are never observed in pk or mwh mutants themselves (Coulson, 1994). These results suggest that fy has a redundant role in tarsal joint development.

Fig. 2.

Examples of the heterogeneity of the fy wing cell hair number phenotype from wings of fy3/Df(2L)N22-14 flies cultured at 18°C. The proximal to distal axis of each wing is from left to right. (a) E region of the dorsal wing surface close to the posterior margin. Most cells retain wild-type hair number but the cells of the row immediately adjacent to the margin secrete two hairs. (B) Region surrounding the junction of the L5 and the posterior cross vein (PCV) on the dorsal surface. Most cells within four rows on both sides of the L5 vein retain wild-type cell hair number and close to wild-type hair polarity. (C) B region of the ventral wing surface. A proportion of cells anterior to the L3 vein display reversed (distal to proximal) hair polarity. The majority of these retain wild-type hair number.

Fig. 2.

Examples of the heterogeneity of the fy wing cell hair number phenotype from wings of fy3/Df(2L)N22-14 flies cultured at 18°C. The proximal to distal axis of each wing is from left to right. (a) E region of the dorsal wing surface close to the posterior margin. Most cells retain wild-type hair number but the cells of the row immediately adjacent to the margin secrete two hairs. (B) Region surrounding the junction of the L5 and the posterior cross vein (PCV) on the dorsal surface. Most cells within four rows on both sides of the L5 vein retain wild-type cell hair number and close to wild-type hair polarity. (C) B region of the ventral wing surface. A proportion of cells anterior to the L3 vein display reversed (distal to proximal) hair polarity. The majority of these retain wild-type hair number.

Clones of fy2 tissue mimic polarity in homozygous fy2 wings

Clones of fy2 mutant tissue marked with forked36a (f36a) were induced in a f+30B background by X-ray-induced mitotic recombination. 20 clones displaying the f36a wing hair phenotype ranging from 20 cells to approximately one third of the dorsal wing surface were examined for changes in hair ori-entation. In each case the changes in planar polarity exhibited by cells within these clones resembles that shown by cells at an equivalent position within a fy2 homozygous wing (Fig. 3). The exceptions were the cells towards the borders of the clones that adopt a polarity that is intermediate between mutant and wild type. It was evident that fy2 clones do not cause domi-neering non-cell autonomy of the degree shown by clones of the fz gene (Gubb and Garcia-Bellido, 1982; Vinson and Adler, 1987), as hair polarity and hair number of fy+ cells surround-ing the clones appeared normal. However, the rescue of the f36a wing hair phenotype by f+30B construct is not complete, which made it difficult to be confident about the precise boundaries of some clones. Therefore it was not possible to rule out a limited degree of non-cell autonomy.

Fig. 3.

Hair polarity of fy2 clone compared to the equivalent region of a fy2 homozygous wing. The proximal to distal axis of both wings is from left to right. (B) f36afy2 clone from a fly cultured at 25°C. The approximate borders of the clone are marked. (A) Equivalent region of the wing of a fy2 homozygous fly cultured at 25°C. Hair polarity within the clone mimics that of the mutant wing with the exception of cells near to the clone boundary which display a polarity intermediate between the mutant and the wild-type orientation.

Fig. 3.

Hair polarity of fy2 clone compared to the equivalent region of a fy2 homozygous wing. The proximal to distal axis of both wings is from left to right. (B) f36afy2 clone from a fly cultured at 25°C. The approximate borders of the clone are marked. (A) Equivalent region of the wing of a fy2 homozygous fly cultured at 25°C. Hair polarity within the clone mimics that of the mutant wing with the exception of cells near to the clone boundary which display a polarity intermediate between the mutant and the wild-type orientation.

The fy locus maps close to gurken

The fy2 allele has a genetic map location of 2-24.1 (Grell, 1969) and had previously been localised to the interval 29C2-29D1.2 using deficiencies of this region recovered by Wustmann et al. (1989; Clare Henchcliffe, personal communication). The cytology of the fy4 translocation (29B4-C2) places the gene towards the distal end of this interval. We have refined this mapping by showing that fy is uncovered by Df(2L)N22-14 (29C1.2;30C8.9), but not by Df(2L)N22-5 (29C3.5;30C8.9). This places the fy locus between the distal endpoints of these two deficiencies (29C1.2;29C3.4), the same interval to which the female-sterile mutation gurken (grk) had previously been mapped. A 48 kb phage walk, initiated from the distal end of YAC clone DY51 and spanning this region, had been under-taken to clone grk (Neumann-Silberberg and Schuepbach, 1993) and was used to assist with the identification of the fy transcript. The fy locus was subsequently localised to the distal end of the grk walk by the identification of a 70 bp deletion on the fy3 chromosome (see below).

