Molecular Genetic Analysis of Drosophila eyes absent Mutants Reveals an Eye Enhancer Element

Corresponding author: Nancy M. Bonini, Department of Biology, 415 S. University Ave., University of Pennsylvania, Philadelphia, PA 19104-6018., nbonini{at}sas.upenn.edu (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The eyes absent (eya) gene is critical for normal eye development in Drosophila and is highly conserved to vertebrates. To define regions of the gene critical for eye function, we have defined the mutations in the four viable eya alleles. Two of these mutations are eye specific and undergo transvection with other mutations in the gene. These were found to be deletion mutations that remove regulatory sequence critical for eye cell expression of the gene. Two other viable alleles cause a reduced eye phenotype and affect the function of the gene in additional tissues, such as the ocelli. These mutations were found to be insertion mutations of different transposable elements within the 5' UTR of the transcript. Detailed analysis of one of these revealed that the transposable element has become subject to regulation by eye enhancer sequences of the eya gene, disrupting normal expression of EYA in the eye. More extended analysis of the deletion region in the eye-specific alleles indicated that the deleted region defines an enhancer that activates gene expression in eye progenitor cells. This enhancer is responsive to ectopic expression of the eyeless gene. This analysis has defined a critical regulatory region required for proper eye expression of the eya gene.

DROSOPHILA eye development is a genetic system in which it is possible to define molecular details of specification and differentiation of a neural structure. Much research has focused on aspects of cell fate specification and pattern formation of the retinal cells during later stages of differentiation (reviewed in ZIPURSKY and RUBIN 1994 ; TREISMAN and HEBERLEIN 1998 ). More recent attention has focused on earlier events of eye specification, which has been shown to involve a set of genes highly conserved in vertebrates. These genes include twin of eyeless (toy) and eyeless (ey), which are PAX-6 homeodomain homologues (QUIRING et al. 1994 ; CZERNY et al. 1999 ), sine oculis (so), which defines the SIX class of vertebrate homeodomain homologies (CHEYETTE et al. 1994 ; SERIKAKU and O'TOUSA 1994 ; OLIVER et al. 1995 ), dachshund (dac) and the vertebrate Dach homologue (MARDON et al. 1994 ; HAMMOND et al. 1998 ), and eyes absent (eya) and the vertebrate Eya1, Eya2, and Eya3 genes (BONINI et al. 1993 ; DUNCAN et al. 1997 ; XU et al. 1997 ; ZIMMERMAN et al. 1997 ).

In Drosophila, ey was the first gene shown to display the dramatic capacity to direct eye formation when ectopically expressed (HALDER et al. 1995 )—a property now known to be shared by dac, eya, and toy, and so when combined with eya (BONINI et al. 1997 ; CHEN et al. 1997 ; PIGNONI et al. 1997 ; SHEN and MARDON 1997 ). dpp, a homologue of transforming growth factor-ß (PADGETT et al. 1987 ), is also involved in appropriate regulation of these eye regulatory pathways and may also be required for ectopic eye induction (CHEN et al. 1999 ). The eya gene appears to be a critical player in eye formation, since it shows functional synergy with both dac and so in directing eye development (CHEN et al. 1997 ; PIGNONI et al. 1997 ). As the EYA protein also physically binds these proteins, EYA, DAC, and SO may form a complex that is critical to eye specification. toy appears to function upstream of ey to activate ey expression (CZERNY et al. 1999 ). For the other genes, it appears that regulatory loops are involved, rather than there being a linear pathway of eye specification, since expression of any one appears to turn on the expression and functionally require the other genes (previous references; DESPLAN 1997 ). Moreover, eya activity has multiple downstream targets and may influence or be required for a number of different processes of eye development (PIGNONI et al. 1997 ; HAZELETT et al. 1998 ). To define additional molecular aspects of the process of eye specification, we have focused on defining molecular mechanisms of proper regulation of the eya gene in eye formation.

The eya gene is essential for the formation of the adult eye in Drosophila, as well as being required during embryogenesis for proper head involution and for proper gonad formation (BONINI et al. 1993 , BONINI et al. 1998 ; BOYLE et al. 1997 ). The eya gene produces two alternatively spliced transcripts called type I and type II, which predict novel proteins that differ only slightly at the N terminus (BONINI et al. 1993 ). These transcripts are both expressed in the same pattern in eye tissue and both function the same in eye development (LEISERSON et al. 1998 ). In eye progenitor cells, eya mutants exhibit hyperproliferation followed by massive programmed cell death anterior to the morphogenetic furrow, which leads to loss of eye progenitor cells (BONINI et al. 1993 ; PIGNONI et al. 1997 ).

We examined the molecular defects in selected alleles of eya to define regions of the gene critical for eye progenitor cell expression. Here we present analysis of the viable alleles of eya, which cause malformation of the adult eye. We show that the eya¹ and eya² mutations define an enhancer element necessary for expression of eya in eye progenitor cells and that the eya^3cs and eya⁴ mutations are due to transposable element insertions that disrupt gene expression in the eye.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks and mutagenesis:
eya alleles have been previously described in BONINI et al. 1993 , BONINI et al. 1998 . Flies were grown on standard corn meal, molasses media, supplemented with dry yeast. All crosses were performed at 25°, unless otherwise noted. For ectopic expression studies, dpp-GAL4 (STAEHLING-HAMPTON et al. 1994 ), UAS-eya (BONINI et al. 1997 ; PIGNONI et al. 1997 ), and UAS-ey (HALDER et al. 1995 ) were used.

Immunocytochemistry and histology:
Immunostaining was performed as described (BONINI et al. 1997 ). Primary antibodies were anti-EYA Mab10H6 (1:10, BONINI et al. 1997 ) and anti-ELAV (1:5, O'NEILL et al. 1994 ; Developmental Studies Hybridoma Bank). All secondary antibodies were from Jackson ImmunoResearch Laboratories, conjugated to either fluorescein or Texas red. Confocal microscopy was performed on a Leica model TCS SP ultraviolet and visible confocal imaging spectrophotometer microscope. X-Gal staining was performed as described (SIMON et al. 1985 ).

