Genetic Analysis of Punt, a Type II Dpp Receptor That Functions Throughout the Drosophila melanogaster Life Cycle

Corresponding author: Anthea Letsou, Department of Human Genetics, 15 N 2030 E RM 2100, University of Utah, Salt Lake City, UT 84112-5330 E-mail: anthea.letsou@genetics.utah.edu.

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

TGF-ß- (transforming growth factor-ß-) mediated signal transduction affects growth and patterning in a variety of organisms. Here we report a genetic characterization of the Drosophila punt gene that encodes a type II serine/threonine kinase TGF-ß/Dpp (Decapentaplegic) receptor. Although the punt gene was originally identified based on its requirement for embryonic dorsal closure, we have documented multiple periods of punt activity throughout the Drosophila life cycle. We demonstrate that potentially related embryonic punt phenotypes, defects in dorsoventral patterning and dorsal closure, correspond to distinct maternal and zygotic requirements for punt. In addition, we document postembryonic requirements for punt activity. The tight correspondence between both embryonic and postembryonic loss-of-function punt and dpp phenotypes implicates a role for Punt in mediating virtually all Dpp signaling events in Drosophila. Finally, our comparison of punt homoallelic and heteroallelic phenotypes provides direct evidence for interallelic complementation. Taken together, these results suggest that the Punt protein functions as a dimer or higher order multimer throughout the Drosophila life cycle.

CYTOKINES of the TGF-ß superfamily evoke a wide range of eukaryotic developmental and physiological responses. These include modulation of cell growth and proliferation, regulation of immune and endocrine function, and control of axial patterning (for review, see MASSAGUE 1996 ). Based on similarities in biological function and sequence, the members of the TGF-ß superfamily have been divided into three subgroups: TGF-ßs, activins, and Dpp/BMPs (bone morphogenetic proteins) (for review, see KINGSLEY 1994 ).

The diversity in responses elicited by the various members of the TGF-ß superfamily derives, at least in part, from heterogeneity in TGF-ß receptor complexes (MATHEWS and VALE 1991 ; ATTISANO et al. 1992 ). Biochemical studies in vitro (WRANA et al. 1994 ) indicate that a dimeric ligand from either the TGF-ß or the activin subgroup forms a complex with two types of transmembrane serine/threonine kinase receptor, the type I and type II receptors. The type II receptor, a constitutively active kinase, is the primary determinant in TGF-ß and activin subgroup binding. Once the type II receptor has bound ligand, it complexes with a type I receptor kinase and activates it by serine/threonine phosphorylation. By virtue of its interaction with Smad proteins, the type I receptor then transduces the TGF-ß signal intracellularly (SEKELSKY et al. 1995 ; MACIAS-SILVA et al. 1996 ; KIM et al. 1997 ).

Studies of the Drosophila melanogaster type II receptor encoded by punt (put) indicated that signaling by the third subgroup of TGF-ß ligands, the Dpp/BMPs, differs only slightly from the paradigm described above: for cytokines in this subgroup, the type II receptor is not the primary determinant in ligand binding. Moreover, these studies revealed dual-ligand specificity by type II receptors to represent yet another mechanism by which TGF-ß superfamily members could elicit diverse cellular responses. Although Punt was characterized originally as an activin receptor in assays of function and sequence homology (CHILDS et al. 1993 ), we observed that Punt binds Dpp/BMP-type ligands in vivo when a type I Dpp/BMP receptor is coexpressed (LETSOU et al. 1995 ). This finding led to our hypothesis that Punt has dual-ligand specificity: Punt binds activin ligands directly and Dpp/BMP ligands in combination with a type I receptor. Consistent with our model for indirect ligand selection by type II Dpp/BMP receptors are two additional observations: (1) binding of BMP 4 by the mammalian type II receptor BRK-3 (BMP Receptor Kinase-3) is enhanced in cultured cells when a type I BMP receptor is coexpressed (NOHNO et al. 1995 ), and (2) binding of BMPs 2 and 7 by the mammalian type II receptor BMPR-II (BMP Receptor Type II) in cultured cells is enhanced by coexpression of various type I BMP or activin receptors (LIU et al. 1995 ).

Three Drosophila gene products [Dpp, 60A, and Screw (Scw)] belong to the Dpp/BMP family of TGF-ßs (PADGETT et al. 1987 ; WHARTON et al. 1991 ; ARORA et al. 1994 ) and can potentially serve as Punt ligands in vivo. The function of Dpp has been well characterized both molecularly and genetically. During oogenesis, the dpp gene product is required for patterning the anterior eggshell (TWOMBLY et al. 1996 ). During embryonic dorsoventral axis formation, Dpp acts as a morphogen; increasing concentrations of the dpp gene product result in the formation of more dorsal structures (FERGUSON and ANDERSON 1992 ; WHARTON et al. 1993 ). Later in embryogenesis, the dpp gene product is required for dorsal closure and for normal constriction of the midgut (IMMERGLUCK et al. 1990 ; PANGANIBAN et al. 1990 ). In addition, Dpp functions postembryonically as a long-range morphogen during imaginal disc development and patterning of the visual centers of the developing brain (HOFFMANN 1991 ; KAPHINGST and KUNES 1994 ; LECUIT et al. 1996 ; NELLEN et al. 1996 ). Specifically, mutations in dpp result in wing, leg, eye, notal, antennal, genital, and anal defects (SPENCER et al. 1982 ). Far less is known of scw and 60A gene functions. Like mutations in dpp, scw mutations disrupt dorsoventral patterning (ARORA et al. 1994 ). The third ligand, 60A, is less well characterized and has yet to be assigned a function.

A definitive role for Punt in mediating certain aspects of Dpp signaling has been documented. Initial insight into the punt gene's function was obtained in analyses of the original punt mutations, punt¹³⁵ and punt ^P1, which were identified in genetic screens for embryonic lethals affecting cuticular pattern. The single ethylmethane sulfate- (EMS-) induced allele, punt¹³⁵, is a recessive lethal that disrupts embryonic processes of dorsal closure and gut development (JURGENS et al. 1984 ; LETSOU et al. 1995 ; RUBERTE et al. 1995 ). The second punt allele, punt ^P1, was identified in a P-element mutagenesis screen and similarly affects embryogenesis; its recessive lethal phenotype is indistinguishable from that of punt¹³⁵ (LETSOU et al. 1995 ; RUBERTE et al. 1995 ). Postembryonic punt functions have been identified in direct assays for function. Both punt¹³⁵ and punt ^P1 exhibit a maternal-effect ventralizing phenotype (LETSOU et al. 1995 ; RUBERTE et al. 1995 ). Whereas embryos derived from wild-type mothers secrete a cuticle that bears distinctive dorsal and ventral pattern elements, embryos derived from mothers homozygous for temperature-sensitive alleles of punt and embryos derived from mothers harboring punt germline clones are severely ventralized and secrete a cuticle that is circumscribed by ventral denticle belts. Finally, clonal analyses of punt function in imaginal discs demonstrated additional roles for punt in postembryonic patterning events that are mediated by dpp (BURKE and BASLER 1996 ; PENTON and HOFFMANN 1996 ).