The fy transcript encodes a novel transmembrane protein

We used the genomic EcoRI restriction fragments from the grk phage walk that flank the site of the fy3 deletion to screen approximately 5×105 plasmid clones from an imaginal disc cDNA library (Brown and Kafatos, 1988). Three cDNA clones spanning the deletion were isolated (fcH, fcI and fcJ) and also clones representing the transcripts lying immediately proximal and distal (Fig. 4). The two shorter clones fcH (1.6 kb) and fcI (1.1 kb) are polyadenylated at almost identical positions. The longer clone fcJ (1.9 kb) is shorter at the 3′ end and has no poly(A) tail but has 344 bp of additional 5′ sequence. The nucleotide sequences of the fcH and fcJ clones are identical in their over-lapping region and the 1880 bp composite cDNA sequence is indicated in Fig. 5. The site of polyadenylation of the shorter cDNAs indicates that the 5′ to 3′ orientation of the transcript is from proximal to distal on the chromosome (Fig. 4).

Fig. 4.

Molecular map of the fy locus at 29C1,2. Transcripts are indicated by hatched boxes. The size of the fy introns is exaggerated for clarity. The intron/exon structure of the flanking genes is not shown. The Crm1 transcripts encode a peptide with high homology to the Chromosomal Region Maintenance protein of yeast (Adachi and Yanagida,, 1989; Toda et al., 1992, Collier et al. unpublished data). fcG is an uncharacterised cDNA clone from an imaginal disc library. The 3.2 kb BamHI fragment sufficient to rescue the fy2 phenotype is indicated by the lightly shaded box. The regions deleted on the fy3 chromosome and by Df(2L)N22-14 are represented by black boxes. The unshaded portion of Df(2L)N22-14 box indicates the degree of uncertainty concerning the distal breakpoint of the deletion. Restriction sites are B; BamHI, E; EcoRI, H; HindIII and X; XhoI. The lower case ‘e’ is a polymorphic EcoRI site present on cn bw sp and derivative chromosomes.

Fig. 4.

Molecular map of the fy locus at 29C1,2. Transcripts are indicated by hatched boxes. The size of the fy introns is exaggerated for clarity. The intron/exon structure of the flanking genes is not shown. The Crm1 transcripts encode a peptide with high homology to the Chromosomal Region Maintenance protein of yeast (Adachi and Yanagida,, 1989; Toda et al., 1992, Collier et al. unpublished data). fcG is an uncharacterised cDNA clone from an imaginal disc library. The 3.2 kb BamHI fragment sufficient to rescue the fy2 phenotype is indicated by the lightly shaded box. The regions deleted on the fy3 chromosome and by Df(2L)N22-14 are represented by black boxes. The unshaded portion of Df(2L)N22-14 box indicates the degree of uncertainty concerning the distal breakpoint of the deletion. Restriction sites are B; BamHI, E; EcoRI, H; HindIII and X; XhoI. The lower case ‘e’ is a polymorphic EcoRI site present on cn bw sp and derivative chromosomes.

Fig. 5.

Nucleotide sequence of the 2.3 kb HindIII/BamHI genomic fragment that spans the fy locus (see Fig. 4). Sequence present in cDNA clones is shown in upper case. The G to A transition introducing the fy2 nonsense mutation, the 70 bp deleted on the fy3 chromosome and the C to T transition introducing the fy5 nonsense mutation are shown in bold type. A one-letter translation of the fy open reading frame is shown below the nucleotide sequence, numbers refer to amino acid positions within the coding sequence. Amino acids highlighted with bold type and underlined are within the putative membrane-spanning regions TM1 to TM4. GenBank accession no. for the cDNA sequence is AF022891.

Fig. 5.