In situ analysis:
In situ analysis was performed as described (BONINI et al. 1993 ). For the I element study, a 0.6-kb HindIII/KpnI fragment of the I element plasmid pY62A, which corresponds to the 3' end of the I element, was used (GEYER et al. 1986 ; plasmid gift of Dr. V. Corces).

Isolation of genomic DNA:
A total of 20–100 flies were placed in a 3-ml Dounce homogenizer on ice, and 1 ml of cold homogenization buffer [(HB) 10 mM Tris-HCl, pH 8.0; 60 mM NaCl; 10 mM EDTA; 150 µM spermidine; and 0.5% Triton X-100] was added. The homogenate was filtered through a fine mesh screen. After centrifugation at 5000 rpm for 5 min at 4°, the supernatant was removed and the pellet was suspended in 1 ml of HB. Centrifugation was repeated and the pellet was suspended in 450 µl HB. Proteinase K and Sarkosyl were added to a final concentration of 200 µg/ml and 2%, respectively, and the reaction was incubated at 50° overnight. Fifty microliters of 3 M sodium acetate, pH 5.5, was added and the reaction was extracted with equal volumes of chloroform/phenol twice and chloroform once. The DNA was precipitated by addition of 1 ml of 100% ethanol and centrifuged at 1000 rpm for 1 min. The genomic DNA pellet was washed with 70% ethanol and suspended in 25–50 µl distilled H₂O. All DNA samples were placed at 60° for 30 min and then stored at -20°.

Defining and sequencing mutations:
The location of deletions in eya¹ and eya² were determined first by genomic Southern analysis using probes to the eya region (see BONINI et al. 1993 ). eya¹ had strain polymorphisms identical to the SM6a balancer chromosome in the eya genomic region, whereas eya² had polymorphisms identical to those of Oregon-R wild-type strain. Comparison of the restriction maps between the control and mutant strains indicated both alleles had deletions as noted.

The eya² deletion was further localized to the 4.5-kb EcoRI/SalI restriction fragment that contains the first exon of the type I transcript using amplification. Amplification products were made from both Oregon-R and eya² genomic DNA using primers NB143 (5'-GGAGGATTCCATGTCCTCGG-3'), which corresponds to sequence just 3' to the SalI site on the genomic map, and NB80 (5'-GAGATATACATCCATTCAAAACCCA-3'), which corresponds to sequence 5' to the first exon. The products were subcloned into the pGEM-T vector (Promega, Madison, WI), pENH-OR, and pENH-PH, respectively, and sequenced. The sequences deleted in the eya² mutation were determined by comparison to the Oregon-R amplification product, pENH-OR, using Seq Ed v1.0 (Applied Biosystems, Foster City, CA). Genomic DNA was sequenced at least twice in independent amplification products.

The eya3-lacZ (eya-enhancer–lacZ) transgene was constructed by using primers 5'-GGATCCAGAGGAGACGAAACTGGC-3' and 5'-TGATCAATTAACTGACCTGCTCAACTC-3' to amplify the 322 bp corresponding to the eya² deletion region, incorporating BamHI and BclI restriction sites at the 5' and 3' ends, respectively. This was sequenced to confirm fidelity, then excised as a BamHI-BclI fragment and inserted into the BamHI and BclI sites within the polylinker of pSL1190 (Stratagene, La Jolla, CA). A concatamer of three tandem repeats was constructed by repeated subcloning of the 322-bp BamHI-BclI fragment into the BclI site of the pSL1190 vector containing the previous 322-bp subclone. This region was then cloned upstream of the hsp43 minimal promoter and the ß-galactosidase gene of the pCasper-hs43-ß-gal vector (THUMMEL et al. 1988 ) to yield eya3-lacZ or eya6-lacZ, respectively, in the forward orientation.

The eya^3cs and eya⁴ insertions were determined from sequence analysis of PCR amplification products using primers NB80 and NB82 (5'-ATTTGGTTGTCTGCAGTGAAAAGCG-3'), which correspond to sequences 3' and 5' of the first exon of the type I transcript, respectively. The sites of insertion for eya^3cs and eya⁴ were determined by aligning the sequences of the PCR products with sequence from eya type I cDNA (GenBank accession no. L08501).

Concentrated and purified PCR products were sequenced using a dye terminator cycle sequencing kit (Perkin-Elmer, Norwalk, CT) and manufacturer's instructions with the exception that all reactions contained DMSO. The program for sequencing was as follows: 95° for 3 min; 25 cycles of 96° for 30 sec, 50° for 1 min, and 60° for 4 min; and a 4° soak. After amplification, unincorporated nucleotides were removed using Centrisep gel filtration columns (Princeton Separations). The reaction was dried down under vacuum. Sequence was run on ABI 377 sequencers with Stretch upgrade using BigDye Taq chemistry. Sequence was analyzed and compared using the DNA Sequencher 3.0 Program (Gene Codes Corporation, Ann Arbor, MI).

Determination of transcription start site:
The start of transcription was determined by ribonuclease protection assay (RPA). Two plasmids, pPST and p278-279, were used for analysis. pPST plasmid contains an 827-bp PstI fragment of the pENH-OR plasmid subcloned into the pBluescript KS II vector. The PstI fragment contains the first exon and sequences 5' to it. Plasmid p278-279 is an amplification product subcloned into the TA cloning vector (Invitrogen, Carlsbad, CA). The amplification product subcloned in p278-279 was produced using the primers NB278 (5'-GCTGAGAAAACTCACTCAAAAGCG-3') and NB279 (5'-TATTTCAGTTTAAGCGCTGGGCCG-3').

The probes for ribonuclease protection were prepared using [³²P]UTP, purified by gel electrophoresis, and extracted using standard procedures, except 20 pmol cold UTP was also added to the transcription reaction (AUSABEL 1994). The ribonuclease protection assay was performed using the Ambion RPA II kit following manufacturer's instructions with the following exceptions: the extracted RNA was recovered by ethanol precipitation and suspended in 50 µl of hybridization buffer, 25 µg of total RNA was dried under vacuum and suspended in 15 µl of hybridization buffer-containing probe, and the hybridization was incubated for 3 min at 95° and transferred to 32°, 37°, 42°, and 47° for overnight incubation. The digested products were electophoresed using a denaturing 8% polyacrylamide gel containing 8 M urea.