In the current article we describe genetic studies that were designed to assess requirements for the punt-encoded type II receptor throughout the Drosophila life cycle. First, we generated new punt alleles. These alleles constitute an allelic series, exhibiting phenotypes of varying severities that range from full viability to embryonic lethality. The punt gene is pleiotropic, and our observation that all punt phenotypes were characterized originally as dpp phenotypes (SPENCER et al. 1982 ) is particularly notable. The tight correspondence between dpp and punt mutant phenotypes prompts us to suggest that (1) Dpp repeatedly functions as the Punt ligand throughout the Drosophila life cycle, and (2) the punt-encoded type II receptor is sufficient to fulfill most type II receptor requirements for Dpp signaling. Second, we exploited the temperature sensitivity of punt alleles to investigate temporal requirements for Punt function. These analyses resolved a long-standing question: whether or not defects in dorsal closure are primary or secondary consequences of mutations in punt. Our studies indicated that rather than being secondary to earlier defects in embryonic dorsoventral patterning, defects in embryonic dorsal closure are a primary consequence of punt mutations. Finally, our genetic interaction studies revealed that the bioactive form of Punt is multimeric.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly maintenance and stocks:
Balancers, marker mutations, and dorsal-open class mutations, including punt¹³⁵, punt ^P1, and tkv⁸ have been described ( JURGENS et al. 1984 ; NUSSLEIN-VOLHARD et al. 1984 ; LINDSLEY and ZIMM 1992 ; LETSOU et al. 1995 ; RUBERTE et al. 1995 ). Flies were maintained on cornmeal/molasses/agar medium at room temperature (22°) unless otherwise stated.

P-element excision:
The punt alleles punt⁵¹, punt ⁶², punt ⁸⁸, and punt¹³⁶ were isolated in a screen for lethal excisions of the PZ transposon p[lacZ ry+] in punt^P1 heterozygotes. The rosy+ (ry+) -marked P element was mobilized by an external transposase source (ROBERTSON et al. 1988 ). Fly lines in which the P-element sequences had been excised were identified by their rosy eye color. The new punt alleles were identified in rosy lines failing to complement the dorsal closure defect of the punt ^P1 mutation. One punt allele (punt¹⁰) was identified in the collection of excision lines that complemented the embryonic lethal punt ^P1 phenotype. Two additional adult-viable punt alleles, punt ²⁴ and punt⁹⁷, were isolated in a second PZ mobilization screen. In this screen of 75 fly lines, imprecise excisions of an unmarked (ry) P element in punt¹³⁶ heterozygotes were identified by polymerase chain reaction (PCR) analyses using punt primers P1 (5' GGGCCTGTTTTCAAGC GAT 3') and P2 (5' GGAATTCATTGTCTCACTACCAGCC 3') to amplify genomic DNA fragments spanning the original transposon insertion site.

Gamma- (-) irradiation:
-radiation was employed to generate deficiencies spanning the punt locus. punt ^P1 ry/TM3, Sb Ser males were exposed to 4000 rads of gamma radiation from a ¹³⁷Cs source. Irradiated males were mated to CxD, ry/MKRS virgin females, and rosy progeny were identified. Deficiencies mapping to the endogenous ry locus were distinguished from those mapping to punt by an analysis of bristle and wing markers.

Temperature shift manipulations:
To determine the temperature-critical period for embryonic Punt function, adult flies were placed in laying blocks at either the permissive (18°) or the restrictive (25°) temperature and induced to lay eggs on grape-juice agar plates supplemented with fresh yeast paste. Since the total duration of Drosophila development is a function of temperature, all experimental time intervals were standardized and expressed as 25° standard time intervals (x 1.75 for 18° and x 1 for 25°; POWSNER 1935 ). Embryos were collected for 2-hr standard time intervals and then transferred to freshly yeasted vials. Embryonic development proceeded initially at the collection temperature and was followed by the indicated shift. After a 24-hr standard time interval, all animals completed development at 18°. To assay postembryonic Punt function, animals underwent embryogenesis at 18° and were shifted up to the restrictive temperature (25°) after a 24-hr standard time interval.

Phenotypic analyses:
Cuticular phenotypes were examined using the Hoyer's mount technique (VAN DER MEER 1977 ). Three classes of dorsal-open, embryonic lethal phenotype were distinguished. Embryos exhibiting single, large holes on their dorsal surface were scored as strongly defective in the process of dorsal closure. Embryos displaying defects that include a tail-up phenotype, increased curvature in abdominal segments, and noticeably reduced dorsal cuticle were scored as moderately defective. Embryos exhibiting more subtle or no visible defects were scored as weakly defective. Postembryonic punt phenotypes were scored in viable and pharate adults, as well as in prepupae. We distinguished prepupae from pupae by scoring for the occurrence of head eversion. Prepupae and pupae were dissected from pupal cases after boiling for 2 min in water. Due to the fragility of the structure, we were unable to examine wing phenotypes at this developmental stage. Both phase and scanning electron microscopy (SEM) were employed to image adult and pupal phenotypes. For SEM, samples were air-dried for 24 hr, attached to mounting stubs with conducting glue, and sputter-coated (Desk-1, Denton, Cherry Hill, NJ) with a ⁶⁰/₄₀ gold/palladium alloy. Micrographs were examined using a S-450 scanning electron microscope (Hitachi, Mountain View, CA) at a working distance of 15 mm at an accelerating voltage of 15 kV.

Molecular analyses:
Molecular lesions in punt alleles were identified by Southern hybridization, PCR, and DNA sequence analyses. Southern hybridization analyses were performed according to published procedures (SAMBROOK et al. 1989 ). Probes included 1) a 2 kb EcoRI genomic fragment that spans the transposon insertion site and 2) pBS5'P (the generous gift of D. MCKEARIN, Southwestern Medical Center, Dallas) that corresponds to 540 bp in the 5'P sequence (RUBIN and SPRADLING 1983 ). The size of insertions in punt¹⁰, punt²⁴, and punt⁹⁷ was determined in PCR analyses using punt primers P1 and P2 to amplify genomic DNA fragments spanning the transposon insertion site. The precise lesions in punt¹⁰, punt²⁴, and punt⁹⁷ were determined by DNA sequence analysis. Briefly, primers P1 and P2 were employed in a PCR reaction using genomic DNA from punt homozygotes as template. The double-stranded, amplified product was gel purified and sequenced on both strands with a DNA sequencer model 373 (Applied Biosystems, Foster City, CA). Sequence reactions were primed with oligos P1 and P2 and terminated with dye-labeled dideoxynucleotides.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

punt mutations disrupt dorsal closure:
Wild-type embryos undergo dorsal closure between 8 and 12 hr after egg lay (AEL), corresponding to embryonic stages 12–15 (CAMPOS-ORTEGA and HARTENSTEIN 1985 ). During this process, epidermal cells change their shape [becoming rectangular in dorsal positions and remaining polygonal in more lateral positions (YOUNG et al. 1993 )], and lose their continuity with the extraembryonic membrane, the amnioserosa. Epidermis and amnioserosa overlap transiently, the latter is incorporated into the embryo, and closure is complete when the edges of the epidermal sheet meet and fuse at the dorsal midline. Since epidermal cells secrete cuticle, wild-type cuticle completely encases the animal and displays characteristic pattern elements (Figure 1A). In contrast, cuticles derived from well-characterized dorsal closure mutants, punt^P1 or punt¹³⁵ homozygotes (JÜR-GENS et al. 1984; LETSOU et al. 1995 ; RUBERTE et al. 1995 ) are grossly abnormal as evidenced by a single, large hole in the dorsal cuticle (Figure 1C). Embryos lacking the type I Dpp receptor encoded by the thick veins (tkv) gene display the same dorsal-open phenotype (Figure 1B).

