Son of Notch, a Winged-helix Gene Involved in Boundary Formation in the Drosophila Wing

Summary

A P element enhancer trap screen was conducted to identify genes involved in dorsal-ventral boundary formation in Drosophila. The son of Notch (son) gene was identified by the son²²⁰⁵ enhancer trap insertion, which is a partial loss-of-function mutation. Based on son²²⁰⁵ mutant phenotypes and genetic interactions with Notch and wingless mutations, we conclude that son participates in wing development, and functions in the Notch signaling pathway at the dorsal-ventral boundary in the wing. Notch signaling pathway components activate son enhancer trap expression in wing cells. son enhancer trap expression is regulated positively by wingless, and negatively by cut in boundary cells. Ectopic Son protein induces wingless and cut expression in wing discs. We hypothesize that there is positive feedback regulation of son by wingless, and negative regulation by cut at the dorsal-ventral boundary during wing development.

Introduction

Development of higher organisms involves formation of compartment boundaries that separate cells into groups with distinct developmental fates (1, 2). Compartment boundaries were discovered because of the inability of genetically marked clones of cells to cross a boundary (3). In Drosophila, several genes that function in dorsal-ventral (DV) and anterior-posterior (AP) boundary formation have been discovered. Some genes function in boundary formation in multiple tissues, indicating that there are conserved genetic pathways that regulate compartmentalization. For instance, the engrailed gene is critical for AP boundary formation in many tissues in the developing Drosophila embryo, and the wingless and Notch signaling pathways are required for DV boundary formation in both the developing eye and wing (1, 2). Formation of the DV compartment boundary in the wing imaginal disc is initiated by wingless in 2nd instar larvae, which restricts expression of the Apterous LIM-homeoprotein to the dorsal half of the wing pouch (4). The boundary forms at the junction of Apterous-expressing and non-expressing cells (5 – 7). Apterous induces expression of Fringe (Fng) protein, and the juxtaposition of Fringe expressing and non-expressing cells activates the Serrate (Ser) Notch receptor ligand, which activates the Notch receptor at the boundary (8). There is feedback between wingless and Notch signaling, because the Notch induction initiated by wingless induces further wingless expression (9, 10).

Notch function at the boundary requires both the Delta (Dl) Notch ligand and the Suppressor of Hairless (Su(H)) DNA-binding protein (11). Notch activates the vestigial gene by binding of the Notch intracellular domain (N^ICD) to Su(H) bound to a vestigial boundary-specific transcriptional enhancer (12). Expression of vestigial helps activate the scalloped gene at the boundary (4), and together these genes activate the cut gene. The cut wing margin enhancer directs cut expression in a narrow band of cells at the boundary beginning in late 3rd instar development (13, 14).

The Scalloped (Sd) protein directly binds the cut wing margin enhancer, and Vestigial (Vg) interacts with Sd to transcriptionally activate cut (15 – 18). In this pathway, wingless has multiple functions, including helping to induce or maintain expression of vestigial and cut (4, 19 – 23). Induction of cut by wingless is presumably indirect, but creates regulatory loop, because cut helps maintain wingless expression at the developing margin (21).

Although many genes that regulate compartment boundary formation have been described, the genetic pathways are not fully defined. It is possible that critical components have not been discovered genetically because they are essential in early development, or because their role in boundary formation is obscured by other mutant phenotypes. We therefore undertook an enhancer trap transposon screen for genes involved in boundary formation. If an enhancer trap transposon inserts near a gene involved in boundary formation, it may be expressed in a pattern consistent with this role. For example, it might show a high level of expression in boundary cells, or in cells flanking a compartment boundary. Here we report identification of the son of Notch (son) gene by an enhancer trap screen, and present evidence that it plays a role in DV boundary formation in the developing Drosophila wing under control of the Notch signaling pathway.