The start site was confirmed by primer extension. Standard procedures were followed with the following modification: after labeling, the oligonucleotides were purified using Probequant G-50 columns (Pharmacia, Piscataway, NJ) prepared according to manufacturer's instructions.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The viable eya alleles show differential protein expression in the eye imaginal disc:
The viable alleles of eya are homozygous recessive mutations that affect the formation of the adult compound eye (Fig 1; Table 1). The eya¹ and eya² mutations are highly specific in that they lead to selective and complete loss of the compound eyes. The other alleles, eya⁴ and eya^3cs, affect the compound eyes in addition to other tissues (BONINI et al. 1998 ). The eya⁴ mutation shows severely reduced eyes and loss of the ocelli, where the eya^3cs mutation shows severely reduced compound eyes, as well as partial to complete loss of the ocelli and female sterility (Fig 1).

View larger version (118K): In this window In a new window Download PPT slide   Figure 1. Effects of eya viable alleles on the eyes and ocelli. Photographs of the eyes and ocelli of the different eya viable alleles. (A and F) Compound eye (A) and ocelli of a wild-type fly (F). (B and G) The eya2 mutant completely lacks the compound eyes (B), whereas the ocelli are normal (G). (C and H) The eya4 mutant has a greatly reduced, although not completely missing, eye (C). In the eya4 mutant, the ocelli are also reduced (H). (D and E) The compound eyes of the eya3cs allele grown at high temperature (D, 30°), and at low temperature (E, 18°).

We addressed whether EYA protein expression was altered in the eya viable alleles in a manner consistent with the observed adult phenotypes. Normally during eye formation, the EYA protein is expressed anterior to the morphogenetic furrow where it has a critical role in eye formation (BONINI et al. 1993 ). EYA expression is maintained posterior to the furrow as development progresses across the eye disc and occurs in progenitor cells of the ocelli (Fig 2A and Fig B). In the eya¹ and eya² mutants, EYA protein was not detectable in the eye progenitor cells (Fig 2C and Fig D). However, the ocellar expression was normal, consistent with the highly specific loss of the adult compound eyes, but with normal ocellar formation in these alleles. In the eya⁴ mutation, EYA expression was greatly reduced in the eye progenitor field and was missing entirely from the ocellar region (Fig 2E and Fig F), consistent with a phenotype in both the eyes and ocelli of eya⁴ adult flies. The eya^3cs allele also showed reduced eye disc expression (data not shown).

View larger version (96K): In this window In a new window Download PPT slide   Figure 2. Eye disc expression of Eya protein in the viable eya mutants. Expression of EYA protein (A, C, and E) was detected with eyaMAb10H6. Neural differentiation was highlighted with an antibody to ELAV (B, D, and F). Normal eye discs (A and B), eya2 mutant eye discs (C and D), and eya4 mutant eye discs (E and F). (A and B) The expression pattern of the EYA protein in normal eye discs (A). ELAV expression (B), which labels neurons, highlights the differentiating photoreceptor neurons posterior to the furrow and in the ocellar region. EYA is expressed both anterior and posterior to the morphogenetic furrow (arrow), as well as in the ocellar progenitors far anterior to the furrow (arrowhead). (C and D) In the eya2 mutant, EYA expression (C) is completely absent from the compound eye field and is present only in the ocellar region (arrowhead). No compound eye development occurs, highlighted by lack of ELAV staining in the developing eye field (D). ELAV expression in the developing ocelli (arrowhead) is normal. In eya2 mutants, the adult compound eyes are entirely missing, whereas the ocelli are normal (see Fig 1). These data indicate that the eya2 allele selectively affects gene expression in the eye progenitor cells. (E and F) In the eya4 mutant, EYA expression occurs, but is greatly reduced (E). A greatly reduced eye field develops, indicated by reduced ELAV expression (F), consistent with the reduced compound eyes that are seen in the adult (see Fig 1). No ocellar expression of EYA or ELAV is apparent; in eya4 adult flies, the ocelli are missing. These data indicate that the eya4 allele reduces EYA expression in both the eye and ocellar progenitors, unlike the eya1 and eya2 alleles that selectively and completely eliminate Eya expression in the eye progenitor cells.

The eya gene produces two alternative transcripts that are both expressed in the eye disc in the same pattern (LEISERSON et al. 1998 ). These and other data indicate that both transcripts function and are regulated in the same manner in eye progenitor cells and that the eye expression of both is affected in the viable mutants (LEISERSON et al. 1998 ; above). The eya¹ and eya² alleles show transvection with eya⁴ (LEISERSON et al. 1994 ), indicating these alleles may affect regulatory regions. Therefore, a molecular characterization of the viable mutations was undertaken to determine whether they defined critical regulatory elements required for proper expression of the gene in the eye.

The eya^3cs and eya⁴ mutations are associated with transposable element insertions:
Analysis of the eya^3cs and eya⁴ mutants, which are not eye specific (see Fig 1 and Table 1), demonstrated that all exons of the gene appeared normal except for the first exon of the type I transcript, which failed to amplify by standard techniques in both mutants. Southern blot analysis confirmed restriction enzyme polymorphisms in the region of the exon and was suggestive of insertional events in both alleles. Amplification of the exon from the alleles using long-range amplification conditions produced products of 9 and 6 kb for the eya^3cs and eya⁴ alleles, respectively, rather than the expected product of 0.8 kb. Sequence analysis and a database search revealed interruption of the eya sequence in the eya^3cs amplification product with a roo element (SCHERER et al. 1982 ; Fig 3). In the eya⁴ allele, sequence analysis indicated interruption of the eya sequence with an I element (BUCHETON et al. 1984 ; FAWCETT et al. 1986 ). Restriction enzyme analysis of the amplification products confirmed insertion of a roo element in the eya^3cs allele and I element in the eya⁴ allele. Both elements inserted into the 5' untranslated region (UTR) of the eya gene downstream of the transcription start site, but upstream of the protein translation initiation codon. The insertion events duplicated 5 bp of genomic sequence in the case of the eya^3cs allele and 8 bp in the eya⁴ allele. The sites of insertion of these two independently generated mutations were within 5 bp (Fig 3C and Fig 4).