View larger version (143K): In this window In a new window Download PPT slide   Figure 1. Dark-field images of cuticles from wild-type and mutant embryos raised at 25°. The cuticle of a wild-type embryo (A) displays characteristic pattern elements. In contrast, cuticles from embryos homozygous for certain alleles of tkv, for example, tkv8 (B), and punt, for example, punt P1 (C) and punt 62 (D) can exhibit severe head defects and large holes in the dorsal hypoderm. punt phenotypes, however, are variably expressed. In contrast to the severe defects seen in (C) and (D), moderately affected cuticles from punt 62/punt 62 embryos exhibit the tail-up phenotype, as well as an increase in the curvature of abdominal segments and a reduction in the dorsal hypoderm (E). Weakly affected punt 62/punt 62 cuticles complete dorsal closure, exhibiting only subtle defects in dorsal cuticle (F). In all panels, dorsal is up, and anterior is to the left.

We generated several new embryonic-lethal punt alleles after mobilization of the P transposon in punt ^P1. In this report, we show that the punt⁵¹/punt⁵¹, punt ⁶²/punt ⁶², and punt ⁸⁸/punt ⁸⁸ homozygous embryonic lethal phenotypes resemble those of punt¹³⁵/punt¹³⁵ and punt ^P1/punt ^P1; each is fully penetrant but variably expressed at all temperatures tested (18°, 25°, and 29°) (Table 1). As an example, at 25°, 17% of punt ⁶² homozy-gotes exhibited the strong dorsal-open phenotype that has been described previously (Figure 1D). Forty-two percent of mutant embryos exhibited an intermediate dorsal-open phenotype, displaying a tail-up phenotype, an increased curvature in abdominal segments, and a noticeably reduced dorsal cuticle (Figure 1E). The remaining 42% of punt ⁶² homozygotes completed the process of dorsal closure and exhibited only minor defects in their dorsal cuticle (Figure 1F). Additional analyses of cuticles derived from all homoallelic combinations of punt alleles revealed the embryonic lethal punt phenotypes to be temperature-sensitive: in all cases the strong dorsal-open punt phenotype was more prevalent at higher temperatures (Table 1).

punt mutations disrupt wing patterning:
We similarly employed transposon mobilization to generate three new adult-viable punt alleles. We had shown previously that in contrast to homoallelic combinations, certain heteroallelic combinations of punt were fully viable at low temperatures. Because ~ 10% of these viable punt/punt adults exhibited defects in wing venation (LETSOU et al. 1995 ), we reasoned that weak hypomorphic punt alleles might be recovered in screens of viable excisions of the mutagenic transposon in punt ^P1. Using wing venation as a criterion for Punt function, we screened our collection of viable excision homozygotes for weak alleles of punt. No visible defects were observed in these fly lines (n > 500 flies/line). One weak punt allele, punt¹⁰, however, was identified in screens of viable alleles in trans to the embryonic lethal allele punt ^P1; although fully viable at 25°, 28% of punt¹⁰/punt ^P1 adults exhibited a defect in wing venation (Table 2A; Figure 2, A and B). The defect, ectopic venation stemming from the posterior cross-vein into the second posterior cell, was recovered as either a unilateral or a bilateral condition. Furthermore, we found punt¹⁰ to be a cold-sensitive allele. Whereas punt¹⁰ homo-zygotes were fully viable at 25°, they exhibited a semilethal phenotype at 18° (P = 0.00015; Table 2, Table 2F and Table 2G). Our molecular characterization of the lesion in punt¹⁰ (see below), as well as our failure to detect cold-sensitive lethality in fly lines in which the P element has excised precisely (data not shown), lead us to conclude that cold-sensitive lethality is a direct consequence of mutation at the punt locus.

View larger version (43K): In this window In a new window Download PPT slide   Figure 2. Wing venation defects in punt mutants. Wild-type wing venation (A) is compared to that in punt10/punt62 mutants (B). Twenty-eight percent of punt10/punt 62 animals exhibit an ectopic vein (arrow) stemming from the posterior cross-vein. In panel (C), the wild-type and punt10 sequences are compared. Imprecise excision of a P-element transposon left a 38-bp insertion in the 5' UTR of punt10. Inserted sequences include: 1) 15-bp inverted repeats (open arrows) that are vestiges of the parental transposon's 31-bp inverted repeats, and 2) an 8-bp sequence (solid arrow) that accompanied the insertion of the parental transposon.

Two additional adult-viable alleles, punt ²⁴ and punt⁹⁷, were recovered in an independent screen for excisions of an unmarked P element in punt¹³⁶ heterozygotes. Molecular characterization of the lesions associated with these alleles revealed an obvious relationship to that in punt¹⁰ (see below); hence, only punt¹⁰ was characterized genetically.

To place the punt alleles in an allelic series, we examined the phenotypes of punt⁵¹/punt¹⁰, punt ⁶²/punt¹⁰, punt ⁸⁸/punt¹⁰, punt¹³⁵/punt¹⁰, and punt ^P1/punt¹⁰ hetero-allelic combinations. Although defects in venation were observed in most classes of punt homozygotes, these defects were recovered at high frequencies only in pun ⁶²/punt¹⁰ and punt ^P1/punt¹⁰ homozygotes (P < 0.00001; Table 2, A–E). In particular, 32% of punt ⁶²/punt¹⁰ adults exhibited defects in wing venation identical to those already identified in punt ^P1/punt¹⁰ animals. Taken together, our phenotypic analyses of punt mutations allowed us to order three classes of punt alleles with respect to decreasing severity as follows: Class I [punt ^P1 = punt ⁶²] > Class II [punt⁵¹ = punt⁸⁸ = punt¹³⁵] > Class III [punt¹⁰ = punt ²⁴ = punt⁹⁷].

punt allele strength is correlated with insertion length:
We had determined previously that the P transposon insertion site in punt ^P1 mapped to the 5' untranslated region (UTR) of the punt gene, 2 bp from its 5' end (LETSOU et al. 1995 ). To examine the molecular nature of the punt mutations that arose after transposon mobilization, we employed Southern hybridization and PCR analyses. These studies revealed a correlation between the molecular and genetic natures of the mutations (Table 3). All embryonic lethal punt alleles arising from mobilization retained P-element sequences in the 5' UTR of the punt gene. Southern analyses revealed that more than 3 kb remained at the insertion site in punt⁵¹ and punt ⁶². PCR analyses revealed the insertion in punt ⁸⁸ to be considerably smaller, ~420 bp in length. The smallest insertions mapped to the 5' UTR of the three adult-viable punt alleles (punt¹⁰, punt ²⁴, and punt ⁹⁷). To precisely define these lesions, we isolated a PCR fragment that contained the insertion and sequenced it. In punt¹⁰ and punt ²⁴, two independently derived alleles, we identified identical 15 bp inverted repeats between the 8-bp target site duplication (Figure 2C). In punt⁹⁷, the inverted repeat unit was increased in length by a single base pair.

