* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download TGF-β/BMP superfamily members, Gbb-60A and Dpp
Survey
Document related concepts
Transcript
2723 Development 125, 2723-2734 (1998) Printed in Great Britain © The Company of Biologists Limited 1998 DEV5189 TGF-β/BMP superfamily members, Gbb-60A and Dpp, cooperate to provide pattern information and establish cell identity in the Drosophila wing Ongkar Khalsa*, Jung-won Yoon*, Sonia Torres-Schumann and Kristi A. Wharton† Department of Molecular Biology, Cell Biology and Biochemistry, Brown University, Providence, RI 02912, USA *These authors have contributed equally †Author for correspondence (e-mail: [email protected]) Accepted 24 April; published on WWW 23 June 1998 SUMMARY Within a developing organism, cells receive many signals which control their proliferation, fate specification and differentiation. One group of such proteins is the TGFβ/BMP class of related signaling molecules. Based on expression studies, multiple members of this class of ligands must impinge upon the same cells of a developing tissue; however, the role that multiple TGF-β/BMP ligands may play in directing the development of such a tissue is not understood. Here we provide evidence that multiple BMPs are required for growth and patterning of the Drosophila wing. The Drosophila BMP gene, gbb-60A, exhibits a requirement in wing morphogenesis distinct from that shown previously for dpp, a well-characterized Drosophila BMP member. gbb-60A mutants exhibit a loss of pattern elements from the wing, particularly those derived from cells in the posterior compartment, consistent with the gbb-60A RNA and protein expression pattern. Based on genetic analysis and expression studies, we conclude that Gbb-60A must signal primarily as a homodimer to provide patterning information in the wing imaginal disc. We demonstrate that gbb-60A and dpp genetically interact and that specific aspects of this interaction are synergistic while others are antagonistic. We propose that the positional information received by a cell at a particular location within the wing imaginal disc depends on the balance of Dpp to Gbb-60A signaling. Furthermore, the critical ratio of Gbb-60A to Dpp signaling appears to be mediated by both Tkv and Sax type I receptors. INTRODUCTION monomers. The diversity of biological processes regulated by TGF-β signaling has often been attributed to the multitude of molecular interactions possible in this signaling pathway, including the potential for different monomer combinations, various associations of multiple type I to type II receptors (Yamashita et al., 1994), and a variety of interactions between an increasing number of intracellular proteins, eg. the Smad proteins (Derynck and Zhang, 1996; Massagué et al., 1997). An additional level of complexity that must be considered is that in many developing tissues multiple TGF-β signals must impinge upon a single cell and possibly signal through the same pathway. It has been observed that a number of members exhibit overlapping expression patterns (e.g. Dudley and Robertson, 1997; Lyons et al., 1995) and the question arises as to whether a cell can distinguish between different TGF-β signals and, if so, how is that information interpreted. During mouse development some degree of functional interchangability has been proposed for different BMP proteins, the largest group of TGF-β superfamily members (Dudley and Robertson, 1997). However, the generality of functional interchangability in other systems has not been determined. The possibility remains that more than one TGFβ signal contributes to a single developmental process and, Intercellular signaling is essential for the proper manifestation of cell growth, fate specification and differentiation during the development of multicellular organisms. The transforming growth factor-β (TGF-β) superfamily of signaling proteins are secreted as dimers and mediate intercellular communication through their association with heteromeric integral membrane receptor complexes consisting of type I and type II serine/threonine kinase receptor proteins (Massagué, 1996). Ligand binding sets off a cascade of biochemical events, which culminates in a change to the target cell ranging from alterations in cell proliferation to the acquisition of specific cell fates (reviewed in Massagué, 1996). Given the plethora of data substantiating the importance of these multifunctional molecules in regulating diverse developmental events in a number of different species (Roberts and Sporn, 1990; Kingsley, 1994a; Hogan, 1996), elucidation of the mechanisms by which TGF-β signaling proceeds and how cells interpret and distinguish between different signals is crucial to our understanding of development. The mature dimer or ligand can exist as a heterodimer as well as a homodimer, consisting of two different TGF-β-related Key words: gbb-60A, dpp, decapentaplegic, TGF-β, BMP, Drosophila, Wing morphogenesis, Imaginal disc, Patterning, Cell proliferation 2724 O. Khalsa and others furthermore, that multiple ligands bind to receptors on the same cell in some way eliciting a different response in that cell. Here we present evidence indicating that Gbb-60A and Dpp, two Drosophila BMP members, are both required for proper wing morphogenesis and that they cooperate to direct normal development. Three BMPs have been identified in Drosophila: Dpp, 60A and Scw (Padgett et al., 1987; Wharton et al., 1991; Doctor et al., 1992; Arora et al., 1994). Dpp and 60A each define a separate subfamily exhibiting greatest sequence similarity to vertebrate BMP2 and BMP4, and BMP5, BMP6, BMP7 and BMP8, respectively (Wharton et al., 1991; Kingsley, 1994a; Hogan, 1996). Dpp has been well-characterized and is known to act as a morphogen during patterning events that take place at multiple stages of development including embryogenesis and imaginal disc development. Cells have been shown to respond to the Dpp signal in a graded manner during the specification of dorsal embryonic cell fates (Ferguson and Anderson, 1992; Wharton et al., 1993) and in the specification of positional values along the anteroposterior (A/P) axis of the future adult wing (Lecuit et al., 1996; Nellen et al., 1996). The precursor cells of imaginal structures are set aside during embryogenesis as separate groups, or discs, of cells that actively proliferate during the larval period (Cohen, 1993). Patterning of the imaginal discs occurs during larval development and differentiation of imaginal structures takes place during the pupal stage. In the wing imaginal disc, dpp RNA is expressed in a localized pattern within a stripe of cells lying just anterior to the anterior/posterior (A/P) compartment boundary (Gelbart, 1989; Masucci et al., 1990; Posakony et al., 1991). Support for the proposal that cells adopt different fates within the wing disc in response to varied levels of a gradient of Dpp signal emanating from the A/P boundary has been provided by analysis of receptor mutants and downstream target genes (Burke and Basler, 1996; Lecuit et al., 1996; Nellen et al., 1996; Singer et al., 1997). In Drosophila, the BMP type I receptors encoded by thick veins (tkv) and saxophone (sax) have been shown to mediate various aspects of Dpp signaling (Affolter et al., 1994; Brummel et al., 1994; Nellen et al., 1994; Penton et al., 1994; Xie et al., 1994). Interestingly, the mutant phenotypes associated with mutations in dpp and the receptor genes, sax and tkv are not equivalent (Brummel et al., 1994; Twombly et al., 1996; Singer et al., 1997), as would be expected if a receptor specifically transmitted a signal unique to that ligand. These discrepancies could be explained by the possibility that Sax and Tkv receptors are activated by a TGF-β family member other than Dpp. A loss of receptor function could then reflect the loss in signaling triggered by more than one ligand, either as a heterodimer or as two homodimers that each bind to the same receptor. Therefore, the function of a particular TGF-β ligand cannot be determined based solely on the examination of receptor function