Mothers against dpp (Mad) is the prototype of a family of genes required for signaling by TGF-β related ligands. In Drosophila, Mad is specifically required in cells responding to Decapentaplegic (DPP) signals. We further specify the role of Mad in DPP-mediated signaling by utilizing tkvQ199D, an activated form of the DPP type I receptor serine-threonine kinase thick veins (tkv). In the embryonic midgut, tkvQ199D mimics DPP-mediated inductive interactions. Homozygous Mad mutations block signaling by tkvQ199D. Appropriate responses to signaling by tkvQ199D are restored by expression of MAD protein in DPP-target cells. Endogenous MAD is phosphorylated in a ligand-dependent manner in Drosophila cell culture. DPP overexpression in the embryonic midgut induces MAD nuclear accumulation; after withdrawal of the overexpressed DPP signal, MAD is detected only in the cytoplasm. However, in three different tissues and developmental stages actively responding to endogenous DPP, MAD protein is detected in the cytoplasm but not in the nucleus. From these observations, we discuss possible roles for MAD in a DPP-dependent serine-threonine kinase signal transduction cascade integral to the proper interpretation of DPP signals.

Intercellular signaling by ligands of the transforming growth factor-β (TGF-β) superfamily are required for control of many aspects of cell growth, patterning and differentiation in many organisms (Kingsley, 1994). Receptor serine-threonine kinases on the surface of responsive cells are activated by these secreted molecules and thereby initiate an intracellular signal transduction cascade (Massagué et al., 1994). Recent studies have begun to identify components of these cascades that relay the signal to the nucleus and elicit appropriate transcriptional responses.

The secreted processed product of the decapentaplegic (dpp) gene in Drosophila melanogaster (Padgett et al., 1987) initiates one of the genetically best characterized TGF-β signaling pathways. The DPP pathway requires both type I and type II receptors (saxophone, thick veins, punt: Brummel et al., 1994; Nellen et al., 1994; Penton et al., 1994; Xie et al., 1994; Letsou et al., 1995; Ruberte et al., 1995) and at least one transcription factor has been suggested as an immediate target of the pathway (schnurri: Arora et al., 1995; Grieder at al., 1995; Staehling-Hampton et al., 1995). We have been focusing on the characterization of another signal transduction cascade component, the MAD protein encoded by Mothers against dpp (Raftery et al., 1995; Sekelsky et al., 1995).

Mad loss-of-function mutants exhibit numerous phenotypes with similarities to those of various dpp mutants and partial reduction in Mad activity exacerbates phenotypes associated with specific dpp mutant genotypes (Raftery et al., 1995; Sekelsky et al., 1995). These genetic parallels and interactions form the basis for our suggestion that Mad encodes a component of a DPP-responsive signal transduction pathway. This idea is further supported by our studies in the embryonic midgut demonstrating that Mad is required in DPP-target cells (Newfeld et al., 1996).

Analyses of Mad activity in adult appendages have shown that heterozygosity for a Mad mutation can suppress gain-of-function phenotypes due to expression of a constitutively active form of the DPP type I receptor kinase thick veins (tkvQ199D: Hoodless et al., 1996). However, inferences about epistatic relationships derived from genotypes in which a gene product is quantitatively reduced but not eliminated are open to multiple interpretations. In the adult appendages, Mad null clones do not proliferate (Wiersdorff et al., 1996) so that epistasis experiments utilizing homozygous Mad null genotypes cannot be conducted. By focusing on DPP signaling in the embryonic midgut, we are able to examine epistatic relationships between Mad and dpp in genotypes that are functionally null for Mad. Here we report that tkvQ199D constitutively induces several DPP-dependent responses during midgut development. These responses are blocked by homozygosity for a Mad null mutation and appropriate tkvQ199D signaling is restored by providing MAD in specific DPP-target cells.

While the predicted MAD polypeptide contains no identifiable motifs (Sekelsky et al., 1995), a series of related genes define a protein family with extensive conserved domains (Smad: Massagué, 1996; Wrana and Attisano, 1996). Transfection studies in mammalian cells determined that cytoplasmic Smad proteins are phosphorylated (Hoodless et al., 1996; Lechleider et al., 1996; Yingling et al., 1996) and accumulate in the nucleus (Hoodless et al., 1996; Liu et al., 1996) in response to ligand. Here we report the ligand-dependent phosphorylation of endogenous MAD in Drosophila cell culture, the nuclear accumulation of cytoplasmic MAD in response to DPP overexpression and efforts to detect nuclear accumulation of MAD in response to endogenous DPP at several developmental stages. Our results are consistent with a pivotal role for MAD in a DPP-dependent serine-threonine kinase signal transduction cascade.

Reporter gene experiments

All Drosophila strains are described in Newfeld et al. (1996) except for P{UAS-tkv.Q199D} (Hoodless et al., 1996) and P{Gal4-Hsp70.PB} (a gift from Andrea Brand via Elizabeth Noll and Norbert Perrimon). Analyses of reporter genes in wild-type, Mad null backgrounds and Mad null backgrounds with tissue-specific expression of MAD are described in Newfeld et al. (1996). For analysis of P{UAS-tkv.Q199D} in wild type and Mad null embryos, strains were constructed which are homozygous for P{UAS-tkv.Q199D} on the X chromosome and P{Gal4-Hsp70.PB} on chromosome 3, with and without Mad12/CyO P{wg-βgal}. Females from both of these strains were crossed to males hemizygous for the lab reporter gene P{3.65lab66A}, heterozygous for Df(2L)JS17 (which deletes Mad)/CyO P{wg-βgal} and homozygous for P{hs-dpp.BP} on chromosome 3. The lab reporter gene contains a DPP-indepen-dent CNS enhancer, which acts as a marker of that chromosome. Females from both strains were also crossed to males heterozygous for Mad12/CyO P{wg-βgal} who were either homozygous for dpp midgut reporter gene P{RD2} or carried an insertion of Scr anterior/posterior midgut reporter gene P{HZR+0.8X/H} on TM6B maintained over Sb gl3. For comparison of the effects of P{UAS-tkv.Q199D} and heat-shock induced DPP on Scr, females homozy-gous for P{hs-dpp.BP} on chromosome 3 were crossed to males het-erozygous for the Scr reporter gene. For analysis of P{UAS-tkv.Q199D} effects on lab in Mad null background embryos with tissue-specific expression of MAD, females homozygous for P{UAS-tkv.Q199D} on the X chromosome and P{Gal4-Hsp70.PB} on chromosome 3 and heterozygous for Mad12/CyO P{wg-βgal} were crossed to males hemizygous for the lab reporter, heterozygous for Df(2L)JS17/CyO P{wg-βgal} and heterozygous for P{hs-dpp.BP} and P{mex1Mad} on chromosome 3. Embryos carrying P{mex1Mad} were identified by observation of an unambiguous positive result, rescue of reporter gene expression in the appropriate tissue predicted from the results of Newfeld et al. (1996). For analyses of P{UAS-tkv.Q199D} effects on dpp and Scr in Mad null background embryos with tissue-specific expression of MAD, males hemizygous for P{UAS-tkv.Q199D}, heterozygous for Mad12/CyO P{wg-βgal} and carrying dpp or Scr reporter genes were crossed to females homozygous for P{UASMad.N} on the X chromosome and P{GawB}24B on chromosome 3 that were also heterozygous for Mad12/CyO P{wg-βgal}. Embryos carrying P{UAS tkv.Q199D} were identified by observation of an unambiguous positive result, rescue of reporter gene expression in the appropriate tissue predicted from the results of Newfeld et al. (1996). All crosses were reared at room temperature (21°C) to avoid low level expression of P{hs-dpp.BP} and P{Gal4-Hsp70.PB} at 25°C. 37°C induction of P{hs-dpp.BP} and P{Gal4-Hsp70.PB} and subsequent analyses of β-galactosidase activity were performed according to Newfeld et al. (1996).

