The Drosophila Genes disconnected and disco-related Are Redundant With Respect to Larval Head Development and Accumulation of mRNAs From Deformed Target Genes

Corresponding author: James W. Mahaffey, Department of Genetics, North Carolina State University, Campus Box 7614, Raleigh, NC 27695-7614., jim_mahaffey{at}ncsu.edu (E-mail)

Communicating editor: A. J. LOPEZ

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

HOM-C/hox genes specify body pattern by encoding regionally expressed transcription factors that activate the appropriate target genes necessary for differentiation of each body region. The current model of target gene activation suggests that interactions with cofactors influence DNA-binding ability and target gene activation by the HOM-C/hox proteins. Currently, little is known about the specifics of this process because few target genes and fewer cofactors have been identified. We undertook a deficiency screen in Drosophila melanogaster in an attempt to identify loci potentially encoding cofactors for the protein encoded by the HOM-C gene Deformed (Dfd). We identified a region of the X chromosome that, when absent, leads to loss of specific larval mouthpart structures producing a phenotype similar to that observed in Dfd mutants. The phenotype is correlated with reduced accumulation of mRNAs from Dfd target genes, though there appears to be no effect on Dfd protein accumulation. We show that these defects are due to the loss of two functionally redundant, neighboring genes encoding zinc finger transcription factors, disconnected and a gene we call disco-related. We discuss the role of these genes during differentiation of the gnathal segments and, in light of other recent findings, propose that regionally expressed zinc finger proteins may play a central role with the HOM-C proteins in establishing body pattern.

HOX genes encode homeodomain-containing transcription factors that specify body pattern during embryogenesis in all metazoans (MCGINNIS et al. 1984 ; SCOTT and WEINER 1984 ; SCOTT et al. 1989 ; MCGINNIS and KRUMLAUF 1992 ; KRUMLAUF 1994 ; MANAK and SCOTT 1994 ). Each individual hox gene is expressed in a specific anterior/posterior domain wherein the encoded protein will specify regional identity through activation of a specific set of target genes (GARCIA-BELLIDO 1977 ; ANDREW and SCOTT 1992 ). Loss of a specific hox gene disrupts pattern formation because appropriate target genes are not activated in the region controlled by that gene. However, the mechanisms underlying target gene selection and activation by hox proteins are unclear because the DNA-binding properties of the proteins encoded by different HOM-C/hox genes are quite similar (HOEY and LEVINE 1988 ; AFFOLTER et al. 1990 ; FLORENCE et al. 1991 ; DESSAIN et al. 1992 ; EKKER et al. 1994 ; WALTER et al. 1994 ; BIGGIN and MCGINNIS 1997 ). In general, hox proteins bind to a consensus sequence with the nucleotides TAAT at the core. Surrounding bases can influence binding strength, but there is significant overlap in the binding abilities of the various hox proteins. However, understanding how hox proteins selectively activate the appropriate target genes is critical to understanding the pattern formation process. In the fruit fly Drosophila melanogaster these genes are located in the Antennapedia and Bithorax complexes and are referred to collectively as the HOM-C genes (LEWIS 1978 ; MCGINNIS and KRUMLAUF 1992 ).

Recent data indicate that interactions with cofactors play an important role in target gene selection. For example, the Extradenticle/Pbx proteins (Exd) participate in cooperative binding with hox proteins, and the heterodimer has a more specific DNA recognition site than the HOM-C/hox protein alone (CHAN et al. 1994 ; CHAN and MANN 1996 ; MANN and CHAN 1996 ). This increases the specificity of DNA binding and thereby could lead to differential activation of specific target genes. Support for this model comes from studies like that of CHAN et al. 1997 , where they show that a small change in the sequence of the heterodimer binding site in a labial response element converts the element into a Dfd response element. Cooperative binding with Exd also increases the binding strength of the protein/DNA interaction as well as influencing hox protein/DNA-binding specificity (CHAN and MANN 1996 ; CHAN et al. 1997 ; PINSONNEAULT et al. 1997 ). Though Exd is a critical component of HOM-C/hox protein function, it seems unlikely that a single cofactor will account for the diversity and specificity needed for target gene selection in all tissues. Yet, to date only the Exd cofactor has been well characterized, though there is evidence that other cofactors exist (LI et al. 1999 ; RYOO et al. 1999 ).

In an attempt to identify potential cofactors functioning with the protein encoded by the HOM-C gene Deformed (Dfd), we carried out a genetic screen using deficiencies available from the Bloomington Drosophila Stock Center. We identified one region of the X chromosome that appears likely to encode such a factor. We show that embryos lacking this region have disruptions of the larval cephalopharyngeal skeleton similar to those seen in Dfd mutants. In addition, Dfd target gene expression is altered in these embryos, though there appears to be no effect on Dfd expression itself. We show that these defects are due to the loss of two functionally redundant, neighboring genes encoding zinc finger transcription factors. One gene, disconnected (disco), has been previously described as a gene necessary for neural connectivity (STELLER et al. 1987 ; HEILIG et al. 1991 ). The second gene, disco-related, encodes a related protein. We discuss the role of these genes in target gene selection during gnathal development and, in a broader sense, as a possible universal mechanism of HOM-C/hox protein function.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks and culture:
Dfd¹⁶, Df(1)4b18, Df(1)19, Dp(1Y)shi⁺1, Df(1)sd72b, and the flies in the deficiency kit were obtained from the Bloomington Indiana Drosophila Stock Center. Df(1)XR14 was a gift from G. Haddad (Yale University). Dp(1;4)81j6e was provided by S. Faulhaber and M. Tanouye (University of California, Berkeley). disco¹ flies were provided by J. Hall (Brandeis University). Flies were reared on standard cornmeal-agar-molasses medium.

Cuticle analysis:
Embryos were collected and prepared for cuticle examination following procedures described in PEDERSON et al. 1996 . Females were allowed to lay eggs for several hours, and a known number of embryos (between 200 and 300) were placed onto a grid in groups of 10 on a new collection plate. The embryos were aged for at least 24 hr and the number and phenotype of the hatched and unhatched larvae were determined. In many experiments the collection plate with the hatched larvae was placed into a bottle containing standard Drosophila medium and the larvae were allowed to develop into adults. The phenotypes of the surviving adults were then determined. The yellow (y) mutation was often used to distinguish between the different classes of larvae. y larvae have lighter colored mouthparts than those carrying the y+ gene.

disco clones:
The disco subclones used in RNAi and whole embryo in situ were derived from a clone of the 14B region (SURDEJ et al. 1990 ; a gift of R. Miassod, Lab. Genet. Biol. Cell., CNRS, Marseille, France). A 4.3-kb EcoRI fragment containing disco was used to generate dsRNA for RNAi, and a 1.8-kb NotI-XhoI fragment from this was used in RNAi and whole embryo in situ experiments. In situ localization of mRNAs followed a modification of the procedure of TAUTZ and PFEIFLE 1989 . In situ hybridization analysis used ribonucleotide probes generated with an RNA transcription kit (Stratagene, La Jolla, CA) and DIG-11-UTP (Boehringer Mannheim, Indianapolis). Anti-DIG-AP (Boehringer Mannheim) was used to detect hybridization.

