Nipped-A, the Tra1/TRRAP Subunit of the Drosophila SAGA and Tip60 Complexes, Has Multiple Roles in Notch Signaling during Wing Development

Northern blot assays.Northern blot assays were performed with formaldehyde-agarose gels and radioactively labeled single-stranded RNA probes as previously described (14). Total RNA was isolated from second-instar larvae with Trizol (Gibco BRL). Ethyl methanesulfonate (EMS)-induced Nipped-A mutations were provided by Steve Myster and Mark Peifer. The mutations were backcrossed to the parental isogenic cn bw chromosome and balanced over CyO Kr-GAL4 UAS-GFP or CyO Df(2R)Kr ⁻ Tp(1;2)y ⁺. Homozygous and hemizygous mutants were selected by the lack of green fluorescent protein expression or reduced pigmentation of larval mouthparts. Hemizygous mutants were generated by crossing mutants to the balanced Nipped-E ⁴³ deficiency that deletes Nipped-A and neighboring genes (45). Radioactively labeled single-stranded RNA probes were prepared by in vitro transcription with T7 RNA polymerase and a vector prepared by cloning a C-terminal fragment of the Nipped-A coding sequence into pGEM-1.

Sequencing of Nipped-A mutations.Total RNA was isolated from second-instar larvae, and 1 μg was used to generate cDNA with random hexamer primers and Superscript III reverse transcriptase (Invitrogen) according to the manufacturer's directions. A series of overlapping fragments of Tra1/TRRAP were amplified by PCR, treated with exonuclease I and shrimp alkaline phosphatase (ExoSAP; U.S. Biochemicals), and sent to Retrogen, Inc., to be sequenced with the PCR primers. Sequences were assembled and compared with CodonCode Aligner software (CodonCode Corp.). Mutations in splice sites and the 3′ untranslated region (UTR) were confirmed by sequencing PCR products generated from cn bw and mutant genomic DNA.

5′ and 3′ rapid amplification of cDNA ends (RACE).The ends of the Tra1/TRRAP RNA were amplified with the BD SMART RACE cDNA amplification kit according to the manufacturer's directions, and the products were cloned into a TOPO TA pCRII vector (Invitrogen) and sent to Retrogen, Inc., for sequencing.

Effects of Nipped-A, Ada2b, and domino mutations on Notch and mastermind mutant phenotypes.The dominant effects of the Nipped-A, Ada2b, and domino mutations on the Notch and mastermind mutant phenotypes were determined by conducting genetic crosses at 25^oC as previously described (45). Balanced mutant males were crossed to females that were homozygous N ^nd-1, homozygous N ^Ax-E2, or mam ^g2/CyO Df(2R)Kr ⁻ Tp(1;2)y ⁺. Male progeny were scored for the severity of the Notch or heterozygous mastermind mutant phenotype as described in Results. Ada2b mutations were provided by Mattias Mannervik and Thomas Kusch, and the dom ¹ allele of domino was provided by Marie-Laure Ruhf and Marie Meister. The dom ¹ mutation is a P element insertion 1 bp downstream of the start of the first intron (47). The dom ^R12 revertant of dom ¹ was generated with P transposase (CyO, HOP2; provided by Ron Blackman) to excise the P element. Reversion was confirmed by restoration of viability and PCR analysis.

Generation of Nipped-A antisera.Two His₆-tagged fragments of Nipped-A encoded by exon 12 (NipA12) and exon 14 (NipA14) were expressed in bacteria with the pMCSG7 vector (50) and purified under denaturing conditions with QIAGEN nitrilotriacetic acid beads. Denatured NipA14 protein was eluted from the resin, precipitated with 30% polyethylene glycol, suspended in phosphate-buffered saline (PBS), and used to immunize a rabbit at the Pocono Rabbit Farm and Laboratory (Canadensis, PA). NipA12 protein was washed with PBS on the beads and used to immunize a guinea pig while attached to the beads. The inserts for protein production were generated by PCR from cn bw cDNA. The primers for NipA12 were 5′-TACTTCCAATCCAATGCAATGATGGTGGAGCCACAAGCTTT-3′ and 5′-TTATCCACTTCCAATGCTAATGAGGATCCATAACTTCAGAAACGG-3′, which amplify sequences encoding amino acids 871 to 1232. The primers for NipA14 were 5′-TACTTCCAATCCAATGCTGAGGCATCGAGTCCATATCGAG-3′ and 5′-TTATCCACTTCCAATGCTAATTTCCATCCTCCATTCGCAACG-3′, which amplify sequences encoding amino acids 1490 to 1870. NipA12 and NipA14 antibodies were affinity purified from serum with antigen on Immobilon membrane (21).

Western blot assays.Twenty microliters of dechorionated 0- to 12-h-old embryos with 80 μl of PBS was lysed with 100 μl of 2× sodium dodecyl sulfate-polyacrylamide gel electrophoresis loading buffer and denatured by boiling for 5 min, and 10 and 20 μl of each sample, containing the denatured protein from ∼115 and 230 embryos, was separated by 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blots were blocked with PBS containing 1% bovine serum albumin and 1% normal donkey serum. The primary antibody was affinity-purified NipA12 diluted 1:100, and the secondary antibody was horseradish peroxidase-conjugated donkey anti-guinea pig antibody diluted 1:10,000 (Jackson Labs), which was detected with chemiluminescent reagents (Perkin-Elmer).