Phosphorylation experiments

Drosophila melanogaster wing imaginal disk cell lines MLDmD8 and MLDmD12 (Ui et al., 1987) were grown at pH 6.5 in M1 media (Sigma) with 166.0 mg/ml yeastolate, 5.0 mg/ml bactotryptone, 10 μg/ml insulin and 12.5% fetal calf serum at 27°C. For in vivo phosphate labeling, the cell lines were treated as described (Wrana et al., 1994). Briefly, cells were incubated at pH 6.5 for 2 hours at 27°C in phosphate-free modified Eagle’s medium containing 20 Mm Hepes and [32P]phosphate at 1 mCi/ml in the presence or absence of 10 nM human recombinant BMP2 in the last 30 minutes of labeling (unless otherwise indicated). Cells were lysed in TNE (50 mM Tris pH 7.4, 150 mM NaCl and 1 mM EDTA) containing 0.5% Triton X-100 in the presence of protease and phosphatase inhibitors (10 μg/ml antipain, 50 μg/ml aprotinin, 100 μg/ml benzamidine hydrochloride, 10 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml pepstatin, 5 μg/ml RNAse A, 25 mM sodium flouride, 1 mM sodium orthovanadate, 10 mM sodium pyrophos-phate and 100 μg/ml soybean trypsin inhibitor). Cell lysates were immunoprecipitated with anti-MAD antibodies or preimmune sera (Newfeld et al., 1996) followed by adsorbtion to protein A sepharose (Pharmacia). Immunoprecipitated complexes were washed twice in TNE containing 0.1% Triton X-100, followed by two washes in RIPA (150 mM NaCl, 1% NP-40, 0.5% deoxychloate, 0.1% sodium dodecyl sulfate in 50 mM Tris at pH 8.0), followed by a final wash in TNE containing 0.1% Triton X-100. Lysate proteins were separated by SDS-PAGE (8%) and visualized by autoradiography. For metabolic labeling, cells were incubated with 100 μCi/ml [35S]methionine for 2 hours at 27°C in methionine-free modified Eagle’s medium containing 20 mM Hepes at pH 6.5. MAD was immunoprecipitated and analyzed as described for phosphate labeling.

MAD subcellular localization experiments

Embryos

For DPP overexpression, a strain containing six copies of P{hs-dpp.BP} and its parental strain y w67c23 were tested in parallel. Embryos from developmental stages corresponding to midgut con-striction formation were heat-shocked and analyzed by confocal microscopy as described in Newfeld et al. (1996). For endogenous DPP, y w67c23 blastoderm embryos and midgut formation stage embryos were analyzed as in Newfeld et al. (1996).

Disks

For endogenous DPP, imaginal disks from the strain P{BS3.0} (Blackman et al., 1991), which carries a dppdisk region reporter gene were tested. Early to mid-third instar larvae were dissected in PBS (20 mM NaH2PO4, 150 mM NaCl, pH 7.2) and fixed in 2% EM-grade formaldehyde (Polysciences) in PEM (100 mM PIPES, 2 mM EGTA, 1 mM MgSO4, pH 6.95) for 30 minutes. Tissue was incubated 18 hours at 4°C in a 1:2000 dilution of rabbit anti-β-galactosidase anti-bodies and/or 1:1000 dilution of guinea pig anti-MAD polyclonal sera. Subsequently, tissue was incubated for 3 hours in a 1:1200 dilution of goat anti-rabbit LSRC and/or a 1:100 dilution of donkey anti-guinea pig-FITC (Jackson Immuno Research). Tissue was placed into 30% glycerol mountant (in 50 mM Tris pH 8.8, 150 mM NaCl, 0.02% NaN3) with 0.5 mg/ml p-phenylenediamine (Sigma) and analyzed by confocal microscopy.

Mad functions downstream of the kinase activity of a DPP type I receptor

To clarify Mad’s function in DPP-responsive cells, we have continued to focus on the embryonic midgut. Here we are able to utilize genotypes that are null for Mad to examine the relationship between Mad and dpp. In the midgut, dpp is expressed in the visceral mesoderm of parasegments (ps) 3 and 7 (Fig. 1). In response to DPP signals, cells expressing dpp in ps3 repress the expression of the homeotic gene Sex combs reduced (Scr). DPP signals are also required to maintain dpp expression in ps3 through an autocrine feedback loop. However, cells in ps4 do not appear to be affected by DPP signals; Scr is expressed while dpp is not (Hursh et al., 1993). In ps7, the homeotic gene Ultrabithorax (Ubx) initiates dpp expression. Subsequently, DPP functions in an autocrine manner to maintain Ubx and thus dpp expression. In ps7, DPP also signals between germ layers to the underlying endoderm. Within the midgut endoderm, which does not express dpp, expression of the homeotic gene labial (lab) is dependent upon DPP signals (Bienz, 1994).

Fig. 1.

MAD is required for proper response to DPP signals. The relationships between several genes known to participate in the DPP signaling pathway and DPP-responsive genes in the embryonic midgut are depicted. Genes in the DPP pathway include dpp, the receptors tkv and punt, and Mad. In ps3 visceral mesoderm, the proper response to DPP signals is the maintenance of dpp expression and the repression of Scr expression. In ps7 visceral mesoderm, Ubx is also involved in the DPP pathway and the proper response to DPP signals is the maintenance of Ubx and dpp expression. In ps7 midgut endoderm, the proper response to DPP signals is the induction of lab expression.

Fig. 1.

MAD is required for proper response to DPP signals. The relationships between several genes known to participate in the DPP signaling pathway and DPP-responsive genes in the embryonic midgut are depicted. Genes in the DPP pathway include dpp, the receptors tkv and punt, and Mad. In ps3 visceral mesoderm, the proper response to DPP signals is the maintenance of dpp expression and the repression of Scr expression. In ps7 visceral mesoderm, Ubx is also involved in the DPP pathway and the proper response to DPP signals is the maintenance of Ubx and dpp expression. In ps7 midgut endoderm, the proper response to DPP signals is the induction of lab expression.

Proper response to DPP requires a heteromeric complex of transmembrane receptors. In the receptor complex, the type I receptor, acting downstream of the type II receptor, initiates the intracellular signaling cascade (Wrana et al., 1994). The cytoplasmic protein MAD is also required for appropriate responses to DPP signals (Newfeld et al., 1996). For our analyses, we have utilized a constitutively active DPP type I serine-threonine kinase receptor, tkvQ199D, that functions in a cell autonomous manner (Hoodless et al., 1996). We have examined Mad function with regard to tkvQ199D in the induction of lab in ps7 midgut endoderm, in dpp maintenance in ps7 and repression of Scr in ps3 visceral mesoderm. If indeed Mad is an essential part of the DPP signal transduction apparatus, the absence of MAD should block any phenotypes due to a constitutive signal from tkvQ199D.

Expression of a lab reporter gene containing a DPP-inde-pendent central nervous system enhancer and a DPP-dependent ps7 midgut endoderm enhancer in a Mad+ background is shown in Fig. 2A. Fig. 2B,D show the anterior expansion of lab midgut endoderm expression in response to ubiquitously expressed tkvQ199D in early and late stage embryos, respec-tively, an effect also seen with ubiquitous DPP (Newfeld et al., 1996). Comparably staged Mad null tkvQ199D embryos (Fig. 2C,E) lack any lab expression in the midgut endoderm, strongly suggesting that Mad is epistatic to tkvQ199D. When we express MAD solely in the midgut endoderm using P{mex1Mad}, in an otherwise Mad null embryo, in the presence of tkvQ199D we observe ectopic lab expression in that tissue, as shown by the staining of cells in the elongating gut (Fig. 2F). The cells showing lab expression derive solely from the lab-independent anterior domain of expression of the mex1 enhancer driving MAD in the endoderm. This result is also seen in a comparable embryo expressing ubiquitous DPP instead of tkvQ199D (Newfeld et al., 1996). Thus, we infer that Mad functions in the DPP signal transduction cascade down-stream of the serine-threonine kinase activity of tkv but upstream of lab induction.