Experimental Procedures

Enhancer Trap Screen

An attached-X chromosome carrying eight P{lacW} constructs (C(1)RM, y¹ P{lacW}5-45fD w* P{lacW}4-5fP P{lacW}3-52d P{lacW}3-76a/0/C(1;Y)13, v¹ f¹, Bloomington stock #16, donated by Mel Green) was used to generate random insertions. These females were crossed to males producing P transposase (y w; Ki P{ry^+t7.2 = Delta2-3}99B, Bloomington stock #4348, donated by Hugo Bellen) males. Female progeny carrying attached-X and transposase were crossed with w¹¹¹⁸ males. Red eye male progeny of this cross were crossed with proper balancers. The balanced lines were stained for β-galactosidase activity using 3rd instar imaginal discs.

Staining of Imaginal Discs for β-galactosidase Activity

Crawling 3rd instar larvae were dissected in phosphate-buffered saline (PBS). Head cuticle regions were fixed in 1% glutaraldehyde in PBS for 15 min at room temperature, and washed twice for 10 min with PBTr (1% Triton X-100 in PBS). The washed discs were stained with 1 mg/ml 5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside (X-gal) in X-gal staining solution (10 mM NaH₂PO₄/Na₂HPO₄ (pH 7.2), 150 mM NaCl, 1 mM MgCl₂, 3.1 mM K₄[Fe₂(CN)₆], 3.1 mM K₃[Fe₃(CN)₆], 0.3% Triton X-100) overnight at 37°C. Discs were washed twice for 10 min and mounted in 50% glycerol for microscopy.

In situ Hybridization to Whole Mount Imaginal Discs

In situ hybridization to whole-mount discs was performed using digoxigenin-labeled RNA probes (visualized as a blue alkaline phosphatase precipitate) based on previous reports (24 – 26).

RNA Probe Synthesis

1-804 bp region of CG17383 ORF was amplified by PCR using the following primer pair, son surgery-1: 5′-GGA ATT CAT GGA GAA CCT GTC GCA C-3′, son surgery-4: 5′-GCT CTA GAC TAC GGT GGC AAG GGG GCT AG-3′, and LD33329 as template. Resulting PCR product was subcloned into pGEM-Teasy vector (Promega). For sense strand, this plasmid was linearized by digestion with NcoI and transcribed with Sp6 RNA polymerase. For antisense strand, plasmid was linearized by digestion with SpeI and transcribed with T7 RNA polymerase. DIG-labeled UTP and DIG RNA labeling mix (Roche) was used to label the RNA probe. Finally, the probe was resuspended in 50% deioinized formamide, 5X SSC, 100 μg/ml salmon sperm DNA, 50 μg/ml heparin, 0.1% Tween20, and stored at −77°C until use.

Immunostaining of Imaginal Discs

Larvae at the appropriate stages were dissected in PBS. The dissected head cuticles were incubated in the fixation buffer (0.1 M PIPES pH 6.9, 1 mM EDTA, 1.0% Triton X-100, 2 mM MgSO₄, 1% formaldehyde (EM grade)) for 15 minutes, and then blocking solution (50 mM Tris-HCl pH 6.8, 150 mM NaCl, 0.5% NP-40, 5 mg/ml bovine serum albumin (BSA), 0.01% Thimerosal). The cuticles were transferred to the appropriate primary antibody in wash solution (50 mM Tris-HCl pH 6.8, 150 mM NaCl, 0.5% NP-40, 1 mg/ml BSA, 0.01% Thimerosal) and incubated at 4°C overnight. After washing three times with wash solution, they were incubated with the appropriate secondary antibody at room temperature for 2 h. After washing three times with wash solution, the cuticles were dissected and mounted in Antifade solution (Molecular Probes). Mounted tissues were photographed by confocal microscopy (Zeiss LSM510).

Antibodies

The primary antibodies used in these experiments were (dilution in parentheses): mouse anti-Ct (1:25), mouse anti-Wg (1:100), and mouse anti-Dl (1:1000) antibodies obtained from the University of Iowa hybridoma bank, rabbit anti-β-galactosidase (1:1000) (Sigma), and rat anti-HA (1:1000) (Roche) antibodies. Various fluorescently-labeled secondary antibodies were purchased from Jackson Immunology and used at 1:200 dilution.