Immunostaining of larval tissues.Larval tissues were fixed and immunostained as previously described (3), with purified NipA12 (1:50) or NipA14 (1:100) antibody, followed by Cy3-labeded donkey anti-guinea pig (1:400) or fluorescein isothiocyanate-conjugated goat anti-rabbit (1:200) secondary antibody. Samples were mounted in Prolong Antifade (Invitrogen) containing 1 μg of 4′,6′-diamidino-2-phenylindole (DAPI) per ml.

Immunostaining of salivary gland polytene chromosomes.Salivary glands were fixed for 30 s in 2% formaldehyde and then in 45% acetic acid and 2% formaldehyde for 3 min as previously described (31) before storage in 67% glycerol-33% PBS at −20°C. Mouse monoclonal antibody against mastermind, provided by Spyros Artavanis-Tsakonas, and rabbit polyclonal serum against Gcn5, provided by Thomas Kusch and Jerry Workman, were both used at a 1:100 dilution. The NipA12 and NipA14 antisera were used at 1:100 and 1:200 dilutions. Alexa Fluor 488 (Molecular Probes)-conjugated goat anti-mouse and Cy3-labeled goat anti-rabbit (Sigma) antibodies were used at a 1:200 dilution. Fluorescein isothiocyanate-conjugated goat anti-rabbit antibody (Sigma) was used at 1:100, and Cy3-labeled donkey anti-guinea pig antibody (Jackson Labs) was used at 1:200.

Nucleotide sequence accession numberThe sequence of the major Nipped-A/Tra1/TRRAP transcript determined in this work has been given GenBank accession no. DQ352451.

Identification of Nipped-A mutations in the Tra1/TRRAP gene.The Tra1/TRRAP protein and the HAT complexes (SAGA and Tip60/NuA4) that contain it have been well characterized biochemically. Little is known, however, about the roles of Tra1/TRRAP in metazoan gene expression, although it has been implicated in signaling pathways that control development. Similar to Notch, Caenorhabditis elegans Tra1/TRRAP was recently found to repress ras signaling in vulval development, although the mechanism is unclear (4, 9). Drosophila Nipped-A mutations were shown to affect Notch functions in wings (45), and a recent molecular analysis of two EMS-induced Nipped-A mutations strongly suggested that Nipped-A encodes Tra1/TRRAP (38). We undertook a molecular genetic analysis to confirm that Nipped-A is the Tra1/TRRAP gene and gain insights into the roles that Tra1/TRRAP complexes play in Notch signaling during wing development.

Northern blot assays of RNAs from several Nipped-A mutants provides further evidence that Nipped-A encodes Tra1/TRRAP (Fig. 1). A mutant allele generated by γ rays, Nipped-A ³²³ (45), produces no detectable Tra1/TRRAP transcript (lane 2), and an allele generated by EMS mutagenesis, Nipped-A ^NC105 (38), produces a transcript shorter than that of the wild type (Fig. 1, lane 17). Several other EMS-generated Nipped-A alleles (38) produce wild-type-sized Tra1/TRRAP transcripts in greater or lesser amounts (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Tra1/TRRAP transcripts are altered in size and level in Nipped-A mutants. The panels show three independent Northern blot assays of total RNA isolated from either hemizygous or homozygous Nipped-A mutant second-instar larvae, as indicated above the lanes. The wild-type Tra1/TRRAP transcripts are 11.8 kb in size. The Northern blots were stripped and reprobed for rp49 as a control for gel loading. Arrows indicate mutants used in the genetic analysis of Notch and mastermind function. Tra1/TRRAP transcripts in these mutants were quantified by phosphorimager, and the levels were normalized to the amount of rp49 transcript. The levels of the mutant transcripts, with the amount of wild-type transcript set to 100%, are shown in the bar graph in the bottom panel.

We sequenced Tra1/TRRAP transcripts by using reverse transcription (RT)-PCR to identify EMS-generated Nipped-A mutations. The wild-type transcript in the isogenic cn bw chromosome (Fig. 2) used for mutagenesis (38) encodes a Tra1/TRRAP protein of 3,790 residues (Fig. 3). The splicing pattern of this major transcript differs from that of the originally reported mRNA (Fig. 2) (29). The transcript we detect, however, is the major transcript in embryos and second-instar larvae of both cn bw and Oregon R. We did not detect the previously reported variant (29) in total RNA from embryos or second-instar larvae by RT-PCR with various primers specific for its unique exons, so we conclude that it is a rare variant.

• Open in new tab
• Download powerpoint

FIG. 2.

The Tra1/TRRAP gene and locations of Nipped-A mutations. Shown is a diagram of the Tra1/TRRAP gene, starting with the most 5′ RACE product and ending with the polyadenylylation site. The numbering is in kilobase pairs. The exons present in the major Tra1/TRRAP transcript are shown and numbered consecutively on the line, and the exons in the previously reported (29) rare transcript are shown below, numbered consecutively (1a, 2a, etc.). While most of the exons that overlap in the major and minor transcripts are identical, the positions of the splice acceptor or donor sites differ in three, as indicated by asterisks. Locations of the Nipped-A mutations listed in Table 1 are indicated by vertical lines.

• Open in new tab
• Download powerpoint

FIG. 3.

Tra1/TRRAP proteins encoded by Nipped-A mutants. The diagrams show the proteins encoded by the major Tra1/TRRAP transcript in the indicated Nipped-A mutants (Fig. 2; Table 1). The locations of the PI3 kinase motif, the FAT and FATC motifs associated with phosphatidylinositol-3 kinases, and two HEAT repeats are indicated by boxes. The Nipped-A ^NC105 and Nipped-A ^NC106 mutations cause frame shifts that add nonnative residues at the ends of the truncated proteins.