Fig. 2.

MAD functions downstream of tkvQ199D in lab induction. (A) Wild-type expression of a lab reporter gene that contains a DPP-independent central nervous system (CNS) enhancer and a DPP-dependent midgut endoderm enhancer in a stage 13 Mad+ embryo. lab expression in the head and midgut endoderm (square bracket) are clearly seen. (B) Anterior expansion of lab midgut endoderm expression (extended square bracket) in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 13 Mad+ embryo. The effect of tkvQ199D on lab is virtually indistinguishable from that seen with P{hs-dpp.BP} (Newfeld et al., 1996). (C) lab CNS expression is unaffected while no expression from the lab midgut endoderm enhancer is detectable in a stage 13 Mad null tkvQ199D embryo expressing heat-shock dpp. (D) Expanded expression from the lab midgut endoderm enhancer continues in a Mad+ stage 17 embryo expressing tkvQ199D. (E) tkvQ199D expression has no effect on lab midgut endoderm expression in a Mad null stage 17 embryo. (F) Providing MAD only in the midgut endoderm (ME) in an otherwise Mad null embryo rescues the induction of lab expression by tkvQ199D in that tissue. lab-expressing cells (arrow) are seen in the elongating midgut of a stage 17 embryo.

Fig. 2.

MAD functions downstream of tkvQ199D in lab induction. (A) Wild-type expression of a lab reporter gene that contains a DPP-independent central nervous system (CNS) enhancer and a DPP-dependent midgut endoderm enhancer in a stage 13 Mad+ embryo. lab expression in the head and midgut endoderm (square bracket) are clearly seen. (B) Anterior expansion of lab midgut endoderm expression (extended square bracket) in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 13 Mad+ embryo. The effect of tkvQ199D on lab is virtually indistinguishable from that seen with P{hs-dpp.BP} (Newfeld et al., 1996). (C) lab CNS expression is unaffected while no expression from the lab midgut endoderm enhancer is detectable in a stage 13 Mad null tkvQ199D embryo expressing heat-shock dpp. (D) Expanded expression from the lab midgut endoderm enhancer continues in a Mad+ stage 17 embryo expressing tkvQ199D. (E) tkvQ199D expression has no effect on lab midgut endoderm expression in a Mad null stage 17 embryo. (F) Providing MAD only in the midgut endoderm (ME) in an otherwise Mad null embryo rescues the induction of lab expression by tkvQ199D in that tissue. lab-expressing cells (arrow) are seen in the elongating midgut of a stage 17 embryo.

Further experiments utilize the dpp reporter gene P{RD2}, which accurately reflects dpp expression in the visceral mesoderm of ps3 and ps7 (Hursh et al., 1993). In both regions, the maintenance of dpp expression is controlled by an autocrine signaling pathway requiring dpp, tkv and Mad. In early stage Mad+tkvQ199D embryos, dpp expression is expanded to include all of the intervening parasegments (Fig. 3A). This result is also seen in embryos expressing DPP throughout the mesoderm (Staehling-Hampton and Hoffmann, 1994). In early stage Mad null tkvQ199D embryos, the initiation of dpp expression is unaffected. dpp ectopic expression in ps4 through ps6, mediated by the activity of tkvQ199D in Mad+ embryos, does not occur in Mad null embryos (Fig. 3D). In late stage Mad+tkvQ199D embryos (Fig. 3B), the expanded domain of dpp expression is maintained at very high levels. In late stage Mad null tkvQ199D embryos, this is not observed (Fig. 3E). dpp expression begins to diminish shortly after initiation and we believe that the perdurance of β-galactosidase is responsible for residual staining in these embryos. In mid-stage Mad null background tkvQ199D embryos that also express MAD specifically in the visceral mesoderm (Fig. 3C), there is main-tenance of dpp expression in ps3 and ps7 and ectopic expression of dpp in the intervening parasegments (due to tkvQ199D). In a comparable embryo lacking tkvQ199D, only the maintenance of ps3 and ps7 dpp expression occurs (Fig. 3F). Thus, we infer that Mad functions in the DPP signal transduc-tion cascade downstream of the serine-threonine kinase activity of tkv but upstream of elements required for maintenance of dpp expression in the visceral mesoderm of ps3 and ps7. Mad also functions downstream of the serine-threonine kinase activity of tkvQ199D but upstream of ectopic dpp expression in the intervening parasegments.

Fig. 3.

MAD functions downstream of tkvQ199D in dpp maintenance. (A) Expression of a dpp reporter gene which contains visceral mesoderm enhancers expressed in ps3 and ps7, in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 13 Mad+ embryo. High levels of expression in ps3 and ps7 as well as ectopic expression in the intervening parasegments (extended square bracket) are clearly seen. (D) Initiation of dpp expression in ps3 and ps7 (arrowheads) is unaffected in a comparably staged Mad null tkvQ199D embryo as expected since the initiation of dpp expression is the result of direct activation by UBX. The ectopic expression of dpp in the intervening parasegments, due to tkvQ199D, is not seen in this embryo. Comparison of B with E shows that dpp expression is maintained for an extended period at very high levels throughout the ps3 to ps7 region in response to tkvQ199D in a Mad+ embryo (B) but not in a Mad null stage 17 embryo (E). The Mad null background embryo in C is expressing P{UASMad.N} and tkvQ199D in the visceral mesoderm (VM) in response to P{GawB} 24B. The comparable Mad null background embryo in F does not contain P{UAS-tkv.Q199D}. (C) Ectopic expression of dpp between ps3 and ps7 in response to tkvQ199D is restored by providing MAD in the VM. Providing MAD in the VM without tkvQ199D does not result in ectopic expression of dpp (F).

Fig. 3.

MAD functions downstream of tkvQ199D in dpp maintenance. (A) Expression of a dpp reporter gene which contains visceral mesoderm enhancers expressed in ps3 and ps7, in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 13 Mad+ embryo. High levels of expression in ps3 and ps7 as well as ectopic expression in the intervening parasegments (extended square bracket) are clearly seen. (D) Initiation of dpp expression in ps3 and ps7 (arrowheads) is unaffected in a comparably staged Mad null tkvQ199D embryo as expected since the initiation of dpp expression is the result of direct activation by UBX. The ectopic expression of dpp in the intervening parasegments, due to tkvQ199D, is not seen in this embryo. Comparison of B with E shows that dpp expression is maintained for an extended period at very high levels throughout the ps3 to ps7 region in response to tkvQ199D in a Mad+ embryo (B) but not in a Mad null stage 17 embryo (E). The Mad null background embryo in C is expressing P{UASMad.N} and tkvQ199D in the visceral mesoderm (VM) in response to P{GawB} 24B. The comparable Mad null background embryo in F does not contain P{UAS-tkv.Q199D}. (C) Ectopic expression of dpp between ps3 and ps7 in response to tkvQ199D is restored by providing MAD in the VM. Providing MAD in the VM without tkvQ199D does not result in ectopic expression of dpp (F).

Analysis of an Scr visceral mesoderm reporter gene in Mad null tkvQ199D embryos provides further evidence that Mad functions downstream of tkvQ199D. Scr is normally expressed in ps4 visceral mesoderm (Fig. 4A) immediately posterior to the dpp-expressing cells in ps3. Failure to initiate dpp expression in ps3 in certain dpp mutants results in an anterior expansion of Scr expression into ps3 (Hursh et al., 1993), indicating a role for DPP in repressing Scr in ps3. Mad null embryos show the same anterior expansion of Scr (Newfeld et al., 1996) demonstrating that Mad is required for this activity of DPP. Alternatively, ubiquitous expression of tkvQ199D or DPP represses Scr in ps4 (Fig. 4B,C). In an Mad null tkvQ199D embryo, the expression of Scr (Fig. 4E) clearly resembles that seen in the Mad null embryo (Fig. 4D) and not the tkvQ199D embryo (Fig. 4B). Mad function is epistatic to tkvQ199D in the repression of Scr in the visceral mesoderm of ps3. This result is supported by examination of Scr expression in Mad null background tkvQ199D embryos that also express MAD specifically in the visceral mesoderm. These embryos display an initial anterior expansion of Scr into ps3 comparable to a Mad null embryo (data not shown). The expansion likely occurs prior to the accumulation of sufficient MAD to allow tkvQ199D to repress Scr expression. In older Mad null background tkvQ199D embryos expressing MAD specifically in the visceral mesoderm (Fig. 4F), upon restoration of tkvQ199D activity by visceral mesoderm-specific MAD, there is clear repression of Scr in both ps3 and ps4. Thus, we infer that Mad functions in the DPP signal transduction cascade downstream of the serine-threonine kinase activity of tkv but upstream of the repression of Scr in the visceral mesoderm of ps3. In summary, proper responses to DPP signals during midgut development require MAD activity in a DPP-dependent signal transduction cascade.