Plasmid Rescue of Enhancer Trap P Element Insertions

Five micrograms of genomic DNA prepared from the enhancer trap lines were digested with EcoRI or SacII, and then ligated and used to transform XL-1 Blue E. coli by electroporation and ampicillin selection. DNA flanking the P insertion in the rescued plasmids was determined using two primers (IR primer: 5′-CGA CGG-GAC CAC CTT ATG-TTA TTT CAT CAT G-3′, A primer: 5′-GAG TTA ATT CAA ACC CCA CGG-3′).

Yeast Two Hybrid Screen

The full length of the CG17383 open reading frame was amplified from the LD33329 EST clone (Research Genetics) and used as bait by subcloning it as an EcoRI-BamHI fragment into the pBCT vector (Panbionet), which fuses it to the Gal4 DNA binding domain. The pBCT-son vector was used to transform the PBN203 yeast strain (MATα, ura-52, his3-200, ade2-101, trp1-901 leu2-3,112, gal4d, gal80d, met-, ura3d::pGAL1-lacZ ura3:: kanMX6pGAL1-URA3 Panbionet). A Drosophila embryonic cDNA library in the pACT2 vector (Clontech) was used to transform the PBN203 yeast containing the CG17383 bait. 8 × 10⁵ colonies were screened by selecting for growth on media lacking uracil and lacZ expression. Plasmids were recovered from positive colonies as described elsewhere (27) and sequenced.

Reverse Transcription (RT)-PCR

son transcripts were detected and quantified using RT-PCR. Total RNA was extracted from adult female flies using Trizol (GIBCO). For RT-PCR, the RNA was reverse transcribed using MMLV reverse transcriptase (Promega) and oligo-dT (18mer) primers. The resulting cDNA was amplified by PCR (rp49 primer set : 5′-CAC CAG TCG GAT CGA TAT GC-3′, 5′-CAC GTT GTG CAC CAG GAA CT-3′, son primer set: 5′-AAC TTC ACC AAA CAG AAC CG-3′, 5′-AAA ATG AGC GGA AAA TGC CC-3′).

P Element Excision

Enhancer trap lines were crossed to y w; Ki P{ry^+t7.2 = Delta2-3}99B flies. Progenies with mosaic eye colors were crossed to w;; TM3/TM6B. Progeny flies with white eyes were selected, and each of them was crossed to w;; TM3/TM6B to establish individually separate lines.

Ectopic Expression of N^ICD, Ser, Dl, HA-Son, H, Wg, and Su(H)

The Gal4-UAS driver system (28) was used to ectopically express several proteins in the wing imaginal discs by crossing flies containing P element insertions with the appropriate UAS constructs to flies containing appropriate Gal4 drivers. The Gal4 drivers used were Ser-Gal4 (w*; P{w^+mC = Ser-GAL4.GF}1 P{Ser-GAL4.GF}2, Bloomington stock #6791, donated by Lucy Cherbas and Robert Fleming) and ptc-Gal4 (w*; P{w^+mW.hs = GawB}ptc^559.1, Bloomington stock #2017, donated by Norbert Perrimon). Jaeseob Kim provided the UAS-N^ICD (29), UAS-Ser (30) and UAS-Dl (31) expression constructs. The UAS-Wg construct (32) was provided by Roele Nusse, and the UAS-H construct was obtained from Yuh Nun Jan. To create a UAS-HA-Son vector, the son open reading frame was amplified by PCR from the LD33329 EST (primer sequences 5′-GGA ATT CAT GGA GAA CCT GTC GCA C, 5′-GCT CTA GAC TAC TCG TCC TCC TCC TG) and cloned into the EcoRI and XbaI sites of the pUAST-HA vector, which contains three copies of the HA antigen coding sequence upstream of the multiple cloning site sequence. This construct was introduced into the germline of w¹¹¹⁸ mutant flies by P element-mediated germline transfer (33).

