Transgenic Analysis of the Smad Family of TGF-ß Signal Transducers in Drosophila melanogaster Suggests New Roles and New Interactions Between Family Members

Corresponding author: Stuart J. Newfeld, Department of Biology, Arizona State University, Tempe, AZ 85287-1501., newfeld{at}asu.edu (E-mail)

Communicating editor: S. YOKOYAMA

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Smad signal transducers are required for transforming growth factor-ß-mediated developmental events in many organisms including humans. However, the roles of individual human Smad genes (hSmads) in development are largely unknown. Our hypothesis is that an hSmad performs developmental roles analogous to those of the most similar Drosophila Smad gene (dSmad). We expressed six hSmad and four dSmad transgenes in Drosophila melanogaster using the Gal4/UAS system and compared their phenotypes. Phylogenetically related human and Drosophila Smads induced similar phenotypes supporting the hypothesis. In contrast, two nearly identical hSmads generated distinct phenotypes. When expressed in wing imaginal disks, hSmad2 induced oversize wings while hSmad3 induced cell death. This observation suggests that a very small number of amino acid differences, between Smads in the same species, confer distinct developmental roles. Our observations also suggest new roles for the dSmads, Med and Dad, in dActivin signaling and potential interactions between these family members. Overall, the study demonstrates that transgenic methods in Drosophila can provide new information about non-Drosophila members of developmentally important multigene families.

SECRETED proteins of the conserved transforming growth factor-ß (TGF-ß) family regulate developmental events in many species. Two subfamilies of TGF-ß molecules have been defined by amino acid sequence conservation. These are the decapentaplegic/bone morphogenetic protein subfamily (Dpp/BMP) and the TGF-ß/Activin subfamily. Within the Dpp/BMP subfamily, the Drosophila protein Dpp shows greater amino acid similarity to human BMP2 and BMP4 than to any other TGF-ß molecule (GELBART 1989 ). Significant functional conservation has been identified between these three family members. Human BMP2 and BMP4 can rescue dpp mutant phenotypes in Drosophila (PADGETT et al. 1993 ) and Dpp can induce bone morphogenesis in mammalian cells (SAMPATH et al. 1993 ).

Members of both TGF-ß subfamilies influence developmental events via conserved signaling pathways (NEWFELD et al. 1999 ). TGF-ß molecules stimulate responses in target cells through a complex of transmembrane receptors. Upon TGF-ß binding, one partner in the receptor complex initiates a signal transduction cascade by phosphorylating a member of the Smad family of cytoplasmic signal transducers. A multi-Smad complex is then formed that translocates to the nucleus and participates in the transcription of target genes (RAFTERY and SUTHERLAND 1999 ).

There are seven known Smad proteins in humans (hSmads). These have been characterized biochemically in cell culture. hSmad1 and hSmad5 can transduce Dpp/BMP subfamily signals. hSmad2 and hSmad3 can transduce TGF-ß/Activin subfamily signals. hSmad4 can transduce signals of both subfamilies. hSmad6 and hSmad7 can antagonize signals of both subfamilies (MASSAGUE and CHEN 2000 ). However, the developmental roles (what signal is transduced and what target genes are activated in a specific developmental event) of hSmads are largely unknown. Some information has been inferred from mouse "knockouts." Smad2, Smad4, Smad5, and Smad6 knockouts have multiple embryonic defects (e.g., CHANG et al. 1999 ), making the identification of specific TGF-ß signals or target genes difficult. Smad3 knockout mice as well as Smad4 and Apc compound heterozygous mice have adult colon tumors (TAKAKU et al. 1998 ; ZHU et al. 1998 ). This suggests that Smad3 and Smad4 regulate colon cell proliferation but provides no information on specific signals or target genes. Null mutations in hSmad2 and hSmad4 are found in human adult pancreatic and colon tumors and in one form of childhood colon cancer (RIGGINS et al. 1997 ; HOWE et al. 1998 ). This suggests that hSmad2 and hSmad4 have tumor suppressor activity but again the relevant signals and target genes are unknown.

There are four Smad proteins in Drosophila melanogaster (dSmads). Two dSmads have been extensively analyzed, revealing a variety of developmental roles. Mothers against dpp (Mad) and Medea (Med) transduce Dpp signals during numerous developmental events. For many of these Dpp-dependent developmental events the target genes are also known. Less is known about two other dSmads, Daughters against dpp (Dad) and dSmad2. Dad antagonizes Dpp signals and dSmad2 transduces dActivin signals during limb development (RAFTERY and SUTHERLAND 1999 ).

Structurally, Smad family members contain conserved N-terminal Mad-homology (MH1) and C-terminal MH2 domains (SEKELSKY et al. 1995 ). The MH1 domain is necessary for transcriptional activity and the MH2 domain is necessary for forming multi-Smad complexes (LAGNA et al. 1996 ). A phylogenetic analysis of these domains showed that the following Smads have very similar sequences: (1) Mad, hSmad1, and hSmad5; (2) dSmad2, hSmad2, and hSmad3; (3) Med and hSmad4; and (4) Dad, hSmad6, and hSmad7 (NEWFELD et al. 1999 ).