Nipped-A ^NC105, which produces a short transcript (Fig. 1), has a mutant splice acceptor site (Table 1) that causes skipping of exon 17 (Fig. 2). The resulting open reading frame encodes a protein lacking much (46%) of the C-terminal half of Tra1/TRRAP, including the putative ATM-PI3 kinase motif (Fig. 3). Nipped-A ^NC105 produces nearly 70% of the wild-type level of Nipped-A transcript (Fig. 1).

The Nipped-A ^NC194 mutation creates a stop codon in exon 14 (Fig. 2; Table 1), truncating the protein after residue 1500 (Fig. 3). The level of Nipped-A ^NC194 transcript is nearly the same as that of the wild type (Fig. 1), suggesting that it is not subject to nonsense-mediated decay.

The Nipped-A ^NC106 mutation affects the exon 5 splice acceptor site (Fig. 2; Table 1). The lack of splicing causes a frame shift, resulting in a truncated 265-residue protein (Fig. 3). Nipped-A ^NC106 produces only 44% of the expected level of transcript (Fig. 1), suggesting that it may be reduced by nonsense-mediated decay.

Nipped-A ^NC96 has a T-to-C nucleotide change in the 3′ UTR (Table 1; Fig. 2), three residues upstream of the polyadenylylation site. This mutant produces more transcript than expected (Fig. 1, bottom panel) but, as described below, is a weak loss-of-function allele. Combined with the Northern blot analysis, identification of point mutations in multiple Nipped-A mutants demonstrates conclusively that Nipped-A encodes Tra1/TRRAP.

Tra1/TRRAP supports mastermind and Notch function in wing margin and wing vein development.The various Nipped-A mutations differ in their effects on Notch signaling (Fig. 4). We measured their dominant effects on the mutant phenotypes displayed by the N ^nd-1 and N ^Ax-E2 alleles of Notch and the heterozygous mam ^g2 allele of mastermind. The N ^nd-1 hypomorphic mutant causes gaps in the wing margin and reduces blade size (Fig. 4). This phenotype was quantified by measuring the anterior-posterior width of the wing midway between the two crossveins (Fig. 4). A narrower wing indicates lower Notch activity. The N ^Ax-E2 hypermorphic mutant has shortened L4 and L5 veins (Fig. 4). This was quantified by measuring the length of the L4 vein from the posterior crossvein to the tip of the wing, and a longer vein indicates reduced Notch signaling (Fig. 4). The heterozygous mam ^g2 mutation, which was induced by γ rays and behaves genetically as a null allele (42), causes rare nicks (gaps in the bristle rows) in the wing margin. This phenotype was quantified by counting the nicks, with more nicks indicating reduced mastermind activity.

• Open in new tab
• Download powerpoint

FIG. 4.

Effects of heterozygous Nipped-A mutations on Notch and mastermind mutant phenotypes. The left panel shows that the heterozygous mam ^g2 null mastermind mutation (45) causes rare wing margin nicks and some Nipped-A mutations increase the frequency of these nicks, indicating reduced mastermind function. The wild-type Nipped-A allele is the parental allele in the cn bw stock used to generate the EMS-induced Nipped-A mutations (NC). The error bars indicate the 95% confidence intervals. The middle panel shows box plots of the distribution of wing widths of N ^nd-1 hypomorphic Notch mutant flies relative to the wild type (wt) in the presence of the indicated Nipped-A alleles. Wing widths were measured from the anterior to the posterior margin between the anterior and posterior crossveins, as shown at the bottom. Some Nipped-A mutations decrease the width, indicating reduced Notch function. The median width is indicated by the black line within each box, the ends of each box indicate the 25th and 75th percentiles, and the ends of the “whiskers” indicate the 10th and 90th percentiles. Circles show the lowest and highest values. The right panel shows the relative lengths of the L4 vein in N ^Ax-E2 hypermorphic Notch mutant flies in the presence of the Nipped-A alleles. As shown below, the veins were measured from the posterior crossvein toward the tip of the wing. Some Nipped-A mutations increase the length of the vein, indicating reduced Notch activity. The error bars show the 95% confidence intervals.

As observed previously (45), the Nipped-A ³²³ null allele had significant effects in all three genetic tests, increasing the severity of the N ^nd-1 mutant phenotype, decreasing the severity of the N ^Ax-E2 gain-of-function phenotype, and increasing the number of margin nicks caused by mam ^g2 (Fig. 4). All these effects indicate reduced Notch activity.

The Nipped-A ^NC106 mutation, predicted to truncate Tra1/TRRAP near the N terminus, had substantial effects on N ^nd-1 and mam ^g2, similar to the null allele, but little or no effect on N ^Ax-E2 (Fig. 4). We cannot quantitatively compare the effects of Nipped-A ^NC106 to those of Nipped-A ³²³ because they were generated in different genetic backgrounds. Because other Nipped-A mutations generated in the same isogenic chromosome as Nipped-A ^NC106 do affect N ^Ax-E2, however, we postulate that Nipped-A ^NC106 is not a null mutation and retains some activity. Because Nipped-A ^NC106 is a splice site mutation, it is possible that an alternatively spliced product provides some Tra1/TRRAP activity. In support of this idea, the splice site affected by Nipped-A ^NC106 is not used in the originally reported transcript (Fig. 2) (29), demonstrating that at least one known splicing pattern can produce a full-length Tra1/TRRAP protein in this mutant. The recessive lethality of Nipped-A ^NC106 and its effects on N ^nd-1 and mam ^g2, however, also demonstrate that the alternative transcript is not sufficient to fulfill some critical Tra1/TRRAP functions.