Su(H)^D47 Somatic Mosaic Clones

The FLP-FRT system (34) was used to generate Su(H)^D47 mutant somatic mosaic clones. FRT40A Su(H)^D47/Bc Gla; son¹⁸⁸³/TM6B flies were crossed to FRT40A, ubi-GFP/Bc Gla; hs-FLP, MKRS/TM6B flies (w¹¹¹⁸; MKRS, P{ry^+t7.2 = hsFLP}/TM6B, Tb¹, Bloomington stock #279 and w¹¹¹⁸; P{w^+mC = Ubi-GFP(S65T)nls}, P{ry^+t7.2 = neoFRT} 40A/CyO, Bloomington stock #5629, donated by Bruce Edgar were balanced and crossed to make this line) and progeny were heat shocked for 1 h at 37°C from 48 – 60 hour after egg-laying, and then dissected and immunostained as described at the 3rd instar stage. The FRT40A Su(H)^D47 chromosome (35) was provided by Jaeseob Kim.

Results

Identification of son (son of Notch), a Novel Gene Expressed along the DV Boundary in Wing Imaginal Discs

To identify novel genes potentially involved in dorsal-ventral boundary formation, we performed an enhancer trap screen with P{lacW} transposon (36). β-Galactosidase expression was examined in larval imaginal discs of nearly 7,500 enhancer trap lines. Two lines, 1883 and 2205, revealed virtually identical patterns in eye and wing discs of 3rd instar larvae (Fig. 1A, B). Notably, the expression in wing discs was detected along the DV boundary, suggesting that the gene(s) represented by these enhancer traps might play functional roles in this region during wing development. We identified the insertion site of each P element by plasmid rescue and both were located in the CG17383 gene (Fig. 1E). Based on the results presented in this study, we named this gene ‘son of Notch (son)’ and the enhancer trap lines will be referred to as son¹⁸⁸³ and son²²⁰⁵. The son¹⁸⁸³ insertion is in the first exon of three of the four predicted transcripts, and the son²²⁰⁵ insertion is in the first intron of all four transcripts. Both are upstream of the coding regions. The open reading frame is the same in all predicted transcripts, and it encodes a putative winged-helix DNA-binding protein of 336 amino acids (Fig. 1F).

Figure 1.

Structure and expression of the son of Notch (son) gene. (A, B) X-gal staining for β-galactosidase (β-gal) activity: (A) 3rd instar wing disc, (B) 3rd instar eye-antenna disc. (C, D) In situ hybridization of son transcripts: (C) 3rd instar wing disc, (D) 3rd instar eye-antenna disc. (E) Map of the son gene (CG17383), enhancer trap P element insertions, and four transcripts described by the Drosophila genome project. All transcripts encode the same protein (coding regions are denoted by filled boxes). (F) Map of the encoded Son protein showing the putative winged-helix and bipartite nuclear localization domains. (G) son transcript levels in wild-type (WT), homozygous son²²⁰⁵, and son¹⁸⁸³ adult female flies measured by RT-PCR. Virtually identical results were obtained using primers that detect exon 1 in the CG17383-RB transcript or exon 4 to 5 in the CG17383-RA transcript (data not shown). rp49 transcripts detected by RT-PCR provide an internal standard.

In situ hybridization of son in wild-type imaginal discs revealed expression patterns identical to that of β-Gal in son¹⁸⁸³ and son²²⁰⁵ (hereafter son-lacZ¹⁸⁸³ and son-lacZ²²⁰⁵, respectively), showing that the enhancer trap lines can serve as reliable markers of son expression (Fig. 1C, D). To check whether son expression is affected by these P insertions, the transcript level of son was examined by semi-quantitative RT-PCR. In both fly lines, the transcript level of son was significantly reduced, but not abolished (Fig. 1G), suggesting that these alleles are hypomorphic mutants of son.

Son Encodes a Putative DNA Binding Protein that Self Interacts

We conducted a yeast two-hybrid screen to find proteins that interact with the Son protein. Using Son as bait, and screening some 800,000 colonies transformed with a Drosophila embryonic cDNA library (Clontech), we recovered five interacting clones, two of which encoded N terminal portions of the Son protein, and three that encoded the entire Son protein. The Son self-interaction was also reported in a large scale genome-wide yeast two-hybrid screen (37). These results imply that Son functions as a homo-multimer.

