We report the identification of a Drosophila Pax gene, eye gone (eyg), which is required for eye development. Loss-of-function eyg mutations cause reduction or absence of the eye. Similar to the Pax6 eyeless (ey) gene, ectopic expression of eyg induces extra eye formation, but at sites different from those induced by ey. Several lines of evidence suggest that eygand ey act cooperatively: (1) eyg expression is not regulated by ey, nor does it regulate ey expression, (2) eyg-induced ectopic morphogenetic furrow formation does not require ey, nor does ey-induced ectopic eye production require eyg, (3) eyg and ey can partially substitute for the function of the other, and (4) coexpression of eyg and ey has a synergistic enhancement of ectopic eye formation. Our results also show that eyg has two major functions: to promote cell proliferation in the eye disc and to promote eye development through suppression of wg transcription.

Development of the Drosophila compound eye begins in the cellular blastoderm embryo, when about 6-23 cells are allocated as the eye-antennal disc primordium. These cells proliferate without differentiation to form the larval eye-antennal disc, from which most of the adult head structures will derive. At early third instar, cells at the posterior of the eye disc start to differentiate, and the differentiation progresses in a posterior-to-anterior direction (Wolff and Ready,1993). The front of differentiation is marked by an indent called the morphogenetic furrow (MF). Cells are undifferentiated anterior to the propagating furrow, and become progressively differentiated posterior to the furrow.

In recent years, a small number of genes, all encoding nuclear proteins,have been identified as be required and sufficient (in certain contexts) for the initiation of eye development in Drosophila (for reviews, see Desplan, 1997; Treisman, 1999; Kumar and Moses, 2001). Loss-of-function mutations of toy, ey, dac, eya and so can lead to reduction or absence of the adult eye, while ectopic expression of these genes, either alone or in combination, can lead to ectopic eye formation(Halder et al., 1995; Shen and Mardon, 1997; Bonini et al., 1997; Chen et al., 1997; Pignoni et al., 1997; Czerny et al., 1999; Kronhamn et al., 2002),suggesting that they act at the very early steps of eye development. Two other nuclear factor genes, tsh and optix, can also induce ectopic eye upon ectopic expression (Pan and Rubin, 1998; Seimiya and Gehring, 2000) [see also Singh et al.(Singh et al., 2002) for an additional role of tsh].

Epistasis analysis in loss-of-function mutants and in misexpression situations suggests that toy, ey, so, eya, dac and tsh form a complex regulatory network (Desplan,1997; Shen and Mardon,1997; Bonini et al.,1997; Chen et al.,1997; Pignoni et al.,1997; Halder et al.,1998; Pan and Rubin,1998; Czerny et al.,1999). toy acts upstream to regulate eyexpression but is not regulated by ey(Czerny et al., 1999). Several lines of evidence suggest that ey acts upstream and regulates the expression of dac, eya, so and tsh. (1) Its ability to induce ectopic eyes is the strongest. (2) Its normal expression in the eye disc starts earlier (in the embryonic eye disc primordia) than eya, so,dac and tsh. (3) The normal expression of ey does not require dac, eya and so, while the normal expression of eya and so requires ey. (4) Ectopic eyexpression can induce the expression of dac, eya, so and tsh. (5) Ectopic eye induction by ey requires dac,eya and so. However, the regulation is not a simple linear pathway, because (1) ectopic expression of dac, eya, so and tsh can also induce ey expression, at least in the antennal disc, (2) ectopic expression of dac and eya can induce each other, and (3) so and eya may up-regulate the expression of the other (when ectopically induced by ey). These relationships suggest a positive feedback regulation among these genes. These feedback regulations are important for their function, because ectopic eye formation by eya, so/eya and dac/eya also requires the upstream ey gene (Bonini et al.,1997; Pignoni et al.,1997; Chen et al.,1997).

Although ey has been called a master regulator of eye development(Halder et al., 1995), eye development does not occur in every cell induced to express ey or its downstream genes dac, eya and so(Halder et al., 1995; Shen and Mardon, 1997; Bonini et al., 1997; Pignoni et al., 1997; Chen et al., 1997; Halder et al., 1998). Other genes must collaborate with ey in the induction of eye development. For example, DPP and HH signaling collaborates with EY for ectopic eye induction (Chen et al., 1999; Kango-Singh, 2003). Therefore,the identification of genes that are required for eye development but not directly under ey regulation will lead to a better understanding of the mechanism of eye induction. The optix gene of the Six/sogene family is such a gene. It is capable of inducing ectopic eye development but its expression is not under regulation by ey(Seimiya and Gehring,2000).

eye gone (eyg) is another gene required for eye development. The first mutation identified, eyg1, causes the eye to become significantly smaller(FlyBase, 2003). It also shows genetic interaction with ey, as mutants doubly homozygous for hypomorphic viable eyg and ey alleles are not viable(Hunt, 1970). The lethal pharate adults have severely reduced head and complete absence of eyes. These results suggest that eyg may act in the early stages of eye development and interact with ey.

In this report, we identify eyg as a Pax gene. It is expressed in the embryonic eye disc primordium and in the larval eye imaginal disc. We characterize its function in eye development and show that ectopic eyg expression can induce ectopic eye formation. eyg is different from eya, so and dac in that its expression is not primarily regulated by ey, nor does it regulate eyexpression. Its ability to induce ectopic eyes does not require ey,nor is it required for the ability of ey to induce ectopic eye. Therefore eyg acts neither upstream nor downstream of ey. In addition, coexpression of eyg and ey causes a synergistic ectopic eye formation. Thus, eyg appears to act cooperatively with ey in eye development. As both genes encode Pax proteins with a homeodomain, this is suggestive of a molecular interaction between the two gene products. We also show that the mechanism by which eyg affects eye development may be through suppressing the expression of wg,which is known to inhibit MF initiation.

Fly stocks

Several dpp-GAL4 lines(Staehling-Hampton et al.,1994) were used: w; CyO/Sp;dpp-GAL4c40.1/TM6B, from Patrick Callaerts (University of Basel, Switzerland), w; Sp/CyO;P[w+m.hs,GAL4-dpp.blk1]40C.6/TM6B, from the Bloomington Stock Center, and w; dpp-lacZ (BS3.0)H1-1;dpp-GAL4c40.6/SM6-TM6B, from Jessica Treisman (Skirball Inst.,NY). The one from Bloomington gave a weaker phenotype in ectopic eye induction, and was used to show the synergistic effect of eyg and ey. E132-GAL4, UAS-ey and ey2 flies(Halder et al., 1995) were from Patrick Callaerts. Two ey-GAL4 lines were from Uwe Walldorf(University of Hohenheim, Stuttgart, Germany). Other fly stocks were obtained from the Bloomington Drosophila Stock Center and the Mid-America Drosophila Stock Center (Bowling Green, Ohio).