Nipped-A ^NC96, which has a point mutation in the 3′ UTR just upstream of the polyadenylation site, behaves as a hypomorph. Nipped-A ^NC96 increased the severity of the N ^nd-1 wing phenotype but had no significant effect on N ^Ax-E2 or mam ^g2. Nipped-A null mutants die at the second- to third-instar molt (45), but flies hemizygous for Nipped-A ^NC96 survive at a significant frequency to the pupal stage (Table 2). Third-instar larvae hemizygous for Nipped-A ^NC96 have smaller or missing imaginal disks, and acridine orange staining suggests that there is apoptosis in the mutant disks (not shown). The Nipped-A ^NC186 mutant, which encodes a Tra1/TRRAP protein with a missense mutation (38), also survives as a hemizygote past pupariation, suggesting that it is also a hypomorph.

Intriguing results were obtained with Nipped-A ^NC105 and Nipped-A ^NC194. Although Nipped-A ^NC105 produces reduced levels of a short transcript lacking exon 17 (Fig. 1; Fig. 2; Table 1) and encodes a protein lacking most of the C-terminal half of Tra1/TRRAP (Fig. 3), it had little dominant effect on Notch signaling in any of the three genetic assays (Fig. 4). None of the known or theoretical splicing patterns could generate a full-length protein, and thus one possible interpretation is that the C-terminal half of Tra1/TRRAP is not required for Notch and mastermind activity in the developing wing margin and veins. This is surprising, because the available evidence indicates that the missing region, which contains the ATM-PI3 kinase motif (Fig. 3), is required for association of Gcn5 and Tip60 with mammalian TRRAP (41) and is also involved in interactions with acidic activator proteins (6). An alternative is that the mutant allele somehow increases the activity of the remaining wild-type Nipped-A gene or Tra1/TRRAP protein.

Nipped-A ^NC194, which has a nonsense mutation in exon 14 (Fig. 2; Table 1) and encodes a shorter protein than Nipped-A ^NC105 (Fig. 3), had strong effects on both the hypomorphic wing margin and hypermorphic wing vein Notch mutant phenotypes but no significant effect on the mam ^g2 wing-nicking phenotype (Fig. 4). Thus, heterozygous Nipped-A ^NC194 reduces Notch, but not mastermind, activity. Nipped-A ^NC194, which supports fewer functions than Nipped-A ^NC105, produces more mRNA (Fig. 1, bottom panel). We cannot exclude the possibility, however, that the Nipped-A ^NC194-encoded proteins are less stable than the Nipped-A ^NC105-encoded proteins, and thus the difference between the two alleles may be caused by a difference in the levels of the two mutant proteins. As discussed below, other evidence suggests that Nipped-A-containing complexes have multiple roles in Notch signaling, and thus it is also possible that Nipped-A ^NC194 is a separation-of-function allele that does not disrupt the ability of Nipped-A to support mastermind activity.

We think it unlikely that the ability of the Nipped-A ^NC105 allele to support mastermind and Notch functions in wing development or the ability of the Nipped-A ^NC194 allele to support mastermind function is caused by linked second-site mutations that counteract or mask the effects of these Nipped-A mutations on the Notch and mastermind mutant phenotypes. The Nipped-A ^NC105 and Nipped-A ^NC194 chromosomes were backcrossed to the parental cn bw chromosome during balancing to reduce the number of unlinked mutations. Of the many reported genetic effects on the wing margin phenotype of N ^nd-1 (http://flybase.bio.indiana.edu/.bin/fbidq.html?FBal0012890&content=geneinter ), most enhance the phenotype, and only mutations in Nipped-B and Notchless (Nle) reduce its severity (45, 46). Nipped-B is tightly linked to Nipped-A, but complementation tests show that the Nipped-A ^NC105 and Nipped-A ^NC194 chromosomes do not contain Nipped-B mutations. Nle is located near the tip of 2L (21C) and thus is essentially unlinked from Nipped-A, which is on 2R. Similarly, most mutations that affect N ^Ax-E2 suppress the vein-shortening phenotype, as do Nipped-A mutations (http://flybase.bio.indiana.edu/.bin/fbidq.html?FBal0012858&content=geneinter ). The chromosome 2 mutations with opposite effects are (paradoxically) particular Su(H) mutations and Suppressor of deltex [Su(dx)] mutations. Su(dx) is also located near the tip of 2L (22C) and thus is essentially unlinked to Nipped-A. Many other Su(H) mutations suppress the N ^Ax-E2 phenotype, and a chromosome with both Su(H) and Nipped-A mutations, Nipped ^226.1, displays a dominant wing-nicking phenotype (45). The Nipped-A ^NC105 and Nipped-A ^NC194 chromosomes do not show such a phenotype. Thus, it is also unlikely that the lack of a significant effect of Nipped-A ^NC105 on the N ^Ax-E2 phenotype is caused by a second-site mutation.

Finally, screens of deficiencies covering much of the genome and known Notch and wingless signaling pathway mutants for modifiers of a phenotypes caused by expression of a dominant-negative mastermind protein identified many enhancers of the mutant phenotypes and few suppressors (22, 53). None of the suppressors map to chromosome 2. Thus, it is also improbable that both the Nipped-A ^NC105 and Nipped-A ^NC194 chromosomes contain linked second-site mutations that suppress the mastermind mutant phenotype. In the case of Nipped-A ^NC105, it would likely require multiple linked second-site mutations to mask effects on all three phenotypes.