PCR amplification and cloning of disco-r:
The 5' and 3' coding domains of disco-r were amplified from genomic DNA using Taq DNA polymerase (QIAGEN, Valencia, CA). The 5' primers used were ATCAGCACCACCCACATTTGC and TCTCTGCAGATAATCCTGTCC. The 3' primers used were ATGAGATTGATAGCATGCCGC and ATCAGTGGATAACCAGCGTGG. Primers were obtained from Operon Technologies (Alameda, CA). Amplification was carried out using the following parameters: 95° for 40 sec, 55° for 1 min, 72° for 2 min for 40 cycles. The two reactions each yielded a product of 1.5 kb. PCR products were cloned into pCRII using a TA cloning kit (Invitrogen, San Diego, CA).

Isolation of disco-r cDNAs:
The disco-r cDNA clone CK00522 (Research Genetics, Huntsville, AL) was used to make a fluorescein-labeled probe for the screening of an embryonic cDNA library (ZINN et al. 1988 ). Detection of the hybridized probe was carried out using anti-fluorescein-AP and CDP-Star (New England Nuclear, Boston). Approximately 1.5 x 10⁵ clones were screened and four disco-r cDNAs were recovered. Currently, only the largest (3.1 kb) has been characterized.

Mapping of Df breakpoints:
DNAs from single mutant and wild-type embryos or larvae were prepared using the method of GLOOR et al. 1993 . Mutant larvae were recognized by the altered mouthpart structures at the terminal developmental stage. Primer pairs capable of amplifying small stretches of genomic DNA flanking the deficiency breakpoints were obtained from Operon Technologies. The following amplification parameters were used: 95° for 40 sec, 55° for 1 min, 72° for 2 min for 40 cycles.

RNAi:
Preparation of dsRNA and injection of embryos followed the procedure of BROWN et al. 1999 , except that embryos were collected for only 20 min prior to dechorionation and injection. We have found that injection at this earlier stage yields a higher percentage of phenocopy larvae (J. W. MAHAFFEY, unpublished observation). For disco-r, a 2.5-kb EcoRI fragment from the 3' portion of the 3.1-kb cDNA was used to prepare the dsRNA. For disco we used both genomic clones described above. Preparation and examination of larval cuticles followed the procedure of PEDERSON et al. 1996 .

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Screen for potential cofactors of Dfd:
To identify genomic regions potentially encoding cofactors, we screened chromosome deficiencies available from the Bloomington Drosophila Stock Center using three criteria: (1) disruption of maxillary and/or mandibular cephalopharyngeal components of the first instar larva, (2) normal expression of Dfd, and (3) alteration of Dfd target gene expression. We identified two deficiencies of the X chromosome that met these criteria. One, Df(1)sd72b, is known to remove the gene exd (PEIFER and WIESCHAUS 1990 ; RAUSKOLB et al. 1993 ). As mentioned above, exd is a cofactor required for hox protein function, so it was not unexpected that loss of this chromosome interval would affect head development. In contrast, the phenotype of Df(1)4b18 (reported to remove 14B8-14C1, STANEWSKY et al. 1993 ) could not be attributed to loss of a known gene.

The segmental origin of larval cephalopharyngeal structures has been well defined (JURGENS et al. 1986 ; CAMPOS-ORTEGA and HARTENSTEIN 1997 ). (Here we will use the terminology found in FlyBase for descriptions of the components of the cephalopharyngeal skeleton, see http://firefly.bio.indiana.edu) Formation of many of these structures requires the action of three HOM-C genes, Sex combs reduced (Scr) for labial-derived structures (PATTATUCCI et al. 1991 ; PEDERSON et al. 1996 ), Dfd for structures originating in the maxillary and mandibular segments (MERRILL et al. 1987 ; REGULSKI et al. 1987 ), and labial for structures derived from the intercalary segment (MERRILL et al. 1989 ). Embryos lacking any one of these HOM-C genes have characteristic defects in the cephalopharyngeal skeletal and sensory structures that arise from the affected segments. For example, embryos lacking Dfd (Fig 1B) are missing the maxillary cirri, the ventral organ, the dental sclerite, and the lateral bar of the H-piece from the maxillary segment (MERRILL et al. 1987 ; REGULSKI et al. 1987 ). The mouth hooks, composite structures derived from the maxillary and mandibular segments, are also absent, and the lateral process is truncated anteriorly. Though not entirely removed, a portion of the maxillary sense organ is missing and the remaining structure does not align properly with the antennal sense organ.

View larger version (143K): In this window In a new window Download PPT slide   Figure 1. Df(1)4b18 potentially removes a Dfd cofactor. (A) Mouthpart structures from a wild-type first instar larva. (B) The terminal phenotype of an embryo homozygous for Dfd16. Absence of the cirri and mouth hooks are most notable, but the lateral process is truncated and the lateral bar of the H-piece is missing. (C) The terminal phenotype of an embryo hemizygous for Df(1)4b18. Note that the bases of the mouth hooks are missing and the H-piece and median tooth are faint. The number of cirri is reduced, and the remaining cirri are disorganized and misshapen. (D) Accumulation of mRNA from the Dfd target gene 1.28 in a wild-type, stage 14 embryo. 1.28 mRNA accumulates along the posterior edge of the maxillary lobe. (E) Similarly staged embryo hemizygous for Df(1)4b18. Note the reduction of 1.28 mRNA accumulation along the posterior edge of the maxillary lobe. In this embryo, slight staining is observed in a few cells along the edge of the lobe. Expression of 1.28 is not altered in other areas of the embryo. In D and E, anterior is to the left and dorsal upward. ci, cirri; H, H-piece; lp, lateral process; mh, mouth hooks; mt, median tooth.

Half the male progeny from a cross of Df(1)4b18/FM7c to FM7c/Y lack a portion of the 14B region of the X chromosome. (In our descriptions below we refer to such embryos and larvae as hemizygous, for, though they lack the 14B region, they are hemizygous for the X chromosome.) Many of the structures missing or disrupted in unhatched larvae hemizygous for Df(1)4b18 are the same as those altered in embryos homozygous for mutations in Dfd (Fig 1). The terminal larvae lack the base of the mouth hooks and dental sclerites. The lateral process is truncated near the H-piece, which is also disrupted. The few remaining cirri are misshapen and disorganized. The maxillary portion of the maxillary sense organ does not fuse with the antennal portion. Overall, this phenotype is similar to that of embryos homozygous for strong hypomorphic mutations of Dfd (MERRILL et al. 1987 ).

Loss or disruption of many Dfd-specific structures in Df(1)4b18 embryos coincides with reduced mRNA accumulation from at least some Dfd target genes. Though many genes are likely to be regulated by Dfd, currently only four genes have been identified as strong candidate target genes: Dfd (through autoactivation), Distal-less (Dll), 1.28, and perhaps Serrate (Ser; KUZIORA and MCGINNIS 1988 ; O'HARA et al. 1993 ; MAHAFFEY et al. 1993 and PEDERSON et al. 2000 ; WIELLETTE and MCGINNIS 1999 , respectively). Expression of Dll and Dfd are not altered in embryos hemizygous for Df(1)4b18; however, maxillary expression of 1.28 is reduced. Normally, 1.28 mRNA accumulates in cells along the posterior edge of the maxillary lobe (Fig 1D). In embryos hemizygous for Df(1)4b18, 1.28 mRNA is reduced below the level of detection in most maxillary cells, though we occasionally observe some accumulation in a few cells (Fig 1E). 1.28 accumulation in other tissues is unaffected.