Screening and genetic characterization of eyg mutants

Several approaches were used to generate eyg mutations.γ-irradiation was used to induce chromosomal aberrations in Eq-1, Eq-2 and Eq-3, which have P-element insertions close to the eyg locus(Sun et al., 1995). The males were irradiated with γ-rays (4000R) and mated to w; DH/TM3, Serfemales. The progeny were screened for loss of the Equatorial eye pigmentation pattern, i.e. loss of the P[lacW] insertion in 69C. C1 was found to be a large deficiency (69A4-5; 69D4-6; cytology determined by Adelaide Carpenter) and used as a reference for null. eygM3-12 was generated from mobilization of P[lacW] from Eq-1. Complementation over the eyg1 allele gave a strong eyg phenotype. eygM3-12 homozygotes are pupal lethal. The pharate adults have a headless phenotype (see text), which is similar to that of eygM3-12/C1, suggesting that eygM3-12is close to a functionally null allele. Similarly, eygEq-2-d2-2, derived from Eq-2 by P mobilization, was defined as close to a functionally null allele. In(3LR)gvu/eyg1 flies showed a small eye phenotype, indicating that In(3LR)gvu (from Bloomington Stock Center) is a weak eyg allele. EM458 (kindly provided by Leslie Vosshall, Columbia University) carries a P[GawB] insertion 527 bp upstream of the first ATG represented in the Lune transcript. It is homozygous viable and exhibits no apparent phenotype. eyg37-1, eyg22-2 and eyg94-4 were independently derived from mobilizing the P[GawB] in EM458. The severity of defect of eyg alleles was tested over null alleles (C1 or eygM3-12) and over weak alleles (eyg1or In(3LR)gvu). eygM3-12 mutant clones were generated using the hs-FLP/FRT method(Xu and Rubin, 1993). hs-FLP22 and FRT(whs)2A(FlyBase, 2003) were used. Heat shock induction of hs-FLP was at 37°C for 1 hour at the indicated time after egg laying (AEL). The eyg mutant clones in the adult are marked by the loss of pigmentation, dependent on the mini-white marker. In eye discs, the heterozygous cells are marked by one copy of the Ubi-GFP-nls, which encodes a nuclear GFP(Davis et al., 1995). The mutant cells should have no GFP expression, while the wild-type twin-spot should have twice the GFP level. However, the twin-spots have a much stronger GFP expression.

Transgene constructs and germline transformation

The 2.7 kb full-length Lune cDNA was cloned into the NotI site of the P[CaSpeR-hs] and P[UAST] vectors(Thummel and Pirrotta, 1992; Brand and Perrimon, 1993). These constructs were used in germline transformation as described previously(Rubin and Spradling,1982).

5′-RACE

Embryos were collected 0-16 hours after egg laying for mRNA purification by the CLONTECH mRNA purification kit. The 5′ end of eygtranscript was amplified by the SMARTTM RACE cDNA Amplification kit using two eyg-specific primers for nested PCR. The two primers were 5′-CTAGCAACTTGGAGACAGCTCC-3′ and 5′-GCCAGAATTAGCGACAGTAAG-3′ respectively. The PCR products were cloned and multiple clones were sequenced.

Molecular analysis of mutants

For the analysis of eygM3-12 which has a P[lacW]insertion, plasmid rescue was performed from genomic DNA. A 10 kb SacII rescued plasmid was sequenced from the termini. The 3′primer read into an opus retrotransposon [also known as nomad and yoyo(Whalen and Grigliatti, 1998; FlyBase, 2003)], which is present in the original Eq-1 fly but not in the w1118parental line. The 5′ primer read into a sequence 13 kb downstream of eyg. Sequence flanking the other side of the P[lacW] in eygM3-12, was rescued in a 12.5 kb BglII fragment. The 5′ primer read into a sequence 24 bp upstream of the eyg transcription start site. Thus the eygM3-12is likely to have a deletion starting 23 bp upstream of eyg and extending 13 kb downstream of eyg. This was confirmed by genomic Southern analysis of eygM3-12 homozygotes(Tb+ larvae and pupae selected from an eygM3-12/TM6B, Tb stock) with EcoRI digestion. For the eyg37-1, eyg22-2 and eyg94-4 mutants that carried a P[GawB] insert, the genomic region flanking the P insert was amplified by PCR and analyzed by sequencing.

In situ hybridization and histochemical staining

Digoxigenin-labeled antisense RNA probes were used for in situ hybridization experiments as described previously(Tautz and Pfeifle, 1989). Alkaline phosphatase histochemical staining was used to visualize in situ hybridization signals. DNA template for the ey probe was derived from ey exon 9 and was recovered by PCR from genomic DNA. The eygprobe was described previously (Jones et al., 1998). The toe probe was transcribed from an EcoRI-linearized toe EST clone pOT2a-toe (kindly provided by Jyh-Lyh Juang, NHRI, Taiwan). The stringent hybridization condition excluded cross hybridization. w1118 was used as the wild-type control. Embryos were photographed using a Leica DMRB microscope with differential interference contrast (DIC) optics. For the double-labeling RNA in situ hybridization, the fluorescein-labeled eyg antisense RNA probes and the biotin-labeled ey antisense RNA probes were transcribed by the T7, T3 promoter (Boehringer Mannheim). The detection was first with HRP-conjugated anti-fluorescein antibody (1:200) and amplified by the Cy-3-tyramides TSA (NEN Life Sciences, UK), followed by inactivation for 15 minutes at 70°C, and then by HRP-conjugated streptavidin and amplified with FITC-tyramid TSA (NEN Life Sciences, UK). X-gal staining of lacZexpression was done according to the method of Sun et al.(Sun et al., 1995). mAb22C10(1:100) (Fujita et al., 1982; Zipursky et al., 1984) was from Seymor Benzer (Caltech). X-gal and antibody double staining was modified from the procedure of Kobayashi and Okada(Kobayashi and Okada, 1993),namely primary and secondary antibody incubation were performed first,followed by X-gal staining, then the peroxidase color development. Incorporation of 5-bromo-2′-deoxyuridine (BrdU) into dividing cells in imaginal discs was done according to the method of Baker and Rubin(Baker and Rubin, 1992). Acridine Orange staining of apoptotic cells was done according to the method of Spreij (Spreij, 1971). Confocal microscopy was performed on a Zeiss LSM510.

eyg is required for early eye disc development

We have identified two enhancer trap lines (Eq-1, Eq-2)(Sun et al., 1995) with the P[lacW] construct inserted near the eye gone (eyg) gene in 69C on the third chromosome. The two lines showed no eyg phenotype. Starting from these lines, a large number of derivative lines were generated(by γ-irradiation induced chromosomal aberrations, P-element imprecise excisions and local transpositions), some of which showed eye reduction phenotypes and failed to complement eyg1.

Weak loss-of-function eyg mutations resulted in the reduction or absence of the adult eyes (Fig. 1A,B). In late third instar larvae of the hypomorphic eyg1 mutant, the eye discs were significantly reduced in size, while the antennal discs appeared normal(Fig. 1E). In strong loss-of-function mutants, the adults failed to emerge from the pupal case. Their heads were severely reduced in size, but the appeared normal(Fig. 1C). In a null mutant eygM3-12, the adults have a headless phenotype(Fig. 1D) and all structures derived from the eye-antennal discs were missing. The prominent remaining structure was the labellum (Fig. 1D) derived from the labial discs. The fish-trap bristles, derived from the clypeo-labral disc, were also present (not shown). The headless phenotype is similar to those reported for ey and toy null mutants (Jiao et al., 2001; Kammermeier et al., 2001; Kronhamn et al., 2002). In flies with eyg alleles of different strengths, the size reduction of third instar eye discs was proportional to the severity of the adult eye phenotype. Strong alleles did not affect the morphology and size of the antennal discs (Fig. 1J). However, no eye-antennal discs could be found in eygM3-12larvae. These observations suggested that the antennal disc only requires a low EYG level or activity, so the antennal phenotype is manifested only in the null mutant, a situation similar to those of ey and toymutants (Kronhamn et al.,2002). Alternatively, the effect of eyg may be specific to the eye disc, and the loss of antennal disc-derived structures in null mutants may be secondary effect due to the missing eye disc.