Effects of Nipped-A mutations on Tra1/TRRAP proteins.We developed antisera against Nipped-A to evaluate the effects of the Nipped-A mutations on Tra/TRRAP protein. Antisera were generated against polypeptides encoded primarily by exon 12 (NipA12) and exon 14 (NipA14). As expected, both antisera recognize a major band larger than 400 kDa in Western blot assays of various developmental stages and cultured cell extracts. The highest levels of Nipped-A protein are found in embryonic nuclear extract, most of which is provided maternally. We examined the levels of Nipped-A protein in nuclear extracts of embryos produced by wild-type mothers and mothers heterozygous for Nipped-A ^NC105 and Nipped-A ^NC194. As expected, the mutant mothers both produced embryos with roughly half the amount of full-length Nipped-A protein compared to the wild type (Fig. 5). This confirms that the major band recognized by the antibodies is Nipped-A (Tra1/TRRAP). It also indicates that there is not substantial overexpression of the wild-type Nipped-A allele in these mutants, which could potentially mask the effects of these mutations on Notch and mastermind mutant phenotypes.

• Open in new tab
• Download powerpoint

FIG. 5.

Western blot assay of Nipped-A mutants. Each lane contained 10 or 20 μl of extract of 0- to 12-h-old embryos from mothers of the indicated genotypes: +/+, wild type; 105/+, Nipped-A ^NC105/+; 194/+, Nipped-A ^NC194/+. The amounts loaded are the equivalent of 115 and 230 whole embryos. The blot was probed with affinity-purified NipA12 antibody. Densitometry indicates that the mutants contain 35 to 53% of the full-length Nipped-A protein present in the wild type.

Based on the abilities of the Nipped-A ^NC105 and Nipped-A ^NC194 genes to support or partially support Notch and mastermind activities in wing development, we thought it possible that truncated mutant proteins could be detected in mutant embryo extracts. Although both mutant proteins should contain the polypeptide used to generate the NipA12 serum, we did not detect truncated proteins of the expected sizes (234 and 171 kDa for Nipped-A ^NC105 and Nipped-A ^NC194, respectively) (Fig. 5). We considered the possibility that the proteins encoded by Nipped-A ^NC105 and Nipped-A ^NC194 might be less stable in embryos than at other developmental stages. We were also unable, however, to detect the expected truncated proteins in Western blots of heterozygous mutant larvae (not shown). It should be noted, however, that the levels of full-length protein are much lower in larvae than in embryos and near the limit of detection by Western blot assay with either antibody. It is also feasible that proteins encoded by the mutant alleles are modified after synthesis and lack the critical epitopes.

We also examined protein expression by immunostaining, which can be more sensitive than Western blot assays. The Nipped-A ^NC105, Nipped-A ^NC194, and Nipped-A ^NC106 mutants all die at the second- to third-instar molt as hemizygotes, so we immunostained several second-instar larval tissues of these mutants for Nipped-A and compared them to hemizygous wild-type Nipped-A (Fig. 6). Wild-type larvae show cytoplasmic and brighter nuclear staining in multiple tissues, as exemplified by the polytene cells of the second-instar proventriculus shown in Fig. 6. Staining above the background was also observed in imaginal disks and several other tissues, including the gut and salivary glands, and the same staining pattern was seen with both the NipA12 and NipA14 antibodies (not shown). No staining above the background was seen with hemizygous Nipped-A ^NC106, demonstrating that the staining is specific for the Nipped-A protein (Fig. 6).

• Open in new tab
• Download powerpoint

FIG. 6.

Fluorescent immunostaining of larval tissues for Nipped-A. The examples shown were performed with affinity-purified NipA14 antibody, and essentially identical results were obtained with NipA12 antibody. The genotypes are indicated on the right: +/+, wild type; 106/Df, Nipped-A ^NC106/Nipped-E ⁴³; 105/Df, Nipped-A ^NC105/Nipped-E ⁴³; 194/Df, Nipped-A ^NC194/Nipped-E ⁴³; +/Df, +/Nipped-E ⁴³; 96/Df, Nipped-A ^NC96/Nipped-E ⁴³. The micrographs were taken with a 60× objective. All second-instar micrographs were taken with identical exposure times, and both third-instar micrographs were identical exposures.

The NipA12 antibody is predicted to detect the Nipped-A ^NC105 and Nipped-A ^NC194 truncated proteins, and the NipA14 antiserum should detect the Nipped-A ^NC105 protein, but we did not see staining above the background with either antiserum in either mutant (Fig. 6). Thus, although the genetic analysis indicates that these mutant alleles retain the ability to support some Notch and mastermind functions in wing development, we were unable to detect protein products. It is possible that mutant proteins are not recognized by the antisera or that they are present at levels below the detection limit.

The genetic analysis indicates that Nipped-A ^NC96 is a hypomorph, even though it produces more mRNA than the wild type. We immunostained larval tissue hemizygous for Nipped-A ^NC96 and wild-type Nipped-A and observed that staining is reduced in the mutant compared to the wild type (Fig. 6). The wild type and mutant were stained simultaneously, and the photographic exposure times were identical. Nipped-A ^NC96 does not contain mutations in the open reading frame but has a point mutation in the 3′ UTR. Because the immunostaining indicates that this mutant allele produces less protein, we speculate that the point mutation reduces translation. Because Nipped-A ^NC96 is a hypomorph and, unlike a null allele, affects only the N ^nd-1 phenotype in the functional genetics tests, we conclude that the N ^nd-1 phenotype is more sensitive to Nipped-A levels than are the N ^Ax-E2 and mastermind phenotypes.