Nucleotide sequence of the 2.3 kb HindIII/BamHI genomic fragment that spans the fy locus (see Fig. 4). Sequence present in cDNA clones is shown in upper case. The G to A transition introducing the fy2 nonsense mutation, the 70 bp deleted on the fy3 chromosome and the C to T transition introducing the fy5 nonsense mutation are shown in bold type. A one-letter translation of the fy open reading frame is shown below the nucleotide sequence, numbers refer to amino acid positions within the coding sequence. Amino acids highlighted with bold type and underlined are within the putative membrane-spanning regions TM1 to TM4. GenBank accession no. for the cDNA sequence is AF022891.

The longest open reading frame (ORF) within the composite cDNA sequence is 416 amino acids in length (Fig. 5). The putative start codon is not embedded within a good Drosophila consensus sequence (Cavener and Ray, 1991), but alignment with the ORF encoded by a homologous Drosophila virilis locus suggests that it is indeed the start site for translation (S. C., unpublished results). The encoded polypeptide is predicted by the PHDhtm neural network system to have four membrane-spanning helices (Rost et al., 1995, see Fig. 5), but otherwise contains no recognisable functional motifs. The primary sequence of the fy gene product has been compared to protein databases using BLAST (Altschul et al., 1990) and FASTA (Pearson and Lipman, 1988) homology searches, but no sig-nificant matches were found. However, the conceptual transla-tion of a human-expressed sequence tag (EST, I.M.A.G.E. clone 45228) from an infant brain cDNA library gives a peptide with 32% identity to a 119 amino acid portion of the fy sequence between the first (TM1) and the third (TM3) putative membrane-spanning regions (Fig. 6). A mouse embryonic cDNA (Life Tech clone 555546) encodes a peptide with similar degree of homology. We conclude that the fy gene product is a novel transmembrane protein that contains at least one domain that is conserved in vertebrate proteins.

Fig. 6.

Alignment by homology of the primary peptide sequences encoded by the fy gene and by Expressed Sequence Tags from human (EST, I.M.A.G.E. clone 45228) and mouse (Life Tech clone 555546) using the ClustalW program. Identical residues are shaded and homologous residues boxed. Our preliminary nucleotide sequence analysis of clone 45228 has identified an error in the sequence of the database entry that results in a reading frame shift and that has been corrected to create this alignment.

Fig. 6.

Alignment by homology of the primary peptide sequences encoded by the fy gene and by Expressed Sequence Tags from human (EST, I.M.A.G.E. clone 45228) and mouse (Life Tech clone 555546) using the ClustalW program. Identical residues are shaded and homologous residues boxed. Our preliminary nucleotide sequence analysis of clone 45228 has identified an error in the sequence of the database entry that results in a reading frame shift and that has been corrected to create this alignment.

The 1.9 kb fy transcript maps entirely within a 2.3 kb BamHI/HindIII genomic restriction fragment. The alignment of the nucleotide sequence of this fragment with that of the composite cDNA has identified two small introns, of 58 and 59 bp, that lie towards the 5′ and 3′ ends of the open reading frame, respectively. The nucleotide sequence of the BamHI/HindIII fragment is presented in Fig. 5.

A small genomic region completely rescues the fy2 phenotype

The 3.2 kb genomic BamHI fragment that spans the fy transcript (Fig. 4) was cloned into the wRabbit transformation vector and introduced into flies by P-element-mediated trans-formation. A single copy of the construct on the third chromosome (fyB1) has proved sufficient to rescue completely the fy2 homozygous mutant phenotype. Individual rescued flies, most noticeably those cultured at 18°C, can display occasional wing cells with doubled hairs, but this is not consistent from animal to animal. In addition to the fy transcription unit, the construct contains just 50 bp downstream of the site of polyadenylation and the small region (around 600bp) of untranscribed sequence between the 5′ ends of the fy and Crm1 loci. The complete rescue of the fy2 phenotype by a small genomic fragment may mean that all elements necessary for the wild-type expression of the fy transcript are contained within this region. However, as the fz, dsh and in phenotypes can be rescued by uniform expression driven by the hsp70 promotor (Krasnow and Adler, 1994; Klingensmith et al., 1994; Park et al., 1996), there are precedents for tissue polarity gene function not being dependent upon wild type expression.