The punt locus is haplo-insufficient:
Because punt maps to a region of the Drosophila genome for which a deficiency has never been recovered, we screened genetically for deletions of the locus. First, we screened 200 rosy lines that were generated by mobilization of the ry⁺-bearing P transposon in punt ^P1 and identified no deletions that extend into the punt locus. Second, we irradiated punt ^P1 ry/++ males and screened 12,641 punt ^P1* ry/+ ry progeny for loss of the ry⁺ eye color marker carried by the mutagenic transposon. Although we recovered deletions of the endogenous ry⁺ gene that is carried by the balancer chromosome (4 ry mutations in 2921 irradiated ry⁺ chromosomes), we recovered no ry deletions at the punt locus in viable adults (P = 2.4 x 10^-5). Intriguingly, this mutagenesis protocol produced two rosy inviable adults; these animals failed to emerge from their pupal cases and exhibited an array of phenotypes, including medial notal clefts and eye and leg defects, which we have documented in viable punt homozygotes (see below). Although not proven through molecular analyses, we believe that these pharate adults display the haplo-insufficient punt phenotype.

Dorsal closure defects are a primary consequence of mutations at the punt locus:
We showed previously that heteroallelic combinations of either punt⁵¹ or punt ⁸⁸ with punt¹³⁵ result in a temperature-sensitive embryonic lethality. Whereas punt⁵¹/punt¹³⁵ and punt ⁸⁸/punt¹³⁵ animals abort development midway through embryogenesis at 25° due to defects in dorsal closure, adult-viable punt⁵¹/punt¹³⁵ and punt ⁸⁸/punt¹³⁵ homozygotes are recovered at 18° at the expected Mendelian frequency (LETSOU et al. 1995 ). We exploited this temperature sensitivity of punt heteroalleles to determine the temperature-critical period for embryonic punt gene function. punt ⁸⁸/punt¹³⁵ embryos were generated in matings of punt ⁸⁸/+ and punt¹³⁵/+ heterozygotes at 25°. Next, F₁ embryos were shifted from the restrictive temperature (25°) down to the permissive temperature (18°) at discrete times, either 2, 8, 10, 12, or 24 hr AEL. Viable punt ⁸⁸/punt¹³⁵ adults were recovered when downshifts were performed at 2 or 8 hr AEL, but no punt homozygotes were recovered in later shifts at 10, 12, or 24 hr AEL (Figure 3, A–E). These results indicate that zygotic Punt is not essential during the first 8 hr of embryogenesis. In a reciprocal set of upshift experiments, we produced viable punt ⁸⁸/punt¹³⁵ adults when shifts from 18° to 25° were performed at 12 or 24 hr AEL, but not in earlier shifts at 2 or 8 hr AEL (Figure 3, F–I). These results indicate that zygotic Punt is essential for development prior to 12 hr AEL. Taken together, results from the temperature-shift experiments demonstrate an absolute requirement for zygotic Punt between 8 and 12 hr of embryogenesis, coincident with the process of dorsal closure and well after dorsoventral cell fate has been established.

View larger version (11K): In this window In a new window Download PPT slide   Figure 3. Temperature-critical period determination for punt. In the first experiment (A–E), embryonic development was initiated at the restrictive temperature (25°, bar), and at successive embryonic time points, including 2, 8, 10, and 12 hr AEL, embryos were moved to the permissive temperature (18°, horizontal line). In the second experiment (F–I), embryonic development was initiated at the permissive temperature, and at successive embryonic time points, including 2, 8, and 12 hr AEL, embryos were moved to the restrictive temperature. In all trials, postembryonic development was completed at 18°. Because the rate of Drosophila development is reduced as temperature decreases, time intervals spent at 18° were multiplied by 1.75, the equivalency factor of POWSNER 1935 . Data are thus integrated into a single schematization with shift points indicated as 25° standard. Dashed vertical lines delimit the punt temperature-critical period. In adjacent columns, the number of viable punt homozygotes that survived to adulthood, the total number of adult progeny that were produced in each experiment, as well as the percentage of expected punt homozygotes, are indicated.

As a further test of our hypothesis that Punt functions in midembryogenesis to effect dorsal closure, we examined the mutant cuticular phenotype of punt homozygotes in both 8-hr upshift and 12-hr downshift experiments. Our observation that all mutant embryos exhibited an intermediate dorsal-open phenotype (data not shown) is consistent with our hypothesis that Punt plays a primary role in the process of dorsal closure.

Post-embryonic punt phenotypes mimic dpp phenotypes:
Temperature-shift manipulations were also employed to investigate postembryonic requirements for Punt. As we had shown previously to be the case for two insertion alleles of punt (punt⁵¹ and punt ⁸⁸), the heteroallelic combination of punt ⁶² with punt¹³⁵ results in a temperature-sensitive embryonic lethality due to defects in dorsal closure: punt ⁶²/punt¹³⁵ homozygotes are inviable at 25° and fully viable at 18° (data not shown). The progeny of punt ⁶²/+ and punt¹³⁵/+ heterozygotes were allowed to complete embryogenesis at the permissive temperature, 18°. At 24-hr standard AEL, animals were shifted to the restrictive temperature, and thus larval and pupal development proceeded to completion at 25°. punt ⁶²/punt¹³⁵ homozygotes generated by this protocol exhibited a marked reduction in viability. In a cross expected to produce 156 punt ⁶²/punt¹³⁵ homozygotes, representing 25% of the total progeny, we recovered only eight homozygotes, corresponding to less than 2% of the total progeny. Further examination of all developmental stages revealed that punt-dependent lethality occurred postembryonically; 131 of the missing punt homozygotes were identified as dead prepupae (36) and pupae (95). Lethal prepupae exhibited defects in head eversion (data not shown).