and/or activation of downstream elements, as such results could incorrectly attribute functions to that ligand alone, as well as confound the actual molecular mechanism of signal transduction. As yet it has not been definitively demonstrated that Sax and Tkv receptors can mediate other signals such as those elicited by Scw or 60A. However, mutant analysis and the coexpression of dpp and scw in the Drosophila embryo has led to the proposal that Scw signals as a Dpp/Scw heterodimer with a higher ‘activity’ than the Dpp homodimer (Arora et al., 1994; Raftery et al., 1995) but data indicating through which receptor such a signal is mediated has not been reported. In this paper, we provide evidence that 60A (henceforth referred to as gbb-60A) is required for the growth of imaginal tissues and for patterning of the adult wing. We examine the expression of gbb-60A in the imaginal discs and show that high levels of Gbb-60A protein expression is in a pattern complementary to dpp expression. We demonstrate that Gbb60A and Dpp signal together to accomplish the wild-type pattern of the adult Drosophila wing. The Tkv and Sax receptors are implicated as mediators of this synergistic signaling. Furthermore, several pieces of evidence suggest that Tkv may directly mediate the Gbb-60A signal. These data provide the first functional evidence that concerted signaling from two different TGF-β/BMP ligands is required in the control of a specific developmental process. MATERIALS AND METHODS Drosophila melanogaster strains and crosses Mutant alleles used in this study are described in Flybase (1996). gbb60A1 is a null allele and gbb-60A4 is a hypomorphic allele (K. A. W., K. deCastro, E. Borod and J. Cook, unpublished data). Recombinant lines were generated that removed a secondary mutation present on the original gbb-60A4 chromosome. All lines behave similarly; line 10 (gbb-60A4 R10) was used in experiments presented here. All recombinants between alleles of dpp, sax, tkv and gbb-60A were tested for the presence of each allele independently. All crosses were done on a cornmeal-sugar media at 25˚C. Percent viability was determined by dividing the number of individuals in the Cy+ class by half of those in the Cy balancer class and multiplying by 100. Wings and legs were mounted in DPX mountant (EM Sciences). Generation of gbb-60A1 mutant clones Clones of genetically marked cells were made using the FLP/FRT system (Golic and Lindquist, 1989; Xu and Harrison, 1994). gbb-60A null clones were generated in the progeny of a cross of: yw1118/Y; P{>whs>}G13 sha1 gbb-60A1/SM6a males to yw1118 P{hsFLP}; P{>whs>}G13 P{y+} sha+ gbb-60A+ females. Adults were transferred every 12 hours. FLP recombinase was induced by heat shock (37˚C) for 50 minutes at 36 hours, 48 hours and 60 hours after egg lay (AEL). Clones were identified based on the shavenoid (sha) recessive marker, sha1, which removes or reduces the tricomes on the wing blade (Lawrence et al., 1986). A total of 158 wings were examined. Construction of UAS-gbb-60A transgenic lines A 1.6 kb EcoRI fragment from cDNA16 which contains the entire gbb-60A coding region (Wharton et al., 1991) was cloned into the ApoI site of pUAST (Brand and Perrimon, 1993). Four transgenic lines were established by injection of UAS-gbb-60A and, when driven by GAL4 24B and hsGAL4 (Brand and Perrimon, 1993), each line was tested for its ability to rescue gbb-60A mutant phenotypes. RNA in situ hybridization and immunohistochemistry Whole-mount in situ hybridization to RNA in imaginal discs using digoxigenin-labeled RNA probes was done as described in Tautz and Pfeifle (1989) with the following modifications. Tissues were treated with proteinase K (50 µg/ml) for 3.5 minutes at room temperature and samples were hybridized O/N 55˚C. Sense and antisense RNA probes were made using cDNA16 (Wharton et al., 1991) as a template. Immunohistochemistry with polyclonal rabbit anti-60A (Doctor et al., 1992) and monoclonal mouse anti-β-galactosidase (Sigma) was done Gbb-60A to Dpp ratio critical for patterning 2725 by the following procedure: imaginal discs were fixed in 2% formaldehyde in PBS for 30 minutes and blocked in 0.2% saponin, 1% normal goat serum in PBS. Imaginal discs were incubated overnight at 4˚C with anti-60A (7.7:1000). In double labelling experiments, dpp expression was detected by mouse anti-βgalactosidase (1:1000) in H1-2 dpp-lacZ/+ (Blackman et al., 1991) larvae. Secondary antibodies were obtained from Jackson Immunological Laboratories (Cy3 and FITC) and used at a dilution of 1:500. Discs were mounted in 80% glycerol, 100 mM Tris (pH 8.0), 0.5% n-propyl galate. Signals were detected by epifluoresence using a Nikon FXA or by laser scanning microscopy on a Zeiss LSM410. Wild-type (Canton S) and gbb-60A mutant third instar larvae were dissected and stained with DAPI in order to visualize imaginal versus polyploid larval tissue. Everted larvae were fixed in a 3:1 ethanol:acetic acid. Samples were stained in 0.05% DAPI (in 0.1M Tris buffer, pH7.5) 10 minutes RT and mounted in 50% glycerol (in PBS). RESULTS Mutations in gbb-60A result in reduced imaginal discs In order to gain a full understanding of the function of the 60A gene during development and to investigate the potential for combined signaling from two different TGF-β/BMP family members, we had previously isolated mutations in the 60A gene (see Materials and Methods). Individuals homozygous for null alleles of the 60A gene, such as gbb-60A1, exhibit embryonic defects in gut morphogenesis and result in early larval lethality (K. A. W., K. deCastro, E. Borod and J. Cook, unpublished data). Allelic combinations between a hypomorphic 60A allele, gbb-60A4, and a null allele or a deficiency deleting the 60A locus result in later larval lethality with only rare adult escapers (<1%). Third instar larvae of such a genotype (gbb-60A4/gbb-60A1) appear transparent and smaller than wild-type larvae (Fig. 1A,B). These larvae develop more slowly than wild type and never attain wild-type size. The transparency appears to be due to a defect in both the quantity and quality of the fat body, as well as a dramatic reduction in imaginal tissues (Fig. 1C,D). In addition to a reduction in imaginal disc tissue, other tissues that normally proliferate during larval development, such as areas of the brain (Fig. 1D), are also reduced. The 60A locus has been named glass bottom boat (gbb) in light of the remarkable transparency of the mutant larvae and we will refer to it as gbb-60A in this paper. Mutations in gbb-60A affect imaginal disc derivatives The rare adult escapers of the genotype gbb-60A4/gbb-60A1 exhibit small misshapen wings, which lack the posterior cross vein (PCV), much of longitudinal vein L5, distal portions of L4 and the posterior half of the anterior cross vein (ACV) (Fig. 2C). The mutant wings are small, narrow and pointed with a loss of intervein material, especially between veins L2/L3, L4/L5 and posterior to L5. In addition to the change in wing shape and the loss of veins, a thinning of vein L2 and some ectopic vein material in the intervein region flanking L2 is common. No abnormalities are found along the margin in the position or type of wing margin bristles, however, ectopic margin bristles were often seen along a vein or in intervein tissue in the distal regions of the wing (not shown). Defects in Fig. 1. gbb-60A is required for proliferation of imaginal tissues. (A) Wild-type third instar larva. (B) gbb-60A1/ gbb-60A4 mutant larva. DAPI staining reveals a dramatic reduction in the size of imaginal discs in dissected gbb-60A mutant (D) as compared to wildtype (C) larvae. A reduction in the size of the gbb-60A1/ gbb-60A4 larval brain is also evident (*, ventral ganglia). Arrows indicate wing and eye/antennal imaginal discs. these mutant flies are not limited to the wings: a reduction in the size of the eye, with the presence of supernumerary vibrissae, an increase in the number of thoracic bristles, the presence of misshapen leg segments and female sterility have all been observed (S.T.