Fig. 4.

MAD functions downstream of tkvQ199D in Scr repression. (A) Wild-type expression of an Scr reporter gene that contains a DPP-responsive anterior midgut enhancer in a stage 14 embryo. Scr expression in ps4 visceral mesoderm (arrows) is clearly seen. (B) Repression of Scr ps4 expression in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 15 embryo. A distinctly reduced amount of Scr expression is seen (arrows). (C) The effect of tkvQ199D on Scr is virtually indistinguishable from that seen with P{hs-dpp.BP}. (D) Scr expression is expanded anteriorly to include ps3 (square bracket) in a stage 15 Mad null embryo. This result is also seen in embryos lacking dpp ps3 expression (Hursh et al., 1993). (E) A stage 15 Mad null tkvQ199D embryo. The anterior midgut expression of Scr resembles the pattern detected in Mad null embryos (anterior expansion – square bracket) and not the pattern seen in tkvQ199D embryos (reduced ps4 expression). The Mad null background embryo in F is expressing P{UASMad.N} and tkvQ199D in the visceral mesoderm (VM) in response to P{GawB}24B. In this stage 16 embryo, the anterior expression of Scr presumably occurred prior to the accumulation of sufficient MAD to allow tkvQ199D to reduce Scr expression. The repression of ectopic Scr expression (square bracket), by visceral mesoderm-specific MAD and tkvQ199D is evident when comparing F with E.

Fig. 4.

MAD functions downstream of tkvQ199D in Scr repression. (A) Wild-type expression of an Scr reporter gene that contains a DPP-responsive anterior midgut enhancer in a stage 14 embryo. Scr expression in ps4 visceral mesoderm (arrows) is clearly seen. (B) Repression of Scr ps4 expression in response to P{Gal4-Hsp70.PB} driven P{UAS-tkv.Q199D} in a stage 15 embryo. A distinctly reduced amount of Scr expression is seen (arrows). (C) The effect of tkvQ199D on Scr is virtually indistinguishable from that seen with P{hs-dpp.BP}. (D) Scr expression is expanded anteriorly to include ps3 (square bracket) in a stage 15 Mad null embryo. This result is also seen in embryos lacking dpp ps3 expression (Hursh et al., 1993). (E) A stage 15 Mad null tkvQ199D embryo. The anterior midgut expression of Scr resembles the pattern detected in Mad null embryos (anterior expansion – square bracket) and not the pattern seen in tkvQ199D embryos (reduced ps4 expression). The Mad null background embryo in F is expressing P{UASMad.N} and tkvQ199D in the visceral mesoderm (VM) in response to P{GawB}24B. In this stage 16 embryo, the anterior expression of Scr presumably occurred prior to the accumulation of sufficient MAD to allow tkvQ199D to reduce Scr expression. The repression of ectopic Scr expression (square bracket), by visceral mesoderm-specific MAD and tkvQ199D is evident when comparing F with E.

Ligand-dependent phosphorylation of endogenous MAD

Studies using transiently transfected cell lines have shown that mammalian Smad proteins become phosphorylated upon induction of TGF-β signaling pathways. For example, Smad1 becomes phosphorylated upon BMP2 addition and Smad2 is phosphorylated in response to TGF-β (Hoodless et al., 1996; Eppert et al., 1996). Thus, we sought to determine whether endogenous Drosophila MAD might be regulated in a similar manner. Since MAD appears to mediate DPP signaling in imaginal disks (Sekelsky et al., 1995), we analyzed two wing imaginal disk cell lines (MLDmD-8 and MLDmD-12: Ui et al., 1987). We stimulated DPP signaling in these cells using human BMP2, a homolog of DPP (Padgett et al., 1993; Sampath et al., 1993) and the cell lines responded to this treatment.

To determine whether MAD phosphorylation is regulated by ligand addition, we immunoprecipitated endogenous MAD from MLDmD-8 and MLDmD-12 cell lines using anti-MAD sera (Newfeld et al., 1996). Examination of immunoprecipi-tates from cells metabolically labeled with [35S]methionine revealed a protein of approximately 50×103Mr (Fig. 5A lower panels) that is close to the predicted size of MAD (50.5×103Mr). This protein was absent in immunoprecipitates prepared using preimmune sera, thus confirming its identity as MAD. In both imaginal disk cell lines, MAD protein levels were unaf-fected by treatment of cells with BMP2 prior to lysis. However, treatment with BMP2 for 30 minutes prior to lysis yielded a MAD protein that displayed a slightly slower mobility on SDS-PAGE gels (MAD*), a characteristic often observed for phos-phorylated proteins.

Fig. 5.

BMP2-dependent in vivo phosphorylation of endogenous MAD.(A) Ligand-dependent phosphorylation of MAD. Cells from DmD12 (left panels) or DmD8 (right panels) imaginal disk cell lines were labeled with either [32P]phosphate (upper panels) or [35S]methionine (lower panels) and incubated in the presence (+) or absence (−) of human BMP2. Lysates were immunoprecipitated with anti-MAD antibodies (α-MAD) or preimmune sera (PI). The migration of phosphorylated MAD is indicated and the approximate positions of molecular mass markers (×10−3) are shown. Note that phosphorylated MAD (MAD* in the lower panels) migrates slightly more slowly than MAD. (B) Kinetics of MAD phosphorylation. DmD12 cells were labeled with [32P]phosphate and stimulated with BMP2 for increasing times as indicated. Labeling of MAD is detected in the first sample and appears complete by 30 minutes of stimulation.

Fig. 5.

BMP2-dependent in vivo phosphorylation of endogenous MAD.(A) Ligand-dependent phosphorylation of MAD. Cells from DmD12 (left panels) or DmD8 (right panels) imaginal disk cell lines were labeled with either [32P]phosphate (upper panels) or [35S]methionine (lower panels) and incubated in the presence (+) or absence (−) of human BMP2. Lysates were immunoprecipitated with anti-MAD antibodies (α-MAD) or preimmune sera (PI). The migration of phosphorylated MAD is indicated and the approximate positions of molecular mass markers (×10−3) are shown. Note that phosphorylated MAD (MAD* in the lower panels) migrates slightly more slowly than MAD. (B) Kinetics of MAD phosphorylation. DmD12 cells were labeled with [32P]phosphate and stimulated with BMP2 for increasing times as indicated. Labeling of MAD is detected in the first sample and appears complete by 30 minutes of stimulation.

To directly examine MAD phosphorylation, MAD was immunoprecipitated from [32P]phosphate-labeled cells treated with or without BMP2. Phosphorylation of MAD was almost undetectable in immunoprecipitates in the absence of BMP2 treatment (Fig. 5A upper panels). Treatment of cells with BMP2 prior to lysis induced phosphorylation of MAD. Ligand-induced phosphorylation of MAD was observed in both cell lines. To characterize the kinetics of MAD phosphorylation, [32P]phosphate-labeled MLDmD-12 cells were treated with BMP2 for varying times. Phosphorylation of MAD was detected within 15 minutes of BMP2 addition, appeared to plateau at 30 minutes and remained elevated for at least 60 minutes (Fig. 5B). Together these findings indicate that MAD is rapidly phosphorylated in response to BMP2 and that MAD is a direct downstream component of the DPP signal trans-duction pathway.