Significant functional conservation has also been identified between Smad family members. In cross-species experiments similar to those described for Dpp/BMP proteins, ectopic Smads phenocopy or substitute for their endogenous relatives. In Xenopus, injected Drosophila Mad mimicked Smad1 in Dpp/BMP signaling (NEWFELD et al. 1996 ). In Drosophila, injected hSmad4 rescued Med mutant phenotypes (HUDSON et al. 1998 ). These experiments suggest the hypothesis that vertebrate Smads have developmental roles analogous to those of the most similar dSmad.

Here we report an experimental test of this hypothesis. We examined gain-of-function phenotypes generated by expressing six hSmad and four dSmad transgenes in Drosophila. A comparative analysis of the phenotypes suggests that the hypothesis is true. In contrast, we noted that two nearly identical hSmads generated distinct phenotypes.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mad clonal analysis:
The P{>w^hs>}G31 FRT cassette is described in WILDER and PERRIMON 1995 . The Mad¹² allele is described in SEKELSKY et al. 1995 . The stubby chaetae (stc) mutation (JIANG and STRUHL 1995 ) was used to mark clones. stc causes multiple trichomes to be produced by a single wing blade cell. w¹¹¹⁸ P{hsFLP}/Y males that were also either Mad⁺ stc P{>w^hs>}G31/CyO or Mad¹² stc P{>w^hs>}G31/CyO were crossed to y w¹¹¹⁸; M(2L)24F¹ P{>w^hs>}G31/CyO females. Clones were induced by 37° heat shock late in the third larval instar (96 ± 12 hr after egglay) and analyzed according to SINGER et al. 1997 .

UAS.hSmad transgenes:
pUAST is as described (BRAND and PERRIMON 1993 ). hSmad nomenclature follows the convention of DERYNCK et al. 1996 . hSmad1 is as described (HOODLESS et al. 1996 ). A ClaI to BamHI fragment from pCMv5B-FlagMadR-1 was cloned into pBluescript. A XhoI to XbaI fragment was then inserted into these sites in pUAST. hSmad2 is as described (EPPERT et al. 1996 ). A ClaI to XbaI fragment from pCMv5B-FlagMadR-2 was cloned into pBluescript. A XhoI to XbaI fragment was then inserted into these sites in pUAST. hSmad3 is as described (RIGGINS et al. 1996 ). A BssHI to XbaI fragment was isolated from JV15-2. The BssHI site was filled in with T4 DNA polymerase. This fragment was cloned into the NotI and XbaI sites in pUAST after the NotI site was filled in with T4 DNA polymerase. hSmad4 is as described (HAHN et al. 1996 ). A NotI to KpnI fragment from pDPC4-wt3 was cloned into these sites in pUAST. hSmad6 and hSmad7 are as described (TOPPER et al. 1997 ). A BamHI to XhoI fragment from pBS-Smad6 was cloned into BglII and XhoI sites in pUAST. An EcoRI to XhoI fragment from pBS-Smad7 was cloned into these sites in pUAST. Multiple independent strains were generated for each transgene.

Strains and mating schemes:
The Gal4 strains 24B.Gal4, 32B.Gal4, 69B.Gal4, and UAS.lacZ are as described (BRAND and PERRIMON 1993 ). A9.Gal4 is as described (HAERRY et al. 1998 ). C765.Gal4 is as described (NELLEN et al. 1996 ). MS1096. Gal4 is as described (MILAN et al. 1998 ). T80.Gal4 is as described (WILDER and PERRIMON 1995 ). ap.Gal4 and dll.Gal4 are as described (CALLEJA et al. 1996 ). dpp-blink.Gal4 is as described (STAEHLING-HAMPTON et al. 1994 ). ptc.Gal4 is as described (SPEICHER et al. 1994 ). en.Gal4 is as described (TABATA et al. 1995 ). The strains UAS.Dad, UAS.Mad, and UAS.Med are as described (RAFTERY and SUTHERLAND 1999 ). UAS.dSmad2 is as described (BRUMMEL et al. 1999 ). All UAS.Smad strains carry homozygous viable insertions. Several of the Gal4 insertions were inviable as homozygotes and were maintained over balancer chromosomes. Progeny with the experimental genotype, derived from crosses of UAS.Smad-bearing males to females carrying Gal4 transgenes, were identified using appropriate phenotypic markers. Complete genotypes for individual UAS.Smad and Gal4 lines as well as the phenotypic markers used in specific matings are available upon request. For UAS.Smad/Gal4 phenotypes affecting wing size, a wing's surface area was estimated as the product of its maximum length and maximum width. The wing surface area for a particular genotype was generated from measurements of 10 individuals. Within a genotype, individual variation was typically ±5%. A table summarizing all observed phenotypes is also available.