The effect of eyg is already apparent in earlier stages. Strong eyg alleles result in no eye disc or only a rudimentary stub of a disc in early third instar (not shown). Weaker allelic combinations produce a smaller eye disc than normal in early third instar (not shown), and in mid-third instar excessive cell death could be detected anterior to the morphogenetic furrow by staining with the dye Acridine Orange(Fig. 1F). Similar apoptosis anterior to the furrow has been observed in other small eye or eyeless mutants, e.g. eya, so, dac and ey(Bonini et al., 1993; Cheyette et al., 1994; Mardon et al., 1994; Halder et al., 1998). In late third instar, there is no more excessive apoptosis.

The small size of early third instar mutant eye discs indicated either that early cell proliferation is affected, or that there is excessive apoptosis prior to third instar, or both. Misexpression of the anti-apoptosis baculoviral protein P35 (Hay et al.,1994), driven by the dpp-GAL4 or ey-GAL4, failed to rescue the `no eye' phenotype in the eyg1/eygM3-12 mutant(Fig. 1I and data not shown). The eyg1/eygM3-12 mutant had complete absence of the adult eyes (Fig. 2H) and had rudimentary eye discs (not shown). The adult eyes and the larval eye discs were not rescued by misexpression of P35, suggesting that apoptosis is not the major cause of the eye phenotype.

The effect on proliferation is restricted to the early eye disc, since cell proliferation in late third instar mutant eye discs seemed normal, as the two normal mitotic waves could be detected in the eye discs(Fig. 1G). A role for eyg in cell proliferation in the eye disc is supported by the finding that ectopic eyg expression can induce local overgrowth in eye discs(see below). We used the hs-FLP/FRT method(Xu and Rubin, 1993) to generate eyg homozygous mutant clones in an otherwise heterozygous eyg mutant fly. When the clones were induced at 24-48 or 48-72 hours after egg lay (AEL), eygM3-12 clones were not detected in the adult (not shown) nor in the eye discs(Fig. 1J), while the twin spots were large and appeared in different regions in the eye discs(Fig. 1J). This result suggests that eyg has a strong effect on the growth of the eye discs and the effect has no strong regional preference.

eyg encodes a Pax protein

A Pax-like cDNA [originally named Lune(Jun et al., 1998)] that also mapped at 69C was independently isolated based on the presence of a homeobox encoding a Paired-type homeodomain (HD) with a characteristic serine at position 50 (Burglin, 1994). This residue is only found when the HD is associated with a paired domain(PD). The PD is the defining character of Pax proteins(Noll, 1993). It consists of two subdomains, the N-terminal PAI and the C-terminal RED subdomains, that can both bind DNA (Czerny et al.,1993; Epstein et al.,1994; Xu et al.,1995; Jun and Desplan,1996). Interestingly, the PD encoded by the Lune cDNA contains only a partial PAI subdomain and a complete RED subdomain(Fig. 2A)(Jun et al., 1998). eyg has an open reading frame of 670 amino acids(Fig. 2A) rather than the 523 amino acids originally reported (Jun et al., 1998). These changes involved both N-terminal and C-terminal sequences. 5′-RACE identified two splicing isoforms with their 5′ends identical to that of the Lune cDNA, indicating that the 5′ end of the cDNA clone represents the transcription start site. Thus the Lune cDNA represents the full-length transcript, and there is no additional upstream exon to provide a functional PAI subdomain. The two isoforms differ in a 67 bp segment (intron I; Fig. 2B),which does not affect the coding region. Five introns were identified(Fig. 2B). All exon-intron junctions conform to the consensus splice site.

In situ hybridization showed that in the eye disc of late third instar larvae, Lune is expressed in the central anterior region, well ahead of the morphogenetic furrow (Fig. 2C). Expression is stronger dorsal to the equator. In the early eye disc, the expression domain is broader (Fig. 2D,E). It is also expressed in the central region of the antennal disc (Fig. 2C), in the anterior notum, dorsal hinge and in an arc at the posterior periphery of the wing pouch of the wing disc (Fig. 2F), and is weakly expressed in several arcs in the leg discs(Fig. 2G). In the embryo it is expressed in the eye-antennal disc primordium(Fig. 7B)(Jones et al., 1998), similar to ey and toy (Quiring et al., 1994; Czerny et al.,1999). In the embryo, it is also expressed in the antennal organ,salivary gland, and in a segmentally repeated lateral pattern(Jones et al., 1998).

Since the Lune cDNA mapped to the same chromosomal region as the eyg mutation, it is expressed in the developing eye, and the genes ey and toy, also Pax genes, have similar `eyeless' mutant phenotype, we speculated that the Lune cDNA might correspond to the eyg gene. We tested this hypothesis using rescue experiments. The Lune cDNA was linked to the hsp70 promoter and the hs-Lunetransgene was tested for its ability to rescue the eyg1/eygM3-12 mutant phenotype, which is characterized by the complete absence of the eyes(Fig. 2H). Heat shock once every 12 hours throughout development resulted in a number of flies with partial or fully rescued eyes. A single heat pulse could also rescue the phenotype. The frequency of full rescue was 4% (n=50 eyes) at 12-24 hours AEL, 5% (n=20) at 24-48 hours AEL, 24% (n=50) at 60-72 hours AEL, 10% (n=60) at 84-96 hours AEL, with peak rescue efficiency at 60-72 hour AEL (about late second instar larva).

A second Pax gene was identified about 30 kb downstream of eyg,based on the fly genome sequence (Adams et al., 2000). The predicted gene is represented by an EST clone. Its encoded protein is most homologous to EYG in its paired domain and homeodomain, so we named the gene twin of eyg (toe). In the eye disc, the expression pattern of toe(Fig. 3B) is very similar to that of eyg (Fig. 2C, Fig. 3B). Three independent lines, 37-1, 22-2 and 94-4, were generated from mobilization of P[GawB]EM458, which has the insertion 124 bp upstream of the Lune transcription start site and 527 bp upstream of the first ATG. The P[GawB] has transposed 8 bp downstream in all three lines, and is accompanied by an 86 bp, 159 bp and 224 bp deletion, respectively, of the flanking genomic region (Fig. 3A). Thus eyg22-2 and eyg94-4 have deletions extending into the 5′ untranslated region. eyg37-1is homozygous viable and results in a very weak small eye phenotype(Fig. 3B). eyg22-2 is homozygous viable and produces a small eye(about 500 ommatidia) phenotype (Fig. 3B). The eyg94-4 homozygote dies at the pharate adult stage. The phenotypes of pharates ranged from nearly headless(Fig. 3B; 25/57 flies=44%), to complete absence of eye (similar to Fig. 1B; 5/114 eyes=4%), to small eyes (about 300-400 ommatidia; 59/114 eyes=52%). These mutations failed to complement eyg1, so they are eyg alleles as defined by genetic complementation. In these mutants, eyg mRNA was strongly reduced in the eye disc and in the antenna disc, while toe mRNA level was not significantly affected(Fig. 3B). The eygM3-12 mutant has a large deletion starting at 23 bp upstream of the eyg transcription start site and extending to about 13 kb downstream of eyg (Fig. 3C). The molecular nature of these mutations, together with the rescue results, strongly suggest that the eyg gene is a Pax gene represented by the Lune cDNA.

eyg expression in early eye disc is functionally important

Since the eye phenotype discussed above suggests an early role for eyg in eye disc development, we checked the expression of eyg in earlier eye discs (Fig. 2D,E). Expression was detectable at least as early as late second instar eye disc, in a pattern slightly broader than in the late third instar. Expression in the antennal disc was detectable beginning in mid-third instar(data not shown).