MAD dramatically alters its subcellular distribution in response to DPP overexpression

In Drosophila embryos and salivary glands, staining with anti-MAD sera showed that endogenous MAD protein maintains a predominantly cytoplasmic subcellular distribution (Newfeld et al., 1996). Transfection studies in mammalian cell lines demonstrated that providing ligand or an activated type I receptor causes epitope-tagged Smad proteins to become completely or predominantly nuclear (Hoodless et al., 1996; Liu et al., 1996). We examined embryos containing six copies of a heat-shock-inducible dpp transgene to determine if endoge-nous MAD was altered in subcellular distribution in response to DPP overexpression.

MAD appears unaffected by the presence of the heat-shock-inducible dpp transgenes at a noninductive temperature (Fig. 6A) and is indistinguishable from the control embryo (Fig. 6F). The midgut endoderm of transgenic embryos stained with anti-MAD sera immediately after a one hour heat-shock display a uniform distribution of MAD throughout the cell, staining both the nucleus and cytoplasm (Fig. 6B). A comparable control embryo shows no change in MAD subcellular localization in response to the heat-shock (Fig. 6G). After a half-hour recovery from heat-shock, the experimental embryo again shows predominantly cytoplasmic anti-MAD staining (Fig. 6C). However, the distinction between nuclear and cytoplas-mic staining is not as pronounced as in the preheat-shock embryo of the same genotype (Fig. 6A) or the comparable control embryo (Fig. 6H). In addition, a single bright spot of anti-MAD staining is visible within each nucleus (Fig. 6C), unlike either the preheat-shock embryo (Fig. 6A) or the post-heat-shock control (Fig. 6H). After a one hour recovery, the experimental embryo shows predominantly cytoplasmic anti-MAD staining (Fig. 6D) and the ratio of nuclear to cytoplas-mic staining is indistinguishable from the preheat-shock condition (Fig 6A) and the comparable control embryo (Fig. 6I). The bright spots of anti-MAD nuclear staining are still visible. After a two hour recovery (Fig. 6E), MAD immuno-localization is indistinguishable from preheat-shock embryos (Fig. 6A) and the comparable control embryo (Fig. 6J). MAD’s subcellular distribution changes dramatically in response to DPP overexpression and upon cessation of the overexpressed signal returns to its presignaling state.

Fig. 6.

DPP overexpression causes dynamic changes in MAD subcellular distribution in the midgut endoderm. (A-E) Midgut endoderm from stage 16 embryos carrying six copies of P{hs-dpp.BP} stained with MAD polyclonal sera. (F-J) Similarly staged and stained embryos without P{hs-dpp.BP}. (A,F) Embryos have not been heat-shocked and MAD appears to be a predominantly cytoplasmic protein in both. (B,G) Embryos stained immediately after a one hour heat-shock (a standard method for expressing transgenes under the control of the hsp70 promoter). (B) MAD appears to be uniformly distributed throughout the cell, staining the nucleus and cytoplasm equally. (G) No change in MAD subcellular distribution. (C,H) Embryos after one-half hour (+.5 hours) room temperature recovery from a one hour heat-shock. (C) Predominantly cytoplasmic anti-MAD staining. However, the distinction between nuclear and cytoplasmic MAD staining is not as pronounced as in the preheat-shock embryo of the same genotype (A) or the comparable y w embryo (H). In addition, bright spots of anti-MAD staining are visible within the nucleus that are not visible in the preheat-shock embryo of the same genotype (A) or the comparable y w embryo (H). (D,I) Embryos after a one hour (+1 hour) room temperature recovery. (D) Predominantly cytoplasmic anti-MAD staining. The ratio of nuclear to cytoplasmic staining is indistinguishable from the preheat-shock embryo of the same genotype (A) and the comparable y w embryo (I). However, the bright spots of nuclear staining are still visible. (E,J) Embryos after a two hour (+2 hours) room temperature recovery. In both of these embryos, MAD appears indistinguishable from preheat-shock embryos.

Fig. 6.

DPP overexpression causes dynamic changes in MAD subcellular distribution in the midgut endoderm. (A-E) Midgut endoderm from stage 16 embryos carrying six copies of P{hs-dpp.BP} stained with MAD polyclonal sera. (F-J) Similarly staged and stained embryos without P{hs-dpp.BP}. (A,F) Embryos have not been heat-shocked and MAD appears to be a predominantly cytoplasmic protein in both. (B,G) Embryos stained immediately after a one hour heat-shock (a standard method for expressing transgenes under the control of the hsp70 promoter). (B) MAD appears to be uniformly distributed throughout the cell, staining the nucleus and cytoplasm equally. (G) No change in MAD subcellular distribution. (C,H) Embryos after one-half hour (+.5 hours) room temperature recovery from a one hour heat-shock. (C) Predominantly cytoplasmic anti-MAD staining. However, the distinction between nuclear and cytoplasmic MAD staining is not as pronounced as in the preheat-shock embryo of the same genotype (A) or the comparable y w embryo (H). In addition, bright spots of anti-MAD staining are visible within the nucleus that are not visible in the preheat-shock embryo of the same genotype (A) or the comparable y w embryo (H). (D,I) Embryos after a one hour (+1 hour) room temperature recovery. (D) Predominantly cytoplasmic anti-MAD staining. The ratio of nuclear to cytoplasmic staining is indistinguishable from the preheat-shock embryo of the same genotype (A) and the comparable y w embryo (I). However, the bright spots of nuclear staining are still visible. (E,J) Embryos after a two hour (+2 hours) room temperature recovery. In both of these embryos, MAD appears indistinguishable from preheat-shock embryos.

No alteration in MAD subcellular distribution detected in response to endogenous DPP

To ascertain if MAD nuclear accumulation occurs in response to endogenous levels of DPP, we analyzed several developmental stages characterized by DPP-dependent events. Based upon our studies of DPP overexpression in the midgut endoderm, we examined additional embryos from the time of midgut constriction formation. Using the constrictions as landmarks for DPP-target cells, we cannot detect MAD nuclear accumulation (data not shown). We also examined blastodermstage embryos and third instar wing imaginal disks. In these stages of development, clearly defined morphogenetic gradients of DPP activity have been documented providing morphological and molecular landmarks for cells exposed to the highest levels of endogenous DPP. For our analysis of blastoderm-stage wild-type embryos, we used confocal microscopy to generate a series of optical sections along the dorsal-ventral axis (Fig. 7A-E). Here the highest level of DPP activity occurs in the dorsalmost cells (Ferguson and Anderson, 1992; Wharton et al., 1993) and one might predict the nuclear accumulation of MAD in these cells. However, MAD subcellular localization revealed no alterations in MAD’s primarily cytoplasmic distribution in any cell type.

Fig. 7.

MAD nuclear accumulation is undetectable in cells responding to endogenous DPP signals in blastoderm embryos. (A-E) A series of optical cross-sections along the dorsal-ventral axis of a blastoderm-stage wild-type embryo stained with anti-MAD polyclonal sera. The highest level of DPP activity at this stage occurs in the dorsalmost cells (Ferguson and Anderson, 1992; Wharton et al., 1993). Thus, the two surfaces of this embryo (A,E) depict cells exposed to maximum and minimum levels of endogenous DPP signals. No alteration in MAD’s primarily cytoplasmic subcellular localization is seen.

Fig. 7.

MAD nuclear accumulation is undetectable in cells responding to endogenous DPP signals in blastoderm embryos. (A-E) A series of optical cross-sections along the dorsal-ventral axis of a blastoderm-stage wild-type embryo stained with anti-MAD polyclonal sera. The highest level of DPP activity at this stage occurs in the dorsalmost cells (Ferguson and Anderson, 1992; Wharton et al., 1993). Thus, the two surfaces of this embryo (A,E) depict cells exposed to maximum and minimum levels of endogenous DPP signals. No alteration in MAD’s primarily cytoplasmic subcellular localization is seen.