Phenotype and mapping of deficiencies removing 14B:
We used other chromosomal aberrations with breaks near 14B to further map the position of the gene or genes responsible for the head defects. Fig 2 shows the larval head phenotype and the mapping data for these deficiencies. Note, since Df(1)sd72b and Df(1)19 delete exd (PEIFER and WIESCHAUS 1990 ), we crossed these deficiencies to Dp(1Y)shi⁺1. This duplication covers exd but extends only as far as 14A (PEIFER and WIESCHAUS 1990 ; RAUSKOLB et al. 1993 ) and does not rescue the head defects of Df(1)4b18.

View larger version (56K): In this window In a new window Download PPT slide   Figure 2. Deficiency mapping of the 14B region. (A–E) Terminal stage cuticle preparations showing the phenotypes of the larval heads produced by embryos hemizygous for the various deficiencies. (A) Wild type. (B) Df(1)sd72b. Note the reduced mouth hook bases. The median tooth and the dental sclerite are faint. (C) Df(1)4b18. The phenotype of this deficiency is described in Fig 1. Note that this phenotype is more severe than that of Df(1)sd72b in B. (D) Df(1)XR14. The mouth hooks, dental sclerite, and H-piece are absent, the median tooth is faint, and the lateral process is shortened. We occasionally see small, partially sclerotized structures that may be remnants of the mouth hooks in a few embryos (arrows) and one or two cirri-like structures as well. (E) Df(1)19. The phenotype is very similar to that seen in embryos hemizygous for Df(1)XR14, but slightly more severe. All that remains of the affected structures is the truncated lateral process. Labels are as in Fig 1 with the addition of the dental sclerite (ds). The map shows the positions of the deficiency breakpoints in 14B. Proximal and distal are in reference to the centromere. Lines below the chromosomal map indicate the deleted region with the arrow pointing in the direction of the deletion. The small bars at the breakpoints indicate the region of uncertainty of the break position. The positions of disco and the two exons of disco-r are indicated; eas is provided for reference. The 0-kb map position is that of SURDEJ et al. 1990 . The letter in parentheses following the deficiency name corresponds to the cuticle image above. Dp(1;4)81j6e was created by a subsequent deletion of the 14B5 to 15A interval from a fourth chromosome duplication of 13F to 16A (FALK et al. 1984 ). Df(1)19 removes 13F through 14E (STELLER et al. 1987 ; PEIFER and WIESCHAUS 1990 ), while Df(1)sd72b is smaller, extending from 13F1 to 14B1. Df(1)XR14 removes the interval between 14B1-2 and 14D1-2 (STANEWSKY et al. 1993 ).

Larval mouthpart structures appear normal in hemizygous Df(1)4b18 embryos when Dp(1;4)81j6e also is present (data not shown). This indicates that the locus responsible for the head defect lies distal to the break of Dp(1;4)81j6e but within the region removed by Df(1)4b18. A slight larval head defect is observed in Df(1)sd72b/Dp(1Y)shi⁺1 embryos (Fig 2B). The bases of the mouth hooks are slightly reduced as are the H-piece and dental sclerites, but the cirri appear to be complete, and the lateral process is not truncated.

The mouthparts of Df(1)XR14 and Df(1)19/Dp(1Y)shi⁺1 embryos (Fig 2D and Fig E) are more severely disrupted than are those of embryos hemizygous for Df(1)4b18. In these terminal larvae, the mouth hooks and cirri are absent, and the lateral process is reduced further than that in Df(1)4b18 larvae. In addition, the H-piece and the hypostomal sclerites (structures are derived from the labial segment) also are absent. In addition, the antennal sense organ is usually not observed. The phenotype of Df(1)19/Dp(1;Y)shi⁺1 embryos usually appears more severe than that of Df(1)XR14. For instance, though the cirri are usually absent in Df(1)XR14 hemizygous larvae, we occasionally observe what might be remnants of the cirri and small bits of partially sclerotized material that may be remnants of the mouth hooks.

We mapped the positions of the 14B breakpoints of these deficiencies using genomic Southern blotting and PCR analyses (see map in Fig 2). Having a molecular map of the region (SURDEJ et al. 1990 ) as well as information from the Berkeley Drosophila Genome Project (ADAMS et al. 2000 ; RUBIN et al. 2000 ) facilitated this mapping. The distal break of Df(1)4b18 lies within a 10-kb fragment including the disco coding region to about 8 kb distal of disco. Therefore, this deficiency removes disco and extends proximally to the gene no on or off transient A (STANEWSKY et al. 1993 ). The 14B breakpoint of Dp(1;4)81j6e had been mapped previously to a fragment about 20 kb distal of the gene easily shocked (eas) (HEILIG et al. 1991 ; PAVLIDIS et al. 1994 ), and we confirmed this location. We mapped the 14B break of Df(1)sd72b to a 15-kb fragment 20–35 kb distal of disco. Therefore, the disco gene is intact, and only genes distal to disco are removed. We were unable to find the distal endpoint of Df(1)XR14 within the interval examined; however, genetic tests indicate that it does not extend as far as exd. Taking the mapping and phenotypic data together we conclude that the gene or genes whose loss leads to the larval head defects reside between the proximal break of Df(1)sd72b and the distal break of Dp(1;4)81j6e. Thus, the region of interest is within a 75-kb interval centered on the gene disco.

Identification of disco-related:
Two groups have searched for transcribed regions in the 14B interval (SURDEJ et al. 1990 ; HEILIG et al. 1991 ), and the only gene they identified that is expressed during embryogenesis is disco. Though null alleles of disco are semiviable and do not cause defects in larval head development (STELLER et al. 1987 ; HEILIG et al. 1991 ), disco is expressed in the gnathal lobes during embryogenesis (LEE et al. 1991 ). Therefore, it seemed possible that the lack of disco somehow was involved in causing the larval head defects. The discrepancy between the phenotype of disco mutations and the embryonic expression pattern was resolved by finding two previously unidentified open reading frames (ORF) encoding peptides related to Disco within the 14AB region. Using information obtained from the Berkeley Drosophila Genome Project (ADAMS et al. 2000 ) and generating and sequencing fragments to span existing gaps in the contigs, we established that these ORFs are located 90–95 kb distal of disco (see map in Fig 2). We isolated a cDNA that spans the intervening sequence between these two ORFs, demonstrating that they are two exons of a single gene. We refer to this gene as disco-related (disco-r). Both disco and disco-r are transcribed from the same strand of DNA, proximal to distal along the chromosome. The putative Disco-r protein contains two pairs of zinc fingers, each related to the single pair in Disco, but there is little or no similarity outside these domains (Fig 3A). Of particular note is the sequence of the first pair of zinc fingers in Disco-r, which is nearly identical to the sequence of the zinc finger pair in Disco (Fig 3B). The amino acids forming the DNA recognition domains are identical, indicating that these proteins could bind to the same DNA sequence. The second pair of zinc fingers in Disco-r, though related to Disco, is somewhat more divergent.