-S. and K. A. W., unpublished data). Subsequent to our initial observations of the lethality and phenotype of the gbb-60A hypomorphic allele, gbb-60A4, we have removed by recombination a secondary mutation present on the gbb-60A4 chromosome (see Materials and Methods). This secondary mutation exacerbated only the lethality associated with the gbb-60A4 mutation and had no effect on adult phenotypes. Once the secondary mutation was removed, it was possible to produce viable adults that were homozygous for the gbb-60A4 mutation at a significantly higher frequency than that observed for the gbb-60A4/gbb-60A null genotype (see Fig. 5I). We have verified by sequence analysis that the viable adults are indeed homozygous for the gbb-60A4 mutation (data not shown). The mutant wing phenotype of the gbb-60A4 homozygous adult is similar to, but less severe than, that observed for the rarer gbb-60A4/Df or gbb-60A4/gbb-60A1 adults (Fig. 2B,C). In gbb-60A4 homozygotes, the overall shape of the wing is broader and less pointed, more like wild type. The ACV is often complete (54% of gbb-60A4 individuals versus 0% of gbb-60A4/gbb-60A1) and longitudinal veins L4 and L5 are longer. Given that gbb-60A4 retains some gbb-60A function, we examined the affect of complete loss of gbb-60A function on wing patterning and morphogenesis through the production of 2726 O. Khalsa and others Fig. 2. gbb-60A signaling is required for normal wing morphogenesis. (A) Wild-type wing. Longitudinal veins L2, L3, L4, L5, anterior and posterior crossveins (ACV and PCV, respectively) and the anterior/posterior (A/P) boundary are denoted. (B,C) Adult viable gbb-60A mutants exhibit a loss of both vein (indicated by arrows) and intervein material. (B) gbb-60A4 R10/ gbb-60A4. (C) gbb-60A4 R10/ gbb-60A1 results in greater loss of L5, L4 and the posterior portion of ACV. Some variability is seen in the degree of loss for L5: 43% have a longer L5 as shown in Fig. 6A. (D-I) Adult wings containing sha gbb-60A1clones. Clone boundaries on the dorsal wing surface (dotted line) and the ventral wing surface (dashed line) are demarcated. Large clones (D) or multiple clones (E) in the anterior and posterior compartments with significant dorsal/ventral overlap produce phenotypes similar to B and C, except distal gaps in L3 are often observed (see rightmost vertical arrow). (F) Ectopic vein material (arrow) is observed when a small dorsoventral clone is located anterior to L3. (G) An example of the non-autonomous effect resulting from gbb-60A clones. All wings are shown with anterior to the top and distal to the right. null clones. The FLP-FRT-mediated recombination system (Golic and Lindquist, 1989; Xu and Harrison, 1994) was used to generate clones of cells null for gbb-60A function in the imaginal discs (Fig. 2D-G). Consistent with a potential role for gbb-60A in proliferation, gbb-60A null clones are never recovered in the adult wing if induced during the early period of cell proliferation (before 36 hours AEL); however, clones in all parts of the wing were recovered from later inductions. Clones located in both the anterior and posterior compartments, as well as clones limited to the posterior compartment, exhibited defects in the same pattern elements affected in hypomorphic wings: loss of the PCV, and portions of L5 and L4 (Fig. 2D,E). Clones often also resulted in the loss of distal L3 at the margin. The majority (>80%) of the clones that exhibited a mutant phenotype covered a portion of the wing where vein material would normally be found. Clones strictly limited to the anterior compartment or along the A/P boundary exhibited no wing defects with the exception of very small clones that were located near the ACV (7% of the total clones restricted to the anterior compartment). These clones resulted in ectopic vein material anterior to L3 (Fig. 2F). In several wings vein loss was observed in wild-type tissue at a distance from the gbb-60A− patch (Fig. 2G). In each case that this non-autonomous effect was found, multiple clones were present in the wing, one in the posterior compartment and another along the A/P boundary, sites that otherwise did not produce abnormalities. This type of vein loss was never observed in the generation of control clones. Expression of gbb-60A in the imaginal discs is complementary to that of dpp The expression of gbb-60A in imaginal discs was examined by both whole-mount RNA in situ hybridizations (Fig. 3A,C,E) and antibody staining using a Gbb-60A antibody (Fig. 3B,D,F). gbb-60A RNA is expressed in the wing disc (Fig. 3A) mainly in the posterior compartment in the pteropleural and medial regions extending into the progenitors of the scutellum. High levels are found within the wing pouch in cells belonging to both the posterior and anterior compartments with higher levels in the posterior. A small amount of expression is found within the hinge region. In the eye/antennal disc, gbb-60A RNA is highest anterior to the morphogenetic furrow and in the medial regions with lower levels of expression posterior to the morphogenetic furrow (Fig. 3C). gbb-60A is expressed throughout the posterior compartment of the leg imaginal discs and within the ventral anterior compartment (Fig. 3E). Gbb-60A protein is generally detected (Fig. 3B,D,F) at locations coincident with gbb-60A RNA with one major exception: the presence of a stripe of markedly lower Gbb-60A protein running through the middle of the wing pouch (Fig. 3B). This ‘stripe’ of reduced expression is coincident with the prominent stripe of dpp RNA expression in the cells anterior to the anterior/posterior (A/P) boundary of the wing imaginal disc (Fig. 4A-C). We also observe a weak but consistent reduction in Gbb-60A expression along the dorsal/ventral boundary from which the cells of the wing margin are derived (Fig. 4A). High levels of Gbb-60A expression exists in a pattern complementary to the localized expression of dpp in the other imaginal discs as well. For example, Gbb-60A expression is absent or at very low levels in the morphogenetic furrow of the eye disc, the site of dpp expression (Figs 3, 4DF). In addition, the regions of low or absent Gbb-60A expression in the antennal and leg imaginal discs are the sites of dpp expression (Fig. 4D-F; data not shown). It seems Gbb-60A to Dpp ratio critical for patterning 2727 therefore appropriate that the abbreviation of glass bottom boat (gbb) is a mirror image of the letters dpp, thus reflecting the observed complementary pattern of gbb-60A and dpp expression. We do not intend to suggest that gbb-60A and dpp expressing cells are mutually exclusive, especially since the reagents available to assay for the presence of these BMP ligands do not allow such resolution. However, based on the reagents available our data indicate that the predominant RNA and protein expression of gbb-60A and dpp appear to be in different regions of the imaginal discs with some cells in the overlapping areas in which the co-expression of gbb-60A and dpp may occur. While in most discs the expression of gbb-60A RNA and protein does not differ significantly, we cannot exclude the possibility that in the wing disc the apparent absence of Gbb60A along the A/P boundary reflects the inability of a Gbb60A polyclonal antibody to recognize the Gbb-60A protein in these cells perhaps due to some particular structural conformation of Gbb-60A, such as a heterodimer, in these cells. We believe this is very unlikely since a second polyclonal antibody directed at a different portion of the protein, the amino-terminal pro domain as opposed to the ligand domain (Doctor et al., 1992), also reveals the same reduced staining along the A/P boundary (data not shown). Fig. 3. gbb-60A is expressed in a discrete pattern in imaginal discs. RNA in situ hybridizations performed on imaginal discs from third instar larvae using a gbb-60A antisense probe are shown (A,C,E). In general, Gbb-60A protein expression (B,D,F) follows the RNA pattern except that it appears more diffuse. A stripe (bracket) of reduced or absent Gbb-60A protein is apparent anterior to the anterior/posterior boundary of the wing disc (B). Wing and leg imaginal