We also analyzed early to mid-third instar wing imaginal disks from a strain carrying the dppdisk reporter gene P{BS3.0}. We double-labeled disks with anti-MAD polyclonal sera and for dpp expression. Consistent with our demonstration that MAD is ubiquitously expressed in embryos (Newfeld et al., 1996), MAD is expressed in all cells of the wing disk (Fig. 8A). At this stage, the highest level of DPP activity occurs in the region expressing DPP (Nellen et al., 1996; Lecuit et al., 1996; Singer et al., 1997) and one might expect nuclear accumulation of MAD in the DPP-expressing cells. However, MAD appears to be cytoplasmic in all cells of the disk. Even upon closer examination of double-labeled cells, those expressing DPP and potentially exposed to the highest levels of en-dogenous DPP (Fig. 8B,C), the cytoplasmic subcellular distribution of MAD appears unaltered.

Fig. 8.

MAD nuclear accumulation is undetectable in cells responding to endogenous DPP signals in wing imaginal disks. (A-C) Two different early to mid-third instar wing imaginal disks from a strain carrying the dppdisk reporter gene P{BS3.0}. Disks are stained with anti-MAD polyclonal sera (green) and for dpp expression (red; visualized with anti-β-galactosidase). The highest level of DPP activity at this stage occurs in the region expressing dpp (Nellen et al., 1996; Lecuit et al., 1996; Singer et al., 1997). In the wing disk shown at low magnification (A), note the ubiquitous expression of MAD in all cells of the disk. In the other wing disk, shown at two magnifications (B,C), note the primarily cytoplasmic subcellular localization of MAD. MAD appears cytoplasmic even within and adjacent to cells expressing DPP (yellow), that ought to be receiving the highest levels of endogenous DPP signals.

Fig. 8.

MAD nuclear accumulation is undetectable in cells responding to endogenous DPP signals in wing imaginal disks. (A-C) Two different early to mid-third instar wing imaginal disks from a strain carrying the dppdisk reporter gene P{BS3.0}. Disks are stained with anti-MAD polyclonal sera (green) and for dpp expression (red; visualized with anti-β-galactosidase). The highest level of DPP activity at this stage occurs in the region expressing dpp (Nellen et al., 1996; Lecuit et al., 1996; Singer et al., 1997). In the wing disk shown at low magnification (A), note the ubiquitous expression of MAD in all cells of the disk. In the other wing disk, shown at two magnifications (B,C), note the primarily cytoplasmic subcellular localization of MAD. MAD appears cytoplasmic even within and adjacent to cells expressing DPP (yellow), that ought to be receiving the highest levels of endogenous DPP signals.

Embryonic midgut formation provides a tractable system for placing Mad in the DPP signaling pathway. Unlike Mad’s contribution to DPP signaling in dorsal-ventral patterning of the blastoderm embryo, Mad’s role in midgut development is based purely upon zygotic expression. Thus, Mad null mutant genotypes can be readily analyzed. Also, investigation of one of the roles of visceral mesoderm DPP signaling (induction of lab in the midgut endoderm) permits a clear distinction between signaling cells and target cells.

By examining various aspects of embryonic midgut formation in genetic backgrounds containing a constitutively signaling tkv mutation, we are able to establish a clear epistatic relationship between MAD and the TKV type I DPP receptor. We demonstrate that tkvQ199D cannot induce lab expression in the endoderm, cannot repress Scr or maintain dpp in the visceral mesoderm in Mad null embryos, unambiguously placing MAD downstream of the serine-threonine kinase activity of TKV. Adding a transgene expressing MAD only in the endoderm restores the ability of tkvQ199D to induce lab expression and supplying MAD only in the mesoderm restores tkvQ199D effects on Scr and dpp. These observations place MAD upstream of additional functions necessary for Scr repression, dpp maintenance and lab induction. Taken together, these observations are completely consistent with a role for MAD in mediating signal transduction in DPP-responsive cells.

These results beg the question of how MAD contributes to signal transduction. Is MAD modified during DPP signaling? Consistent with observations in mammalian cell lines (Hoodless et al., 1996; Lechleider et al., 1996; Yingling et al., 1996), in Drosophila cell culture MAD is rapidly phosphory-lated in a ligand-dependent manner. Is MAD’s subcellular dis-tribution altered during DPP signaling? As seen in vertebrate studies (Hoodless et al., 1996; Liu et al., 1996), a heat-shock promoter driven dpp transgene causes a substantial fraction of the normally cytoplasmic MAD protein to accumulate in the nucleus of cells in the midgut endoderm. Recent observations that MAD accumulates in the nucleus of follicle cells in response to ectopic DPP (L. Dobens and L. Raftery, personal communication) suggests that our observations of MAD sub-cellular translocation are not specific to the endoderm but represent a general feature of MAD’s role in DPP signal transduction. Thus, data for MAD are consistent with the current working hypothesis for the function of proteins of the Smad family, developed largely through studies of cultured mammalian cells. Upon activation of the DPP receptor, MAD is in turn activated through phosphorylation and translocates to the nucleus where it serves to regulate directly or indirectly the transcription of target genes (Chen et al., 1996, Macías-Silva et al., 1996). However, it should be noted that, in all of the experiments reporting nuclear accumulation, signals are likely to be at levels far higher than maximal endogenous signaling.

To date, in living animals, we have only observed the nuclear accumulation of MAD in response to DPP overexpression. We have looked extensively for evidence of MAD nuclear accu-mulation in three different tissues known to have active DPP signaling: the blastoderm embryo in which dorsal-ventral fates in the dorsal half of the embryo are being established according to a graded DPP signal, the midgut endoderm during stages when lab induction is occurring and third instar wing imaginal disks in which anterior-posterior fates are being established according to a graded signal emanating from a stripe of DPP expression bisecting the disk. In each case, we observe an identical pattern of MAD subcellular distribution in DPP-responsive and nonresponsive cells: a primarily cytoplasmic distribution. Further, single-labeling experiments were conducted on ovarian follicle cells from a myc-Mad transgenic strain expressing a spatially restricted DPP-dependent enhancer trap. myc-epitope-tagged MAD cannot be detected in nuclei of the follicle cell subpopulation that expresses the MAD-dependent DPP target gene (L. Dobens and L. Raftery, personal communication).

The absence of nuclear MAD under native signaling con-ditions may be explained in two ways. The possibility that we favor is that there is a considerable reservoir of cytoplasmic MAD, even under conditions of maximal endogenous signaling, such that the fraction of MAD translocated to the nucleus is too small to be detected by our techniques. The alter-native, however, is that MAD nuclear accumulation is an artifact of overstimulation of the signaling pathway and that MAD’s biological role is independent of the observed subcellular translocation. This situation needs to be addressed by more sensitive methods such as the development of antisera specific to the phosphorylated form of MAD.

Changes to the subcellular distribution of Smad proteins upon cessation of ligand overexpression have not been examined in vertebrate studies. After withdrawal of DPP overexpression, MAD’s subcellular distribution returns to its presignaling state through a transitional phase characterized by a bright spot within each nucleus. How is the cytoplasmic localization restored? Perhaps MAD is recycled out of the nucleus to the cytoplasm for further use or nuclear MAD is targeted for rapid degradation. Consistent with several features of Mad, we favor the second possibility. First, Mad RNA and protein are ubiquitously expressed (Sekelsky et al., 1995; Newfeld et al., 1996) suggesting ongoing production of MAD in every cell. Further, the MAD polypeptide contains motifs characteristic of rapid protein turnover. In the prolinerich region between the highly conserved N-terminal and C-terminal domains, there are small islands of amino acid identity between subsets of Smad proteins. One of these, a hexaamino acid block (TPPPA/GY: Mad #220-225; Sekelsky et al., 1995) is found in several human and Xenopus Smads (Graff et al., 1996; Zhang et al., 1996) but not in C. elegans (Savage et al., 1996) nor human Smad4 (Hahn et al., 1996). In all Smad proteins that contain this hexaamino acid block, it is located in a potential PEST sequence. PEST sequences are signals for rapid intracellular proteolysis and PEST-FIND scores above 5 are considered significant (Rechsteiner and Rogers, 1996). In MAD, the PEST sequence (amino acid #191-272) has a PEST-FIND score of 6.3. In addition, the PEST sequence in MAD is similar to the PEST sequence which controls signal-independent degradation of unbound Cactus protein in Drosophila (Belvin et al., 1995). Both sequences contain several casein kinase II phosphorylation sites. When the casein kinase II sites of the Cactus mammalian homolog IκBα are mutated, the protein is twice as stable as wild-type (Schwartz et al., 1996).