View larger version (77K): In this window In a new window Download PPT slide   Figure 3. Putative sequence of the protein encoded by disco-r and alignment of the zinc finger motifs. (A) Conceptual translation of the Disco-r protein from the genomic ORFs and partial sequence of the 3.1-kb cDNA. The translation begins at the first Met-initiated open reading frame in the 5' exon and extends to the stop codon in the 3' exon. The splice between the two ORFs was determined from the cDNA sequence. Each boldface region contains one pair of zinc fingers. The predicted Mr is about 140 kD. (B) Alignment of the Disco zinc finger motifs with those from Disco-r. Dots indicate identity with the Disco sequence and dashes indicate gaps. The expected Zn-binding residues are in boldface type. The putative DNA-binding domain is in italics. Note the identity between the DNA-binding domains of Disco and the first pair of zinc fingers of Disco-r. The DNA-binding domain of the second pair from Disco-r is less similar. The numbers refer to the amino acid position in the putative protein shown in A.

disco and disco-r are redundant genes together responsible for the larval head defect:
That disco and disco-r encode related proteins and knowing the phenotypes and positions of the deficiency breakpoints in the 14B region indicate that one or both of these genes could be involved in morphogenesis of the larval head. However, since mutations in the disco gene are viable, we suspected that the genes might have redundant functions. Recently we and others have used double-stranded (ds) RNA interference (RNAi) to generate null phenocopies of specific genes in Drosophila (KENNERDELL and CARTHEW 1998 ; BROWN et al. 1999 ; MISQUITTA and PATERSON 1999 ). We prepared dsRNA from our PCR clone of disco and from the 2.5-kb disco-r cDNA fragment (see MATERIALS AND METHODS). Injecting either of these dsRNAs into wild-type embryos had little or no effect on development. Most injected embryos hatched and were found wiggling in the halocarbon oil. For example, using the disco-r cDNA as the template, of the 56 embryos that developed, 50 embryos were wild type, and 6 had general head defects not resembling the defects observed in the deficiency embryos (data not shown).

To remove the functions of both genes, we injected dsRNA synthesized from the disco-r cDNA into embryos homozygous for the mutation disco¹. Homozygous disco¹ embryos develop into normal larvae with an occasional, slight reduction in the mouth hook base (Fig 4A). However, injecting disco-r dsRNA into disco¹ homozygous embryos caused the majority of these embryos to fail to hatch, and the mouthparts of the unhatched larvae were disrupted in a manner similar to those observed in the deficiencies described above (Fig 4C and Fig D). Of the 44 developed larvae from one experiment, 3 were similar to disco¹, 34 resembled embryos hemizygous for the deficiencies, and 7 had general head defects not resembling the deficiency embryos. Of the 34 larvae appearing similar to the deficiencies, a few resembled larvae hemizygous for Df(1)4b18 (Fig 4C), while most had more severe head defects (Fig 4D). This indicates that the head defects associated with deficiencies of 14B are due to the loss of these two genes that have redundant functions during gnathal lobe development.

View larger version (139K): In this window In a new window Download PPT slide   Figure 4. RNAi phenocopy of the deficiency phenotypes. (A) Homozygous disco1 first instar larval cuticle. Although the mouthparts are usually complete, we occasionally note slight reductions in the base of the mouth hooks (arrowhead). (B) Terminal phenotype of the hemizygous Df(1)XR14 embryo as shown in Fig 2D. (C and D) Homozygous disco1 embryos injected with dsRNA synthesized from the 2.5-kb disco-r cDNA fragment. Note the nearly complete absence of mouth hooks, shortened lateral process, and absence of cirri as seen in hemizygous Df(1)XR14 embryos. Abbreviations are as in Fig 1.

Comparison of disco and disco-r expression during embryogenesis:
The distributions of disco mRNA and protein have been described (LEE et al. 1991 ). Here we compare disco-r mRNA accumulation with that of disco using in situ localization (Fig 5). disco mRNA is first detected during the late syncytial blastoderm stage in a cap of cells at the posterior end of the embryo, excluding the pole cells (late stage 4; stages according to CAMPOS-ORTEGA and HARTENSTEIN 1997 ). As gastrulation begins the posterior disco-expressing cells invaginate, and new accumulation of disco mRNA is detected in two bands of cells anterior and posterior to the dorsal portion of the cephalic furrow (Fig 5A and Fig B). The invaginating cells will form the posterior midgut rudiment and the amneoproctodeal invagination. By stage 10, when the germ band reaches full extension, disco mRNA is no longer observed in the posterior midgut, but accumulates in the proctodeum, along the cephalic furrow and on the dorsal side of the clypeolabrum. Formation of the gnathal lobes (mandibular, maxillary, and labial) is preceded by expression of disco in the three lobe primordia (Fig 5C) but not in the ventral region of these head segments. We note that there is a gap of one or two nonstaining cells between each lobe primordium, so that, at least during this stage, not all cells of the lobe accumulate disco mRNA. disco mRNA also is present in cells along the lateral edge of the acron and in the proctodeum, the optic lobe, and the clypeolabrum.

View larger version (95K): In this window In a new window Download PPT slide   Figure 5. In situ localization of mRNA from disco and disco-r. A–C and G–I are disco; D–F and J–L are disco-r. All embryos except in I and L are oriented anterior to the right, dorsal up. (A) disco mRNA in an early embryo after cellularization of the blastoderm. Note the cap at the posterior pole and the two stripes flanking the dorsal cephalic furrow. (B) As the germ band begins to extend, disco mRNA is detected in the clypeolabrum at the anterior-dorsal tip of the embryo. (D and E) disco-r is not detected in similarly aged embryos. (C) As the germ band reaches full extension (late stage 10) disco mRNA accumulates in the gnathal lobe primordia. Note the few nonstained cells between the stained regions. Staining is also detected in the optic lobe and in the posterior-lateral acron, near the remnant of the cephalic furrow. (F) disco-r is first detected at early stage 11 in the proctodeum and clypeolabrum. (G and J) As the germ band begins to contract and the gnathal lobes form, the distributions of disco (G) and disco-r (J) appear to be identical. Both genes are expressed in the gnathal lobes and in the leg disc primordia. Staining resembling that in the disc primordia extends through the abdominal segments, but as shown in H and K, the abdominal expression does not persist through germ band contraction (stage 12). The cells of the visceral mesoderm and dorsal vessel also express both genes. (I and L) Ventral views of embryos undergoing head involution stained to detect disco and disco-r mRNA, respectively. mRNA continues to accumulate in the gnathal region during head involution (arrow). The location of the stained cells anticipates the mouthpart defects of the deficiencies. mRNA from both genes is still present in the visceral mesoderm. cf, cephalic furrow; cl, clypeolabrum; dv, dorsal vessel; vm, visceral mesoderm; mn, mandibular lobe; mx, maxillary lobe; lb, labial lobe; T# and A#, thoracic and abdominal segments, respectively.