discs (A,B,E,F) are shown dorsal up and posterior to the right. Eye/antennal discs are shown lateral to the left and posterior down (C,D). The location of the morphogenetic furrow is indicated by an arrowhead (C,D). Fig. 4. High levels of Gbb-60A are complementary to the expression of dpp. Confocal images of a wing disc (A-C) and epifluoresence images of an eye/antennal disc (D-F) demonstrate that low levels of Gbb-60A protein (red) are coincident with the localized expression of dpp RNA (green) along the A/P boundary (bracket) and in the morphogenetic furrow (arrowhead). Arrow marks the dorsal/ventral boundary in (A). (A-C) Wing discs: dorsal up and posterior right. (D-F) Eye/antennal discs: lateral left and posterior down. Dpp and Gbb-60A signaling are required together for normal wing morphogenesis The structures in the wing that are affected most dramatically by mutations in gbb-60A, the PCV and L5 are those that are least sensitive to the reduction or absence of dpp (Posakony et al., 1991; deCelis, 1997; Sturtevant et al., 1997). The ACV and longitudinal veins L2 and L4, lying on either side of the A/P boundary, have been shown to be most sensitive to the loss of Dpp signaling (Segal and Gelbart, 1985; Posakony et al., 1991; deCelis, 1997) consistent with the proposal that Dpp organizes wing pattern via a morphogen gradient emanating from the A/P boundary. (It has been proposed that longitudinal vein L3 is fated by a different mechanism and as a result is not sensitive to the level of Dpp signaling (Sturtevant, 1997).) 2728 O. Khalsa and others I 80 , % viability 60 40 20 0 + gbb4 gbb4 dppd12 gbb4 + dppd5 gbb4 gbb4 gbb4 (763) gbb4 + gbb4 dppd12 gbb1 + (718) dppd5 gbb1 gbb4 gbb1 gbb1 (>1000) Df (536) (537) (646) (106) Fig. 5. gbb-60A and dpp interact synergistically. (A-D) Genetic combinations involving the dppd alleles, dppd5 and dppd12, with gbb-60A4 homozygotes or the heteroallelic gbb-60A4/ gbb-60A1 result in enhanced wing phenotypes. (A) dppd5 gbb-60A4/ + gbb-60A4R10. (B) dppd12 gbb60A4 / + gbb-60A4 R10. The complete loss of the ACV is indicated by an arrow. (C) dppd5gbb-60A1/ + gbb-60A4 R10. (D) dppd12 gbb-60A1/ + gbb60A4 R10. (E) Wild-type male prothoracic leg. (F) Male prothoracic leg from dppd12 gbb-60A1/ + gbb-60A4 R10. The five tarsal segments (bracket) in the wild-type male prothoracic leg are indicated (E). Loss of the ACV (arrow) in the dpp heteroallelic combination dpps4/ dppd6 (G) is suppressed by gbb-60A null allele, dpps4 + / dppd6 gbb-60A1 (H). All wings were taken from females and photographed at the same magnification. (I) The percentage viability of dppd gbb-60A/ + gbb-60A genotypes is indicated. gbb-60A is abbreviated gbb. The total number of individuals scored in the viaibility crosses is shown in parenthesis. Since mutations in the gbb-60A locus indicate that gbb-60A is essential for establishing cell identity in the wing, especially within the posterior compartment, we investigated the possibility that Gbb-60A and Dpp signal together to provide positional information for the entire wing. No dominant genetic interactions were observed between alleles of gbb-60A and dpp, therefore, recombinant chromosomes were constructed using alleles of dpp which result in an overall lowering of dpp expression in the imaginal discs (dppd5, dppd6 and dppd12) (Masucci et al., 1990), and gbb-60A alleles, gbb60A4 and gbb-60A1. Individuals heterozygous for dppd alleles (dppd/+) are phenotypically wild type. As homozygotes or in various heteroallelic combinations, these dpp alleles can be lethal or can generate adults with appendage defects ranging from minor loss of vein material to truncations of the entire appendage (Spencer et al., 1982). An individual heterozygous for a dppd allele and homozygous or transheterozygous for gbb-60A alleles (dppd gbb-60A4/ + gbb-60A4 or dppd gbb-60A4/ + gbb-60A1) exhibits wing phenotypes qualitatively different than that observed in the gbb-60A mutant alone (Fig. 5). gbb-60A mutants heterozygous for dppd5 exhibit a significant loss of L4 (>60% lack more than half of L4) and a more frequent loss of the ACV. gbb-60A mutants heterozygous for a more severe dpp allele, dppd12, not only show a greater loss of L4, ACV and L2 vein material (>90% lack more than half of L4 and 100% lack the entire ACV including a thinning of L2) but also a loss of intervein tissue, especially between L2/L3 and L4/L5 as exhibited by a reduction in the overall size of the wing (Fig. 5B,D). In a small number of individuals (3%), a gap at the distal end of L3 is apparent. Interestingly, in the dpp gbb-60A/ + gbb-60A genotypes, we do not observe a greater loss of L5 but in fact a reduction in the loss of this posteriormost vein. 100% of dppd12 gbb-60A4/ + gbb-60A1 exhibit only a gap in Gbb-60A to Dpp ratio critical for patterning 2729 L5 with >K of the vein being complete, while 100% of gbb60A4/gbb-60A1 lack >K of L5 with >60% lacking >G of L5. These phenotypes are not only observed with dpp regulatory mutants (dppd alleles) but also with alleles that affect activity of the dpp protein (Wharton et al., 1993). For example, dpphr56 gbb4/ + gbb4 exhibits wing phenotypes similar to dppd12 gbb60A4/ + gbb-60A4. In addition to defects in wing patterning, abnormalities were noted in the proximal/distal organization of the legs (Fig. 5F), a process in which dpp is known to play a central role (Jiang and Struhl, 1996; Lecuit and Cohen, 1997). The common defects exhibited by the dppd12 gbb-60A1/ + gbb-60A4 genotype are truncations and/or apparent fusions of the distal most tarsal segments of the male prothoracic leg (Fig. 5E,F). A decrease in the viability of dpp gbb-60A/ + gbb-60A individuals as compared to that observed for gbb-60A mutations alone was also observed (Fig. 5I). The adult viability of gbb-60A4 homozygotes decreases from 74.5% to 10.1% of the expected size class with the inclusion of a single mutant allele of dpp (dppd12 gbb-60A4/ + gbb-60A4). These data indicate that heteroallelic combinations of gbb60A and dpp result in phenotypes more pronounced and distinct from phenotypes observed as a result of mutations in either gene alone. These new phenotypes are not simply additive. Individual pattern elements are affected differently as a result of these heteroallelic combinations, for example, a greater loss of L4 is seen while the loss of L5 is reduced. In an attempt to test further the apparent synergistic, as well as antagonistic, nature of the gbb-60A/dpp interaction, we examined several other allelic combinations. In one case, we investigated the effect of lowering gbb-60A function in a viable dpp mutant. Wings derived from dpps4/dppd6 individuals lack only the ACV (Fig. 5G) presumably due to the fact that, in this mutant, dpp expression is absent specifically from the primordial cells for that wing vein (deCelis, 1997). In this genotype lowering the dose of gbb-60A by one copy (dpps4 +/dppd6 gbb-60A1) results in a restoration of the ACV (Fig. 5H) not a further loss of other vein material as seen above with other dpp/gbb-60A combinations (Fig. 5A-D). However, similar to the other dpp/gbb-60A interactions, the size of these wings is reduced and the viability of these individuals is significantly reduced (15% of expected as compared to control crosses). These results suggest that the requirement for both dpp and gbb-60A in earlier stages of larval development, presumably at times of high cell proliferation, is not met in the dpps4 +/dppd6 gbb-60A1 genotype. Yet the suppression of the ACV defect indicates that gbb-60A does not solely act to modulate levels of Dpp signaling but rather as the ratio of gbb-60A to dpp changes across the wing imaginal disc the specification of a pattern element is affected. Different relative levels of Gbb-60A to Dpp signaling would result in different positional information. At some points within the developing wing, the readout may be synergistic while at other sites antagonistic. Correct balance of Gbb-60A and Dpp signaling is essential to normal patterning If cells across the wing disc indeed respond to different levels of Gbb-60A versus Dpp