Both strong and weak fy alleles encode truncated proteins

Phenotypically the fy2 and fy3 alleles behave as amorphs as they are not appreciably stronger in combination with a deficiency of the region than when homozygous. The fy5 allele is an extremely weak hypomorph that displays almost wild-type wing hair polarity and number. Molecularly the fy3 chromosome lacks an EcoRI site present in its progenitor strain (cn bw sp) and BamHI and HindIII restriction fragments spanning this site are slightly reduced in size. Direct sequencing of a PCR product spanning the missing EcoRI site amplified from fy3 homozygote DNA, has identified a deletion of 70 bp in comparison with cn bw sp. The deletion is entirely within the ORF of the cDNAs (Fig. 5) and is predicted to result in a protein product in which the 210 C-terminal amino acids are replaced with just two residues, a proline and an isoleucine. The fy2 and fy5 lesions were identified by sequencing clones of PCR products spanning the fy coding region that had been amplified from homozygote genomic DNA. All fy2 clones contained a G to A transition within the ORF that introduces a nonsense mutation (Fig. 5) and is predicted to result in a protein lacking the 220 C-terminal amino acids. Conse-quently the fy2 and fy3 alleles encode similar truncated proteins that lack the two putative C-terminal membrane-spanning domains and interrupt the region of homology shared with products encoded by human and mouse ESTs (Fig. 6). The fy5 allele has a C to T transition at codon 331 that introduces a nonsense mutation resulting in a protein lacking the fourth putative membrane domain. The very weak phenotype of the fy5 allele suggests that this part of the protein, which is not conserved in the human I.M.A.G.E. clone 45228 (S. C., unpub-lished results), is not required for the greater part of the fy gene’s role in tissue polarity specification.

Expression of the fy transcript is developmentally regulated

Hybridisation of the longest fy cDNA (fcJ) to a northern blot of poly(A)+ RNA isolated from wild-type (Canton-S) animals at a series of developmental stages is shown in Fig. 7. An mRNA species of just under 1.9 kb was identified consistent with the length of the composite cDNA sequence presented in Fig. 5. A low level of expression of this transcript is detectable at all stages of development, with a relatively strong peak of expression in the 2-day-old pupa that coincides with the period of prehair initiation. Significant expression is also detectable in the embryo and female, but not male, adult flies suggesting that the embryonic expression has a maternally derived component. The maternal and embryonic transcripts appear slightly shorter than the pupal transcript (Fig. 7). As no evidence has been found for alternatively spliced fy tran-scripts, this size difference may reflect differences in polyadenylation.

Fig. 7.

(A) Hybridisation of a radiolabeled fy cDNA (fcJ) probe to a northern blot of poly(A)+ RNA from Drosophila developmental stages. Stages are; E; embryo, L1; first larval instar, L2; second larval instar, L3; third larval instar (ea; early, la; late), pP; prepupal, P1; first day pupa, P2; second day pupa, P3; third day pupa; M; adult male, F; adult female. Sizes refer to migration of RNA marker bands. Hybridisation of an Rp49 probe to the same blot is shown below as an indication of the relative amounts of RNA loaded. (B) Hybridisation of a digoxigenin-labeled fy antisense riboprobe to a pupal wing 31 hours after pupal formation. The distal ends of the L2, L3, L4 and L5 veins are indicated.

Fig. 7.

(A) Hybridisation of a radiolabeled fy cDNA (fcJ) probe to a northern blot of poly(A)+ RNA from Drosophila developmental stages. Stages are; E; embryo, L1; first larval instar, L2; second larval instar, L3; third larval instar (ea; early, la; late), pP; prepupal, P1; first day pupa, P2; second day pupa, P3; third day pupa; M; adult male, F; adult female. Sizes refer to migration of RNA marker bands. Hybridisation of an Rp49 probe to the same blot is shown below as an indication of the relative amounts of RNA loaded. (B) Hybridisation of a digoxigenin-labeled fy antisense riboprobe to a pupal wing 31 hours after pupal formation. The distal ends of the L2, L3, L4 and L5 veins are indicated.