In contrast to disco, disco-r mRNA is not detected until late stage 10 to early stage 11 when it accumulates in the proctodeum and the dorsal clypeolabrum (Fig 5F). disco-r mRNA is not detected in the gnathal region until after segmentation is apparent (stage 11, Fig 5J). From this point onward the distribution of disco-r is nearly indistinguishable from that of disco, though we have not examined expression in the nervous system in detail. By the end of stage 11 both genes are expressed in the mandibular, maxillary, and labial lobes and in primordia of the leg discs. Low levels of mRNA also accumulate in each of the abdominal segments, in a position analogous to the leg disc primordia, but this staining soon disappears (visible in Fig 5G and Fig J, but absent in H and K; see COHEN et al. 1991 ). As the germ band continues to contract, disco and disco-r mRNAs accumulate in the visceral mesoderm and slightly later in the dorsal vessel (Fig 5H and Fig K). During stage 14 and later, disco mRNA is detected in the peripheral nervous system (PNS). We have not detected disco-r in the PNS, though transcripts from both genes are detected later in a few cells of each neuromere along the ventral nerve cord. Finally, we can detect mRNA from both genes in cells of the gnathal lobes during head involution until accessibility of the mRNA is blocked by cuticle synthesis (Fig 5I and Fig L).

disco-r expression in Df(1)4b18 embryos:
The results from our RNAi analyses indicate that both disco and disco-r functions must be removed to disrupt larval head development. Therefore, it is surprising that embryos hemizygous for Df(1)4b18 develop with mouthpart defects, because this deficiency removes only disco (see above). One possible explanation is that the deficiency alters expression of disco-r without removing the gene. To determine whether or not this was the case, we examined the accumulation of disco-r transcripts in embryos hemizygous for Df(1)4b18 (Fig 6). disco-r mRNA is first detected about stage 11 as in wild-type embryos; however, in the mutant embryos disco-r transcripts are not distributed throughout the gnathal lobes. Instead, mRNA accumulates in small clusters of cells in the ventral-posterior region of each lobe (Fig 6B and Fig D). In addition, where there is normally only a low level of transient disco-r mRNA in the abdominal segments, these regions stain more intensely, and staining persists for a longer period in Df(1)4b18 hemizygous embryos (Fig 6B). As germ band contraction continues, the abdominal staining subsides, but staining remains restricted to the small clusters of cells in the gnathal lobes. During later development, when labial and maxillary lobes have migrated to the edges of the stomodeum, the mRNA is no longer detectable in the gnathal lobes (Fig 6F). Accumulation in other areas appears to be unaffected by the deficiency. This altered distribution of disco-r mRNA is caused by the deletion and not by the loss of disco alone, as embryos homozygous for disco¹ appear to have normal disco-r mRNA distribution (data not shown).

View larger version (83K): In this window In a new window Download PPT slide   Figure 6. Distribution of disco-r mRNA in embryos hemizygous for Df(1)4b18. (A) Stage 11 wild-type embryo. (B) Similarly staged Df(1)4b18 hemizygous embryo. Note the reduced staining in the mandibular, maxillary, and labial lobes where staining is present in ventral-posterior clusters of cells instead of throughout the lobes. (C) Wild-type stage 14 embryo. (D) Similarly staged Df(1)4b18 hemizygous embryo. (E) Wild-type embryo during head involution, ventral view. (F) Head involution in Df(1)4b18 hemizygous embryo. At this stage, disco-r mRNA is no longer detectable in the gnathal region of the hemizygous Df(1)4b18 embryo. Abbreviations are as in Fig 5.

Dfd target gene expression in mutants lacking both disco and disco-r:
Since Df(1)XR14 removes both disco and disco-r, we looked at Dfd target gene expression in embryos hemizygous for this deficiency (Fig 7). As mentioned above, there are four potential target genes of Dfd: 1.28, Dll, Dfd, and perhaps Ser (MAHAFFEY et al. 1993 and PEDERSON et al. 2000 ; O'HARA et al. 1993 ; KUZIORA and MCGINNIS 1988 ; WIELLETTE and MCGINNIS 1999 , respectively). We can identify Df(1)XR14 hemizygous embryos after germ band contraction because the labial lobes fail to migrate ventrally, the mandibular lobes do not fuse with the maxillary lobes, and the maxillary lobes do not fully rotate. Note, these are also characteristics of mutant embryos lacking the HOM-C genes Dfd and Scr (MERRILL et al. 1987 , MERRILL et al. 1989 ).

View larger version (139K): In this window In a new window Download PPT slide   Figure 7. Target gene expression in embryos lacking disco and disco-r. (A, C, and E) Wild-type embryos. (B, D, and F) Hemizygous Df(1)XR14 embryos. (A and B) 1.28 mRNA accumulation. We did not detect any 1.28 mRNA in the maxillary lobe of this mutant embryo, though in a few others we noted a low level of staining along the posterior edge of the lobe, particularly in the midlateral cells. In A, the arrow points to the posterior maxillary cells that accumulate 1.28 mRNA in the wild-type embryo. (C and D). Serrate mRNA accumulation. In wild-type embryos (C) Ser mRNA accumulates throughout the mandibular lobe and in the anterior and posterior of the maxillary and labial lobes. In hemizygous Df(1)XR14 embryos (D), Ser mRNA was not detected in the maxillary and labial lobes, but was present in a few cells of the mandibular lobe. The slight darkening of the gnathal lobes in D is due to staining below the lobes that is not altered in the mutants. (E and F) Distal-less mRNA accumulation in the ventral spot of the maxillary lobes. Note the reduced staining of the ventral spot in the Df(1)XR14 embryo (F). Abbreviations are as in Fig 5 with the addition of the Dll ventral spot (vs).

1.28 mRNA is not detected in the maxillary lobes of most Df(1)XR14 hemizygous embryos (Fig 7A and Fig B). Occasionally we note slight staining in a few cells, mainly along the posterior edge of the lobes, but this often approaches background levels (data not shown).

In wild-type embryos, Ser mRNA accumulates throughout the mandibular lobes and along the lateral anterior and posterior edges of the maxillary and labial lobes (Fig 7C). The proteins encoded by the HOM-C genes Dfd and Scr are necessary for this expression (WIELLETTE and MCGINNIS 1999 ). In the absence of Dfd, Ser mRNA does not accumulate in the mandibular lobes or in the anterior portion of the maxillary lobes; absence of Scr causes loss of Ser mRNA in the posterior maxillary-anterior labial and posterior labial domains. Staining in embryos lacking both disco and disco-r appears similar to a combination of the HOM-C mutant patterns; Ser mRNA is not detected in the maxillary and labial regions, though a few cells stain in the mandibular lobe (Fig 7D).

We also examined Dll mRNA distribution in embryos hemizygous for Df(1)XR14. In wild-type embryos Dll mRNA accumulates in a large number of cells in the anterior-lateral portion of the maxillary lobe and in a smaller group of cells more ventrally located (Fig 7E). Dll mRNA accumulation in the ventral cells requires Dfd while expression in the anterior-lateral region does not (O'HARA et al. 1993 ). In embryos lacking disco and disco-r the ventral Dll spot forms but is smaller than that in wild-type embryos (Fig 7E and Fig F). We conclude from these observations that the loss of the disco and disco-r can have varying effects on accumulation of mRNAs from Dfd target genes. Ser and 1.28 are absent or occasionally detected at low levels, and Dll, although reduced, is present.