Analysis of Son Expression in Wing Imaginal Discs

In 2nd instar wing discs, son-lacZ²²⁰⁵ was expressed in a stripe at the presumptive DV boundary, weakly at the hinge region, and in a circle just outside of the presumptive wing pouch. The DV band of son-lacZ²²⁰⁵ coincided with wingless (wg) expression at the DV boundary (Fig. 2A – C). In 3rd instar wing discs, the DV band of son-lacZ²²⁰⁵ resolved into two narrowly separated bands, and double immunolabeling showed that cut (ct) is expressed in the narrow gap between the two bands (Fig. 2D – F). These bands coincided with the expression domain of Delta (Dl), the Notch receptor ligand, that flanks the DV boundary (Fig. 2G – I). At this stage, the AP band of son-lacZ²²⁰⁵ expression lied between the AP bands of Dl expression. Such expression patterns of son further support the possibility that this gene might play a role in DV boundary formation by being involved with wg or Notch signaling pathways.

Figure 2.

Expression of the son enhancer traps in wing imaginal discs. (A – C) 2nd instar wing imaginal disc: (A) β-gal immunostaining, (B) Wingless (Wg) immunostaining, (C) merge of panels A and B. (D – F) 3rd instar wing imaginal disc: (D) β-gal immunostaining, (E) Cut (Ct) immunostaining, (F) merge of panels D and E. (G – I) Late 3rd instar wing imaginal disc: (G) β-gal immunostaining, (H) Delta (Dl) immunostaining, (I) merge of panels G and H.

The Son²²⁰⁵ Insertion alters Eye and Wing Development

We observed that the ventral posterior section of eyes in flies homozygous for the son²²⁰⁵ insertion was reduced in size (Fig. 3B). To test if the mutant phenotype was caused by the son²²⁰⁵ insertion, we carried out precise-excision of the P element. We recovered several viable excisions, and none displayed the mutant phenotype. For ten of these excisions, we sequenced the genomic region covering the original insertion site and confirmed that they were reverted to wild-type or contained small duplications of the insertion site. These results suggest that the son²²⁰⁵ insertion causes the mutant eye phenotype.

Figure 3.

Effects of the son²²⁰⁵ mutation on eye and wing development. Adult eyes from (A) wild type, (B) homozygous son²²⁰⁵ mutants, (C) N^nd-1 (nd¹) mutants and (D) nd¹; son²²⁰⁵ double mutants. Adult wings from (E) wild type, (F) nd¹ mutants, (G) nd¹; son²²⁰⁵/+ double mutants, and (H, I) wg^Gla-1/+; son²²⁰⁵ homozygous mutants. (I) Rare extreme notching phenotype of wg^Gla-1/+; son²²⁰⁵ homozygous mutants. Heterozygous or homozygous son²²⁰⁵ mutants do not display a mutant wing margin phenotype in the absence of nd¹ or wg^Gla-1 mutations (data not shown).

The Notch receptor plays multiple roles in eye development, including definition of the DV boundary, and cell differentiation within each ommatidium. Some mutations affecting Notch function can cause a preferential loss of the ventral domain (38). We observed that N^nd-1 (nd¹), the temperature-sensitive hypomorphic Notch allele, has a synergistic genetic interaction with son²²⁰⁵ (Fig. 3C, D). At 18°C, nd¹ eyes are slightly smaller and rougher than wild-type (39), but when combined with homozygous son²²⁰⁵, eyes were strongly reduced in size with virtually complete loss of the ventral half. These results suggest that son interacts with the Notch signaling pathway in the developing eye.

The son²²⁰⁵ insertion also provided genetic evidence demonstrating that son participates in the DV boundary formation in the developing wing. The slight adult wing margin-nicking phenotype displayed by nd¹ mutants (Fig. 3F) was dominantly enhanced by son²²⁰⁵ (Fig. 3G). Moreover, combining the gain-of-function wg^Gla-1 allele of wingless with homozygous son²²⁰⁵ caused a nicking phenotype in some 2/3 of the adult flies (Fig. 3H). In rare cases, these flies display a strong notching phenotype (Fig. 3I). The wg^Gla-1 and son²²⁰⁵ homozygotes by themselves did not display loss of wing margin (data not shown). The dramatic genetic interactions between son²²⁰⁵ and mutant alleles of Notch or wg strongly support the idea that this gene has functional roles in DV boundary formation by interacting with Notch or wg signaling pathways.