We have hybridised pupal wings at the time of highest fy expression with sense and antisense riboprobes. The pattern shown by the antisense strand is shown in Fig. 7B, the sense strand did not produce significant hybridisation. At this stage the fy transcript is present in all wing cells consistent with the fy gene having a cell-autonomous phenotype that affects all regions of the wing blade. The transcript is relatively abundant adjacent to the wing margin and in bands of cells flanking each of the wing veins with the exception of L2. This pattern may simply reflect cell densities in the pupal wing at this stage. The lack of staining surrounding L2, however, is reminiscent of the BrdU-staining patterns at 15-20 hours after pupation (Schuebiger and Palka, 1987; Milan et al., 1996), raising the possibility that fy expression is related to the exit of pupal wing cells from their last cell division.

The fy wing hair polarity phenotype

Mutations in the fy gene result in new patterns of wing hair polarity (see Fig. 1B,C). As a rule, hairs in the anterior third of the fy mutant wing point more anteriorly than wild type and those in the remainder of the wing point more posteriorly than normal. Similar patterns are produced by the other tissue polarity mutations (with the exception of the pk single mutant where the pattern is almost reversed) and also by the double mutant combinations of tissue polarity genes (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993; Coulson, 1994). It seems, therefore, that there is an underlying pattern that results from reduced tissue polarity gene activity. The existence of such a ‘default’ pattern would imply that mutant cells still express a planar polarity that depends upon their position within the wing blade, even without information from the tissue polarity gene pathway. It is not clear, however, what information the mutant cells are responding to. Tissue polarity mutant hair patterns do not appear to be influenced by the anterior-posterior compartment boundary of the wing and the domineering non-cell-autonomy associated with fz clones can cross the compartment boundary (Gubb and Garcia-Bellido, 1982). These observations suggest that the orientation of a wing cell’s planar polarity is independent of its lineage.

It is clear that the fy phenotype is influenced by the presence of wing veins. Veins often correspond to discontinuities in the polarity pattern (see Fig. 1B,C) and mutations that remove wing veins also remove the associated discontinuities in the fy wing hair pattern (S. C., unpublished observations). Cells sur-rounding the wing veins often show close to wild-type polarity (see Fig. 2B), implying that veins can direct polarity in the absence of fy gene function. However, wing vein differen-tiation is not required to align cell polarity in the wing as the veinlet (ve) vein (vn) double mutant combination, which elim-inates all veins but L1, displays normal hair polarity.

The fy wing hair number phenotype

One property of the fy wing phenotype that is shared by in and mwh, is a high incidence of cells producing two or more hairs. As a rule, the further a mutant cell’s planar polarity is from wild type, the more likely it is to produce multiple hairs (see for instance Fig. 1B). There are, however, two pieces of evidence from the fy mutant wing that suggest that mutant cell hair number is separable from hair polarity. First, the cells that are most strongly mutant for polarity, i.e. those producing hairs pointing from distal to proximal, usually retain wild-type hair number (Fig. 2C). Second, although the cells close to the posterior wing margin show a comparable hair polarity, only those in the two rows immediately adjacent to the margin display an abnormal hair number (Fig. 2A). Indeed, these are the cells that are most likely to show multiple hairs on fy5 mutant wings that retain close to wild-type hair polarity. In addition, some combinations of fy alleles display an incidence of split, rather than doubled, wing hairs. Since wing hairs have been proposed to elongate from the tip (Eaton et al., 1996), the formation of split hairs suggests that fy has a role in main-taining the integrity of the hair as well as controlling its initiation. Pupal wings treated with microtubule antagonists phenocopy the multiple hair phenotype of fy wing cells (Chris Turner and Paul Adler, personal communication), suggesting that fy is required for the integrity of the microtubule arrays that direct prehair development. The cold-sensitivity of amorphic fy alleles is consistent with a role in stabilising microtubules as microtubules are inherently cold sensitive. A similar mechanism has been previously proposed to explain the cold-sensitivity of in alleles (Adler et al., 1994). Flies mutant for the fritz (frtz) gene have a similar phenotype to fy and in (Gubb, 1993). Strong frtz alleles are also cold sensitive (S. C., unpublished observations) supporting a general role for the group II genes in a temperature-sensitive developmental process.