During the animal’s lifetime, individual cells and their descendants are subjected to multiple rounds of DPP signaling. For example in the wing imaginal disk, DPP signals are required for cell proliferation and/or cell survival (Burke and Basler, 1996; Singer at al., 1997), for anterior/posterior pat-terning (Posakony et al., 1991; Singer at al., 1997) and for vein formation (Sturtevant and Bier, 1995). Perhaps the restoration of MAD’s cytoplasmic localization reflects a DPP-target cell’s return to non-signaling status. This suggests that, in DPP-responsive cells, there is an equilibrium of MAD between the nucleus and cytoplasm. Observations of Smad2 subcellular localization in unstimulated versus ligand-stimulated mammalian cells led Macías-Silva et al. (1996) to propose a similar model of MAD nuclear-cytoplasmic equilibrium. From this perspective, it is tempting to speculate that a gradient of MAD nuclear accumulation is the basis for how individual cells interpret the morphogenetic gradient of DPP which impacts their cell surfaces. Is MAD a direct readout of the DPP gradient such that certain thresholds of nuclear MAD elicit distinct responses or does MAD nuclear accumulation simply relay the level of receptor activation for interpretation by other signaling pathway components? Addressing this question will require additional experiments including the detection of MAD nuclear accumulation in response to endogenous DPP signaling.

Other aspects of DPP signal transduction currently under investigation are whether there is a cytoplasmic sequestering activity holding MAD and its potential transcriptional partners outside the nucleus in the absence of DPP signals and what are the transcriptional partners for MAD? Schnurri (Arora et al., 1995; Grieder at al., 1995; Staehling-Hampton et al., 1995) and Extradenticle (Mann and Abu-Shaar, 1996) are two potential transcriptional partners for MAD. Our laboratory is currently conducting genetic screens that may identify molecules that participate in both of these aspects.

In summary, our data on MAD are consistent with a dynamic role in a DPP-dependent serine-threonine kinase signal trans-duction cascade. An attractive possibility is that one function of MAD is to interpret the morphogenetic gradient of DPP signals. A nuclear-cytoplasmic equilibrium model of MAD subcellular localization provides a mechanistic basis for DPP’s ability to induce multiple changes in a cell’s fate throughout development.

We are very grateful to Rosa Tarng for assistance in subcellular translocation experiments and to Laurel Raftery and Len Dobens for communicating data prior to publication. We thank Peter and Lucy Cherbas for cell lines, Mike O’Connor and Beth Noll for fly stocks and Vickie Rosen for recombinant BMP2. Supported by the NIH (W. M. G.) and the MRC (Canada; J. L. W. and L. A). J. L. W. and L. A. are MRC scholars and A. M. is supported by an MRC studentship.