Notch Receptor Signaling Induces Son Enhancer Trap Expression

In order to examine the relationship between son and the Notch signaling pathway, we checked the effects of Notch signaling activity on son expression. When a Notch ligand binds to the Notch receptor, the intracellular fragment of Notch (N^ICD) is produced in response and functions as a transcriptional activator (40). To ectopically activate Notch signaling, N^ICD was expressed in the entire dorsal wing compartment using a Ser-Gal4 driver, and as a result, son-lacZ¹⁸⁸³ expression was induced in the corresponding region (Fig. 4A). Such effect on son-lacZ²²⁰⁵ expression was also observed (data not shown). Ectopic Ct expression was detected in this region as well, showing induction of Notch signaling activity (Fig. 4B).

Figure 4.

Activation of son-lacZ²²⁰⁵ by Notch signaling in wing discs. (A – I) 3rd instar wing imaginal discs immunostained for β-gal and Ct proteins. (A – C) Ectopic expression of the Notch receptor intracellular domain (N^ICD) in the dorsal compartment using a Ser-Gal4 driver. (D – F) Ectopic expression of the Serrate Notch ligand (Ser) at the anterior-posterior (AP) compartment boundary using a patched (ptc)-Gal4 driver. (G – I) Ectopic expression of the Delta Notch ligand (Dl) at the AP boundary using the ptc-Gal4 driver. (J – L) 3rd instar wing imaginal disc with a Su(H)^D47 mutant mosaic clone: (J) GFP expression showing the Su(H) mutant clone lacking GFP, (K) β-gal immunostaining, (L) merge of panels J and K. The clone border is marked by a white dashed line.

Ectopic expression of Notch ligands also induced son enhancer trap expression. Panel D – F of Figure 4 shows that ectopic expression of Ser by a patched (ptc)-Gal4 along the anterior-posterior (AP) boundary induced son-lacZ¹⁸⁸³ expression along the boundary, along with ct expression. A similar induction occurred when Dl ligand was expressed along the AP boundary (Fig. 4G – I).

N^ICD interacts with the Su(H) DNA-binding protein to activate gene transcription (41). We tested if Su(H) is required for induction of son-lacZ¹⁸⁸³ by generating mitotic clones homozygous for the Su(H)^D47 null allele (35) in wing imaginal discs. Su(H) mutant clones in the wing imaginal disc lost expression of son-lacZ¹⁸⁸³ (Fig. 4J – L). Thus, son expression in the DV boundary region is likely induced by a N^ICD-Su(H) complex.

Wg Potentiates Son Enhancer Trap Expression

The wg signaling pathway potentiates and refines expression of genes required for DV boundary formation in the wing. Ectopic expression of Wg protein in the dorsal compartment using a Ser-Gal4 driver by itself did not induce son-lacZ¹⁸⁸³ or ct expression (Fig. 5A – C). Also, ectopic expression of the Su(H) protein alone in the dorsal compartment did not induce son-lacZ¹⁸⁸³ or ct expression (Fig. 5D – F). In combination, however, ectopic dorsal expression of both Su(H) and Wg induced son-lacZ¹⁸⁸³ throughout most of the dorsal compartment, and ct expression in some dorsal cells neighboring the DV boundary (Fig. 5G – I). These results imply that ectopic Wg enhances the induction of the son enhancer trap by the Notch signaling pathway.

Figure 5.

Coexpression of wg and Su(H) induces son-lacZ¹⁸⁸³ expression. Regulatory interactions between wg, Su(H) and son. (A – C) Wg protein ectopically expressed in the dorsal compartment of a 3rd instar wing imaginal disc using the Ser-Gal4 driver: (A) β-gal immunostaining, (B) Cut (Ct) immunostaining, (C) merge of panels A and B. (D – F) Su(H) ectopically expressed in the dorsal compartment: (D) β-gal immunostaining, (E) Ct immunostaining, (F) merge of panels D and E. (G – I) Ectopic expression of Su(H) and Wg in the dorsal compartment: (G) β-gal immunostaining, (H) Ct immunostaining, (I) merge of panels G and H. Anterior is to the left and dorsal is to the top.