The Ada2b subunit of SAGA is required for mastermind activity.Tra1/TRRAP (Nipped-A) is a component of the SAGA complex that contains the Gcn5 HAT protein. To determine if SAGA is required for Notch and mastermind activity, we tested the effects of mutations affecting the Ada2b subunit of SAGA on the Notch and mastermind mutant phenotypes. Biochemical analysis indicates that Ada2b is found primarily in SAGA (29). All three Ada2b mutations tested increased the number of wing margin nicks displayed by heterozygous mam ^g2 but did not have any significant effects on the N ^nd-1 and N ^Ax-E2 mutant phenotypes (Fig. 7). All three mutant alleles were generated by P element excision and were compared to the original viable P element-bearing chromosome [EP(3)3412] used to generate them (44; T Kusch, personal communication). The Ada2b ¹ and Ada2b ² alleles have large deletions (1.1 and 2.8 kbp) that remove the transcription start site and much of the gene, and homozygotes die during pupariation (44). The Ada2b ^HD-5 allele, an unpublished allele provided by Thomas Kusch and Jerry Workman (Stowers Institute), appears to be hypomorphic because a small percentage of homozygotes eclose but die shortly thereafter. Consistent with this idea, the effect of Ada2b ^HD-5 on mam ^g2 was not as strong as the effects of the other alleles (Fig. 7).

• Open in new tab
• Download powerpoint

FIG. 7.

Effects of Ada2b mutations on Notch and mastermind mutant phenotypes. The effects of the indicated Ada2b mutations and the parental P insertion [EP(3)3412] used to generate them on the mam ^g2 wing nicking phenotype, the N ^nd-1 wing width phenotype, and the N ^Ax-E2 vein-shortening phenotype were determined and diagrammed as described in the legend to Fig. 1. Homozygous Ada2b ^HD-5 mutations occasionally survive to eclosion, indicating that they are hypomorphic, while the Ada2b ¹ and Ada2b ² alleles both delete significant portions of the transcribed regions and die during pupal development (44).

Compared to the Nipped-A ^NC194 mutation, the Ada2b mutations have the opposite specificity in their effects on Notch signaling: Nipped-A ^NC194 affected the Notch mutants but had no detectable effect on mam ^g2 (Fig. 4), while the Ada2b mutations affected mam ^g2 but not the Notch mutants (Fig. 7). We conclude, therefore, that only a mastermind-dependent function of Nipped-A requires the Ada2b subunit of SAGA. This also indicates that Nipped-A has functions in Notch signaling that are not Ada2b dependent.

The domino subunit of the Tip60 complex is required for Notch and mastermind activities.Domino is a Tip60 (NuA4) component in drosophila and mammals (15, 28), but it is unknown if this is the only complex that contains domino in vivo. The domino gene encodes two proteins, domino-A and domino-B, which have significant sequence similarity to the mammalian p400 and SRCAP proteins (47). Both are putative ATPase nucleosome-remodeling enzymes. Domino-A is more restricted in expression than domino-B, and domino-B is likely the major domino product expressed in wing disks (47). Previous studies indicated that a domino mutation increased the severity of the N ^nd-1 mutant phenotype, suggesting that domino-B participates in Notch signaling at the wing margin, and cultured cell transfection experiments with mammalian SRCAP suggest that it supports transcriptional activation by Notch (16).

Figure 8 shows that, similar to the previously reported effect of the dom ⁷ mutation, the dom ¹ mutation increased the severity of the N ^nd-1 mutant phenotype. The dom ¹ mutation had only a slight effect on the length of the N ^Ax-E2 wing veins but strongly increased the number of wing margin nicks displayed by heterozygous mam ^g2. The dom ¹ mutation is a recessive lethal P element insertion allele, and these effects were measured relative to a viable revertant, dom ^R12, generated by excision of the P insertion. We conclude, therefore, that domino supports Notch and mastermind activities in the developing wing margin and has little effect in the wing veins. In this respect, dom ¹ behaves similarly to the Nipped-A ^NC106 mutation (Fig. 4). Because the effects of the dom ¹ mutation are similar to those observed with a specific Nipped-A mutant allele, we postulate that domino functions as a component of the Tip60 complex in Notch signaling, although we cannot rule out the possibility that it can also act independently of Tip60.

• Open in new tab
• Download powerpoint

FIG. 8.

Effects of domino mutations on Notch and mastermind mutant phenotypes. Dominant effects of the recessive lethal dom ¹ P insertion allele of domino (44) and a viable revertant (dom ^R12) generated for this work on the mam ^g2, N ^nd-1, and N ^Ax-E2 phenotypes were measured as described in the legend to Fig. 1, except that the wing widths and vein lengths are presented in micrometers instead of relative values.

Nipped-A is required for binding of mastermind to chromosomes.As a test of the possibility that the effect of Nipped-A on mastermind activity is direct, we looked to see if they bind to the same sites on salivary gland polytene chromosomes. The NipA12 antisera generated in a guinea pig and the NipA14 antisera generated in a rabbit show virtually identical staining patterns, confirming that the staining is specific (Fig. 9). The secondary antibodies do not show species cross-reactivity.

• Open in new tab
• Download powerpoint

FIG. 9.