Arora
,
K.
,
Dai
,
H.
,
Kazuko
,
S.
,
Jamal
,
J.
,
O’Connor
,
M.
,
Letsou
,
A.
and
Warrior
,
R.
(
1995
).
The Drosophila schnurri gene acts in the Dpp/TGF-β pathway and encodes a transcription factor homologous to the human MBP family
.
Cell
81
,
781
790
.
Belvin
,
M.
,
Jin
,
Y.
and
Anderson
,
K.
(
1995
).
Cactus protein degradation mediates Drosophila dorsal-ventral signaling
.
Genes Dev
.
9
,
783
793
.
Bienz
,
M
(
1994
).
Homeotic genes and positional signaling in the Drosophila viscera
.
Trends Genet
.
10
,
22
26
.
Blackman
,
R.
,
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
665
.
Brummel
,
T.
,
Twombly
,
V.
,
Marques
,
G.
,
Wrana
,
J.
,
Newfeld
,
S.
,
Attisano
,
L.
,
Massagué
,
J.
,
O’Connor
,
M.
and
Gelbart
,
W.
(
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
.
Chen
,
X.
,
Rubock
,
M.
and
Whitman
,
M.
(
1996
).
A transcriptional partner for MAD proteins in TGF-β signaling
.
Nature
383
,
691
696
.
Eppert
,
K.
,
Scherer
,
S.
,
Ozcelik
,
H.
,
Pirone
,
R.
,
Hoodless
,
P.
,
Kim
,
H.
,
Tsui
,
L-C.
,
Bapat
,
B.
,
Gallinger
,
S.
,
Andrulis
,
I.
et al. 
., (
1996
).
MADR2 maps to 18q21 and encodes a TGF-β regulated MAD-related protein that is functionally mutated in colorectal carcinoma
.
Cell
86
,
543
552
.
Ferguson
,
E.
and
Anderson
,
K.
(
1992
)
decapentaplegic acts as a morphogen to organize dorsal-ventral pattern in the Drosophila embryo
.
Cell
71
,
451
461
.
Graff
,
J.
,
Bansal
,
A.
and
Melton
,
D.
(
1996
).
Xenopus MAD proteins transduce distinct subsets of signals for the TGF-β family
.
Cell
85
,
479
487
.
Grieder
,
N.
,
Nellen
,
D.
,
Burke
,
R.
,
Basler
,
K.
and
Affolter
,
M.
(
1995
).
schnurri is required for Drosophila dpp signaling and encodes a zinc finger protein similar to the mammalian transcription factor PRDII-BF1
.
Cell
81
,
791
800
.
Hahn
,
S.
,
Schutte
,
M.
,
Hoque
,
A.
,
Moskaluk
,
C.
,
da Costa
,
L.
,
Rozenblum
,
E.
,
Weinstein
,
C.
,
Fischer
,
A.
,
Yeo
,
C.
,
Hruban
,
R.
et al. 
., (
1996
).
DPC-4, candidate tumor suppressor gene at human chromosome 18q21
.
Science
271
,
350
353
.
Hoodless
,
P.
,
Haerry
,
T.
,
Abdollah
,
S.
,
Stapleton
,
M.
,
O’Connor
,
M.
,
Attisano
,
L.
and
Wrana
,
J.
(
1996
).
MADR1, a MAD-related protein that functions in BMP2 signaling pathways
.
Cell
85
,
589
500
.
Hursh
,
D.
,
Padgett
,
R.
and
Gelbart
,
W.
(
1993
).
Cross regulation of decapentaplegic and Ultrabithorax transcription in the embryonic visceral mesoderm of Drosophila
.
Development
117
,
1211
1222
.
Kingsley
,
D.
(
1994
).
The TGF-β superfamily: new members, new receptors, and new genetic tests of function in different organisms
.
Genes Dev
.
8
,
133
146
.
Lechleider
,
R.
,
de Caestecker
,
M.
,
Dehejia
,
A.
,
Polymeropoulos
,
M.
and
Roberts
,
A.
(
1996
).
Serine phosphorylation, chromosomal localization, and TGF-β signal transduction by human bsp-1
.
J. Biol Chem
.
271
,
17617
17620
.
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
.
Letsou
,
A.
,
Arora
,
K.
,
Wrana
,
J.
,
Simin
,
K.
,
Twombly
,
V.
,
Jamal
,
J.
,
Staehling-Hampton
,
K.
,
Hoffmann
,
F.
,
Gelbart
,
W.
,
Massagué
,
J.
et al. 
., (
1995
).
Drosophila DPP signaling is mediated by the punt gene product: a dual ligand-binding Type II receptor of the TGF-β receptor family
.
Cell
80
,
899
908
.
Liu
,
F.
,
Hata
,
A.
,
Baker
,
J.
,
Doody
,
J.
,
Cárcamo
,
J.
,
Harland
,
R.
and
Massagué
,
J.
(
1996
).
A human MAD protein acting as a BMP-regulated transcriptional activator
.
Nature
381
,
620
623
.
Macías-Silva
,
M.
,
Abdollah
,
S.
,
Hoodless
,
P.
,
Pirone
,
R.
,
Attisano
,
L.
and
Wrana
,
J.
(
1996
)
MADR2 is a substrate of the TGF-β receptor and its phosphorylation is required for nuclear accumulation and signaling
.
Cell
87
,
1215
1224
.
Mann
,
R.
and
Abu-Shaar
,
M.
(
1996
).
Nuclear import of the homeodomain protein Extradenticle in response to Wg and Dpp signaling
.
Nature
383
,
630
633
.
Massagué
,
J.
(
1996
).
TGF-β signaling: receptors, transducers and MAD proteins
.
Cell
85
,
947
950
.
Massagué
,
J.
,
Attisano
,
L.
and
Wrana
,
J.
(
1994
).
The TGF-β family and its composite receptors
.
Trends Cell Biol
.
4
,
172
178
.
Nellen
,
D.
,
Affolter
,
M.
and
Basler
,
K.
(
1994
).
Receptor serine/threonine kinases implicated in the control of Drosophila body pattern by dpp
.
Cell
78
,
225
237
.
Nellen
,
D.
,
Burke
,
R.
,
Struhl
,
G.
and
Basler
,
K.
(
1996
).
Direct and long-range action of a DPP morphogen gradient
.
Cell
85
,
357
368
.
Newfeld
,
S.
,
Chartoff
,
E.
,
Graff
,
J.
,
Melton
,
D.
and
Gelbart
,
W.
(
1996
).
Mothers against dpp encodes a conserved cytoplasmic protein required in DPP/TGF-β responsive cells
.
Development
122
,
2099
2108
.
Padgett
,
R.
,
St. Johnston
,
R.
and
Gelbart
,
W.
(
1987
).
A transcript from a Drosophila pattern gene predicts a protein homologous to the TGF-β family
.
Nature
325
,
81
84
.
Padgett
,
R.
,
Wozney
,
J.
and
Gelbart
,
W.
(
1993
).
Human BMP sequences can confer normal dorsal-ventral patterning in the Drosophila embryo
.
Proc. Natl. Acad. Sci. USA
90
,
2905
2909
.
Penton
,
A.
,
Chen
,
Y.
,
Staehling-Hampton
,
K.
,
Wrana
,
J.
,
Attisano
,
L.
,
Szidonya
,
J.
,
Cassill
,
J.
,
Massagué
,
J.
and
Hoffmann
,
F.
(
1994
).
Identification of two BMP type I receptors in Drosophila and evidence that Brk25D is a decapentaplegic receptor
.
Cell
78
,
239
250
.
Posakony
,
L.
,
Raftery
,
L.
and
Gelbart
,
W.
(
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 the decapentaplegic pathway in Drosophila
.
Genetics
139
,
241
254
.
Rechsteiner
,
M.
and
Rodgers
,
S.
(
1996
).
PEST sequences and regulation by proteolysis
.
Trends Biochem. Sci
21
,
267
271
.
Ruberte
,
E.
,
Marty
,
T.
,
Nellen
,
D.
,
Affolter
,
M.
and
Basler
,
K.
(
1995
).
An absolute requirement for both the type II and type I receptors, punt and thick veins for DPP signaling in vivo
.
Cell
80
,
889
897
.
Sampath
,
T.
,
Rashka
,
K.
,
Doctor
,
J.
,
Tucker
,
R.
and
Hoffmann
,
F.
(
1993
).
Drosophila TGF-β superfamily proteins induce endochondrial bone formation in mammals
.
Proc. Natl. Acad. Sci. USA
90
,
6004
6008
.
Savage
,
C.
,
Das
,
P.
,
Finelli
,
A.
,
Townsend
,
S.
,
Sun
,
C.
,
Baird
,
S.
and
Padgett
,
R.
(
1996
).
C. elegans genes sma-2, sma-3, and sma-4 genes define a conserved family of TGF-β pathway components
.
Proc. Natl. Acad. Sci. USA
93
,
790
794
.
Schwartz
,
E.
,
Van Antwerp
,
D.
and
Verma
,
I.
(
1996
).
Constitutive phosphorylation of IκBα by casein kinase II occurs preferentially on serine 293: requirement for degradation of free IκBα
.
Mol. Cell Biol
.
16
,
3554
3559
.
Sekelsky
,
J.
,
Newfeld
,
S.
,
Raftery
,
L.
,
Chartoff
,
E.
and
Gelbart
,
W.
(
1995
).
Genetic characterization and cloning of Mothers against dpp, a gene required for decapentaplegic function in Drosophila melanogaster
.
Genetics
139
,
1347
1358
.
Singer
,
M.
,
Penton
,
A.
,
Twombly
,
V.
,
Hoffmann
,
F.
, and
Gelbart
,
W.
(
1997
).
Signaling through both type I DPP receptors is required for anterior-posterior patterning of the entire Drosophila wing
.
Development
124
,
79
89
.
Staehling-Hampton
,
K.
and
Hoffmann
,
F.
(
1994
).
Ectopic decapentaplegic in the Drosophila midgut alters the expression of five homeotic genes, dpp and wingless causing specific morphological defects
.
Dev. Biol
.
164
,
502
512
.
Staehling-Hampton
,
K.
,
Laughon
,
A.
and
Hoffmann
,
F.
(
1995
).
A Drosophila protein related to the human zinc finger transcription factor PRDII/MBP/HIV-EP1 is required for dpp signaling
.
Development
121
,
3393
3403
.
Sturtevant
,
M.
and
Bier
,
E.
(
1995
)
Analysis of the genetic hierarchy guiding wing vein development in Drosophila
.
Development
121
,
785
801
.
Ui
,
K.
,
Ueda
,
R.
and
Miyake
,
T.
(
1987
).
Cell lines from imaginal disks of Drosophila melanogaster
.
In Vitro Cell. Dev. Biol
.
23
,
707
711
.
Wharton
,
K.
,
Ray
,
R.
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
.
Wiersdorff
,
V.
,
Lecuit
,
T.
,
Cohen
,
S.
and
Mlodzik
,
M.
(
1996
).
Mad acts downstream of DPP receptors, revealing a differential requirement for dpp signaling in initiation and propagation of morphogenesis in the Drosophila eye
.
Development
122
,
2153
2162
.
Wrana
,
J.
,
Attisano
,
L.
,
Wieser
,
R.
,
Ventura
,
F.
and
Massagué
,
J.
(
1994
).
Mechanism of activation of the TGF-β receptor
.
Nature
370
,
341
347
.
Wrana
,
J.
and
Attisano
,
L.
(
1996
).
MAD-related proteins in TGF-β signaling
.
Trends Genetics
12
,
493
496
.
Xie
,
T.
,
Finelli
,
A.
and
Padgett
,
R.
(
1994
).
The Drosophila saxophone gene: a serine-threonine kinase receptor of the TGF-β superfamily
.
Science
263
,
1756
1759
.
Yingling
,
J.
,
Das
,
P.
,
Savage
,
C.
,
Zhang
,
M.
,
Padgett
,
R.
and
Wang
,
X.
(
1996
).
Mammalian dwarfins are phosphorylated in response to TGF-β and are implicated in control of cell growth
.
Proc. Natl. Acad. Sci USA
93
,
8940
8944
.
Zhang
,
Y.
,
Feng
,
X.
,
Wu
,
R.
and
Derynck
,
R.
(
1996
).
Receptor-associated Mad homologues synergize as effectors of the TGF-β response
.
Nature
383
,
168
171
.