HA-Son Protein Induces ct, Dl, and wg Expression

The above results suggest that son is downstream of Notch and wg during wing margin development. To examine the function of son in these pathways, we ectopically expressed Son protein in the wing disc. We used a fusion protein containing the hemagglutinin epitope (HA-Son) to allow detection of the protein by immunostaining. Interestingly, ectopic expression of HA-Son along the AP boundary using the ptc-Gal4 induced expression of ct, Dl, and wg (Fig. 6A – I), suggesting that son might be involved in supporting the activation of signaling pathways involved in DV boundary formation.

Figure 6.

Crossregulation between ct, Dl, wg, and son. (A – J) HA-Son protein ectopically expressed at the AP boundary using the ptc-Gal4 driver: (A, D, G) HA-Son immunostaining, (B) Ct immunostaining, (C) merge of panels A and B, (E) Dl immunostaining, (F) merge of panels D and E, (H) Wg immunostaining, (I) merge of panels H and I, (J) β-gal immunostaining. (K – N) Ectopic expression of Su(H) and HA-Son at the AP boundary. (K) HA-Son immunostaining, (L) Ct immunostaining, (M) merge of panels K and L, (N) β-gal immunostaining.

Ectopic expression of HA-Son alone was insufficient to cause misexpression of son-lacZ¹⁸⁸³ (Fig. 6J). In Figure 5, we showed that coexpression of Wg and Su(H) leads to son-lacZ¹⁸⁸³ expression. Since ectopic HA-Son can cause wg expression, it seemed probable that coexpression of Su(H) with HA-Son could induce endogenous son expression. As expected, combined ectopic expression of Su(H) and HA-Son along the AP boundary induced expression of son-lacZ¹⁸⁸³ (Fig. 6N). Collectively, these results imply that son, wg, and the Notch signaling pathway support activation of each other.

Ct Negatively Regulates Son Enhancer Trap Expression

The Hairless (H) protein interacts with Su(H) and recruits co-repressor proteins such as Groucho and dCtBP (42). We reasoned that ectopic expression of H along the AP boundary might therefore reduce son enhancer trap expression at the intersection where the DV boundary crosses the AP boundary. However, while ptc-Gal4-driven expression of H caused downregulation of ct at the intersection, the expression of son-lacZ¹⁸⁸³ was not reduced (Fig. 7A – C). Notably, son enhancer trap expression appeared to increase in the cells where ct expression was reduced, and the two bands of son-lacZ¹⁸⁸³ expression that flank the DV boundary, which was marked by Ct staining, were fused at this intersection.

Figure 7.

Ct functions to downregulate the transcriptional activity of son. (A – C) Hairless protein ectopically expressed at the AP boundary: (A) β-gal immunostaining, (B) Ct immunostaining, (C) merge of panels A and B. (D – G) Hairless and HA-Son protein ectopically expressed at the AP boundary. (D) β-gal immunostaining, (E) Ct immunostaining, (F) merge of panels D, E, and G, (G) HA-Son immunostaining. (H-J) ct^2s mutant discs lacking ct gene expression at the dorsal-ventral boundary (wing margin). (H) β-gal immunostaining, (I) Ct immunostaining, (J) merge of panels H and I.

The loss of ct expression caused by ectopic H expression along the AP boundary was partially reversed by simultaneous ectopic expression of HA-Son (Fig. 7D – G). The ct expression was not fully restored, and it appeared that the son-lacZ¹⁸⁸³ expression remained as a single band at the intersection. The AP bands of HA-Son and ct expression overlapped along the AP boundary, as when HA-Son was expressed alone (Fig. 6A – C). These results suggest that in the DV boundary cells, where ct expression is the highest, son expression may be repressed by ct activity. Indeed, in ct^2s mutant discs, which lack ct expression in the boundary cells, the two bands of son-lacZ¹⁸⁸³ expression were fused into one band (Fig. 7H – J). Thus, it appears that high levels of ct expression in the DV boundary cells inhibit son expression. Notably, when ectopic expression of Ct or Son was induced in cells outside the DV boundary in previous results, ct expression and son expression were not always mutually exclusive. Thus, it is likely that high levels of Ct and/or Ct in combination with other factors that function in the DV boundary are responsible for repressing son expression in the cells at the DV boundary.