signaling, we reasoned that by changing the level of one ligand we should be able to generate a patterning mutant phenotype. We should then be able to potentially rescue that phenotype by changing the level of the Table 1. UAS-Gbb-60A rescues gbb-60A mutants UAS-gbb-60A GAL4 24B hsGAL4 Genotype % Rescue gbb-60A1/gbb-60A4 gbb-60A1/gbb-60A4 70.9 42.4 other ligand in an attempt to restore the balance of signaling elicited by the two ligands to that found in wild type. To test this hypothesis, we first made a UAS-Gbb-60A construct which contained the entire gbb-60A coding region and generated multiple transgenic lines. We have demonstrated that these UAS-Gbb-60A lines produce functional Gbb-60A since gbb-60A mutations can be rescued to adulthood when the expression of UAS-Gbb-60A is driven by hsGAL4 or by GAL4-24B (Table 1). GAL4-24B directs expression mainly in the embryonic mesoderm (Brand and Perrimon, 1993), the site of high levels of gbb-60A expression in the embryo (Doctor, 1992; K. A. W., unpublished data). In addition to the rescue of lethality, the mutant wing phenotype associated with gbb-60A viable mutants (Fig. 2) was rescued when hsGAL4 was used to direct expression. Using the same experimental conditions, we were unable to rescue gbb-60A mutants (lethality or wing phenotype) with previously reported UAS-60A lines (Staehling-Hampton et al., 1994). Since Gbb-60A is normally expressed at a very reduced level along the A/P boundary, we investigated the effect of overexpressing Gbb-60A in this region. Wings from UAS-Gbb60A/ptc-GAL4 exhibit ectopic vein material anterior to the A/P boundary (Fig. 6B). If dpp is also mis-expressed (UAS-gbb60A/UAS-dpp/ptc-GAL4) this mutant phenotype is rescued (Fig. 6C). The effect of raising Gbb-60A signaling along the A/P boundary where Dpp is normally very high results in mispatterning. If the original balance of higher Dpp to Gbb-60A along the A/P boundary is re-established by adding more Dpp the wild-type positional information is restored. When ectopically expressed, the UAS-dpp (gift from Mike O’Connor) used in these studies fails to produce any mutant phenotypes. Other examples of the effects of manipulating the relative levels of Gbb-60A and Dpp is shown when ectopic expression is more widespread using 71B-GAL4 (Fig. 6E,F) or A9-GAL4 (Fig. 6H,I). In each case some but not all of the defects in pattern elements caused by ectopic expression of gbb-60A are rescued by also expressing dpp. Given that our loss-of-function interaction data fail to demonstrate a simple antagonistic relationship between gbb-60A and dpp (indeed quite the opposite), we must conclude that these ectopic expressions reflect a change in the balance of dpp to gbb-60A activity or signaling. Which specific elements of the ectopic Gbb-60A mutant phenotype are rescued clearly depends on the pattern of ectopic expression during the entire period of wing patterning as well as the endogenous levels of Dpp and Gbb-60A. gbb-60A alleles genetically interact with mutations in BMP type I receptor genes, tkv and sax The Dpp signal has been shown to be mediated by two different BMP type I receptors, Tkv and Sax, during wing morphogenesis as well as during other stages of development (Xie et al., 1994; Brummel et al., 1994; Nellen et al., 1994; Penton et al., 1994; Twombly et al., 1996; Singer et al., 1997). To address whether either or both of these receptors are important in mediating the signals resulting from the actions of 2730 O. Khalsa and others Fig. 6. The relative levels of gbb-60A and dpp are critical to wing development. Expression pattern of UAS-lacZ directed by ptc-GAL4 (A), GAL4-71B (D) and GAL4-A9 (G) in the wing imaginal disc as assayed by X-gal staining. Wings exhibit dominant phenotypes when UAS-gbb60A is ectopically expressed using ptc-GAL4 (B), GAL4-71B (E) and GAL4-A9 (H). A decrease in intervein tissue between L3/L4 accompanied by ectopic vein material is seen when Gbb-60A is misexpressed along the A/P boundary (B, arrow). Wing blisters and ectopic vein material are apparent when UAS-gbb-60A is ectopically expressed in the wing pouch (E). Excessive vein formation is seen in both the anterior and posterior compartments of the wing (H). Aspects of the dominant phenotypes shown in (B,E,H) are rescued or suppressed by simultaneous expression of UAS-dpp (C,F,I). Gbb-60A and Dpp, we investigated the possibility of a genetic interaction between alleles of gbb-60A and alleles of tkv or sax. Recombinants were constructed between gbb-60A4 or gbb-60A1 and several alleles of tkv and sax. The addition of Df(2L)tkv2, a chromosomal deficiency that removes the tkv locus (Penton et al., 1994), into a gbb mutant background results in a severe mutant wing phenotype with a dramatic loss of both the PCV and ACV and most of L4 and L5 (Fig. 7A, B). In addition, distal gaps are present in L2 and L3. A less extreme phenotype is seen with tkv6, a hypomorphic allele that retains significant receptor function (data not shown). Interestingly, the more severe phenotype of tkv2 gbb-60A1/ + gbb-60A4 has qualities that are distinct from those observed in dpp gbb/ + gbb individuals, notably the effect on L5. Furthermore, unlike the genetic interactions observed with dpp alleles, the viability of tkv gbb60A1/ + gbb-60A4 is not reduced but in fact is considerably increased when compared to gbb-60A1/gbb-60A4 (Fig. 7E). These data indicate that the interaction that we observe between tkv and gbb-60A cannot be explained solely as a secondary consequence of lowering Dpp signaling readout by mutating a receptor that mediates Dpp signaling. These data suggest that Tkv is able to mediate Gbb-60A signaling and that it may do so in different ways at different times during development. We also analysed the effect of reducing the gbb-60A copy number in an individual compromised for functional Tkv receptor. tkv427 behaves as a hypomorphic allele. As a homozygote, tkv427 is phenotypically wild type and exhibits no defects in the wings; however, in trans to a null, Df(2L)tkv2, a slight thickening of wing veins is observed (Fig. 7C). This ectopic vein phenotype is thought to reflect a feedback loop in which dpp fails to be downregulated properly by tkv in the vein primordia resulting in the response of boundary intervein cells to the vein-promoting Dpp signal, and thus the formation of vein outside of the vein proper (deCelis, 1997). We observe that reducing gbb-60A in this tkv mutant background (tkv427 +/Df(2L)tkv2 gbb-60A1) produces a further thickening of wing veins (Fig. 7D). This result suggests that gbb-60A may play a role in vein differentiation itself and/or in the tkv/dpp feedback loop important in defining the boundaries of the vein. Genetic combinations used to investigate the potential interaction between gbb-60A and sax alleles indicate a reduction in viability for gbb-60A mutant genotypes containing a single copy of a sax null allele (sax gbb-60A1/ + gbb-60A4) (Fig. 7E). This reduction in viability suggests that lowering both gbb-60A and sax compromises development. The wing phenotype of the few viable adults recovered was similar to a very severe gbb-60A mutant wing phenotype shown in Fig. 2C, with a substantial loss of L5, complete loss of the PCV and ACV and loss of half of L4. Clearly the levels of Gbb-60A signaling and Sax function are dependent on one another. DISCUSSION gbb-60A is required for growth and patterning of imaginal tissues Our gbb-60A mutant analysis indicates that the gbb-60A gene is required for proper wing morphogenesis, a process requiring Gbb-60A to Dpp ratio critical for patterning 2731 E E 50 % viability 40 30 20 10 + gbb4 tkv2 gbb1 (407) + gbb4 (283) tkv6 gbb1 + gbb4 (618) sax5 gbb1 (718) + gbb4 (497) sax4 gbb1 gbb1 both growth and patterning of wing imaginal tissues. The requirement for gbb-60A in growth of imaginal discs is exemplified by the dramatic reduction of imaginal tissues observed in gbb-60A mutant larvae (Fig. 1D), which contributes to a transparent phenotype for which we have named the 60A gene glass bottom boat (gbb). Our mutant analysis further demonstrates that