fy specifies tissue polarity cell autonomously

The wing hair polarity within fy2 clones mimics that at the equivalent position on the wing of a fy2 homozygous mutant fly, implying that the planar polarity of a fy cell is expressed cell autonomously and is dependent upon its position within the wing blade. The exceptions to this observation are the cells at the borders of fy2 clones which display a polarity between mutant and wild type (see Fig. 2B). This could be due to partial rescue of the mutant cell polarity phenotype by wild-type tissue. Alternatively, there may be an mechanism, independent of fy function, by which hair polarity is aligned with neigh-bouring cells (Adler et al., 1987). Such an activity would explain why adjacent hairs on wings mutant for fy wing are ori-entated to form smooth, rather than irregular, curves (see Fig. 1B,C). The fy2 clones examined did not affect the polarity or hair number shown by adjacent wild-type cells over a signifi-cant distance. Therefore, whilst the fy gene is required for cells to interpret polarity information correctly, it is probably not needed for the transmission of such information between cells. This is consistent with the proposed role of the fy gene down-stream of fz and dsh in a cell-autonomous pathway that specifies prehair initiation (Krasnow et al., 1995). The in gene is proposed to function at the same step in this pathway and has a similar cell-autonomous phenotype to fy (Gubb and Garcia-Bellido, 1982; Park et al., 1996). The finding that both fy and in encode putative transmembrane proteins suggests their gene products may be co-localised within the cell.

fy expression coincides with polarity determination

There is an abrupt peak of fy expression in the 2-day-old pupa. At this time, cell division in the wing is complete (Schuebiger and Palka, 1987; Milan et al., 1996) and the formation of wing cell prehairs, which is characterised by the localisation of F-actin to the distal vertex of the cell (Wong and Adler, 1993). The pattern of fy expression in the pupal wing at this time (Fig. 7B) resembles the pattern of incorporation of BrdU in the final cell divisions of the wing (Schuebiger and Palka, 1987; Milan et al., 1996). As the patterns of last cell divisions in the wing do not appear to be as strictly regulated as they are, for example, in the Drosophila eye, a reference point for gene acti-vation would be the time at which a cell becomes mitotically quiescent. The transient expression of fy after cell division prior to cell differentiation suggests that the fy gene product is expressed specifically to interpret polarity information. This raises the possibility that one of the components of the putative tissue polarity signal transduction pathway will be a transcrip-tion factor that promotes fy expression. None of the tissue polarity genes characterised so far encodes a transcription factor, but it has been reported that the fz protein is required for normal expression of the nemo gene and in the eye (Zheng et al., 1995). A simple model of fz activating fy transcription seems unlikely, however, as data on epistatic interactions have suggested that the fz gene regulates the activity of fy and in (Wong and Adler, 1993). It is possible, therefore, that fy expression is activated by a mechanism that promotes cell differentiation at the time the cell has completed division and that the fy gene product is regulated by fz through protein-protein interactions rather than transcriptional control.

The embryonic expression of the fy transcript is intriguing, as fy mutants have no known embryonic phenotype. However, the tissue polarity genes fz and in are also expressed in the embryo and have no associated phenotype (Adler et al., 1990; Park et al., 1996). This suggests either that the putative tissue polarity signaling pathway is redundant during embryogenesis, or that it plays a role in establishing planar polarity in the embryo that is reflected in a less conspicuous way than the ori-entation of bristles and hairs on the adult cuticle.

Two roles for the fuzzy gene in wing hair development

In summary, the fy gene encodes a novel four-pass transmem-brane protein that plays two roles in the development of hairs on the Drosophila wing. The first is to specify the correct ori-entation of the hair by restricting its initiation to the distal vertex of the cell. This activity is proposed to be directed by polarity information received by the Fz receptor, possibly through Wnt signaling, and transmitted by the Dsh protein. The second role of fy is to permit the development of just a single cell hair by maintaining the integrity of the F-actin and micro-tubule arrays that are required for hair development, a process that may also require the activity of the small GTPase Rac1.

We thank Clare Henchcliffe for use of the fy alleles created in her X-ray screen and her preliminary deficiency mapping data of the fy locus; Shira Neumann-Silberberg and Trudi Schüpbach for providing deficiencies and the phage walk to grk; José Felix de Celis for help with pupal wing stainings and for providing fly stocks for clonal analysis; Michael Ashburner and Adelaide Carpenter for confirming the cytology of chromosomal aberrations; Paul Adler, Michael Ashburner, José Felix de Celis, Rachel Drysdale and Randi Krasnow for discussion and critical reading of the manuscript and the members of the Ashburner laboratory for support and encouragement. This work was funded by an MRC programme grant to Michael Ashburner and David Gubb.