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. The eya genomic region, with mutation sites of the viable alleles. (A) The eya genomic region. The positions of the eya1 and eya2 deletions are indicated, as well as the insertion sites of the transposable elements in the eya3cs and eya4 alleles. E, EcoRI; B, BamHI; S, SalI; X, XbaI; N, NotI. (B) Greater magnification schematic view of the genomic region of interest, with the mutations noted. The eya1 deletion is estimated to be 1.5 kb from genomic Southern blots; the position was not resolved beyond defining that the genomic SalI sites are deleted in the mutant. (C) Detail of the sequence of the 5' UTR of the eya transcript, indicating the sites of the eya4 and eya3cs transposable element insertions. Genomic sequence is capitalized, whereas sequence of the transposable elements is in lower case. Both mutations show a duplication (indicated in boldface type) at the site of the insertions. For each mutant, the top line indicates the sequence upstream of the site of the insertion, and the bottom line indicates the sequence downstream of the insertion site, with the normal genomic sequence in the middle line. The insertions in the two different mutants occurred in adjacent regions of genomic DNA. The 45-bp genomic sequence shown is 284 bp downstream from the transcriptional start site and 131 bp upstream of the protein initiation site of the type I transcript. The 5' UTR is 463 bp in total. Only the ends of the transposable element insertions were sequenced; restriction enzyme digests confirmed the size and identity of the elements.

View larger version (35K): In this window In a new window Download PPT slide   Figure 4. Sequence of the genomic region 5' of the eya gene. The SalI sites of the Fig 3A genomic map are underlined. The eya2 deletion is boxed in. Sites of the I and roo transposable element insertions in the eya4 and eya3cs alleles, as well as the transcription start and protein initiation sites are indicated. Noted by a star (*) are sites of DNA polymorphisms found in the eya2 sequence at the 3' end of the deletion (bp 390, C instead of A; bp 392, G instead of T). Some polymorphisms in the Oregon-R wild-type strain were found within the deletion region (bp 215, C/T; bp 281, G/A; the A stretch starting at bp 220 ranged from 13 to 15); the sequence presented is that used in subsequent analysis of enhancer activity. Sequence submitted to GenBank (accession no. AF190902).

The insertion of the transposable elements into the 5' UTR likely disrupts proper transcription and translation of the eya gene, thereby leading to the phenotypes of the respective alleles. We predicted that, in vivo, the I element inserted in the eya⁴ allele might be expressed under control of eya regulatory sequences. Indeed, in situ hybridization to eye-antennal imaginal discs of the eya⁴ mutant revealed an I element expression pattern in the eye progenitor cells and ocelli that mimicked the normal expression pattern of the eya gene (Fig 5). Note that both the ocellar and eye progenitor cells in the eya⁴ allele lack normal EYA protein and are thus reduced or missing in the eya⁴ allele (Fig 1 and Fig 2).

View larger version (74K): In this window In a new window Download PPT slide   Figure 5. The eya4 mutant shows I element expression in the pattern of the eya gene in eye progenitor cells. In situ hybridization to eye imaginal discs of normal (A) and eya4 mutant (B) larvae. (A) In normal eye discs, no expression of I element sequences was seen. (B) The eya4 mutant showed an I element expression pattern resembling that of the normal eya eye disc expression pattern: expression occurred broadly within the eye disc, as well as in the ocellar progenitors (arrow). Compare the normal pattern of the EYA protein and the greatly diminished expression in the eya4 mutant (see Fig 2).

The eya¹ and eya² eye-specific alleles are deletion mutations:
The eya¹ and eya² alleles are eye specific, affecting no other functions of the eya gene. Southern analysis revealed the eya¹ and eya² mutations were deletions of approximately 1.5 and 0.3 kb, respectively, within the genomic restriction fragment containing the first exon of the type I transcript (Fig 3). In the eya¹ mutant, a SalI site upstream of the exon was deleted; further analysis of the eya² mutation placed the two deletions in the same region of the genomic DNA upstream of this exon. We then attempted to define the deletion regions in greater detail. This was successful for the eya² allele, but not for the eya¹ allele. As the two alleles were overlapping deletions that act similarly by genetic analysis (see Table 1), we thus focused exclusively on the eya² deletion, which was the smaller of the two deletions. The eya² deletion was further defined by designing primers to the genomic restriction fragment within which the deletion was found, amplifying the region from Oregon-R and the mutant, and sequencing the amplified products. This analysis revealed that the eya² mutation was a 322-bp deletion 5' to the first exon of the type I transcript, located 587 bp upstream of the longest cDNA previously isolated (Fig 3B and Fig 4).

The position of the overlapping deletions in the eya¹ and eya² alleles indicated that the eyeless phenotype observed in these mutants could be due to deletion of regulatory sequences specific for eye progenitor cell expression of the gene, deletion of transcriptional start site sequences, or a combination of both. Primer extension and RNAase protection assays were performed to define the start of transcription of the type I transcript. This showed that the transcriptional start site is downstream of the deletion in the eya² mutant by 581 bp, suggesting that the eya¹ and eya² alleles delete genomic DNA that is regulatory in nature (Fig 3B and Fig 4). This molecular analysis is consistent with the genetic nature of the alleles, as the eya¹ and eya² mutations are eye specific and show pairing-dependent complementation, also called transvection, with other alleles of the gene (LEISERSON et al. 1994 ). Models for transvection predict that the mutations in the eya¹ and eya² alleles may be or may encompass an eye-specific enhancer for the gene; our data suggest that those mutations delete critical sequences required for eye progenitor cell expression of the gene.