Dfd is also a target through autoactivation (KUZIORA and MCGINNIS 1988 ), and an explanation for the phenotype we observe in larvae lacking disco and disco-r might be that the encoded proteins are required for normal Dfd expression. Failure of the autoactivation process results in loss of Dfd protein from the maxillary cells after stage 10 (PINSONNEAULT et al. 1997 ). (It is unlikely that disco and disco-r are involved in activating Dfd, as Dfd mRNA and protein accumulate prior to disco-r expression, and our results indicate that disco-r is sufficient for normal cephalopharyngeal development.) We collected Df(1)XR14 embryos and stained these embryos with antibodies recognizing the Dfd protein (MAHAFFEY et al. 1989 ). Dfd protein accumulates in the maxillary cells throughout development of the mutant embryos (data not shown) indicating that Disco and Disco-r are not likely required for autoactivation of Dfd.

disco and disco-r are not targets of Dfd:
Another possible cause of the larval head defect might be that disco and disco-r are themselves Dfd target genes that, once activated by Dfd, are necessary for further development of the gnathal lobes. If this is the case we would not expect to see disco and disco-r mRNAs accumulate in embryos lacking Dfd. We, therefore, collected embryos from a cross of Dfd¹⁶ heterozygous parents and detected disco and disco-r mRNA accumulation by in situ hybridization. Dfd¹⁶ has a nonsense mutation before the homeobox (ZENG et al. 1994 ), and the phenotype of homozygous Dfd¹⁶ embryos is identical to that of embryos carrying deficiencies of Dfd, indicating that this is a null allele (MERRILL et al. 1987 ). We observe no difference in disco or disco-r mRNA accumulation between embryos homozygous for Dfd¹⁶ and wild-type embryos, indicating that Dfd is not required to activate disco or disco-r (data not shown).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The two genes, disco and disco-r, appear to have redundant functions during development of the Drosophila larval head. Presence of either gene product is sufficient for normal development of the mandibular, maxillary, and labial lobes, but absence of both gene products disrupts development in these regions. The phenotype of terminal larvae lacking these two genes is strikingly similar to that of larvae lacking the HOM-C genes Dfd and Scr. disco was identified earlier as encoding a protein required for the formation of certain neural connections during embryonic and adult development of Drosophila (STELLER et al. 1987 ). This does not appear to be a redundant function, because the phenotype was no more severe in Df(1)19 hemizygous embryos that lack both disco and disco-r (STELLER et al. 1987 ). At present, we do not know whether disco-r also has an independent role.

disco and disco-r encode proteins containing paired zinc finger domains, Disco with one pair while Disco-r has two pairs. The near identity of the Disco zinc finger pair and the first pair in Disco-r indicates that these proteins may bind to the same DNA sequence. This, along with overlapping distribution of mRNAs, likely explains the redundancy. However, the putative Disco-r protein contains a second pair of zinc fingers, and it is possible that these also influence DNA binding. If so, there may be some differences in the recognition site of these two proteins and, possibly, differences in their roles during development. It is worth noting that a mammalian gene, basonuclin, has been identified that encodes a protein with zinc finger domains similar to those in Disco (TSENG and GREEN 1992 ); Basonuclin contains three pairs of zinc fingers, so in this respect it is more similar to the Disco-r protein. We also identified an ORF in the Caenorhabditis elegans genome that encodes a peptide containing a single pair of zinc fingers quite similar to those in Disco; however, at this time little is known of the gene. Finding similar proteins in animals widely divergent from Drosophila indicates that at least some functions of Disco and/or Disco-r may be conserved during evolution.

basonuclin mRNA and protein accumulate in some cells that have the potential to divide, leading to the prediction that the protein is involved in regulation of cell proliferation (TSENG and GREEN 1994 ), though the protein is found in nonproliferating cells as well (YANG et al. 1997 ). Perhaps reduced cell proliferation in the gnathal lobes could cause the phenotype we observe, but we find no evidence that cell proliferation is altered in embryos lacking disco and disco-r. The maxillary lobes of embryos hemizygous for Df(1)XR14 are nearly equal in size to those of wild-type embryos, and 4',6-diamidino-2-phenylindole staining reveals that there are comparable numbers of nuclei in mutant and wild-type lobes (data not shown). Further, Dfd autoactivation and 1.28 expression occur at about the same time and in the same cells along the posterior-lateral edge of the maxillary lobes. In the absence of disco and disco-r, Dfd autoactivation occurs, but there is little or no accumulation of 1.28 mRNA. We conclude that the cells along the posterior edge of the maxillary lobes are viable and possess the correct homeotic identity, but there is a defect in mRNA accumulation from at least some Dfd target genes (1.28 and Ser).

Other redundant gene pairs have been identified in Drosophila. Of particular note are the head gap gene buttonhead (btd) and the gene D-Sp1 (WIMMER et al. 1993 , WIMMER et al. 1996 ; SCHOCK et al. 1999 ). Like disco and disco-r, these two genes encode C₂ H₂ zinc finger proteins, but these proteins each have triplet zinc finger domains instead of paired domains as in Disco and Disco-r. btd and D-Sp1 also are closely linked on the X chromosome, in division 9A. However, unlike disco and disco-r, the redundancy between btd and D-Sp1 appears in their roles during neural development, not during segmentation (SCHOCK et al. 1999 ).

With this in mind, one may wonder whether the disco and disco-r are also head gap genes. The early distribution of disco mRNA may be suggestive, but we think that it is unlikely for the following reasons. Loss of disco and disco-r does not appear to cause a gap phenotype. We observe no loss of segments; the gnathal lobes form as expected. In addition, we do not observe a change in the distribution of the engrailed protein in the gnathal cells until head involution is underway, and then the changes appear to be due to improper migration of the gnathal lobes in the mutant embryos (data not shown). Further, disco-r function is sufficient for normal gnathal development, yet accumulation of disco-r mRNA in gnathal cells occurs well after segmentation. Finally, the process of segmentation in the gnathal region follows that of the trunk, relying on the gap, pair rule, and segment polarity functions (LEHMANN and NUSSLEIN-VOLHARD 1987 ; MOHLER et al. 1989 ; KRAUT and LEVINE 1991 ; GALLITANO-MENDEL and FINKELSTEIN 1998 ; SCHOCK et al. 2000 ), though we note that buttonhead is required for development of the mandibular segment (WIMMER et al. 1993 ; SCHOCK et al. 2000 ). Taking this into consideration, it seems unlikely that disco and disco-r are head gap genes.

However, we suggest that disco/disco-r and btd may have similar roles. Recently, SCHOCK et al. 2000 presented evidence that the Btd protein is required along with the homeodomain-containing protein Empty spiracles (Ems) to specify intercalary identity. Ectopic Ems is capable of transforming regions only where Btd is present, indicating that Btd is necessary for Ems activity. SCHOCK et al. 2000 go on to demonstrate that Btd and Ems proteins can interact, and this can occur at the Btd zinc finger domain as well as elsewhere in the protein. From their studies and those of others, SCHOCK et al. 2000 conclude that Btd and Ems together specify intercalary identity, and that Btd represses phenotypic suppression of Ems. They go on to state that this supports the contention that Ems is an escaped HOM-C gene as proposed by MACIAS and MORATA 1996 .