Adachi
,
Y.
and
Yanagida
,
M.
(
1989
).
Higher order chromosome structure is affected by cold-sensitive mutations in Schizosaccharomyces pombe gene crm1+ which encodes a 115-kD protein preferentially localized in the nucleus and its periphery
.
J. Cell Biol
.
108
,
1195
1207
.
Adler
,
P. N.
(
1992
).
The genetic control of tissue polarity in Drosophila
.
BioEssays
14
,
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.
,
Charlton
,
J.
and
Vinson
,
C.
(
1987
).
Allelic variation at the frizzled locus of Drosophila
.
Dev. Genet
.
8
,
99
119
.
Adler
,
P. N.
,
Vinson
,
C.
,
Park
,
W. J.
,
Conover
,
S.
and
Klein
,
L.
(
1990
).
Molecular structure of frizzled, a Drosophila tissue polarity gene
.
Genetics
126
,
401
416
.
Altschul
,
S. F.
,
Gish
,
W.
,
Millar
,
W.
,
Myers
,
E. W.
and
Lipman
,
D. J.
(
1990
).
Basic local alignment search tool
.
J. Mol. Biol
.
215
,
403
410
.
Aviv
,
H.
and
Leder
,
P.
(
1972
).
Purification of biologically active globin mRNA by chromatography on oligothymidylic acid-cellulose
.
Proc. Natl. Acad. Sci. USA
69
,
1408
1412
.
Bhanot
,
P.
,
Brink
,
M.
,
Samos
,
C. H.
,
Hsieh
,
J.-C.
,
Wang
,
Y.
,
Macke
,
J. P.
,
Andrew
,
D.
,
Nathans
,
J.
and
Nusse
,
R.
(
1996
).
A new member of the frizzled family from Drosophila functions as a Wingless receptor
.
Nature
382
,
225
230
.
Brown
,
N. H.
and
Kafatos
,
F. C.
(
1988
).
Functional cDNA libraries from Drosophila embryos
.
J. Mol. Biol
.
203
,
425
437
.
Cavener
,
D. R.
and
Ray
,
S. C.
(
1991
).
Eukaryotic start and stop translation sites
.
Nucleic Acids Res
.
19
,
3185
3192
.
Chomczynski
,
P.
and
Sacchi
,
N.
(
1987
).
Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction
.
Anal. Biochem
.
162
,
156
159
.
Collier
,
S.
,
Tassabehji
,
M.
and
Strachan
,
T.
(
1992
).
A method for specific amplification and PCR sequencing of individual members of multigene families: application to the study of steroid 21-hydroxylase deficiency
.
PCR Meth. Applic
.
1
,
181
186
.
Coulson
,
D.
(
1994
).
The genetic analysis of prickle and spiny-legs; two cuticular polarity mutants of Drosophila melanogaster
.
PhD Thesis
,
University of Cambridge
.
Dunin-Borkowski
,
O. M.
and
Brown
,
N. H.
(
1995
).
Mammalian CD2 is an effective heterologous marker of the cell surface in Drosophila
.
Dev. Biol
.
168
,
689
693
.
Eaton
,
S.
,
Auvinen
,
P.
,
Luo
,
L.
,
Jan
,
Y. N.
and
Simons
,
K.
(
1995
).
CDC42 and Rac1 Control Different Actin-dependant Processes in the Drosophila Wing Disc Epithelium
.
J. Cell Biol
.
131
,
151
164
.
Eaton
,
S.
,
Wepf
,
R.
and
Simons
,
K.
(
1996
).
Roles for Rac1 and Cdc42 in Planar Polarization and Hair Outgrowth in the Wing of Drosophila
.
J. Cell Biol
.
135
,
1277
1289
.
Grell
,
E. H.
(
1969
).
New Mutants
.
Dros. Inf. Serv
.
44
,
46
.
Gubb
,
D.
(
1993
).
Genes controlling cellular polarity in Drosophila
.
Development
1993
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
.
Held
,
L. I.
,
Duarte
,
C. M.
and
Derakhshanian
,
K.
(
1986
).
Extra tarsal joints and abnormal cuticular polarities in various mutants of Drosophila melanogaster
.
Roux’s Arch. Dev. Biol
.
195
,
145
157
.
Klingensmith
,
J.
and
Nusse
,