The deletion in the eya² eye-specific allele region defines an eye enhancer:
To test whether the region deleted in the eya² allele, which overlaps the domain missing from the eya¹ allele, could define an element important for eya eye expression, we determined whether this DNA element had the capacity to activate expression of a marker gene in eye progenitor cells. To do this, three or six tandem copies of the element were subcloned into a transformation vector upstream of a minimal promoter and the gene for ß-galactosidase. The constructs were then injected into flies to obtain transgenic lines that were analyzed by staining for ß-galactosidase expression in eye-antennal discs. This analysis revealed that the element displayed the capacity to activate expression in eye progenitor cells (Fig 6A). Expression was sufficiently broad within the eye disc that we could not distinguish whether expression was also separately activated in the ocellar progenitor cells. The broad expression pattern in the eye disc is similar to that of the EYA protein, which is broadly expressed both anterior and posterior to the furrow (BONINI et al. 1993 ). Greater analysis revealed that expression was present by the second instar larval stage, which is when EYA protein expression is first apparent in the eye field (data not shown; BONINI et al. 1993 ). To address specificity of expression to eye progenitor cells, we determined whether expression elsewhere in the animal was observed. EYA is expressed in select cells of the brain and ventral nerve cord and the wing and haltere discs (BONINI et al. 1998 ). Analysis of these tissues and other tissues of the animal revealed that the enhancer appeared to be selectively activated in the eye progenitor field (Fig 6B and Fig C). These data indicate that the region deleted in the eya² mutation serves as an enhancer element that activates expression of eya in eye progenitor cells.

View larger version (141K): In this window In a new window Download PPT slide   Figure 6. The eya2 deletion region defines an eye enhancer element. ß-Galactosidase expression in imaginal discs from larvae bearing transgenic constructs with three or six tandem copies of the region deleted in the eya2 mutant allele, upstream of a minimal promoter and the ß-galactosidase gene. (A) Expression is activated in the eye progenitor field, both before and after the furrow (arrowheads), similar to normal EYA expression (see Fig 1). (B and C) Expression of the enhancer is not activated elsewhere that the EYA protein is expressed, such as (B) the wing disc and (C) brain and ventral nerve cord, or leg discs. EYA protein is expressed in the wing disc, brain, and ventral nerve cord (BONINI et al. 1998 ).

We analyzed the sequence of the region to determine whether any striking motifs could be identified. Possible binding sites for various transcription factors included one consensus site for the SO homeodomain (GATAC; HAZBURN et al. 1997 ), one ETS domain site (TTCCNGGAA; MACLEOD et al. 1992 ), and three sites for MAD (GCCGnCGc; KIM et al. 1997 ), a transcription factor involved in transforming growth factor ß signaling. There were no good consensus binding sites for the EY protein (CZERNY et al. 1999 ), which is known to activate expression of the EYA protein in ectopic eye development (BONINI et al. 1997 ; HALDER et al. 1998 ).

Regulation of the eya enhancer region by ectopic Ey expression:
Ectopic expression of EY induces expression of EYA protein (BONINI et al. 1997 ; HALDER et al. 1998 ). We therefore addressed whether EY would activate expression of this defined eya enhancer. To do this, we generated recombinant flies with an eya enhancer transgene in the background of the dpp-GAL4 transgene, which will direct ectopic UAS-transgene activity to the antennal disc. In the background of dpp-GAL4, eya enhancer expression, detected by ß-galactosidase activity, was unaltered, being limited to the eye field of the eye-antennal disc (Fig 7A). When crossed to lines bearing UAS-ey, ectopic eye development occurs in the antennal portion of the eye-antennal imaginal disc with ectopic Eya expression (BONINI et al. 1997 ). We observed that the eya enhancer was indeed ectopically activated in the antennal portion of the disc upon ectopic ey expression, detected by ectopic ß-galactosidase expression (Fig 7B, arrow). We also tested whether ectopic expression of EYA itself would activate expression of the eya enhancer. In this case, we failed to detect ectopic ß-galactosidase expression (Fig 7C), suggesting that eya apparently cannot activate its own expression, at least through this defined DNA element.

View larger version (73K): In this window In a new window Download PPT slide   Figure 7. eya enhancer expression is activated by the ey gene. ß-Galactosidase expression in an eye-antennal imaginal disc from larvae bearing a transgenic construct with six tandem copies of the region deleted in the eya2 mutant allele, upstream of a minimal promoter and the ß-galactosidase gene, in the dpp-GAL4 transgenic background. (A) In dpp-GAL4 transgenic flies and in absence of UAS-transgene activity, the eya enhancer activates gene expression selectively in the eye portion of the eye-antennal imaginal disc. (B) Upon crossing the dpp-GAL4 line with the enhancer insertion to UAS-ey to direct ectopic expression of ey in the antennal disc, ectopic activation of the eya enhancer occurs in the antennal disc (arrow). This ectopic activation is consistent with observed EYA protein expression upon ectopic ey gene expression (BONINI et al. 1997 ; HALDER et al. 1998 ). (C) Upon crossing the dpp-GAL4 line with the enhancer insertion to a UAS-eya transgenic line, ectopic ß-galactosidase expression in the antennal disc is not observed. This suggests that the eya gene itself does not autoregulate through this DNA element.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To define domains within the eya gene critical for directing expression to eye progenitor cells, we have defined the nature of the four viable mutations in the gene. The viable alleles of eya share a common phenotype of loss of the compound eye. The eya⁴ and eya^3cs mutations also affect the expression of the gene in the ocelli, so are not selective for eye expression of the gene. These two mutations, although independently isolated, share the same mutational mechanism in that both are insertional mutations of transposable elements into the 5' UTR of the gene, which is predicted to disrupt proper eya transcription. The eya¹ and eya² mutations are highly specific for loss of eya expression in the eye progenitor cells. In these mutations, the eye-specific phenotype is due to loss of regulatory sequences present in the regions deleted in these mutations. The fact that both of these mutants display only an eye phenotype and are otherwise normal for eya expression (LEISERSON et al. 1998 ) indicates that the deleted sequence is critical for eye cell expression. Indeed, our studies show this region can direct expression to eye progenitor cells.

eya^3cs and eya⁴ mutations are due to transposable element insertions:
These two viable mutations in eya are not eye specific, although they affect expression of the gene in the eye. Our analysis revealed that both are insertions of transposable elements into the same region of the gene in the 5' UTR. Although they are not disruptions within the open reading frame of the protein, they disrupt proper eya gene function. Many spontaneous Drosophila mutations are due to transposable element insertion, which can have a variety of effects on the gene depending upon the site of insertion and nature of the element (SMITH and CORCES 1991 ). In the case of the eya^3cs mutation, the mutational mechanism displays temperature sensitivity, with the mutant phenotype being more severe at low temperature than at high temperature. For the eya⁴ allele, the I element inserted into the gene is subject to eya regulation in eye progenitor cells, showing expression in both the eye region and the ocellar progenitors. This suggests that, whereas the I element has come under control of eya regulatory sequences, the I element has disrupted these normal controls on eya expression, leading to the mutant effects. EYA protein expression is greatly reduced, leading to the mutant phenotype (see Fig 2).