To more precisely examine the cause of death in punt homozygotes, we dissected the 131 dead animals from their pupal cases. The presumed punt ⁶²/punt¹³⁵ genotype of dissected animals was confirmed by analysis of bristle markers. All punt homozygotes, including the eight fully viable adults, were grossly deformed (Table 4). All punt homozygotes exhibited notal defects; these were almost exclusively medial notal clefts (Table 4 and Figure 4). Ninety-nine percent of punt homozygotes exhibited leg defects, including truncations, bifurcations, and abnormal twists (Table 4 and Figure 5). Distal pattern elements were deleted in at least one limb in 126 of 127 animals examined. All six legs were rarely affected to the same degree, and defects could be unilateral or bilateral with respect to each pair of legs. Although the posterior leg pair was more frequently affected than were the anterior pairs, duplicated sex combs were readily discernible on several of the mutant's forelegs (data not shown). Sex comb duplication is a hallmark of ventralization of patterning in imaginal discs (THEISEN et al. 1996 ). A very large fraction (97%) of punt homozygotes exhibited gross eye and antennal defects (Table 4 and Figure 6). The eyes of punt homozygotes were highly disorganized and showed striking reductions in ommatidial number. Antennal defects, like leg defects, included deletions and duplications of distal pattern elements. Both eye and antennal defects could be unilateral or bilateral. None of these defects were identified in the 469 wild-type siblings that were produced by the initial mating of punt heterozygotes. With one exception, the bifurcation of legs in punt mutant animals, all of these defects were identified in dpp mutant animals (SPENCER et al. 1982 ). With respect to the single, apparantly unique punt phenotype, it is notable that a role for Dpp signaling in this type of limb patterning has been suggested by THEISEN et al. 1996 . These investigators observed a high rate of bifurcations in the ventral leg after disruption of the type I Dpp receptor that is encoded by tkv.

View larger version (92K): In this window In a new window Download PPT slide   Figure 4. Scanning electron microscopy of adult nota. A wild-type notum (A) is compared to that of a punt 88/punt135 mutant (B). The punt mutant exhibits a deep medial cleft; its head and wings were removed for imaging. In both A and B, dorsal views are presented, and anterior is up. Magnification is x100.

View larger version (66K): In this window In a new window Download PPT slide   Figure 5. Scanning electron microscopy of adult hindlimbs. A wild-type hind limb (A) is compared to those of two punt 88/punt135 mutants (B and C). Although punt limb phenotypes were highly penetrant (>99%), they were variably expressed. Most mutants exhibited shortened and bent femurs, as well as deleted and deformed tarsi (B). In addition, a significant fraction (59%) exhibited pattern duplications, including duplicated metatarsi (C) and duplicated sex combs (data not shown). Wild-type magnification is x63; mutant magnification is x150.

View larger version (108K): In this window In a new window Download PPT slide   Figure 6. Scanning electron microscopy of adult eyes. A wild-type eye (A) displays a regular pattern of ommatidia and interommatidial bristles. In contrast, punt 88/punt135 mutants exhibited a reduced number of ommatidia and bristles, and consequently were highly disorganized overall (B). Magnification is x300.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Although Punt was originally identified as a type II receptor for activin-type and not Dpp/BMP-type ligands, we and others have suggested that Dpp indeed functions as the physiologically relevant Punt ligand in vivo. This hypothesis is based primarily upon four lines of evidence. First, zygotic loss-of-function mutations in the punt gene result in a dorsal-open embryonic phenotype, indistinguishable from that seen in loss-of-function alleles of tkv, the type I Dpp receptor. Second, maternal loss-of-function mutations in the punt gene result in a ventralized embryonic phenotype, indistinguishable from that seen in loss-of-function dpp alleles. Third, under certain conditions, Punt can bind Dpp in vitro. And fourth, as yet no activin-type ligand has been identified in Drosophila (LETSOU et al. 1995 ; RUBERTE et al. 1995 ). In the current article, we describe additional genetic tests of this hypothesis. Our studies indicate that Punt regulates multiple developmental decisions in response to the Dpp signal.

Punt function is required for dorsal closure between 8 and 12 hr AEL:
The possibility that defects in dorsal closure result secondarily from defects in dorsoventral patterning has often been invoked to explain how mutations in punt and other Dpp signaling molecules, including tkv and schnurri (shn), first affect dorsal patterning, and subsequently, dorsal closure (NELLEN et al. 1994 ; PENTON et al. 1994 ; TERRACOL and LENGYEL 1994 ; ARORA et al. 1995 ). We considered the alternative hypothesis that defects in dorsoventral patterning and dorsal closure are each primary consequences of the mutation and result from independent requirements for the punt gene during two different stages of embryogenesis. Using temperature-sensitive punt alleles, we demonstrated that the embryonic requirements for punt correspond to distinct time intervals; thus, the temporal requirement for punt during dorsal closure is separable from its earlier requirement for dorsoventral pattern formation. Our demonstration of a dual requirement for Punt during embryogenesis indicates that the dorsal-open punt phenotype is a primary effect of the zygotic mutation and is consistent with the documented roles for Dpp and Tkv in dorsal closure. In addition to the recent demonstration by several groups that dpp is a transcriptionally regulated target of Drosophila Jun-amino-terminal kinase (DJNK) signaling in leading-edge dorsal epithelial cells (RIESGO-ESCOVAR and HAFEN 1997 ; HOU et al. 1997 ; GLISE and NOSELLI 1997 ), RIESGO-ESCOVAR and HAFEN 1997 have shown that ectodermal expression of dpp or tkv transgenes partially rescues the dorsal-open phenotype in a DJNK signaling-deficient embryo.

Targets of Punt-mediated signaling during dorsal closure have yet to be identified. Several likely candidates, however, include cytoskeletal components of the dorsal epithelium. Some of these proteins play structural roles in transforming epithelial cell shape and consequently effect dorsal closure. Specifically, mutations in the nonmuscle myosin heavy chain (YOUNG et al. 1993 ), the Drosophila homologue of band 4.1 (FEHON et al. 1994 ), and the canoe gene product (MIYAMOTO et al. 1995 ) all produce the same gross defects in dorsal closure that we identify in punt mutants.

Punt functions postembryonically as a Dpp receptor:
We additionally exploited temperature-sensitive punt alleles to demonstrate a clear requirement for punt postembryonically. Since the phenotypes of punt animals that we generated by manipulation of temperature mimicked the array of dpp syndromes documented by GELBART and colleagues over a decade ago (SPENCER et al. 1982 ), we conclude that Dpp functions as a Punt ligand repeatedly throughout the Drosophila life cycle. Not only are both Punt and Dpp required for embryonic patterning events, including dorsoventral axis formation and dorsal closure and gut development, but both gene products are similarly required for imaginal patterning events in the leg, wing, thoracic, and eye/antennal discs (see Figure 4 –6).