lowering gbb-60A function results in the reduction of the size of the adult wing (Fig. 2). The reduction in wing tissue is suggestive of a failure of imaginal cell proliferation. These results are similar to what has been found for mutations in dpp, as well as in a number of other genes that encode BMP signaling components in Drosophila (Spencer et al., 1982; Raftery et al., 1995; Brummel et al., 1994; Burke and Basler, 1996; Singer et al., 1997). These data taken together suggest that BMP signaling in Drosophila plays a role in the positive control of cell proliferation, opposite to the negative effect on cell growth attributed to TGF-β itself (Massagué et al., 1990; Roberts and Sporn, 1990). Studies in progress will delineate in which cells and at what times during development gbb-60A is required for proliferation, as well as how the signal is mediated. With the identification of a gbb-60A hypomorphic allele, we have been able to partially bypass the requirements for gbb60A in early stages of development which when not met lead to lethality. Thus, we have been able to elucidate a requirement for gbb-60A in wing morphogenesis (Fig. 2). The generation of clones null for gbb-60A function reaffirmed this requirement for gbb-60A in the specification of pattern elements within the wing. Our analysis of loss-of-function mutations demonstrates that the posterior portion of the wing is most sensitive to the level of gbb-60A function with the PCV and vein L5 most sensitive to a reduction of gbb-60A, followed by L4, the ACV, L3 and least sensitive, L2 (Fig. 2). In addition, the nonautonomous effect of gbb-60A− clones suggests that disrupting gbb-60A levels in one part of the disc must alter how cells in other regions of the disc interpret their position (Fig. 2G). 0 gbb4 Fig. 7. Alleles of gbb-60A genetically interact with tkv and sax receptor mutant alleles. (A) gbb60A1/ gbb-60A4. (B) Df(2L)tkv2gbb-60A1/gbb-60A4 results in a more severe wing phenotype but less severe lethality (E). Note gaps in L2 and L3 (arrows). (C) Slight thickening of veins (*) in tkv427/Df(2L)tkv2. (D) tkv427+/Df(2L)tkv2gbb-60A1results in an enhancement of vein thickenings (*). (E) Percentage viability of gbb-60A1/gbb-60A4 in conjunction with receptor mutant alleles. The number of individuals scored is shown in parenthesis. Complementary expression pattern of gbb-60A and dpp suggest these ligands exist primarily as homodimers The expression of gbb-60A RNA and protein are consistent with the requirement that we have identified for gbb-60A in the wing. The pattern of Gbb-60A protein expression in the wing imaginal disc as revealed by a Gbb-60A polyclonal antibody (Fig. 3B) is intriguing in light of the highly localized expression of dpp RNA in the cells just anterior to the A/P boundary of the wing imaginal disc (Gelbart, 1989; Masucci et al., 1990; Posakony et al., 1991). The highest levels of expression observed for these two BMPs not only complement one another in the wing imaginal disc but also in the leg and eye/antennal discs (Figs 3, 4). One implication of the mainly non-overlapping expression domains of gbb-60A and dpp is that these two BMPs may exist predominately as homodimers. Heterodimer formation would only take place in cells where Gbb-60A and Dpp were co-expressed and given the localized expression of dpp RNA this could only occur along the A/P boundary. Further evidence of a role for Gbb-60A as a homodimer is provided by the defects that we observe in gbb60A null clones in posterior cells, which do not express dpp and do not exhibit a mutant phenotype when null for dpp. The loss of pattern elements within these clones indicates that Gbb60A must act as a homodimer in these cells. Ratio of Gbb-60A to Dpp signaling is critical to normal wing morphogenesis An increasing number of TGF-β/BMP family members have been implicated as key players in directing a number of developmental processes including cell proliferation, patterning and differentiation. Given that multiple TGFβ/BMP members have been found to be expressed within the same developing tissue (see Hogan, 1996; Kingsley, 1994b), it is of interest to understand the specific function of each of these members as well as the mechanism by which they signal. Thus 2732 O. Khalsa and others far, we know very little about the role of more than one ligand in a tissue. In some cases the expression domains overlap, raising the possibility of heterodimer formation and therefore the potential for eliciting a different response than that acheived by the homodimer. Varied responses within a tissue may not only result from homo/heterodimer differences but also from multiple homodimers signaling in an additive or perhaps antagonistic manner. Furthermore, should the signaling from multiple homodimers be synergistic, other responses could be achieved. Empirical evidence to support this possibility has not previously been presented. Rather, in the mouse, it has been suggested that the loss of BMP7 can be compensated for by BMP2 which is expressed in the same tissue (Dudley and Robertson, 1997). In this paper, we provide data indicating that, in addition to the fact that Gbb-60A has a role in wing morphogenesis distinct from that of Dpp, Gbb60A and Dpp signal together to pattern the adult wing. Both genes are also required earlier in larval development and again the process appears to depend on the presence of both ligands. In order to investigate the potential interplay between Gbb60A and Dpp signaling, we examined the possiblity that alleles of both genes may genetically interact. The genetic interaction that we observed between dpp and gbb-60A resulted in a change in viability as well as a change in the presence or absence of pattern elements in the adult. Given the nature of the molecules encoded by these genes, the genetic interaction that we observe could be interpreted in several ways. One interpretation is that gbb-60A somehow modulates dpp activity such that in the presence of Gbb-60A the output of Dpp signaling is boosted. Given the dramatic morphogenetic organizing properties exhibited by dpp and thus far not observed for gbb-60A, we could speculate that the role of gbb60A may be to merely fine tune the effectiveness of Dpp signaling. A simplistic model to explain such a modulation of Dpp signaling could either be that a different response is attained by the Dpp/Gbb-60A heterodimer versus the Dpp homodimer. Alternatively, the two ligands could signal independently as homodimers and Gbb-60A signaling somehow facilitates Dpp signaling, in which case we would expect the functions of dpp and gbb-60A to be additive. Another interpretation is that for the most part Dpp and Gbb60A signal independently and where the signals overlap cells are able to differentiate between the relative levels of these signals, read the relative levels of these two ligands and respond accordingly. The important issue being that the response elicited is dependent on the balance of Dpp to Gbb60A signaling at that location in the developing tissue. This model would allow for both homodimer and heterodimer signaling as well as the complex response that we observe in the interactions between dpp and gbb-60A loss-of-function alleles, which exhibits aspects of both synergy and antagonism. Dpp and Gbb-60A signaling could be additive in some cells but more importantly a shift in the balance of these ligands would provide new or different positional information. Our data provide strongest support for the second model based on the following reasons. Genetic interactions between alleles of gbb-60A and dpp were identified by mutant wing phenotypes, which include the loss of pattern elements not normally absent in phenotypes produced by mutations in either gene alone or by the copy number of that mutation alone (Fig. 5). In addition to the more severe wing phenotype, we observed an increase in the degree of lethality of dpp gbb-60A/ + gbb60A genotypes (Fig. 5I) beyond what would be expected if their signaling were additive. The effect of dpp− on the percent viability observed for gbb-60A mutants is not limited to alleles of dpp that alter dpp transcriptional levels but is also apparent in genotypes where the reduction of Dpp protein activity is affected (S. S.