Discussion

Son Participates in the Notch Signaling Pathway at the Dorsal-ventral Boundary in Wing Development

We named the gene defined by the son²²⁰⁵ and son¹⁸⁸³ enhancer trap insertions son of Notch (son) because several lines of evidence support the idea that it participates in the Notch signaling pathway at the DV boundary in the developing wing: First, the son enhancer traps are expressed at the dorsal-ventral boundary in wing discs. Second, heterozygous and homozygous son²²⁰⁵ strongly enhance the mutant wing margin phenotype displayed by the hypomorphic N^nd-1 mutation of Notch, and these effects are reverted by precise excision of the P element. Third, son enhancer trap expression in the wing disc is induced by expression of Notch pathway components, including Notch ligands, and the intracellular fragment of Notch. Fourth, the Su(H) DNA-binding protein that forms a transcriptional activator complex with Notch is essential for son enhancer-trap expression at the wing DV boundary, and together with Wg, it can induce son enhancer trap expression. Fifth, ectopic expression of Son protein in the wing disc induces expression of ct, Dl and wg, which are normally induced by Notch signaling at the DV boundary. Sixth, son enhancer trap expression is negatively regulated by ct at the DV boundary. Altogether, these observations provide compelling evidence that son is a component of the genetic pathway induced by Notch at the wing margin, and we hypothesize that son functions downstream of Notch at the wing margin, participating in a positive regulatory loop with wg, and a negative regulatory loop with ct (Fig. 8).

Figure 8.

Possible regulatory interactions between son and components of the Notch signaling pathway at the wing dorsal-ventral boundary. The interactions involving son are hypothesized based on the results presented in this study. Previously known direct interactions are indicated with solid arrows.

Analysis of the roles of son in development would be aided by a null mutant allele of son, but we were unsuccessful in generating such an allele, despite extensive efforts. The Df(3R)Su9 allele that lacks a genomic region containing son is haplolethal (43). It may be that a son null allele may be haploinsufficient or haplolethal, or there may be a haplolethal gene near son, thus making it difficult to generate deletion mutants of this gene.

What is the Function of Son at the Wing Dorsal-ventral Boundary?

The precise role of son at the wing dorsal-ventral boundary is yet unclear, but the pattern of son enhancer-trap expression raises certain possibilities. The expression of son-lacZ¹⁸⁸³ and Dl overlap in cells flanking the DV boundary where ct is expressed. Although misexpression of Son can ectopically induce expression of ct and Dl, the patterns of son-lacZ¹⁸⁸³ and ct expression at the DV boundary appear to be mutually exclusive. Moreover, loss of ct expression augments son-lacZ¹⁸⁸³ expression and leads to fusion of the two son stripes into a single broad band. Thus one possibility is that son and ct function in a loop to control ct levels, in which son helps induce ct expression, but at a high level and/or in combination with other factors, ct represses son expression. We also speculate that son could help maintain Dl expression along the DV boundary, and that ct blocks Dl expression in the boundary cells by repressing son. This scenario could explain why the son²²⁰⁵ mutation enhances the defective wing margin phenotype of a hypomorphic Notch mutation, since reducing Dl levels would further decrease Notch signaling activity. It is also possible that Son supports wg expression at the DV boundary, since wg is also induced by HA-Son. Although the underlying mechanisms are currently unresolved, this study presents clear evidence that son plays a definite role involved with the Notch signaling pathway in regulating DV boundary formation. Moreover, the significant effects between this novel component and several other components in this pathway suggest that further analyses of such cross-regulations will contribute to understand the intricate process of DV boundary formation.

Son encodes a self-interacting protein with a winged-helix DNA-binding domain, and it seems likely that Son acts as a transcriptional activator or repressor. For example, it could directly contribute to activation of Dl or other genes that function at the DV boundary, such as vestigial, scalloped, or wg.

In addition to the eye and wing discs, the son enhancer traps are also expressed in leg discs and the central/peripheral nervous systems of developing embryos, and thus it is likely that son has several other developmental roles.