It is significant that no unique phenotypes were attributed to mutations in punt, and this observation suggests that Punt is a dedicated Dpp receptor. However, we cannot exclude the possibility that Punt mediates signaling by additional ligands because the punt alleles included in this study are not null. In contrast, a slightly wider range of structural abnormalities is evident in some dpp mutant animals. For example, Dpp is required for patterning of adult terminalia (SPENCER et al. 1982 ) and for patterning of the anterior egg shell during oogenesis (TWOMBLY et al. 1996 ). With respect to a potential requirement for Punt in genital morphogenesis, no combination of punt alleles produced the gross defects in terminal structures that are characteristic of certain mutations in dpp (data not shown). Again, this negative result might be misleading because we have been unable to study the punt null phenotype. We are unable to comment upon an oogenic requirement for Punt, because our genetic studies did not permit us to examine this very early developmental stage. Despite these caveats, the overall striking phenotypic similarities in punt and dpp mutants, including embryonic defects in dorsoventral patterning, dorsal closure, and gut development, and disc defects in leg, thoracic, eye, and antennal development, clearly implicate a role for Punt in most, and perhaps all, Dpp-mediated signaling events in Drosophila.

The ability of molecules belonging to the BMP/Dpp subgroup of the TGF-ß cytokine superfamily to form bioactive, disulfide-linked heterodimers in cultured cells indicates that these complexes are likely to have natural biological functions (AONO et al. 1995 ; HAZAMA et al. 1995 ; ISRAEL et al. 1996 ). In addition to Dpp, two other gene products, Scw and 60A, belong to the Drosophila Dpp/BMP family of TGF-ß signaling molecules (PADGETT et al. 1987 ; WHARTON et al. 1991 ; ARORA et al. 1994 ), and hence the ability to form hetero-dimeric ligands might contribute to the diversity of biological responses elicited by Dpp/BMP signaling cascades in flies. The data presented in the current article, that punt functions as a Dpp receptor throughout the Drosophila life cycle, are certainly consistent with this view. The correspondence between dpp and punt loss-of-function phenotypes could reflect Punt function as a receptor for either homodimeric Dpp ligands or heterodimeric Dpp ligands. Scw, like Dpp and Punt, is required for differentiation of dorsal structures (ARORA et al. 1994 ), and perhaps a Dpp/Scw heterodimer is a signal for dorsal cell fate differentiation in vivo. As we showed previously, combined maternal and zygotic loss of punt function results in a more severely ventralized phenotype than that observed in scw null mutants, and thus we think it unlikely that Punt acts solely as a Scw receptor during embryogenesis (LETSOU et al. 1995 ). Since roles for the 60A gene product have been documented in oogenesis and in gut and wing development (K. A. WHARTON, personal communication), it is tempting to speculate that like Dpp and Scw, Dpp and 60A can heterodimerize to form a bioactive cytokine.

punt mutations define an allelic series:
Based upon mo- lecular and genetic analyses, mutant punt alleles were grouped into three distinct classes, Classes I through III, with Class I alleles being defined as strongest and Class III alleles being defined as weakest. Class I and II alleles are genetically very similar; they are heat-sensitive and exhibit a fully penetrant embryonic lethality at all temperatures tested (see Table 1). Class III alleles, in contrast, are cold-sensitive and fully viable at 25°.

The original transposon insertion allele, punt ^P1, as well as three alleles that were generated by P-element mobilization, punt⁵¹, punt ⁶², and punt ⁸⁸, fall into two classes, I and II. Molecular analyses revealed all four to be regulatory mutants, harboring large insertions (>420 bp) between the second and third nucleotides in the 5' untranslated region of punt. The temperature-sensitive nature of all four alleles indicates that Class I alleles, albeit capable of evoking a stronger phenotype than Class II alleles, retain partial function and are hypomorphic. It is likely that the size of the insertions in these alleles dictates the strength of the mutation because 1) the largest insertion (14 kb) maps to a Class I allele, punt ^P1 and 2) the smallest insertion (~420 bp) maps to a Class II allele, punt ⁸⁸. Consistent with this hypothesis is our identification of even smaller insertions in adult-viable punt alleles (see Table 2).

The original EMS-induced allele punt¹³⁵ was identified as a recessive lethal that disrupts the embryonic process of dorsal closure (JURGENS et al. 1984 ) and was reported to represent the null phenotype (RUBERTE et al. 1995 ). Molecular studies (RUBERTE et al. 1995 ), as well as genetic studies reported here and elsewhere (LETSOU et al. 1995 ; RUBERTE et al. 1995 ), have led, however, to our designation of punt¹³⁵ as a temperature-sensitive Class II allele. In punt¹³⁵, a missense mutation (Ala-376Thr) maps to the conserved kinase domain VIII that is required for substrate recognition (RUBERTE et al. 1995 ). The conclusion that the Thr substitution at position 376 does not completely abolish the kinase activity encoded by punt¹³⁵ at any of the temperatures tested is based on the following observations. First, functional serine/threonine kinases harbor a conserved Ala, Ser, or Pro at this position (HANKS et al. 1988 ), and SerThr substitutions can be considered to be conservative. Second, the punt¹³⁵ phenotype is not the strongest in our collection (see Table 3). Finally, our observation that punt¹³⁵ phenotypes are variably expressed, albeit not itself a criterion for a hypomorphic allele, is consistent with our conclusion that the kinase encoded by punt¹³⁵ retains some catalytic activity (see Figure 1).

Class III alleles, punt¹⁰, punt ²⁴, and punt⁹⁷, harbor very small insertions, either 38 or 40 bp in length. In all three alleles, transposon-derived sequences remaining at the insertion site are oriented as inverted repeats and are capable of forming perfectly base-paired stem-loops, either 15 or 16 bp in length. The molecular definition of these mutations suggests that mRNA secondary structure reduces the efficiency of translation of the very weak Class III hypomorphic alleles (for review, see BROWN and SCHREIBER 1996 ). Consistent with this hypothesis is our observation that conditions that stabilize stem-loop structures, such as lower temperatures, enhance the mutant phenotype of punt¹⁰.

It is notable that all punt alleles described to date exhibit a temperature-sensitive phenotype. It is tempting to speculate that the pathway is itself temperature sensitive, but since the number of mutant alleles studied is relatively small, we favor the hypothesis that these are allele-specific temperature sensitivities and that these alleles retain partial function at the permissive temperatures.

Consistent with our hypothesis that the punt alleles described here are hypomorphic, but not null, are genetic studies indicating that the punt null phenotype is haplo-insufficient. First, punt maps to one of the few regions in the Drosophila genome for which a deficiency has never been recovered, and regions such as this are likely to harbor haplo-lethal loci (LINDSLEY et al. 1972 ). Even more meaningful were our specific efforts to generate null punt alleles in screens for deletions of the punt locus; in this regard, our failure to recover a deletion of punt was statistically significant (P = 2.4 x 10^-5). Our conclusion that punt is a haplo-insufficient gene indicates that it is a dosage-sensitive component of the Dpp signaling pathway. While this finding does not rule out the possibility that Dpp is another dosage-sensitive component of the pathway, it is nonetheless intriguing. In standard paradigms for signaling, ligand is the single dosage-sensitive component of the pathway.