Nipped-A and mastermind bind to the same sites on salivary gland chromosomes. The top panels show double immunostaining with the two Nipped-A antibodies, NipA12 and NipA14. The staining patterns are virtually identical, providing evidence additional to the Western blot data (Fig. 5) and the tissue immunostaining (Fig. 6) that the antibodies are specific for Nipped-A. The bottom panels show double immunostaining for Nipped-A and mastermind. As seen in the merge, virtually all sites that stain for mastermind also stain for Nipped-A, and there are several minor Nipped-A staining sites that that do not stain for mastermind. Arrowheads indicate puffs that stain for both Nipped-A and mastermind.

Both Nipped-A antisera show considerable overlap with mastermind staining (Fig. 9). Virtually all mastermind binding sites also bind Nipped-A, but there are many minor Nipped-A sites that do not show significant staining for mastermind. The secondary antibodies do not show cross-species reactivity. Although the functional target genes of Nipped-A and the mastermind target genes in salivary glands are unknown, both proteins associate with some chromosomal puffs, which are highly transcribed (Fig. 9, arrows).

The survival of flies hemizygous for the Nipped-A ^NC96 or Nipped-A ^NC186 allele to the third-instar larval stage of development provided an opportunity to examine the effects of Nipped-A on the binding of SAGA and mastermind to polytene chromosomes. Relative to wild-type control chromosomes, hemizygous Nipped-A ^NC96 mutant polytene chromosomes displayed significantly reduced immunostaining for the Gcn5 HAT component of SAGA and dramatically reduced staining for mastermind (Fig. 10). These effects were observed in multiple nuclei in multiple salivary glands and indicate that Nipped-A is required for efficient binding of mastermind to chromosomes. The reduced association of Gcn5 confirms that Nipped-A, as expected, also affects binding of the SAGA complex. We obtained similar results with Gcn5 and mastermind immunostaining with flies hemizygous for Nipped-A ^NC186 (not shown). We conclude that Nipped-A (Tra1/TRRAP) regulates the association of mastermind with chromosomes. These effects are unlikely to be general effects on chromosome structure. The polytene chromosomes of the Nipped-A mutants are normal in structure and contain puffs indicating that they are transcriptionally active. Moreover, a significant fraction of the Nipped-A ^NC96 hemizygous mutants live beyond the third-instar stage and pupariate (Table 2), indicating that there is unlikely to be a generalized defect in chromosome structure and function.

• Open in new tab
• Download powerpoint

FIG. 10.

The Nipped-A ^NC96 hypomorphic mutation reduces binding of mastermind and Gcn5 to chromosomes. The panels show representative wild-type (Oregon R) and hemizygous Nipped-A ^NC96 mutant (Nipped-A ^NC96/Nipped-E ⁴³) third-instar salivary gland polytene chromosome squashes immunostained for either the mastermind or the Gcn5 protein. For each antibody, the wild-type and mutant glands were immunostained under identical conditions. The digital micrographs were taken with the same illumination and exposure times. Examination of several nuclei in multiple squashes revealed that Gcn5 staining intensity is reproducibly reduced but not abolished, while mastermind staining is severely reduced. Similar results were obtained with hemizygous Nipped-A ^NC186 mutant polytene chromosomes (not shown).

Effects of Nipped-A mutants encoding truncated proteins on Notch and mastermind function.An unexpected finding is that the Nipped-A ^NC105 allele, which encodes the N-terminal 2,048 residues of Tra1/TRRAP, suffices to replace one wild-type copy of Nipped-A to support Notch and mastermind function in vivo. This was unexpected because the protein encoded by Nipped-A ^NC105 lacks the ATM-PI3 kinase motif which, in mammalian cell culture experiments, is required for Tra1/TRRAP to associate with Gcn5 and Tip60 (41). One possible explanation is that the C terminus of the Nipped-A protein is not required for Notch and mastermind function and that the truncated protein can replace the full-length protein. Because we could only assay the effects of the Nipped-A mutations on Notch functions in wing development in the presence of a wild-type allele, it is also possible that a truncated protein somehow increases the activity of the remaining full-length Nipped-A protein. We were unable to detect the truncated protein in Western blot assays of extracts or by immunostaining, suggesting that if this is the case, only a small amount of the mutant protein is sufficient. We consider it improbable that linked second-site mutations are masking effects of Nipped-A ^NC105 on both Notch mutant phenotypes and the mastermind phenotype. Many mutations have effects similar to Nipped-A, and few have opposing effects, and it would likely require multiple mutations to counteract the effects of Nipped-A ^NC105 on all three phenotypes. It is also unlikely that there is a linked second-site mutation that counteracts the effects of Nipped-A ^NC105 by increasing the expression of wild-type Nipped-A, because mutant embryos and larvae show the expected decrease in full-length Nipped-A protein.

The Nipped-A ^NC194 allele, which encodes residues 1 to 1500, had a significant effect on both of the Notch mutant phenotypes but did not increase the severity of the wing-nicking phenotype displayed by mam ^g2. Again, this differs from null alleles of Nipped-A, which affect all three phenotypes, suggesting that Nipped-A ^NC194 retains sufficient activity to replace one copy of the wild type in support of mastermind activity. Again, one possible explanation is that Nipped-A residues 1 to 1500 are sufficient to support mastermind function, although it is conceivable that the truncated protein somehow increases the activity of the remaining wild-type Nipped-A protein. We were also unable to detect this truncated protein, suggesting that if a truncated protein is responsible, only low levels are required. Despite extensive screens with a deficiency collection and candidate genes, no mutations that suppress mastermind mutant phenotypes have been mapped to chromosome 2 (22, 53). Thus, it is unlikely that a linked second-site mutation masks an effect of Nipped-A ^NC194 on the mastermind phenotype. Similar to Nipped-A ^NC105, heterozygous Nipped-A ^NC194 mutants display the expected reduced levels of full-length protein, although we cannot exclude the possibility of a subtle increase in the expression of the wild-type Nipped-A allele that is sufficient to rescue the mastermind phenotype but not the Notch mutant phenotypes.