In preliminary studies, we have isolated several suppressors of the eya⁴ phenotype that subsequent studies have shown also suppress the phenotype of the eya^3cs allele, although they do not suppress the eye-specific alleles (N. BONINI, unpublished observations). This common suppression suggests that the mechanism of gene disruption in the two alleles may be similar, even though different transposable elements are involved. This is consistent with the similar sites of integration into the 5' UTR of the two different mutants. Greater analysis of the mutational mechanisms in eya^3cs and eya⁴, coupled with detailed analysis of such suppressor mutations, will reveal greater insight into the process or processes affected. Such mutations and their modifiers offer an approach to defining molecular mechanisms by which the host organism controls transposable element activity (RUTLEDGE et al. 1988 ; SMITH and CORCES 1991 ).

eya¹ and eya² define an eye enhancer element:
Previous analysis of the eya¹ and eya² alleles revealed that they are highly selective for loss of the eye (BONINI et al. 1993 ). Moreover, they complement all functions of the eya gene except the eye phenotype. Detailed analysis of the expression pattern of eya reveals that they are normal for eya gene expression, except in the eye where they show complete loss of eya expression in progenitor cells for the compound eye, with normal ocellar expression (LEISERSON et al. 1998 ; see Fig 2). Previously, we had found that these alleles show interallelic complementation with other eya alleles, including the eya⁴ allele. This was shown to reflect transvection, also known as pairing-dependent complementation, of the gene (LEISERSON et al. 1994 ). Transvection reflects interactions between the chromosomal homologues, such that two alleles with different types of mutations can partially complement, leading to restoration of gene function (reviewed in LEWIS 1954 ; WU 1993 ; HENIKOFF 1997 ). One model for transvection is that one class of mutations defines a regulatory element required for proper gene expression, whereas the complementing class is in the transcription unit. The eya¹ and eya² alleles, being highly specific for loss of gene function in the eye, were candidates to define regions of the eya gene critical for eye progenitor cell expression.

We have found that these two alleles, both spontaneous alleles that were independently isolated, have deletions within the same region of the eya gene. The eya¹ allele defines an approximately 1.5-kb deletion, which we did not analyze further, whereas the eya² allele was a small 322-bp deletion. The region deleted in eya² activated gene expression in eye progenitor cells, indicating that this is a DNA element that is both necessary and sufficient for expression in eye progenitor cells. The expression pattern in the eye reflects that seen of the EYA protein, being broadly expressed within the entire eye field (BONINI et al. 1993 ). Further analysis of the expression pattern revealed that indeed the element appears to be selective for eye progenitor cells, failing to be expressed in other tissues where the eya gene is expressed.

We tested whether this element displayed a response to directed expression of the ey gene, which has been shown to direct ectopic eye formation and EYA expression (HALDER et al. 1995 , HALDER et al. 1998 ; BONINI et al. 1997 ). This regulatory domain of eya indeed responded to ectopic ey expression, consistent with previous studies of EYA protein. This suggests that the enhancer reflects aspects of the normal regulatory pattern of the eya gene in eye formation. We failed to find evidence that this region responds to ectopic expression of the eya gene itself, however, suggesting that it appears not to be a domain involved in a potential autoregulatory loop. Within this region are several intriguing protein binding elements, including potential ETS binding domains and an SO binding site; their possible significance awaits further analysis.

Of genes involved in eye formation, most information regarding regulatory elements required is known for the ey gene. For ey, the eye-specific alleles have also been shown to be disrupted in an eye enhancer element (HAUCK et al. 1999 ). The region was subsequently found to be a target of the toy gene, indicating that ey is a direct target of toy activity (CZERNY et al. 1999 ). As eya remains expressed in ey mutants (HALDER et al. 1998 ), this indicates that eya is downstream of ey in expression initiation. Expression of Eya1 and Eya2 in the mouse is also downstream of Pax-6 activity (XU et al. 1997 ). Greater analysis of the DNA element defined by the eya eye-specific mutations may reveal whether this region is a direct target of ey gene activity or whether additional or other molecular players are involved. In so¹ mutants, which affect eye and ocellar formation, a region is deleted in an intron that defines a domain important for selective expression of the gene to the eye disc (CHEYETTE et al. 1994 ; SERIKAKU and O'TOUSA 1994 ). The otd^uvi mutant, an eye-specific allele of the otd gene, is also a deletion mutation that defines an element required for proper otd expression in eye cells (VANDENDRIES et al. 1996 ). We hope that the domain defined here in eya will provide greater details of the molecular pathway of the eya gene expression in relation to these and other critical genes of eye formation. Such domains can even reveal mechanisms conserved between flies and vertebrates: the Drosophila ey enhancer can direct expression to eye tissue in the mouse, and regulatory domains of the mouse ey homologue Pax-6 can direct gene expression to the eye field in Drosophila (XU et al. 1999 ).

Analysis of the viable alleles of eya has provided tools to apply toward the greater molecular analysis of gene pathways involved in eye formation with the definition of an eye enhancer by the eya¹ and eya² alleles. In addition, these studies have provided tools for greater analysis of molecular mechanisms of transvection, due to the nature of these mutations, as well as gene disruption due to transposable element insertion, with the second class of eya alleles.

	FOOTNOTES

¹ Present address: Center for Sleep, 3600 Spruce St., Hospital of the University of Pennsylvania, Philadelphia, PA 19104.