R.
(
1994
).
Signaling by wingless in Drosophila
.
Dev. Biol
.
166
,
396
414
.
Klingensmith
,
J.
,
Nusse
,
R.
and
Perrimon
,
N.
(
1994
).
The Drosophila segment polarity gene dishevelled 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
.
Milan
,
M.
,
Campuzano
,
S.
and
Garcia-Bellido
,
A.
(
1996
).
Cell cycling and patterned cell proliferation in the Drosophila wing during metamorphosis
.
Proc. Natl. Acad. Sci. USA
93
,
11687
11692
.
Neumann-Silberberg
,
F. S.
and
Schuepbach
,
T.
(
1993
).
The Drosophila dorsoventral patterning gene gurken produces a dorsally localized RNA and encodes a TGFa-like protein
.
Cell
75
,
165
174
.
Noordermeer
,
J.
,
Klingensmith
,
J.
,
Perrimon
,
N.
and
Nusse
,
R.
(
1994
).
dishevelled and armadillo act in the Wingless signalling pathway in Drosophila
.
Nature
367
,
80
82
.
Park
,
W. J.
,
Liu
,
J.
,
Sharp
,
E. J.
and
Adler
,
P. N.
(
1996
).
The Drosophila tissue polarity gene inturned acts cell autonomously and encodes a novel protein
.
Development
122
,
961
969
.
Pearson
,
W. R.
and
Lipman
,
D. J.
(
1988
).
Improved tools for biological sequence comparison
.
Proc. Natl. Acad. Sci
.
85
,
2444
2448
.
Rost
,
B.
,
Casadio
,
R.
,
Fariselli
,
P.
and
Sander
,
C.
(
1995
).
Transmembrane helices at 95% accuracy
.
Prot. Sci
.
4
,
521
533
.
Schuebiger
,
M.
and
Palka
,
J.
(
1987
).
Changing spatial patterns of DNA replication in the developing wing of Drosophila
.
Dev. Biol
.
123
,
145
153
.
Spradling
,
A.
(
1986
).
P-element mediated transformation
.
In Drosophila; A Practical Approach
(ed.
D. B.
Roberts
), pp.
176
197
.
Oxford
:
IRL Press
.
Sturtevant
,
M. A.
,
O’Neil
,
J. W.
and
Bier
,
E.
(
1993
).
The Drosophila rhomboid gene mediates the localized formation of wing veins and interacts genetically with componenets of the EGF-R signalling pathway
.
Genes Dev
.
7
,
961
973
.
Theisen
,
H.
,
Purcell
,
J.
,
Bennett
,
M.
,
Kansagara
,
D.
,
Syed
,
A.
and
Marsh
,
J. L.
(
1994
).
dishevelled is required during wingless signaling to establish both cell polarity and cell identity
.
Development
120
,
347
360
.
Toda
,
T.
,
Shimanuki
,
M.
,
Saka
,
Y.
,
Yamano
,
H.
,
Adachi
,
Y.
,
Shirakawa
,
M.
,
Kyogoku
,
Y.
and
Yanagida
,
M.
(
1992
).
Fission yeast pap1-dependent transcription is negatively regulated by an essential nuclear protein, crm1
.
Mol. Cell. Biol
.
12
,
5474
5484
.
Vinson
,
C. R.
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
.
Vinson
,
C. R.
,
Conover
,
S.
and
Adler
,
P. N.
(
1989
).
A Drosophila tissue polarity locus encodes a protein containing seven potential transmembrane domains
.
Nature
338
,
263
264
.
Wong
,
L. L.
and
Adler
,
P. N.
(
1993
).
Tissue polarity genes of Drosophila regulate the subcellular location for prehair initiation in pupal wing cells
.
J. Cell. Biol
.
123
,
209
221
.
Wustmann
,
G.
,
Szidonya
,
J.
,
Taubert
,
H.
and
Reuter
,
G.
(
1989
).
The genetics of position-effect variegation modifying loci in Drosophila melanogaster
.
Mol. Gen. Genet
.
217
,
520
527
.
Zheng
,
L.
,
Zhang
,
J.
and
Carthew
,
R. W.
(
1995
).
frizzled regulates mirror-symmetric pattern formation in the Drosophila eye
.
Development
121
,
3045
3055
.