Punt functions as a multimer:
In analyzing the genetic interactions of specific punt alleles, we observed striking differences in the phenotypes of animals harboring homoallelic and heteroallelic punt combinations. Homoallelic combinations of Class I and II punt alleles (punt ^P1, punt⁵¹, punt ⁶², punt ⁸⁸, and punt¹³⁵) produced a fully penetrant embryonic lethality; viable adults were never recovered at any temperature (see Table 1). In contrast, heteroallelic combinations of punt¹³⁵ with alleles arising from mobilization of the P element in punt ^P1 (punt⁵¹, punt ⁶², and punt ⁸⁸) were viable at 18°, and phenotypically normal adults were recovered at the expected Mendelian frequency (see Figure 3). Our observation that heteroallelic punt combinations produced a weaker phenotype than homoallelic combinations led us to speculate that there was a direct interaction between punt gene products in heteroallelic animals.

The specific interaction of punt alleles can be explained if active Punt protein functions as either a dimer or a higher order multimer in the heteromeric complex that also contains its partner and substrate, the type I receptor encoded by the tkv gene. In accordance with all models of interallelic complementation, each punt allele must supply a function in receptor complex activation and signaling that is disrupted in the other.

We suggest that Dpp-mediated signal transduction is diminished in punt¹³⁵ homozygotes due to the defect in the kinase substrate recognition site that was identified by RUBERTE et al. 1995 . In addition, we suggest that signaling is defective in insertion alleles such as punt ⁸⁸ because a 5' UTR sequence insertion results in down-regulation of Punt and its consequent inefficient multimerization. In our model for interallelic complementation in punt ⁸⁸/punt¹³⁵ animals, wild-type quantities of signaling-defective Punt¹³⁵ increase the pool of receptor monomers and consequently increase the probability of forming a functional receptor dimer (or higher order multimer) that contains at least one wild-type Punt⁸⁸ isoform (Figure 7).

View larger version (16K): In this window In a new window Download PPT slide   Figure 7. Theoretical comparison of receptor activity in punt 88/punt 88 and punt 88/punt135 mutants. For this theoretical comparison we have made the following simplifying assumptions: 1) Punt135 has no signaling activity, and 2) a dimer containing a single Punt88 subunit will have full signaling activity. In a monomer model for Punt function (A), the rate of receptor complex formation is linearly related to the relative concentration of Punt88. In a punt 88/punt 88 animal: Fraction(active Punt) = K([Puntrel] + [Puntrel])2K = [Puntrel]. In a punt 88/punt135 animal: Fraction(active Punt) = K([Puntrel] + [0])/2K. Because the punt 88/punt135 line always remains below the punt 88/punt 88 line, this monomeric model for Punt function is inconsistent with our experimental observation of interallelic complementation. In a dimer model for Punt function (B), the fraction of active receptor complex formation is exponentially related to the relative concentration of wild-type punt gene product, [Puntrel], that is produced. In a regulatory mutant (e.g., punt 88/punt 88): Fraction(active Punt) = K([Puntrel]2 + 2[Puntrel][Puntrel] + [Puntrel]2)/4K = [Puntrel]2. In a heteroallelic mutant that harbors a regulatory allele in combination with a kinase-defective allele (e.g., punt 88/punt135): Fraction(active Punt) = K([Puntrel]2 + 2[Puntrel][1] + [0]2)/4K. The dimeric model (or any multimeric model for Punt function) is consistent with our experimental observation of interallelic complementation: when regulatory mutant levels are low, the punt88/punt135 model curve crosses over the punt 88/punt 88 curve. In (C), a schematic representation of a signaling heteromer is presented.

Consistent with our genetic demonstration that type II receptor serine/threonine kinases function as multimers in vivo are the results of several biochemical studies performed in somatic cells. Similarities between the TGF-ß receptor serine/threonine kinases and the well-characterized EGF (epidermal growth factor) homodimeric receptor tyrosine kinases that are capable of autophosphorylation in vivo (LAMMERS et al. 1990 ; CANALS 1992 ) were first observed in structure/function studies (WRANA et al. 1994 ). In biochemical experiments designed to determine whether the stoichiometry of components in the TGF-ß signaling complex is analogous to that in EGF signaling complexes, mink lung epithelial cells and COS-1 and -7 cells were transfected with type I and type II TGF-ß receptors. In all cell types, homomeric receptor dimers were recovered (CHEN and DERYNCK 1994 ; HENIS et al. 1994 ; YAMASHITA et al. 1994 ).

The biological relevance of the type II receptor interaction is revealed in the genetic studies described in this article. That the type I receptor similarly functions as an oligomer in vivo has also been only recently demonstrated (WEIS-GARCIA and MASSAGUE 1996 ). Whereas cotransfection of kinase-defective TßR-I (TGF-ß type I receptor) mutants and activation-defective TßR-I mutants restores TGF-ß responsiveness to R-1B cells, neither construct restores TGF-ß responsiveness to R-IB cells when transfected alone. Taken together, the interallelic complementation studies presented in the current article and elsewhere (WEIS-GARCIA and MASSAGUE 1996 ) indicate that the structure of TGF-ß receptor complexes, such as the Dpp receptor complex in Drosophila, is oligomeric and harbors at least two copies of the type I and II serine/threonine kinase receptors.

In summary, our characterization of three classes of temperature-sensitive punt alleles has led to a fuller understanding of Dpp receptor complex function in vivo. Our molecular and genetic studies revealed essential and repeated roles for the Dpp receptor encoded by the punt gene throughout the Drosophila life cycle. In addition, the shared loss-of-function phenotypes that we documented in punt and dpp mutants implicate Punt as an important mediator of virtually all Dpp signals. Finally, our observation that interallelic complementation depended on the presence of two types of punt mutation, one that specifically reduces kinase function and one that disrupts gene regulation, revealed the multimeric structure of Punt during signaling.

	ACKNOWLEDGMENTS

We thank members of our laboratory for stimulating discussions and A. GODWIN, S. MANGO, and C. THUMMEL for critical reading of the manuscript. We thank KRISTI WHARTON for sharing data prior to publication, MIKE HOFFMANN for tkv mutant fly lines, and ED KING for help with SEM. We are particularly grateful to S. SAKONJU and C. THUMMEL for sharing reagents and expertise. This work was supported by grants from the American Cancer Society to A.L. (DB-84 and JFRA-657) and by a training grant from the National Institutes of Health to K.S. (5T32HD07491-03).

Manuscript received August 25, 1997; Accepted for publication October 29, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AONO, A., M. HAZAMA, K. NOTOYA, S. TAKETOMI, H. YAMASAKI, R. TSUKUDA, and S. SASKI et al., 1995  Potent ectopic bone-inducing activity of bone morphogenetic protein-4/7 heterodimer. Biochem. Biophys. Res. Commun. 210:670-677[Medline].

ARORA, K., M. LEVINE, and M. O'CONNOR, 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[Abstract/Free Full Text].