Though repression of phenotypic suppression may occur, we propose that there is a more fundamental role for the proteins encoded by btd and disco/disco-r. We propose that these zinc finger-containing proteins are required along with the HOM-C proteins to activate the appropriate target genes necessary to establish segment identity. In the case of disco and disco-r, this is with Dfd and Scr during differentiation of the gnathal lobes. disco and disco-r have a lot in common with the HOM-C genes. They encode spatially restricted transcription factors. Absence of these genes causes a similar phenotype to loss of Dfd and Scr, suggesting a loss of segment identity. We suggest that, as with the HOM-C genes, disco and disco-r are needed to establish the appropriate transcriptional environment for gnathal segment identity. In an analogous manner, Btd and Ems are required for intercalary identity. Further, since Btd interacts directly with Ems, it seems possible that similar interactions may occur between other HOM-C proteins and zinc finger cofactors. It is tempting to speculate that this occurs with Disco/Disco-r and Dfd and Scr, but this may be a bit premature. Additional studies are necessary to determine if this model is correct, but the similarity of larvae lacking these genes to those lacking Dfd and Scr implies that the disco and disco-r function is crucial for normal pattern formation in the gnathal lobes.

With regard to general HOM-C/hox gene specification of body pattern, perhaps similar systems are in operation in other regions of the embryo that have gone undetected due to redundancy. There are numerous zinc finger encoding genes within the Drosophila genome (RUBIN et al. 2000 ). Some of these are closely linked as are disco and disco-r and btd and D-Sp1 (J. W. MAHAFFEY, unpublished observation). At present, evidence of such a mechanism involving zinc finger transcription factors has been detected in two regions of the embryonic head. Perhaps further studies will determine whether similar mechanisms are underway in other regions of the embryo. Finally, since genes encoding similar proteins to Disco and Disco-r are found in other animals, perhaps this is a conserved mechanism involved in establishing body pattern in all animals.

	ACKNOWLEDGMENTS

The authors acknowledge the generosity of the many individuals who kindly supplied reagents for this study: The Bloomington Indiana Drosophila Stock Center for many stocks including the deficiency kit stocks, G. Haddad (Yale University) for Df(1)XR14, S. Faulhaber and M. Tanouye (UC Berkeley) for Dp(1;4)81j6e, J. Hall (Brandeis University) for disco¹ flies, and R. Miassod (Lab. Genet. Biol. Cell. CNRS Marseille, France) for the lambda walk clones of the 14B region. We also thank Dr. Dipak Mahato and Mary C. Clark for help with this deficiency screen. We also thank Dr. G. Gibson for critical reading and helpful comments about the manuscript. This work was supported by National Science Foundation Grant IBN-9514246 to J.W.M.

Manuscript received June 23, 2000; Accepted for publication September 12, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, M. D., S. E. CELNIKER, R. A. HOLT, C. A. EVANS, and J. D. GOCAYNE et al., 2000  The genome sequence of Drosophila melanogaster.. Science 287:2185-2195.

AFFOLTER, M., A. PERCIVAL-SMITH, M. MULLER, W. LEUPIN, and W. J. GEHRING, 1990  DNA binding properties of the purified Antennapedia homeodomain. Proc. Natl. Acad. Sci. USA 87:4093-4097.

ANDREW, D. J. and M. P. SCOTT, 1992  Downstream of the homeotic genes. New Biol. 4:5-15.

BIGGIN, M. D. and W. MCGINNIS, 1997  Regulation of segmentation and segmental identity by Drosophila homeoproteins: the role of DNA binding in functional activity and specificity. Development 124:4425-4433.

BROWN, S. J., J. P. MAHAFFEY, M. D. LORENZEN, R. E. DENELL, and J. W. MAHAFFEY, 1999  Using RNAi to investigate orthologous homeotic gene function during development of distantly related insects. Evol. Dev. 1:11-15.

CAMPOS-ORTEGA, J. A., and V. HARTENSTEIN, 1997 The Embryonic Development of Drosophila melanogaster, Ed. 2. Springer Verlag, Berlin.

CHAN, S.-K. and R. S. MANN, 1996  A structural model for a homeotic protein-extradenticle-DNA complex accounts for the choice of HOX protein in the heterodimer. Proc. Natl. Acad. Sci. USA 93:5223-5228.

CHAN, S. K., L. JAFFE, M. CAPOVILLA, J. BOTAS, and R. S. MANN, 1994  The DNA binding specificity of Ultrabithorax is modulated by cooperative interactions with extradenticle, another homeoprotein. Cell 78:603-615.

CHAN, S.-K., H.-D. RYOO, A. GOULD, R. KRUMLAUF, and R. S. MANN, 1997  Switching the in vivo specificity of a minimal Hox-responsive element. Development 124:2007-2014.

COHEN, B., E. A. WIMMER, and S. M. COHEN, 1991  Early development of leg and wing primordia in the Drosophila embryo. Mech. Dev. 33:229-240.

DESSAIN, S., C. T. GROSS, M. A. KUZIORA, and W. MCGINNIS, 1992  Antp-type homeodomains have distinct DNA binding specificities that correlate with their different regulatory functions in embryos. EMBO J. 11:991-1002.

EKKER, S. C., D. G. JACKSON, D. P. VON KESSLER, B. I. SUN, and K. E. YOUNG et al., 1994  The degree of variation in DNA sequence recognition among four Drosophila homeotic proteins. EMBO J. 13:3551-3560.

FALK, D. R., L. ROSELLI, S. CURTISS, D. HALLADAY, and C. KLUFAS, 1984  The characterization of chromosome breaks in Drosophila melanogaster. I. Mass isolation of deficiencies which have an end point in the 14A-15A region. Mutat. Res. 126:25-34.

FLORENCE, B., R. HANDROW, and A. LAUGHON, 1991  DNA-binding specificity of the fushi tarazu homeodomain. Mol. Cell. Biol. 11:3613-3623.

GALLITANO-MENDEL, A. and R. FINKELSTEIN, 1998  Ectopic orthodenticle expression alters Segment polarity gene expression but not head segment identity in the Drosophila embryo. Dev. Biol. 199:125-137.

GARCIA-BELLIDO, A., 1977  Homoeotic and atavic mutations in insects. Am. Zool. 17:613-629.

GLOOR, G. B., C. R. PRESTON, D. M. JOHNSON-SCHLITZ, N. A. NASSIF, and R. W. PHILLIS et al., 1993  Type I repressors of P element mobility. Genetics 135:81-95.

HEILIG, J. S., M. FREEMAN, T. LAVERTY, K. J. LEE, and A. R. CAMPOS et al., 1991  Isolation and characterization of the disconnected gene of Drosophila melanogaster.. EMBO J. 10:809-815.

HOEY, T. and M. LEVINE, 1988  Divergent homeo box proteins recognize similar DNA sequences in Drosophila. Nature 332:858-861.

JÜRGENS, G., R. LEHMANN, M. SCHARDIN, and C. NUSSLEIN-VOLHARD, 1986  Segmental organization of the head in the embryo of Drosophila melanogaster. A blastoderm fate map of the cuticle structures of the larval head. Roux's Arch. Dev. Biol. 195:359-377.

KENNERDELL, J. R. and R. W. CARTHEW, 1998  Use of dsRNA-mediated genetic interference to demonstrate that frizzled and frizzled 2 act in the wingless pathway. Cell 95:1017-1026.

KRAUT, R. and M. LEVINE, 1991  Spatial regulation of the gap gene giant during Drosophila development. Development 111:601-609.

KRUMLAUF, R., 1994  Hox genes in vertebrate development. Cell 78:191-201.

KUZIORA, M. A. and W. MCGINNIS, 1988  Autoregulation of a Drosophila homeotic selector gene. Cell 55:477-485.