Rescue experiments were performed by induced eyg expression to determine the site of functional requirement for eyg. The eyg cDNA was cloned into the pUAST vector behind a tandem array of five GAL4 binding sites (Brand and Perrimon, 1993). Expression of the UAS-eyg transgene is dependent on the presence of the yeast GAL4 transcription factor. Expression of eyg induced by the E132-GAL4 and ey-GAL4(abbreviated E132>eyg and ey>eyg,respectively) could partially or fully rescue the strong eyg1/eygM3-12 mutant: the eyes were morphologically normal and often of normal size(Fig. 4B and data not shown). ey>eyg can rescue the eye in eyg1/eygM3-12 mutants to medium to full size with complete penetrance (Fig. 4D). E132 drives GAL4 expression in the center of the posterior margin in late third instar eye discs(Fig. 4A)(Halder et al., 1995). In second instar larval eye discs E132 drives expression in a broader posterior-ventral domain (Fig. 4A inset) (Pignoni and Zipursky, 1997). ey-GAL4 drives expression from the posterior margin to a few rows of cells anterior of the MF in late third instar eye disc (Fig. 4C), and in the entire eye disc in the early third and late second instar eye disc(Fig. 4C inset). These rescue results suggest that eyg expression in cells common to both E132 and ey domain in early eye discs is important for eye development.

Ectopic expression of eyg can cause ectopic eye formation

eyg and ey are similar in that both are Pax genes and are required for eye formation. Therefore, we asked whether ectopic expression of eyg would be sufficient to direct the development of ectopic eyes, as is the case for ey. Five independent UAS-eyg lines were crossed with many GAL4 lines specifying a variety of expression patterns. Several phenotypes could be observed among the few GAL4 lines that gave surviving adults: missing bristles on the notum, extra scutellar bristles,shortening and malformation of the legs, and unexpanded wings. The most dramatic phenotype was the formation of extra eyes (see below), observed when ectopic expression of UAS-eyg was induced by dpp-GAL4.

In the eye disc, dpp-GAL4 is expressed along the posterior and lateral margins (Treisman and Rubin,1995). The eyg-induced extra eyes were almost always located ventral to the endogenous eyes. They often appeared linked to the endogenous eye (Fig. 5A), but occasionally were well separated and were surrounded by orbital bristles(Fig. 5B), suggesting that they represent intact eye fields. The penetrance of extra eye formation is low. With a strong UAS-eyg line, almost all UAS-eyg/+; dpp-GAL4/+(abbreviated dpp>eyg) flies have some recognizable defect in the ventral head: the ventral-posterior rim of the eye was reduced and replaced by head cuticle, and/or a few extra bristles, only 16% of eyes have extra ommatidia, and only 38% of these have the extra ommatidia as an isolated extra eye.

The formation of an extra eye field could also be detected in the larval eye disc. For these experiments, we used the monoclonal antibody 22C10(Fujita et al., 1982; Zipursky et al., 1984) to detect photoreceptor neurons and lacZ driven by a different dpp enhancer (BS3.0) (Blackman et al., 1991) to mark the morphogenetic furrow. In wild-type eye-antennal discs, dpp-lacZ is expressed in the MF(Fig. 5C, black arrow) and in a small dorsal region in the antennal disc(Fig. 5C, black arrowhead). In dpp>eyg eye disc, ectopic dpp-lacZ-expressing MFs could be seen (Fig. 5C,D),most often at dorsal and ventral sites between the eye and antennal discs, and at the dorsal and ventral poles of the eye disc. These were propagating toward the center of the disc and usually associated with local overgrowth and sometimes with photoreceptor neuronal differentiation. The photoreceptor clusters were sometimes well separated from the endogenous eye(Fig. 5C), corresponding to the extra eyes found in the adults (Fig. 5B). The percentage of ectopic photoreceptors in discs is higher than the percentage of extra eyes in adults, suggesting that as the endogenous and extra eye fields expand, they often fuse. All these observations indicate that ectopic expression of eyg along the posterior and lateral margin of the eye disc can induce ectopic MF initiation. When the ectopic furrow occurs far away from the endogenous furrow, an extra eye may result.

eyg and ey are transcriptionally independent

Ectopic expression of eyg and ey with the same dpp-GAL4 driver produced ectopic eyes at different sites: eyg-induced eyes occurred in the ventral part of the head, while ey-induced eyes occurred at the base of antennae, wings, halteres and on leg segments (Fig. 5E)(Halder et al., 1995). In imaginal discs, dpp>ey caused overgrowth and the formation of MFs and photoreceptors in the antenna, wing, haltere and leg discs (Fig. 5F, and data not shown), consistent with the ectopic eyes seen in adults. No extra photoreceptors were detected at the ventral margin in the eye-antennal disc(Fig. 5F), in contrast to the effect of eyg expression. The preferential effect in the ventral side of the eye disc by dpp>eyg is similar to the effect of dpp>dac+eya (Chen et al., 1997). In the wing disc, dpp>eygoccasionally caused an extra dpp-lacZ-expressing spot at the anterior side of the hinge region (not shown). However, the dpp-lacZspot does not represent an ectopic MF, as no photoreceptors were detected in the wing and leg discs. These results suggest that neither eyg nor ey activate the expression of the other, since their phenotypes are so distinct. We checked this possibility by analyzing ey and eyg expression in mutant backgrounds.

eyg expression was examined in the ey2 mutant. Although ey2 is not a null mutant(Kronhamn et al., 2002), it has no detectable ey transcript in the embryonic eye disc primordium and in the larval eye disc (Quiring et al., 1994). ey2 eye phenotype is variable. For the ey2 stock we used, 40% of the eyes had 300-400 ommatidia, 40% had about 200 ommatidia, and 20% had less than 100 ommatidia. The ey2 eye disc is also variable in size. However, an ey2 eye disc of substantial size still has no eyexpression (Quiring et al.,1994). So we examined ey2 eye discs that were clearly reduced in size (to be sure that it was a mutant disc) but had sufficient eye field present to check for eyg expression. In ey2 late third instar eye disc, the eyg dorsal expression was not affected while the ventral expression was reduced(Fig. 6E). In ey2 embryos, eyg expression was still present in the eye disc primordium (Fig. 6B). Similarly, ey expression in the eye disc primordium(Fig. 6C) was still present in eyg null mutant embryos (C1/C1)(Fig. 6D). The presence of ey expression suggests that the development of the eye disc primordium is not strictly dependent on eyg function. Halder et al.(Halder et al., 1995)indicated that ey expression is not affected in eyg mutant eye discs. These results suggested that, except for a small amount of eyg expression in the ventral part of the eye disc, the expression of neither eyg nor ey is strongly dependent on the other.

We next checked whether ectopic eyg expression could induce the expression of ey, or vice versa. In the imaginal disc, dpp>eyg induced eyg expression at the lateral margins of the eye disc (Fig. 6F, black arrowheads), in the antennal disc(Fig. 6F, black arrow) and in the wing disc (Fig. 6J, black arrowhead), but did not induce ey expression(Fig. 6G,K). dpp>ey induced ey expression in the eye-antennal disc (Fig. 6I, black arrowheads) and the wing disc (Fig. 6L, green), but did not induce eyg expression(Fig. 6H, red in L). The results with embryos were similar; neither ectopic eyg nor ectopic ey (driven in both cases by rho-GAL4) could induce expression of the other (results not shown). These results suggest that neither eyg nor ey is sufficient for the expression of the other. In summary, the two genes appeared to be transcriptionally independent.