-T., unpublished data). Therefore, lowering the level of Dpp signaling in a gbb-60A mutant background results in an exacerbation of the degree of lethality normally observed for either genotype alone. The qualitative changes in the wing phenotype of dpp/gbb-60A interacting genotypes is indicative of the fact that the role of gbb-60A is not merely to enhance the Dpp signal in an additive manner but that the relationship between Dpp and Gbb-60A signaling must be much more complex. If Dpp and Gbb-60A signaling were simply additive, we would predict that in a gbb-60A mutant the effective activity of Dpp would just be lowered and thus the first pattern element to be affected would be those that are normally most sensitive to the loss of dpp, such as the ACV, L2 and L4. This is not the case, as the pattern elements most sensitive to the lowering of gbb-60A are the PCV, L5 and then L4. In addition, as the levels of both dpp and gbb-60A are lowered, we see a restoration of L5 indicating that the ratio of Gbb-60A to Dpp is the important criteria not the absolute levels of both ligands. Furthermore, if Gbb-60A enhanced the Dpp signal, we would expect that increasing both dpp and gbb-60A would lead to phenotypes associated with elevated Dpp signaling. Instead, phenotypes induced by ectopic expression of gbb-60A are partially suppressed by ectopic expression of dpp (Fig. 6) further supporting our proposal that the balance of Gbb-60A and Dpp signaling activity is the mitigating factor in establishing positional information. BMP type I receptors, Tkv and Sax, mediate signaling We investigated the possibility that BMP type I receptors, Tkv and Sax, mediate the signaling by both Gbb-60A and Dpp. Heteroallelic combinations sax gbb-60A1/ + gbb-60A4 or tkv gbb-60A1/ + gbb-60A4 exhibit more extreme mutant phenotypes than those observed in gbb-60A genotypes alone (Fig. 7). Since Dpp and Gbb-60A signaling is linked in some cases, the tkv/gbb-60A or sax/gbb-60A interactions could reflect an indirect effect of compromising the efficiency of Dpp signaling by lowering the pool of receptors that are known to mediate the Dpp signal. Several aspects of the tkv/gbb-60A genetic interactions suggest that this is not wholly the explanation for the enhanced phenotypes. First, in Df(2L)tkv2 gbb-60A1/ +gbb-60A4 individuals, we observe a severe loss of veins including L3, a pattern element that is never affected in dpp mutants and has been postulated to be determined independently of Dpp signaling (Sturtevant et al., 1997). Second, if the interaction that we observe is a secondary consequence of reducing Dpp signaling, we would expect the viability of the Df(2L)tkv2 gbb-60A1/+ gbb-60A4 genotype to be much reduced as is seen in dpp gbb-60A1/+ gbb-60A4 genotypes (Fig. 5I), yet it is significantly increased. Third, a similar enhancement of the tkv427/Df(2L)tkv2 vein thickening phenotype as induced by gbb-60A1/ + has been observed by increasing, not decreasing, dpp+ through the addition of a dpp transgene (Penton et al., 1994). Thus, different levels of Gbb60A and Dpp affect signaling through Tkv in dramatically Gbb-60A to Dpp ratio critical for patterning 2733 different ways at different stages of development. These pieces of data do not negate the possibility that Gbb-60A boosts or enhances the Dpp signal at some level and that this is mediated through Tkv in some fashion, perhaps as a heterodimer with higher affinity or by some other as yet unknown mechanism. This possibility is consistent with data suggesting that Tkv mediates a higher level of Dpp signaling (Singer et al., 1997). However, our results indicate that, for some developmental events, it is likely that Tkv mediates the Gbb-60A signal alone and may be responsible for mediating the antagonism between these two ligands observed in some aspects of their genetic interaction. The mechanism by which this differential mediation of signal remains to be elucidated. Despite the unique roles that dpp and gbb-60A each play, the observed genetic interactions indicate that Dpp and Gbb-60A signaling can be dependent upon one another and that they must signal together in a complex manner involving both synergy and antagonism to achieve normal wing morphogenesis as well as viability. gbb-60A, like dpp, has multiple requirements during development, in addition to its role in growth, patterning and vein promotion in the developing wing. Given the expression pattern of gbb-60A and dpp in other imaginal discs and the phenotypic consequences of their genetic interaction in the adult leg (Fig. 5F), it is likely that gbb-60A and dpp function together in other tissues as well, to direct various developmental programs. The molecular implications of the functional data presented here can be quite straightforward. The complex signaling of gbb-60A and dpp must be mainly accomplished by Gbb-60A and Dpp homodimers, however, some degree of Gbb60A/Dpp heterodimer signaling can not be eliminated by our data. Regardless, cells in the developing wing are clearly able to interpret different levels of Gbb-60A versus Dpp. Numerous protein/protein interactions at different points in the transduction of TGF-β/BMP signals contribute to the diversity of effects elicited by these signaling molecules. The mechanism responsible for the interpretation of different ligands is unknown and remains to be examined. However, at one level both Tkv and Sax receptors appear to mediate Gbb-60A and Dpp signals and in one scenario a difference in binding affinities for the two different ligands could account for different signaling readouts. Experiments using activated and dominant negative receptors support the notion that Tkv and Sax mediate Dpp and Gbb-60A signaling at differential levels (T. E. Haerry, O. K., M. B. O’Connor and K. A. W., unpublished data). TGF-β receptors are thought to form tetrameric signaling complexes consisting of two type I and two type II receptors, however, as yet it is not known whether two different type I receptors can both contribute to that complex. If this were the case, it is conceivable that Gbb-60A and Dpp could signal through complexes containing both Tkv and Sax. Therefore, various combinations of these ligands and type I receptors could form to elicit different types of output. When considering ligand/receptor interactions it is important to remember that the presence of proteases and other binding proteins may dictate the availability of the ligand to bind to a receptor complex, as well as with what affinity it binds (LopezCasillas et al., 1994; Nakato et al., 1995). Having demonstrated that two different BMPs are both required for a single developmental process, we can now begin to dissect the signaling network by which these two ligands accomplish their task. We thank Laura Dickinson for technical assistance with RNA in situ hybridizations. We thank Bruce Reed, Matt Singer and Michael O’Connor for sending fly stocks and John Doctor for Gbb-60A antibodies. O. K. and J. Y. have been supported in part by NIH GM07601 training grant and this work was done in partial fulfillment of the requirements for a PhD. This research was supported in part by NSF IBN-9205808, IBN-9604769 and an ACS Research Grant DB32. K. A. W. was supported by an Established Investigatorship from the American Heart Association and with funds contributed in part by the AHA, Maine Affiliate. REFERENCES Affolter, M., Nellen, D., Nussbaumer, U. and Basler, K. (1994). Multiple requirements for the receptor serine/threonine kinase thick veins reveal novel functions of TGF-ß homologs during Drosophila embryogenesis. Development 120, 3105-3117. Arora, K., Levine, M. and O’Connor, M. (1994). The screw gene encodes a ubiquitously expressed member of the TGF-ß family required for specification of dorsal cell fates in the Drosophila embryo. Genes Dev. 8, 2588-2601. Blackman, R. K., Sanicola, M., Raftery, L., Gillevet, T. and Gelbart, W. M. (1991). An extensive 3′ cis-regulatory region directs the imaginal disk expression of decapentaplegic, a member of the TGF-β family in Drosophila. Development 111, 657-666. Brand, A. H. and Perrimon, N. (1993). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401-415. Brummel, T. J., Twombly, V., Marqués, G., Wrana, J. L., Newfeld, S. J., Attisano, L., Massagué, J., O’Connor, M. B. and Gelbart, W. M. (1994). Characterization and relationship of dpp receptors encoded by the saxophone and thick veins genes in Drosophila. Cell 78, 251-261. Burke, R. and Basler, K. (1996). Dpp receptors are autonomously required for cell proliferation in the entire developing Drosophila wing. Development 122, 2261-2269. Cohen, S. M. (1993). Imaginal disc development. In The Development of Drosophila melanogaster, pp. 747-841. Cold Spring Harbor Laboratory Press deCelis, J. F. (1997). Expression and function of decapentaplegic and thick veins during the differentiation of the veins in the Drosophila wing. Development 124, 1007-1018. Derynck, R. and Zhang, Y. (1996). The mad way to do it. Curr. Biol. 6,12261229. Doctor, J. S., Jackson, D., Rashka, K. E., Visalli, M. and Hoffmann, F. M. (1992). Sequence, biochemical characterization, and developmental expression of a new member of the TGF-ß superfamily in Drosophila melanogaster. Dev. Biol. 151, 491-505. Dudley, A. T. and Robertson, E. J. (1997). Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 deficient embryos. Devel. Dynam. 208, 349362. Ferguson, E. L. and Anderson, K. v. (1992). decapentaplegic acts as a morphogen to organize dorsal-ventral pattern in the Drosophila embryo. Cell 71, 451-461. The Flybase Consortium (1996). Flybase: the Drosophila database. Nucl. Acids Res. 24, 53-56. Gelbart, W. M. (1989). The decapentaplegic gene: a TGF-β homologue controlling pattern formation in Drosophila. Development 1989 Supplement, 65-74. Golic, K. G. and Lindquist, S. (1989). The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 59, 499-509. Hogan, B. L. M. (1996). Bone morphogenic proteins: multifunctional regulators of vertebrate development. Genes Dev. 10, 1580-1594. Jiang, J. and Struhl, G. (1996). Complementary and mutually exclusive activities of Decapentaplegic and Wingless organize axial pattern during Drosophila limb development. Cell 86, 401-409. Kingsley, D. M. (1994a). The TGF-β superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 8, 133-146. Kingsley, D. M. (1994b). What do BMPs do in mammals? Clues from the mouse short-ear mutation. Trends in Genet. 10, 16-21. 2734 O. Khalsa and others Lawrence, P. A., Johnston, P. and Morata, G. (1986). Methods of marking cells. In Drosophila, a Practical Approach (D. B. Roberts). pp. 229-242. Oxford: IRL Press Lecuit, T., Brook, W., Ng, M., Calleja, M., Sun, H. and Cohen, S. (1996). Two distinct mechanisms for long-range patterning by decapentaplegic in the Drosophila wing. Nature 381, 387-393. Lecuit, T. and Cohen, S. M. (1997). Proximal-distal axis formation in theDrosophila leg. Nature 388, 139-145. Lopez-Casillas, F., Patne, H. M., Andres, J. L. and Massague, J. (1994). Betaglycan can act as a dual modulator of TGF-β access to signaling receptors: mapping of ligand binding and GAG attachment sites. J. Cell Biol. 124, 557-568. Lyons, K. M., Hogan, B. L. M. and Robertson, E. J. (1995). Colocalization of BMP7 and BMP2 RNAs suggests that these factors cooperatively mediate tissue interactions during murine development. Mech. Dev. 50, 7183. Massagué, J. (1996). TGF-β signalling: receptors, transducers and mad proteins. Cell 85, 947-950. Massagué, J., Cheifetz, S., Boyd, F. T. and Andres, J. L. (1990). Mediators of TGF-β action: TGF-β receptors and TGF-β binding proteoglycams. Ann. NY Acad. Sci. USA 593, 59-72. Massagué, J., Hata, A. and Liu, F. (1997). TGF-β signalling through the Smad pathway. Trends Cell Biol. 7, 187-192. Masucci, J. D., Miltenberger, R. J. and Hoffmann, F. M. (1990). Patternspecific expression of the Drosophila decapentaplegic gene in the imaginal disks is regulated by 3’ cis-regulatory elements. Genes Dev. 4, 2011-2023. Nakato, H., Futch, T. A. and Selleck, S. B. (1995). The division abnormally delayed, dally, gene: A putative integral membrane proteoglycan required for cell division patterning during postembryonic development of the nervous system in Drosophila. Development 121, 3687-3702. Nellen, D., Affolter, M. and Basler, K. (1994). Receptor serine/threonine kinase implicated in the control of Drosophila body pattern by decapentaplegic. Cell 78, 225-237. Nellen, D., Burke, R., Struhl, G. and Basler, K. (1996). Direct and longrange action of a DPP morphogen gradient. Cell 85, 357-368. Padgett, R., St. Johnston, R. and Gelbart, W. (1987). A transcript from a Drosophila pattern gene predicts a protein homologous to the transforming growth factor-β family. Nature 325, 81-84. Penton, A., Chen, Y., Staehling-Hampton, K., Wrana, J. L., Attisano, L., Szidonya, J., Cassill, J. A., Massagué, J. and Hoffmann, F. M. (1994). Identification of two bone morphogenetic protein type I receptors in Drosophila and evidence that Brk25D is a decapentaplegic receptor. Cell 78, 239-250. Posakony, L. G., Raftery, L. A. and Gelbart, W. M. (1991). Wing formation in Drosophila melanogaster requires decapentaplegic gene function along the anterior-posterior compartment boundary. Mech. Dev. 33, 69-82. Raftery, L., Twombly, V., Wharton, K. and Gelbart, W. (1995). Genetic screens to identify elements of the decapentaplegic signaling pathway in Drosophila. Genetics 139, 241-254. Roberts, A. B. and Sporn, M. B. (1990). The transforming growth factor-βs. In Peptide Growth Factors and Their Receptors I. (ed. Michael B. Sporn and Anita B. Roberts). pp. 419-472. Springer-Verlag. Segal, D. and Gelbart, W. M. (1985). Shortvein, a new component of the decapentaplegic gene complex in Drosophila melanogaster. Genetics 109, 119-143. Singer, M. A., Penton, A., Twombly, V., Hoffmann, F. M. and Gelbart, W. M. (1997). Signaling through both type I DPP receptors is required for anterior-posterior patterning of the entire Drosophila wing. Development 124, 79-89. Spencer, F. A., Hoffmann, F. M. and Gelbart, W. M. (1982). decapentaplegic: A gene complex affecting morphogenesis in Drosophila melanogaster. Cell 28, 451-461. Staehling-Hampton, K., Jackson, P. D., Clark, M. J., Brand, A. H. and Hoffmann, F. M. (1994). Specificity of bone morphogenetic protein-related factors: Cell fate and gene expression changes in Drosophila embryos induced by decapentaplegic but not 60A. Cell Growth, Differen. 5, 585-593. Sturtevant, M. A. B., Marin, E. and Bier, E. (1997). The spalt gene links the A/P compartment boundary to a linear adult structure in the Drosophila wing. Development 124, 21-32. Tautz, D. and Pfeifle, C. (1989). A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals a translational control of the segmentation gene hunchback. Chromosoma 98, 81-85. Twombly, V., Blackman, R. K., Jin, H., Graff, J. M., Padgett, R. and Gelbart, R. W. (1996). The TGF-β signaling pathway is essential for Drosophila oogenesis. Development 122, 1555-1565. Wharton, K. A., Ray, R. P. and Gelbart, W. M. (1993). An activity gradient of decapentaplegic is necessary for the specification of dorsal pattern elements in the Drosophila embryo. Development 117, 807-822. Wharton, K. A., Thomsen, G. H. and Gelbart, W. M. (1991). Drosophila 60A gene, another transforming growth factor-β family member, is closely related to human bone morphogenetic proteins. Proc. Natn. Acad. Sci. USA 88, 9214-9218. Xie, T., Finelli, A. L. and Padgett, R. W. (1994). The Drosophila saxophone gene: a serine-threonine kinase receptor of the TGF-β superfamily. Science 263, 1756-1759. Xu, T. and Harrison, S. D. (1994). Mosaic analysis using FLP recombinase. Vol. 44. Drosophila melanogaster: Practical Uses in Cell and Molecular Biology. (ed. S. Lawrence, B. Goldstein and Eric A. Fyrberg). San Diego: Academic Press. Yamashita, H., ten Dijke, P., Franzen, P., Miyazono, K. and Heldin, C. H. (1994). Formation of hetero-oligomeric complexes of type I and type II receptors for transforming growth factor-β. J. Biol. Chem. 269, 2017220178.