LEE, K. J., M. FREEMAN, and H. STELLER, 1991  Expression of the disconnected gene during development of Drosophila melanogaster.. EMBO J. 10:817-826.

LEHMANN, R. and C. NUSSLEIN-VOLHARD, 1987  hunchback, a gene required for segmentation of an anterior and posterior region of the Drosophila embryo. Dev. Biol. 119:402-417.

LEWIS, E. B., 1978  A gene complex controlling segmentation in Drosophila. Nature 276:565-570.

LI, X., A. VERAKSA, and W. MCGINNIS, 1999  A sequence motif distinct from Hox binding sites controls the specificity of a Hox response element. Development 126:5581-5589.

MACIAS, A. and G. MORATA, 1996  Functional hierarchy and phenotypic suppression among Drosophila homeotic genes: the labial and empty spiracles genes. EMBO J. 15:334-343.

MAHAFFEY, J. W., R. J. DIEDERICH, and T. C. KAUFMAN, 1989  Novel patterns of homeotic protein accumulation in the head of the Drosophila embryo. Development 105:167-174.

MAHAFFEY, J. W., D. F. JONES, J. A. HICKEL, and C. M. GRISWOLD, 1993  Identification and characterization of a gene activated by the deformed homeoprotein. Development 118:203-214.

MANAK, J. R. and M. P. SCOTT, 1994  A class act: conservation of homeodomain protein functions. Development 120S:61-77.

MANN, R. S. and S. K. CHAN, 1996  Extra specificity from extradenticle: the partnership between HOX and PBX/EXD homeodomain proteins. Trends Genet. 12:258-262.

MCGINNIS, W. and R. KRUMLAUF, 1992  Homeobox genes and axial patterning. Cell 68:283-302.

MCGINNIS, W., C. P. HART, W. J. GEHRING, and F. H. RUDDLE, 1984  Molecular cloning and chromosome mapping of a mouse DNA sequence homologous to homeotic genes of Drosophila. Cell 38:675-680.

MERRILL, V. K., F. R. TURNER, and T. C. KAUFMAN, 1987  A genetic and developmental analysis of mutations in the Deformed locus in Drosophila melanogaster.. Dev. Biol. 122:379-395.

MERRILL, V. K., R. J. DIEDERICH, F. R. TURNER, and T. C. KAUFMAN, 1989  A genetic and developmental analysis of mutations in labial, a gene necessary for proper head formation in Drosophila melanogaster.. Dev. Biol. 135:376-391.

MISQUITTA, L. and B. M. PATERSON, 1999  Targeted disruption of gene function in Drosophila by RNA interference (RNA-i): a role for nautilus in embryonic somatic muscle formation. Proc. Natl. Acad. Sci. USA 96:1451-1456.

MOHLER, J., E. D. ELDON, and V. PIRROTTA, 1989  A novel spatial transcription pattern associated with the segmentation gene, giant, of Drosophila. EMBO J. 8:1539-1548.

O'HARA, E., B. COHEN, S. M. COHEN, and W. MCGINNIS, 1993  Distal-less is a downstream gene of Deformed required for ventral maxillary identity. Development 117:847-856.

PATTATUCCI, A. M., D. C. OTTESON, and T. C. KAUFMAN, 1991  A functional and structural analysis of the Sex combs reduced locus of Drosophila melanogaster.. Genetics 129:423-441.

PAVLIDIS, P., M. RAMASWAMI, and M. A. TANOUYE, 1994  The Drosophila easily shocked gene: a mutation in a phospholipid synthetic pathway causes seizure, neuronal failure, and paralysis. Cell 79:23-33.

PEDERSON, J. A., J. W. LAFOLLETTE, JR., C. GROSS, A. VERAKSA, and W. MCGINNIS et al., 2000  Regulation by homeoproteins: a comparison of Deformed-responsive elements. Genetics 156:677-686.

PEDERSON, J. D., D. P. KIEHART, and J. W. MAHAFFEY, 1996  The role of HOM-C genes in segmental transformations: reexamination of the Drosophila Sex combs reduced embryonic phenotype. Dev. Biol. 180:131-142.

PEIFER, M. and E. WIESCHAUS, 1990  Mutations in the Drosophila gene extradenticle affect the way specific homeo domain proteins regulate segmental identity. Genes Dev. 4:1209-1223.

PINSONNEAULT, J., B. FLORENCE, H. VAESSIN, and W. MCGINNIS, 1997  A model for extradenticle function as a switch that changes HOX proteins from repressors to activators. EMBO J. 16:2032-2042.

RAUSKOLB, C., M. PEIFER, and E. WIESCHAUS, 1993  extradenticle, a regulator of homeotic gene activity, is a homolog of the homeobox-containing human proto-oncogene pbx1.. Cell 74:1101-1112.

REGULSKI, M., N. MCGINNIS, R. CHADWICK, and W. MCGINNIS, 1987  Developmental and molecular analysis of Deformed; a homeotic gene controlling Drosophila head development. EMBO J. 6:767-777.

RUBIN, G. M., M. D. YANDELL, J. R. WORTMAN, G. L. GABOR MIKLOS, and C. R. NELSON et al., 2000  Comparative genomics of the eukaryotes. Science 287:2204-2215.

RYOO, H. D., T. MARTY, F. CASARES, M. AFFOLTER, and R. S. MANN, 1999  Regulation of Hox target genes by a DNA bound Homothorax/Hox/Extradenticle complex. Development 126:5137-5148.

SCHÖCK, F., B. A. PURNELL, E. A. WIMMER, and H. JÄCKLE, 1999  Common and diverged functions of the Drosophila gene pair D-Sp1 and buttonhead.. Mech. Dev. 89:125-132.

SCHÖCK, F., J. REISCHL, E. WIMMER, H. TAUBERT, and B. A. PURNELL et al., 2000  Phenotypic suppression of empty spiracles is prevented by buttonhead. Nature 405:351-354.

SCOTT, M. P. and A. J. WEINER, 1984  Structural relationships among genes that control development: sequence homology between the Antennapedia, Ultrabithorax, and fushi tarazu loci of Drosophila. Proc. Natl. Acad. Sci. USA 81:4115-4119.

SCOTT, M. P., J. W. TAMKUN, and G. W. HARTZELL, III, 1989  The structure and function of the homeodomain. Biochim. Biophys. Acta. 989:25-48.

STANEWSKY, R., K. G. RENDAHL, M. DILL, and H. SAUMWEBER, 1993  Genetic and molecular analysis of the X chromosomal region 14B17-14C4 in Drosophila melanogaster: loss of function in NONA, a nuclear protein common to many cell types, results in specific physiological and behavioral defects. Genetics 135:419-442.

STELLER, H., K. F. FISCHBACH, and G. M. RUBIN, 1987  Disconnected: a locus required for neuronal pathway formation in the visual system of Drosophila. Cell 50:1139-1153.

SURDEJ, P., C. GOT, and R. MIASSOD, 1990  Developmental expression pattern of a 800-kb DNA continuum cloned from the Drosophila X chromosome 14B-15B region. Biol. Cell 68:105-118.

TAUTZ, D. and C. PFEIFLE, 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.

TSENG, H. and H. GREEN, 1992  Basonuclin: a keratinocyte protein with multiple paired zinc fingers. Proc. Natl. Acad. Sci. USA 89:10311-10315.

TSENG, H. and H. G