Functional relationship between eyg and ey

We tested whether eyg is functionally dependent on ey,and vice versa. When E132>eyg induction occurred in an ey2 mutant, ectopic ventral eyes could still form(Fig. 7A), suggesting that ey is not required for eyg function. This is in contrast to the situation with eya, so, dac and toy: the ectopic eyes caused by their ectopic expression cannot form in the ey2mutant (Bonini et al., 1997; Chen et al., 1997; Pignoni et al., 1997; Czerny et al., 1999). Similarly, dpp>ey induced ectopic eye formation in eygM3-12 mutants (Fig. 7B), suggesting that eyg is not required for eyfunction. Again, this is in contrast to the requirement for eya, soand dac in ey-induced ectopic eye formation(Bonini et al., 1997; Chen et al., 1997; Shen and Mardon, 1997; Halder et al., 1998). These results suggest that eyg and ey can function independently to induce eye formation.

While the above experiments show that eyg and ey can independently induce eye formation, their coexpression showed synergistic enhancement of ectopic eye formation. The ectopic eyes in the antenna, wing,haltere and legs are larger (Fig. 5G), similar to the effect of dpp>ey at higher temperatures (due to higher GAL4 activity). The enhancement is only evident with a weak dpp-GAL4 line. With stronger dpp-GAL4lines, the ectopic eye phenotype is already strong with UAS-ey alone and cannot be enhanced further by adding UAS-eyg. The enhancement is more evident in the imaginal discs than in the adults. The ectopic eyes in the ventral head are not significantly enhanced in the adults, but are clearly enhanced in the eye discs (Fig. 5H). The difference between the strength of phenotypes in adults and the imaginal discs suggest that there may be some regulative mechanism in the eye field that compensates for the ectopic photoreceptors. The synergistic effect of eyg and ey coexpression was also observed when driven by the E132-GAL4 line (not shown).

Expression of ey induced by ey-GAL4 can partially rescue the eyg1/eygM3-12 mutant eye phenotype(Fig. 4E), indicating that ey can functionally substitute for eyg. Since there is endogenous ey expression, the rescue suggests that EY is required at a level higher than its endogenous expression level in order to compensate for the loss of eyg. Reciprocally, we checked whether expression of eyg could rescue ey mutant phenotype. Since even the strongest ey alleles result in a variable eye phenotype(Kronhamn et al., 2002; Benassayag et al., 2003), we used the eyD/ey2 allelic combination, which results in no eyes and is nearly completely penetrant. eyDhas a chromosomal rearrangement interrupting the ey gene, producing a truncated protein lacking the homeodomain(Kronhamn et al., 2002). ey2 contains a transposon insertion in an eye-specific enhancer and has no detectable RNA and protein expression in the larval eye disc and in the embryonic eye disc primordium(Quiring et al., 1994; Halder et al., 1998). dpp>eyg can partially rescue eyD/ey2 mutants(Fig. 4F), suggesting that eyg can functionally substitute for ey.

eyg suppresses wg transcription

The wingless (wg) gene encodes a secreted signaling protein of the Wnt family. It is expressed in the dorsal and ventral margins of the eye disc (Fig. 8A), and acts to inhibit MF initiation from these sites(Ma and Moses, 1995; Treisman and Rubin, 1995). In eyg1 eye disc, wg expression domain expands toward the posterior margin (Fig. 8C) (Hazelett et al.,1998). However, wg expression is not derepressed in the central region (Fig. 8C) where eyg is normally expressed (Fig. 4), probably because eyg1 is a hypomorphic allele and has sufficient activity in this region to suppress wg. Because the enhancer trap wg-lacZ reporter was used to monitor wg expression, the suppression is at the transcriptional level. In addition, ectopic expression of eyg (dpp>eyg)suppressed wg-lacZ expression in the dorsal and ventral margins of the eye disc and in part of the wing disc (compare Fig. 8B,E with 8A,D).

eyg is an important regulator of eye development

Two Pax genes are located in the chromosomal region of the eygmutation. One of these is represented by the previously reported Lune cDNA(Jun et al., 1998). In the eyg1 mutant, we sequenced the coding region of both Pax genes. Although a few changes were found, none affected the conserved residues(C.-C.J. and Y.H.S., unpublished). Furthermore, the expression level of both genes in eyg1 mutant eye disc appeared normal (C.-C.J. and Y.H.S., unpublished). Thus the analysis of the eyg1 mutant did not clearly indicate which of the two genes corresponds to eyg. We showed that three new eyg alleles, as defined by failure to complement the eyg1 mutation, have molecular defect in the Lune transcription unit and affected Lune RNA level, while the expression level of the adjacent Pax gene is not affected. A fourth mutant, eygM3-12, also does not complement eyg1 and is genetically defined as being functionally null for the eyg eye function. eygM3-12 has a deletion of the Lune transcription unit but is not affected in the downstream Pax transcription unit. In our preliminary analysis of the eyg-toeregion, only one region specifying eye disc expression was identified and it is located downstream of toe and not affected in these four mutants(S.-J. Chiou and Y.H.S., unpublished). The molecular nature of these four mutations, coupled with the rescue results, strongly suggest that the Lune cDNA corresponds to the eyg gene. We named the adjacent Pax gene twin of eyg (toe).

In the embryo eyg transcripts appear in the eye-antennal disc primordium beginning at stage 15. It continues to be expressed as the disc cells proliferate during early larval development and then is expressed in an anterior region overlapping the equator of the eye disc as photoreceptor differentiation occurs. eyg is required for eye development, as loss-of-function mutations lead to the reduction or absence of the eye. It appears to be required for the early proliferation of the eye disc cells, as the early third instar eye disc is significantly smaller. The rescue experiments (hs-eyg, ey>eyg and E132>eyg) suggested that the critical time for eyg function is in the late second instar. Excessive apoptosis occurred in the mid-third instar eye disc, but is not the major cause of the eye phenotype because blocking apoptosis did not rescue the eye phenotype. Ectopic expression of eyg can lead to ectopic MF initiation in the ventral side of the eye disc. Thus, the loss-of-function and gain-of-function phenotypes suggest that eyg acts as an important regulator of eye development.

eyg appears to have two major functions. The first is to promote cell proliferation in the eye disc. eyg loss-of-function mutants have reduced eye discs, already apparent in early third instar, before photoreceptor differentiation. In clonal analysis, eygM3-12 mutant clones induced in first or second instar are undetectable in late third instar eye disc(Fig. 1). Ectopic eygexpression caused local overgrowth (Fig. 6C), a phenotype opposite of the loss-of-function phenotype. The overgrowth does not always develop into photoreceptor cells(Fig. 6C). These results indicate that eyg promotes cell proliferation independent of photoreceptor differentiation. The second function of eyg is to promote eye development or MF initiation. If the eyg-induced proliferation occurs at the ventral margin of the eye disc, ectopic MF can initiate (Fig. 6C,D). The induction of ectopic MF is probably mediated by the suppression of wg(see later), which is known to repress MF initiation along the lateral margins.

eyg and ey act cooperatively

Since eyg is a Pax gene that shares sequence similarity with ey and toy in the PD and HD domains(Jun et al., 1998), its relationship with ey is of particular interest. Our results indicate that eyg and ey are transcriptionally and functionally independent: (1) except for a small amount of eyg expression ventral to the equator of the eye disc, eyg and ey do not regulate each other's expression. In this respect, eyg is different from dac, so and eya, whose expression is strongly regulated by ey (and can induce ey expression in some cases). Thus eyg transcription is neither downstream of ey, nor does eyg participate in the ey/eya/so/dac positive feedback loop. This transcriptional independence is similar to that of optix(Seimiya and Gehring, 2000).(2) eyg and ey can each function (to induce ectopic eyes) in the absence of the other. Again, this is similar to optix, which can induce ectopic eyes in ey2 mutant(Seimiya and Gehring, 2000). Whether optix is required for ey function has not been tested, because of the lack of optix mutants.