Isolation and analysis of additional Nipped-A truncation alleles and development of more sensitive biochemical assays will lead to a fuller understanding of how Nipped-A alleles encoding truncated proteins support Notch signaling.

The roles of Nipped-A (Tra1/TRRAP) in Notch signaling involve both the SAGA and Tip60 complexes.The experiments presented here indicate that the roles of Nipped-A in supporting mastermind function likely involve both the SAGA and Tip60 complexes. The Ada2b protein is specific to SAGA (29), and Ada2b mutations affect the mastermind phenotype but not the two Notch mutant phenotypes. We think it unlikely that the effect of the Ada2b mutations is more specific than Nipped-A mutations because the mastermind phenotype is more sensitive. As shown by the Nipped-A ^NC96 hypomorph, the N ^nd-1 phenotype is more sensitive to the Nipped-A dosage than is mastermind. Moreover, the Nipped-A ^NC194 allele has the opposite specificity to the Ada2b mutations and affects the Notch mutant phenotypes but not the mastermind phenotype. Combined, the contrasts in the effects of Ada2b and various Nipped-A mutations show that Nipped-A and its complexes play multiple roles in Notch signaling. They suggest that the SAGA complex, or at least the Ada2b subunit, is more specific for mastermind function and that Nipped-A has additional functions.

Another possibility raised by the specificity of the effects of Ada2b mutations for effects on mastermind activity in wing margin development is that mastermind may have functions in margin development independent of Notch. For example, mastermind could conceivably function as a coactivator for other activator proteins in addition to Notch. This possibility is consistent with the binding of mastermind to several sites in polytene chromosomes, including the ecdysone-dependent puffs.

The domino protein, a putative ATPase remodeling enzyme, is a subunit of the Tip60 complex (28). The N ^nd-1 and N ^Ax-E2 phenotypes and the mastermind phenotype are modified by domino mutations, although the effect on N ^Ax-E2 is modest. These effects are similar to those of the Nipped-A ^NC106 allele and thus suggest that that the Tip60 complex also supports mastermind function and Notch signaling during wing development. It is possible, however, that domino functions independently of Tip60 and Nipped-A because the human domino homologue SRCAP interacts directly with the CBP HAT enzyme (24) that interacts with mastermind (18). Nevertheless, the likely involvement of the Tip60 complex raises the possibility that histone exchange could facilitate transcriptional activation by Notch because, in addition to acetylating histone H4, Tip60 exchanges histone H2 variants during DNA repair (28).

Is the effect of Nipped-A on mastermind function direct?As revealed by immunostaining of salivary gland polytene chromosomes, at least one function of Nipped-A is to regulate the binding of mastermind to chromosomes. The reduction in binding of mastermind to polytene chromosomes caused by the hypomorphic Nipped-A ^NC96 and Nipped-A ^NC186 alleles is dramatic. Supporting the idea that Nipped-A directly regulates mastermind binding, virtually all sites on polytene chromosomes that bind mastermind also bind Nipped-A. We envision a few possible explanations for these results. The SAGA and Tip60 complexes that contain Nipped-A could acetylate mastermind, proteins in the Notch activator complex, and/or possibly histones to facilitate binding of the Notch activator complex to chromatin. These modifications could be made by Gcn5 and/or Tip60, which acetylate histones H3 and H4, respectively. Alternatively, Nipped-A or its complexes could bind to chromosomes cooperatively with mastermind. This would be consistent with the published observation that the ankyrin repeats of the NICD fragment of Notch, which help recruit mastermind to the Notch activator complex, are also required to recruit Gcn5/PCAF SAGA subunit in transfected mouse cells (27). Both the Ada2b component of SAGA and the domino subunit of Tip60 affect mastermind function, so it is likely that Nipped-A supports mastermind function in more than one way.

Because the evidence suggests that Nipped-A supports mastermind function through both the SAGA and Tip60 chromatin-modifying complexes, we theorize that, in addition to controlling the binding of mastermind to chromosomes, Nipped-A could also cooperate with mastermind to recruit these complexes to facilitate transcriptional activation through chromatin modification.

Other possible functions for Nipped-A in Notch signaling.Our data indicate that the SAGA complex, or at least its Ada2b subunit, is not required for some functions of Nipped-A in Notch signaling. Unlike Nipped-A and domino mutations, Ada2b mutations did not affect Notch mutant phenotypes, while they did enhance the phenotype caused by a mastermind mutation. We postulate, therefore, that the Tip60 complex is also required for functions of Nipped-A beyond controlling the binding of mastermind to chromosomes. The Tip60 complex could affect the expression of Notch activator complex components, or it could modify proteins in the Notch activator complex. It is also possible that Tip60 modifies chromatin to either aid binding of the Su(H) protein to the Notch target genes or, as mentioned above, to aid transcriptional activation by the Notch activator complex. In any case, the evidence indicates that two subunits of Tip60, Nipped-A and domino, play more than one role in Notch signaling during wing development.