	ACKNOWLEDGMENTS

We thank members of the Drosophila community for generously sharing fly lines and reagents. This research was supported in part by a March of Dimes Basil O'Connor Award, the John Merck Fund, and the National Eye Institute (EY11259; to N.M.B.).

Manuscript received July 27, 1999; Accepted for publication September 13, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AUSUBEL, F. M., 1994 Current Protocols in Molecular Biology. J. Wiley & Sons, New York.

BONINI, N. M., W. M. LEISERSON, and S. BENZER, 1993  The eyes absent gene: genetic control of cell survival and differentiation in the developing Drosophila eye. Cell 72:379-395[Medline].

BONINI, N. M., Q. T. BUI, G. L. GRAY-BOARD, and J. M. WARRICK, 1997  The Drosophila eyes absent gene directs ectopic eye formation in a pathway conserved between flies and vertebrates. Development 124:4819-4826[Abstract].

BONINI, N. M., W. M. LEISERSON, and S. BENZER, 1998  Expression and multiple roles of the eyes absent gene in Drosophila.. Dev. Biol. 129:42-57.

BOYLE, M., N. BONINI, and S. DINARDO, 1997  Expression and function of clift in the development of somatic gonadal precursors within the Drosophila mesoderm. Development 124:971-982[Abstract].

BUCHETON, A., R. PARO, H. M. SANG, A. PELISSON, and D. J. FINNEGAN, 1984  The molecular basis of I-R hybrid dysgenesis in Drosophila melanogaster: identification, cloning, and properties of the I factor. Cell 38:153-163[Medline].

CHEN, R., M. AMOUI, Z. ZHANG, and G. MARDON, 1997  Dachshund and Eyes absent proteins form a complex and function synergistically to induce ectopic eye development in Drosophila.. Cell 81:893-904.

CHEN, R., G. HALDER, Z. ZHANG, and G. MARDON, 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[Abstract].

CHEYETTE, B. N. R., P. J. GREEN, K. MARTIN, H. GARREN, and V. HARTENSTEIN et al., 1994  The Drosophila sine oculis locus encodes a homeodomain-containing protein required for the development of the entire visual system. Neuron 12:977-996[Medline].

CZERNY, T., G. HALDER, U. KLOTER, A. SOUABNI, and W. J. GEHRING et al., 1999  twin of eyeless, a second Pax-6 gene of Drosophila, acts upstream of eyeless in the control of eye development. Mol. Cell 3:297-307[Medline].

DESPLAN, C., 1997  Eye development: governed by a dictator or a junta? Cell 91:861-864[Medline].

DUNCAN, M. K., L. KOS, N. A. JENKINS, D. J. GILBERT, and N. G. COPELAND et al., 1997  Eyes absent: a gene family found in several metazoan phyla. Mamm. Genome 8:479-485[Medline].

FAWCETT, D. H., C. K. LISTER, E. KELLETT, and D. J. FINNEGAN, 1986  Transposable elements controlling I-R hybrid dysgenesis in D. melanogaster are similar to mammalian LINEs. Cell 47:1007-1015[Medline].

GEYER, P. K., C. SPANA, and V. G. CORCES, 1986  On the molecular mechanism of gypsy-induced mutations at the yellow locus of Drosophila melanogaster.. EMBO J. 5:2657-2662[Medline].

HALDER, G., P. CALLAERTS, and W. J. GEHRING, 1995  Induction of ectopic eyes by targeted expression of the eyeless gene of Drosophila.. Science 267:1788-1792[Abstract/Free Full Text].

HALDER, G., P. CALLAERTS, S. FLISTER, U. WALLDORF, and U. KLOTER et al., 1998  Eyeless initiates the expression of both sine oculis and eyes absent during Drosophila compound eye development. Development 125:2181-2191[Abstract].

HAMMOND, K. L., I. M. HANSON, A. G. BROWN, L. A. LETTICE, and R. E. HILL, 1998  Mammalian and Drosophila dachshund genes are related to the Ski proto-oncogene and are expressed in the eye and limb. Mech. Dev. 74:121-131[Medline].

HAUCK, B., W. J. GEHRING, and U. WALLDORF, 1999  Functional analysis of an eye specific enhancer of the eyeless gene in Drosophila.. Proc. Natl. Acad. Sci. USA 96:564-569[Abstract/Free Full Text].

HAZBURN, T., F. STABURA, and M. MOSSING, 1997  Site-specific recognition by an isolated DNA binding domain of the sine oculis protein. Biochemistry 36:3680-3686[Medline].

HAZELETT, D., M. BOURBOIS, U. WALLDORF, and J. TREISMAN, 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[Abstract].

HENIKOFF, S., 1997  Nuclear organization and gene expression: homologous pairing and long-range interactions. Curr. Opin. Cell Biol. 9:388-395[Medline].

KIM, J., K. JOHNSON, H. CHEN, S. CARROLL, and A. LAUGHON, 1997  Drosophila Mad binds to DNA and directly mediates activation of vestigial by Decapentaplegic. Nature 388:304-308[Medline].

LEISERSON, W. M., N. M. BONINI, and S. BENZER, 1994  Transvection at the eyes absent gene of Drosophila.. Genetics 138:1171-1179[Abstract].

LEISERSON, W. M., S. BENZER, and N. M. BONINI, 1998  Dual functions of the Drosophila eyes absent gene in the eye and embryo. Mech. Dev. 73:193-202[Medline].

LEWIS, E. B., 1954  The theory and application of a new method of detecting chromosomal rearrangements in Drosophila melanogaster.. Am. Nat. 88:225-239.

MACLEOD, K., D. LEPRINCE, and D. STEHELIN, 1992  The ets gene family. Trends Biochem. Sci. 17:251-256[Medline].

MARDON, G., N. M. SOLOMON, and G. M. RUBIN, 1994  Dachshund encodes a nuclear protein required for normal eye and leg development in Drosophila.. Development 120:3473-3486[Abstract].