However, other evidence indicates that the functions of eyg and ey must converge at some point in the pathway leading to eye development: (1) eyg; ey double hypomorphic mutants showed a much stronger eye-loss phenotype (Hunt,1970), (2) coexpression of ey and eyg caused synergistic enhancement of the ectopic eye phenotype, (3) eyg and ey are able to substitute functionally for each other. Overall, the results suggest that these two Pax genes may act cooperatively. This genetic cooperativity might mean that eyg and ey interact and cooperate as proteins in the same pathway or that they act in parallel pathways. eyg and ey are coexpressed in the eye disc primordium in the embryo (Fig. 7). Their expression domain also overlap in the eye disc,especially in the early eye disc (Fig. 4) when eyg function is critically required. So it is possible that the two Pax proteins act within the same cell, although we do not rule out the possibility they act in different cells to achieve a functional cooperativity.

If eyg and ey are both required for eye development, how could ectopic expression of either one be sufficient for ectopic eye development? One possibility is that the two Pax proteins form heterodimers,directly or indirectly via other proteins, to activate target genes. When the level of either one is low, the target genes that lead to eye formation cannot be induced. However, when either one is strongly expressed ectopically, the high level of homodimer can partially substitute for the heterodimer. Since both genes are required for normal eye formation, this model predicts that the EYG-EY heterodimer is more effective than either homodimer in inducing eye formation. As expected by this model, coexpression of eyg and ey caused enhanced ectopic eye formation.

Possible mechanisms of EYG protein function

The EYG protein has two DNA binding domains: the RED subdomain of its truncated PD and the Prd-class HD. It probably functions as a transcription factor by binding DNA targets through these domains, singly or in combination. In addition, its interaction with other proteins may affect this DNA binding.

The PD consists of two independent subdomains: the N-terminal PAI and the C-terminal RED subdomains. Based on crystal structure of the human Pax6 PD,the linker region connecting the two subdomains also contacts DNA(Xu et al., 1999). In EYG, the PAI subdomain is largely missing and most likely cannot bind DNA. One interesting possibility is that the truncated EYG PD has a dominant negative effect, competing with other PD proteins. In addition, truncation of the PAI subdomain in the Pax6-5a and Pax8(S) isoforms probably exposes the RED subdomain to recognize a distinctly different DNA sequence(Epstein et al., 1994; Kozmik et al., 1998). Thus the EYG PD may bind DNA through its RED domain, similarly to the Pax6-5a and Pax8(S) isoforms (Epstein et al.,1994; Kozmik et al.,1998) and distinct from the Pax6 PD. This prediction was in fact proved by site-selection using the EYG RED domain(Jun et al., 1998). Through its RED domain, EYG can probably regulate different target genes than those regulated by EY. This ey-independent function of eyg is also shown by its involvement in salivary gland development(Jones et al., 1998), and in bristle formation when ectopically expressed (see Results). Vertebrate homologs of EYG have not yet been identified. It is possible that EYG plays a role equivalent to the vertebrate Pax6-5a isoform.

In addition to the PD, many Pax proteins (including EY and EYG) also contain a Prd-class homeodomain. Two Prd-type HDs can bind cooperatively to a palindromic site composed of two inverted TAAT motifs separated by 2 or 3 bps(Wilson et al., 1993). The Prd-type HD of EYG can form heterodimers with the Prd-type HD of Prd upon binding to a consensus DNA target (Wilson et al., 1993; Wilson et al.,1996). It is possible that EY and EYG also form heterodimers via their HDs. This would be consistent with our findings that they act synergistically. However, although the HD of EYG is required for its functions(J. G. Yao and Y.H.S., unpublished), the HD of EY has been shown not to be required for its function in eye development(Punzo et al., 2001). Thus the HD of EYG is required, not for direct interaction with the HD of EY, but may be for DNA binding or for interacting with other proteins.

eyg suppresses wg transcription in the eye disc

Dpp and Wg are two signaling molecules important for the initiation of eye differentiation: Dpp activates MF initiation while Wg suppresses it(Heberlein et al., 1993; Wiersdorff et al., 1996; Chanut and Heberlein, 1997; Pignoni and Zipursky, 1997; Ma and Moses, 1995; Treisman and Rubin, 1995). Does eyg exert its effect on eye development by activating Dpp signaling or by suppressing Wg signaling?

dpp is expressed at two stages in the eye disc: an early expression along the posterior and lateral margins (represented by the dpp-GAL4), and a later expression in the propagating MF (represented by the dpp-lacZ). The early expression in the margins is required for MF initiation (Burke and Basler,1996; Wiersdorff et al.,1996). It was found that dpp expression along the lateral margins is absent in early third instar eyg1 eye disc(Hazelett et al., 1998),suggesting that dpp expression in the lateral margins is regulated by eyg. However, activating DPP signaling at the lateral margin did not rescue the eyg1 phenotype(Hazelett et al., 1998),suggesting that eyg has other functions in addition to activating dpp expression.

wg is expressed uniformly in the eye disc of second instar larvae(Royet and Finkelstein, 1997). In the third instar eye disc, wg is expressed in the lateral margins and acts to prevent MF initiation from the lateral margins(Ma and Moses, 1995; Treisman and Rubin, 1995). The wg-expression domain expands in eyg1 eye discs(Fig. 8C)(Hazelett et al., 1998). Our results further showed that ectopic eyg expression(dpp>eyg) could suppress wg expression at the transcriptional level. The suppression of wg is functionally significant, because expression of the wg-activated omb gene is similarly suppressed in dpp>eyg (J.-L. Chao and Y.H.S., unpublished). Hazelett et al.(Hazelett et al., 1998) have shown that blocking of the Wg signaling pathway can partially rescue the eyg mutant phenotype. These results indicate that the suppression of wg transcription by eyg may be a major mechanism by which eyg induces MF initiation, hence eye development. This is consistent with our finding that ectopic eyg induces ectopic eye formation primarily in the ventral margin of the eye disc, where wg expression is weaker (Fig. 8A) and most easily suppressed by dpp (Pignoni and Zipursky, 1997). wg is normally expressed in the entire eye disc during second instar(Royet and Finkelstein, 1997). It was shown that Wg signaling can suppress the expression of so and eya (Baonza and Freeman,2002). It is possible that in the late second instar eye disc, eyg expression in the central domain of the eye disc suppresses wg expression in the central domain, thus allowing the expression of eya and so, hence eye development.

As predicted by the eyg and ey interaction, eyalso suppresses wg expression (data not shown). Suppression of wg expression by eyg (and ey) is also seen in the wing disc (Fig. 8E). However,suppression does not occur in all cells expressing eyg, suggesting that additional factors are required for the wg suppression. The relationship of eyg/ey and wg may be mutually antagonistic, since ectopic ey cannot induce eya and so expression in regions of high wg expression(Halder et al., 1998).