ARORA, K., H. DAI, S. G. KAZUKO, J. JAMAL, and M. B. O'CONNOR et al., 1995  The Drosophila schnurri gene acts in the Dpp/TGFß signaling pathway and encodes a transcription factor homologous to the human MBP family. Cell 81:781-790[Medline].

ATTISANO, L., J. L. WRANA, S. CHEIFETZ, and J. MASSAGUÉ, 1992  Novel activin receptors: distinct genes and alternative mRNA splicing generate a repertoire of serine/threonine kinase receptors. Cell 68:97-108[Medline].

BROWN, E. J. and S. L. SCHREIBER, 1996  A signaling pathway to translational control. Cell 86:517-520[Medline].

BURKE, R. and K. BASLER, 1996  Dpp receptors are autonomously required for cell proliferation in the entire developing Drosophila wing. Development 122:2261-2269[Abstract].

CAMPOS-ORTEGA, J. A., and V. HARTENSTEIN, 1985 The embryonic development of Drosophila melanogaster. Springer-Verlag, Berlin.

CANALS, F., 1992  Signal transmission by epidermal growth factor receptor: coincidence of activation and dimerization. Biochemistry 31:4493-4501[Medline].

CHEN, R. H. and R. DERYNCK, 1994  Homomeric interactions between type II transforming growth factor-ß receptors. J. Biol. Chem. 269:22868-22874[Abstract/Free Full Text].

CHILDS, S. R., J. L. WRANA, K. ARORA, L. ATTISANO, and M. B. O'CONNOR et al., 1993  Identification of a Drosophila activin receptor. Proc. Natl. Acad. Sci. USA 90:9475-9479[Abstract/Free Full Text].

FEHON, R. G., I. A. DAWSON, and S. ARTAVANIS-TSAKONAS, 1994  A Drosophila homologue of membrane-skeleton protein 4.1 is associated with septate junctions and is encoded by the coracle gene. Development 120:545-557[Abstract].

FERGUSON, E. L. and K. V. ANDERSON, 1992  decapentaplegic acts as a morphogen to organize dorsal-ventral pattern in the Drosophila embryo. Cell 71:451-461[Medline].

GLISE, B. and S. NOSELLI, 1997  Coupling Jun amino-terminal kinase and Decapentaplegic signaling pathways in Drosophila morphogenesis. Genes Dev. 11:1738-1747[Abstract/Free Full Text].

HANKS, S. K., A. M. QUINN, and T. HUNTER, 1988  The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 241:42-52[Abstract/Free Full Text].

HAZAMA, M., A. AONO, N. UENO, and Y. FUJISAWA, 1995  Efficient expression of a heterodimer of bone morphogenetic protein subunits using a baculovirus expression system. Biochem. Biophys. Res. Commun. 209:859-866[Medline].

HENIS, Y., A. MOUSTAKAS, H. LIN, and H. LODISH, 1994  The types II and III transforming growth factor-ß receptors form homo-oligomers. J. Cell. Biol. 126:139-154[Abstract/Free Full Text].

HOFFMANN, F., 1991  Transforming growth factor-ß-related genes in Drosophila and vertebrate development. Curr. Opin. Cell Biol. 3:947-952[Medline].

HOU, X. S., E. S. GOLDSTEIN, and N. PERRIMON, 1997  Drosophila Jun relays the Jun amino-terminal kinase signal transduction pathway to the Decapentaplegic signal transduction pathway in regulating epithelial cell sheet movement. Genes Dev. 11:1728-1737[Abstract/Free Full Text].

IMMERGLÜCK, K., P. A. LAWRENCE, and M. BIENZ, 1990  Induction across germ layers in Drosophila mediated by a genetic cascade. Cell 62:261-268[Medline].

ISRAEL, D. I., J. NOVE, K. M. KERNS, R. J. KAUFMAN, and V. ROSEN et al., 1996  Heterodimeric bone morphogenetic proteins show enhanced activity in vitro and in vivo. Growth Factors 13:291-300[Medline].

JÜRGENS, G., E. WIESCHAUS, C. NÜSSLEIN-VOLHARD, and H. KLUDING, 1984  Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. II. Zygotic loci on the third chromosome. Roux's Archives of Developmental Biology 193:283-295.

KAPHINGST, K. and S. KUNES, 1994  Pattern formation in the visual centers of the Drosophila brain: wingless acts via decapentaplegic to specify the dorsoventral axis. Cell 78:437-448[Medline].

KIM, J., K. JOHNSON, H. J. CHEN, S. CARROLL, and A. LAUGHON, 1997  Drosophila Mad binds to DNA and directly mediates activation of vestigial by Decapentaplegic. Nature 388:304-308[Medline].

KINGSLEY, D. M., 1994  The TGF-ß superfamily: new members, new receptors, and new genetic tests of function in different organisms. Genes Dev. 8:133-146[Free Full Text].

LAMMERS, R., E. VAN OBBERGHEN, R. BALLOTTI, J. SCHLESSINGER, and A. ULLRICH, 1990  Transphosphorylation as a possible mechanism for insulin and epidermal growth factor receptor activation. J. Biol. Chem. 265:16886-16890[Abstract/Free Full Text].

LECUIT, T., W. J. BROOK, M. NG, M. CALLEJA, and H. SUN et al., 1996  Two distinct mechanisms for long-range patterning by Decapentaplegic in the Drosophila wing. Nature 381:387-393[Medline].

LETSOU, A., K. ARORA, J. L. WRANA, K. SIMIN, and V. TWOMBLY et al., 1995  Drosophila Dpp signaling is mediated by the punt gene product: a dual ligand-binding type II receptor of the TGFß receptor family. Cell 80:899-908[Medline].

LINDSLEY, D. L., L. SANDLER, B. S. BAKER, A. T. CARPENTER, and R. E. DENELL et al., 1972  Segmental aneuploidy and the genetic gross structure of the Drosophila genome. Genetics 71:157-184[Abstract/Free Full Text].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, New York.

LIU, F., F. VENTURA, J. DOODY, and J. MASSAGUÉ, 1995  Human type II receptor for bone morphogenic proteins (BMPs): extension of the two-kinase receptor model to the BMPs. Mol. Cell. Biol. 15:3479-3486[Abstract].

MACIAS-SILVA, M., S. ABDOLLAH, P. A. HOODLESS, R. PIRONE, and L. ATTISANO et al., 1996  MADR2 is a substrate of the TGF-ß receptor and its phosphorylation is required for nuclear accumulation and signaling. Cell 87:1215-1224[Medline].

MASSAGUÉ, J., 1996  TGFß signaling: receptors, transducers, and Mad proteins. Cell 85:947-950[Medline].

MATHEWS, L. S. and W. W. VALE, 1991  Expression cloning of an activin receptor, a predicted transmembrane serine kinase. Cell 65:973-982[Medline].

MIYAMOTO, H., I. NIHONMATSU, S. KONDO, R. UEDA, and S. TOGASHI et al., 1