O'NEILL, E., I. REBAY, R. TJIAN, and G. RUBIN, 1994  The activities of two Ets-related transcription factors required for Drosophila eye development are modulated by the Ras/MAPK pathway. Cell 78:137-147[Medline].

OLIVER, G., A. MAILHOS, R. WEHR, N. G. COPELAND, and N. A. JENKINS et al., 1995  Six3, a murine homologue of the sine oculis gene, demarcates the most anterior border of the developing neural plate and is expressed during eye development. Development 121:4045-4055[Abstract].

PADGETT, R. W., R. D. ST. JOHNSTON, and W. M. GELBART, 1987  A transcript from a Drosophila pattern gene predicts a protein homologous to the transforming growth factor-ß family. Nature 325:81-84[Medline].

PIGNONI, F., B. HU, K. H. ZAVITZ, J. XIAO, and P. A. GARRITY et al., 1997  The eye-specification proteins So and Eya form a complex and regulate multiple steps in Drosophila eye development. Cell 91:881-891[Medline].

QUIRING, R., U. WALLDORF, U. KLOTER, and W. J. GEHRING, 1994  Homology of the eyeless gene of Drosophila to the small eye gene in mice and aniridia in humans. Science 265:785-789[Abstract/Free Full Text].

RUTLEDGE, B. J., M. A. MORTIN, E. SCHWARZ, D. THIERRY-MIEG, and M. MESELSON, 1988  Genetic interactions of modifier genes and modifiable alleles in Drosophila melanogaster.. Genetics 119:391-397[Abstract/Free Full Text].

SCHERER, G., C. TSCHUDI, J. PERERA, H. DELIUS, and V. PIRROTTA, 1982  B104, a new dispersed repeated gene family in Drosophila melanogaster and its analogies with retroviruses. J. Mol. Biol. 157:435-451[Medline].

SERIKAKU, M. A. and J. E. O'TOUSA, 1994  Sine oculis is a homeobox gene required for Drosophila visual system development. Genetics 138:1137-1150[Abstract].

SHEN, W. and G. MARDON, 1997  Ectopic eye development in Drosophila induced by directed dachshund expression. Development 124:45-52[Abstract].

SIMON, J., C. SUTTON, R. LOBELL, R. GLASER, and J. LIS, 1985  Determinants of heat shock-induced chromosome puffing. Cell 40:805-817[Medline].

SMITH, P. A. and V. G. CORCES, 1991  Drosophila transposable elements: mechanisms of mutagenesis and interactions with the host genome. Adv. Genet. 29:229-300[Medline].

STAEHLING-HAMPTON, K., P. D. JACKSON, M. J. CLARK, A. H. BRAND, and F. M. HOFFMAN, 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[Abstract].

THUMMEL, C. S., A. M. BOULET, and H. D. LIPSHITZ, 1988  Vectors for Drosophila P-element-mediated transformation and tissue culture transfection. Gene 74:445-456[Medline].

TREISMAN, J. E. and U. HEBERLEIN, 1998  Eye development in Drosophila: formation of the eye field and control of differentiation. Curr. Top. Dev. Biol. 39:119-158[Medline].

VANDENDRIES, E. R., D. JOHNSON, and R. REINKE, 1996  Orthodenticle is required for photoreceptor cell development in the Drosophila eye. Dev. Biol. 173:243-255[Medline].

WU, C., 1993  Transvection, nuclear structure, and chromatin proteins. J. Cell Biol. 120:587-590[Free Full Text].

XU, P.-X., I. WOO, H. HER, D. R. BEIER, and R. L. MAAS, 1997  Mouse Eya homologues of the Drosophila eyes absent gene require Pax6 for expression in cranial placodes. Development 124:219-231[Abstract].

XU, P. X., X. ZHANG, S. HEANEY, A. YOON, and A. M. MICHELSON et al., 1999  Regulation of Pax6 expression is conserved between mice and flies. Development 126:383-395[Abstract].

ZIMMERMAN, J., Q. BUI, E. STEINGRIMMSON, D. L. NAGLE, and W. FU et al., 1997  Cloning and characterization of two vertebrate homologs of the Drosophila eyes absent gene. Genome Res. 7:128-141[Abstract/Free Full Text].

ZIPURSKY, S. L. and G. M. RUBIN, 1994  Determination of neuronal cell fate: lessons from the R7 neuron of Drosophila.. Annu. Rev. Neurosci. 17:373-397[Medline].

This article has been cited by other articles:

				L. R. Braid and E. M. Verheyen Drosophila Nemo Promotes Eye Specification Directed by the Retinal Determination Gene Network Genetics, September 1, 2008; 180(1): 283 - 299. [Abstract] [Full Text] [PDF]

				A. B. Coulthard, N. Nolan, J. B. Bell, and A. J. Hilliker Transvection at the vestigial Locus of Drosophila melanogaster Genetics, August 1, 2005; 170(4): 1711 - 1721. [Abstract] [Full Text] [PDF]

				K. S. Pappu, E. J. Ostrin, B. W. Middlebrooks, B. T. Sili, R. Chen, M. R. Atkins, R. Gibbs, and G. Mardon Dual regulation and redundant function of two eye-specific enhancers of the Drosophila retinal determination gene dachshund Development, June 15, 2005; 132(12): 2895 - 2905. [Abstract] [Full Text] [PDF]

				T. Pauli, M. Seimiya, J. Blanco, and W. J. Gehring Identification of functional sine oculis motifs in the autoregulatory element of its own gene, in the eyeless enhancer and in the signalling gene hedgehog Development, June 15, 2005; 132(12): 2771 - 2782. [Abstract] [Full Text] [PDF]

				H. Ozaki, K. Nakamura, J.-i. Funahashi, K. Ikeda, G. Yamada, H. Tokano, H.-o. Okamura, K. Kitamura, S. Muto, H. Kotaki, et al. Six1 controls patterning of the mouse otic vesicle Development, February 1, 2004; 131(3): 551 - 562. [Abstract] [Full Text] [PDF]

				I. Ruvinsky and G. Ruvkun Functional tests of enhancer conservation between distantly related species Development, November 1, 2003; 130(21): 5133 - 5142. [Abstract] [Full Text] [PDF]