We thank Patrick Callaerts, Jessica Treisman, Uwe Walldorf and the Bloomington Stock Center for providing the fly stocks, Jyh-Lyh Juang for the toe EST clone, and Seymor Benzer for mAb22C10. We are grateful to Cheng-Ting Chien, Don Ready, Jessica Treisman and Gert Pflugfelder for valuable suggestions, and to Jessica Treisman for communicating unpublished results. We thank Adelaide Carpenter for helping with the polytene cytology,to Chien-hui Chuang (Department of Zoology, National Taiwan University),Yu-Chen Tsai (Institute of Genetics, National Yang-Ming University) and Shing-Li Chen (Institute of Life Science, National Defense Medical Center) for eyg mutant screens, to Ya-Chieh Hsu (Taipei Municipal First Girls'Senior High School) for help with the E132-GAL4 induction experiments, to Sue-Ping Li and An-Li Huang for help with the SEM, to Chen-shu Hsieh and Chun-lan Hsu for preparing fly food, to Helen Lo for help with cloning and to Chiou-Yang Tang for maintaining fly stocks. This study was supported by grants to Y.H.S. from the Academia Sinica and the National Science Council (NSC-83-0211-B-001-041, NSC-87-2311-B-001-038-B26,NSC-88-2312-B-001-016) of the Republic of China, to S.K.B. from NIH and to C.D. from HHMI.

Adams, M. D., Celniker, S. E., Holt, R. A. et al.(
2000
). The genome sequence of Drosophila melanogaster.
Science
287
,
2185
-2195.
Baker, N. E. and Rubin, G. M. (
1992
). Ellipse mutations in the Drosophila homologue of the EGF receptor affect pattern formation, cell division, and cell death in eye imaginal discs.
Dev. Biol.
150
,
381
-396.
Baonza, A. and Freeman, M. (
2002
). Control of Drosophila eye specification by Wingless signaling.
Development
129
,
5313
-5322.
Benassayag, C., Plaza, S., Callaerts, P., Clements, J., Romeo,Y., Gehring, W. J. and Cribbs, D. L. (
2003
). Evidence for a direct functional antagonism of the selector genes proboscipedia and eyeless in Drosophila head development.
Development
130
,
575
-586.
Blackman, R. K., Sanicola, M., Raftery, L. A., Gillevet, T. and Gelbart, W. M. (
1991
). An extensive 3′ cis-regulatory region directs the imaginal disk expression of decapentaplegic, a member of the TGF-beta family in Drosophila.
Development
111
,
657
-666.
Bonini, N. M., Leiserson, W. M. and Benzer, S.(
1993
). The eyes absent gene: genetic control of cell survival and differentiation in the developing Drosophila eye.
Cell
72
,
379
-395.
Bonini, N. M., Bui, Q. T., Gray-Board, G. L. and Warrick, J. M. (
1997
). The Drosophila eyes absent gene directs ectopic eye formation in a pathway conserved between flies and vertebrates.
Development
124
,
4819
-4826.
Brand, A. H. and Perrimon, N. (
1993
). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes.
Development
118
,
401
-415.
Burglin, T. R. (
1994
). A comprehensive classification of homeobox genes. In
Guidebook to the Homeobox Genes
(ed. D. Duboule), pp.
25
-246. New York, NY: Oxford University Press.
Burke, R. and Basler, K. (
1996
). Hedgehog-dependent patterning in the Drosophila eye can occur in the absence of Dpp signaling.
Dev. Biol.
179
,
360
-368.
Chanut, F. and Heberlein, U. (
1997
). Role of decapentaplegic in initiation and progression of the morphogenetic furrow in the developing Drosophila retina.
Development
124
,
559
-567.
Chen, R., Amoui, M., Zhang, Z. and Mardon, G.(
1997
). Dachshund and Eyes Absent proteins form a complex and function synergistically to induce ectopic eye development in Drosophila.
Cell
91
,
893
-903.
Chen, R., Halder, G., Zhang, Z. and Mardon, G.(
1999
). Signaling by the TGF-beta homolog decapentaplegic functions reiteratively within the network of genes controlling retinal cell fate determination in Drosophila.
Development
126
,
935
-943.
Cheyette, B. N., Green, P. J., Martin, K., Garren, H.,Hartenstein, V. and Zipursky, S. L. (
1994
). The Drosophila sine oculis locus encodes a homeodomain-containing protein required for the development of the entire visual system.
Neuron
12
,
977
-996.
Czerny, T., Schaffner, G. and Busslinger, M.(
1993
). DNA sequence recognition by Pax proteins: bipartite structure of the paired domain and its binding site.
Genes Dev.
7
,
2048
-2061.
Czerny, T., Halder, G., Kloter, U., Souabni, A., Gehring, W. J. and Busslinger, M. (
1999
). twin of eyeless, a second Pax-6 gene of Drosophila, acts upstream of eyelessin the control of eye development.
Mol. Cell
3
,
297
-307.
Davis, I., Girdham, C. H. and O'Farrell, P. H.(
1995
). A nuclear GFP that marks nuclei in living Drosophila embryos: maternal supply overcomes a delay in the appearance of zygotic fluorescence.
Dev. Biol.
170
,
726
-729.
Desplan, C. (
1997
). Eye development: governed by a dictator or a junta?
Cell
91
,
861
-864.
Epstein, J. A., Glaser, T., Cai, J., Jepeal, L.. Walton, D. S. and Mass, R. L. (
1994
). Two independent and interactive DNA-binding subdomains of the Pax6 paired domain are regulated by alternative splicing.
Genes Dev.
8
,
2022
-2034.
FlyBase Consortium (
2003
). The FlyBase database of the Drosophila genome projects and community literature.
Nucleic Acids Res.
31
,
172
-175.http://flybase.org/
Fujita, S. C., Zipursky, S. L., Benzer, S., Ferrus, A. and Shotwell, S. L. (
1982
). Monoclonal antibodies against the Drosophila nervous system.
Proc. Natl. Acad. Sci. USA
79
,
7929
-7933.
Halder, G., Callaerts, P. and Gehring, W. J.(
1995
). Induction of ectopic eyes by targeted expression of the eyeless gene in Drosophila.
Science
267
,
1788
-1792.
Halder, G., Callaerts, P., Flister, S., Walldorf, U., Kloter, U. and Gehring, W. J. (
1998
). Eyeless initiates the expression of both sine oculis and eyes absent during Drosophila compound eye development.
Development
125
,
2181
-2191.
Hay, B. A., Wolff, T. and Rubin, G. M. (
1994
). Expression of baculovirus P35 prevents cell death in Drosophila.
Development
120
,
2121
-2129.
Hazelett, D., Bourouis, M., Walldorf, U. and Treisman, J. E.(
1998
). decapentaplegic and wingless are regulated by eyes absent and eyegone and interact to direct the pattern of retinal differentiation in the eye disc.
Development
125
,
3741
-3751.
Heberlein, U., Wolff, T. and Rubin, G. M.(
1993
). The TGF beta homolog dpp and the segment polarity gene hedgehog are required for propagation of a morphogenetic wave in the Drosophila retina.
Cell
75
,
913
-926.
Hunt, D. M. (
1970
). Lethal interactions of the eye-gone and eyeless mutants in Drosophila melanogaster.
Genet. Res.
15
,
29
-34.
Jiao, R., Daube, M., Duan, H., Zou, Y., Frei., E. and Noll,M. (
2001
). Headless flies generated by developmental pathway interference.
Development
128
,
3307
-3319.
Jones, N., Kuo, Y. M., Sun, Y. H. and Beckendorf, S. K.(
1998
). The Drosophila Pax gene eye gone is required for embryonic salivary duct development.
Development
125
,
4163
-4174.
Jun, S. and Desplan, C. (
1996
). Cooperative interactions between paired domain and homeodomain.
Development
122
,
2639
-2650.
Jun, S., Wallen, R. V., Goriely, A., Kalionis, B. and Desplan,C. (
1998
). Lune/eye gone, a Pax-like protein, uses a partial paired domain and a homeodomain for DNA recognition.
Proc. Natl. Acad. Sci. USA
95
,
13720
-13725.
Kammermeier, L., Leemans, R., Hirth, F., Flister, S., Wenger,U., Walldorf, U., Gehring, W. J. and Reichert, H. (
2001
). Differential expression and function of the Drosophila Pax6 genes eyeless and twin of eyeless in embryonic central nervous system development.
Mech. Dev.
103
,
71
-78.
Kango-Singh, M., Singh, A. and Sun, Y. H.(
2003
). Eyeless collaborates with Hedgehog and Decapentaplegic signaling in Drosophila eye induction.
Dev. Biol.
256
,
49
-61.
Kobayashi, S. and Okada, M. (
1993
). A double staining technique using 5-bromo,4-chloro-3-indolyl-b-D-galactopyranoside(Xgal) and immunoperoxidase in whole Drosophila embryos.
Biotech. Histochem.
68
,
237
-239.
Kozmik, Z., Czerny, T. and Busslinger, M.(
1998
). Alternatively spliced insertions in the paired domain restrict the DNA sequence specificity of Pax6 and Pax8.
EMBO J.
16
,
6793
-6803.
Kronhamn, J., Frei, E., Daube, M., Jiao, R., Shi, Y. and Noll,M. (
2002
).Headless flies produced by mutations in the paralogous Pax6 genes eyeless and twin of eyeless.
Development
129
,
1015
-1026.
Kumar, J. P. and Moses, K. (
2001
) Eye specification in Drosophila: perspectives and implications.
Semin. Cell Dev. Biol.
12
,
469
-474.
Ma, C. and Moses, K. (
1995
). winglessand patched are negative regulators of the morphogenetic furrow and can affect tissue polarity in the developing Drosophila compound eye.
Development
121
,
2279
-2289.
Mardon, G., Solomon, N. M. and Rubin, G. M.(
1994
). dachshund encodes a nuclear protein required for normal eye and leg development in Drosophila.
Development
120
,
3473
-3486.
Noll, M. (
1993
). Evolution and role of Pax genes.
Curr. Opin. Genet. Dev.
3
,
595
-605.
Pan, D. and Rubin, G. (
1998
). Targeted expression of teashirt induces ectopic eyes in Drosophila.
Proc. Natl. Acad. Sci. USA
95
,
15508
-15512.
Pignoni, F., Hu, B., Zavitz, K. H., Xiao, J., Garrity, P.A. and Zipursky, S. L. (
1997
). The eye-specification proteins So and Eya form a complex and regulate multiple steps in Drosophila eye development.
Cell
91
,
881
-891.
Pignoni, F. and Zipursky, S. L. (
1997
). Induction of Drosophila eye development by Decapentaplegic.
Development
124
,
271
-278.
Punzo, C., Kurata, S. and Gehring, W. J.(
2001
). The eyeless homeodomain is dispensable for eye development in Drosophila.
Genes Dev.
15
,
1716
-1723.
Quiring, R., Walldorf, U., Kloter, U. and Gehring, W. J.(
1994
). Homology of the eyeless gene of Drosophila to the Small eye gene in mice and Aniridia in humans.
Science
265
,
785
-789.
Royet, J. and Finkelstein, R. (
1997
). Establishing primordia in the Drosophila eye-antennal imaginal disc:the roles of decapentaplegic, wingless and hedgehog.
Development
124
,
4793
-4800.
Rubin, G. M. and Spradling, A. C. (
1982
). Genetic transformation of Drosophila with transposable element vectors.
Science
218
,
348
-353.
Seimiya, M. and Gehring, W. J. (
2000
). The Drosophila homeobox gene optix is capable of inducing ectopic eyes by an eyeless-independent mechanism.
Development
127
,
1879
-1886.
Shen, W. and Mardon, G. (
1997
). Ectopic eye development in Drosophila induced by directed dachshundexpression.
Development
124
,
45
-52.
Singh, A., Kango-Singh, M. and Sun, Y. H.(
2002
). Eye suppression, a novel function of teashirt,requires Wingless signaling.
Development
129
,
4271
-4280.
Spreij, T. E. (
1971
). Cell death during the development of the imaginal disks of Calliphora erythrocephala.
Neth. J. Zool.
21
,
221
-264.
Staehling-Hampton, K., Jackson, P. D., Clark, M. J., Brand, A. H. and Hoffmann, F. M. (
1994
). Specificity of bone morphogenetic protein-related factors: cell fate and gene expression changes in Drosophila embryos induced by decapentaplegic but not 60A.
Cell Growth Differ.
5
,
585
-593.
Sun, Y. H., Tsai, C.-J., Green, M. M., Chao, J.-L., Yu, C.-T.,Jaw, T. J., Yeh, J.-Y. and Bolshakov, V. N. (
1995
). white as a reporter gene to detect transcriptional silencers specifying position-specific gene expression during Drosophilamelanogaster eye development.
Genetics
141
,
1075
-1086.
Tautz, D. and Pfeifle, C. (
1989
). A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback.
Chromosoma
98
,
81
-85.
Thummel, C. S. and Pirrotta, V. (
1992
). New pCaSpeR P element vectors.
Dros. Info. Serv.
71
,
150
.
Treisman, J. E. (
1999
). A conserved blueprint for the eye?
BioEssays
21
,
843
-850.
Treisman, J. E. and Rubin, G. M. (
1995
). wingless inhibits morphogenetic furrow movement in the Drosophila eye disc.
Development
121
,
3519
-3527.
Whalen, J. H. and Grigliatti, T. A. (
1998
). Molecular characterization of a retrotransposon in Drosophila melanogaster, nomad, and its relationship to other retrovirus-like mobile elements.
Mol. Gen. Genet.
260
,
401
-409.
Wiersdorff, V., Lecuit, T., Cohen, S. M. and Mlodzik, M.(
1996
). Mad acts downstream of Dpp receptors, revealing a differential requirement for dpp signaling in initiation and propagation of morphogenesis in the Drosophila eye.
Development
122
,
2153
-2162.
Wilson, D., Sheng, G., Lecuit, T., Dostatni, N. and Desplan,C. (
1993
). Cooperative dimerization of paired class homeo domains on DNA.
Genes Dev.
7
,
2120
-2134.
Wilson, D. S., Sheng, G., Jun, S. and Desplan, C.(
1996
). Conservation and diversification in homeodomain-DNA interactions: A comparative genetic analysis.
Proc. Natl. Acad. Sci. USA
93
,
6886
-6891.
Wolff, T. and Ready, D. F. (
1993
). Pattern formation in the Drosophila retina. In The Development ofDrosophila melanogaster (ed. M. Bate and A. Martinez Arias), pp.
1277
-1325. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Xu, T. and Rubin, G. M. (
1993
). Analysis of genetic mosaics in developing and adult Drosophila tissues.
Development
117
,
1223
-1237.
Xu, W., Rould, M. A., Jun, S., Desplan, C. and Pabo, C. O.(
1995
). Crystal structure of a paired domain-DNA complex at 2.5Å resolution reveals structural basis for Pax developmental mutations.
Cell
80
,
639
-650.
Xu, H. E., Rould, M. A., Xu, W., Epstein, J. A., Maas, R. L. and Pabo, C. O. (
1999
). Crystal structure of the human Pax6 paired domain-DNA complex reveals specific roles for the linker region and carboxy-terminal subdomain in DNA binding.
Genes Dev.
13
,
1263
-1275.
Zipursky, S. L., Venkatesh, T. R., Teplow, D. B. and Benzer,S. (
1984
). Neuronal development in the Drosophilaretina: monoclonal antibodies as molecular probes.
Cell
36
,
15
-26.