ß-Galactosidase analyses:
Females carrying homozygous viable insertions of UAS.lacZ on chromosome II or III were crossed to males from each Gal4 line in vials with one exception. Females carrying MS1096.Gal4 were crossed to UAS.lacZ-bearing males. Embryos were collected and analyzed according to NEWFELD et al. 1996 . Wandering third instar larvae were dissected in ice-cold PBS, transferred to ice cold PBSTween, and rinsed. Fixation and staining proceeded according to NEWFELD et al. 1996 . After fixation, carcasses were placed into a solution of 50% glycerol in PBS and incubated at 4° overnight. Wing and leg disks were dissected and mounted on slides in the same solution.

Acridine orange analyses:
Wing disks of wandering third instar larvae were stained similarly to SPREIJ 1971 . Larvae were dissected in PBS, stained 5 min in a 0.6-µg/ml solution of acridine orange in PBS, rinsed 5 min in PBS, and the wing disks were mounted for observation.

Sequence comparison:
Calculations of amino acid identity and similarity were deduced from alignments using the Baylor College of Medicine web site: http://dot.imgen.bcm.tmc.edu:9331.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mad clonal analysis:
Analyses of mitotic clones have shown that dpp is required for three processes during wing development.

1. dpp plays a major role in cell proliferation and/or cell survival (BURKE and BASLER 1996 ). Eliminating dpp expression results in a nearly complete loss of wing tissue (BURKE and BASLER 1996 ) and ectopic dpp expression results in significant wing overgrowth (CAPDEVILA and GUERRERO 1994 ).

2. dpp is required for anterior/posterior patterning (SINGER et al. 1997 ).

3. dpp is required for vein formation (STURTEVANT and BIER 1995 ).

Clonal analysis has also shown that the Dpp signaling pathway components thickveins (tkv) and schnurri (shn) are also required for vein formation (BURKE and BASLER 1996 ). To date, Mad's role in vein formation has not been formally documented because Mad null clones induced early in the third instar do not survive (WIERSDORFF et al. 1996 ). However, hemizygous-viable Mad alleles show vein defects (SEKELSKY et al. 1995 ) and nuclear-localized Mad is detected at the future site of cross veins in wing disks (CONLEY et al. 2000 ), suggesting a role for Mad in this process. If Mad has a direct role in vein formation it supports the possibility that expression of UAS.Mad, and perhaps other UAS.Smad transgenes, will generate vein phenotypes.

To formally demonstrate that Mad has a role in vein formation we examined Mad null clones induced late in the third instar that were marked with stc (JIANG and STRUHL 1995 ). stc mutant cells have multiple trichomes. A stc control clone had no effect on vein formation (Fig 1A). A Mad clone caused the third longitudinal vein (L3) to bifurcate or terminate at the clone boundary (Fig 1B). A smaller Mad clone caused L2 to loop around the mutant tissue (Fig 1C). Over 50 Mad clones were examined. In every case, Mad mutant clones disrupted vein formation in a cell-autonomous manner. No Mad mutant cells formed veins, even those adjacent to wild-type cells. We conclude that Mad plays an essential role in vein formation.

View larger version (76K): In this window In a new window Download PPT slide   Figure 1. Mad mutant clones disrupt vein formation. (A) Control clone in an adult wing. The multiple trichome phenotype of stc does not effect longitudinal vein 2 (L2; between arrowheads). (B) A clone marked by stc that is also homozygous for the null allele Mad12. The clone disrupts formation of L3 in a cell-autonomous manner. The vein either bifurcates (left arrowhead) or terminates (right arrowhead) at the clone boundary. (C) A small Mad12 clone causes L2 to loop (arrowheads) around the mutant tissue.

UAS.Smad/Gal4 phenotypes:
To test our hypothesis that an hSmad performs developmental roles analogous to those of the most similar dSmad, we assembled a collection of 4 UAS.dSmad, 6 UAS.hSmad, and 12 Gal4 strains. We histochemically examined embryos, wing, and leg imaginal disks from each Gal4 line to determine their complete expression pattern (Fig 2). Then we conducted inter se crosses between all UAS.Smad and all Gal4 lines. Many of the resulting UAS.Smad/Gal4 genotypes generated limb phenotypes. Other combinations were lethal. To test our hypothesis that the developmental roles of each hSmad are analogous to those of its closest dSmad relative we compared limb phenotypes generated by Smads with similar sequences. Here we describe representative examples of our comparisons.

View larger version (65K): In this window In a new window Download PPT slide   Figure 2. Gal4 expression patterns. Each row shows the expression of Gal4-driven UAS.lacZ in embryos, wing, and leg imaginal disks. (A–K) Embryos, stage 10 or later, histochemically stained for lacZ activity driven by the indicated Gal4 line. (A'–K') Stained third instar wing disks. (A''–K'') Stained third instar leg disks.

UAS.Mad, UAS.hSmad1, UAS.Med, and UAS.hSmad4: hSmad1 and Mad can transduce Dpp/BMP signals. hSmad4 and possibly Med can transduce signals for both TGF-ß subfamilies (NEWFELD et al. 1999 ). hSmad4 forms complexes with hSmad1 and Med forms complexes with Mad (RAFTERY and SUTHERLAND 1999 ; WRANA 2000). These relationships suggest that these Smads will produce similar phenotypes. One copy of UAS.Mad or UAS.hSmad1 did not generate many phenotypes. We then utilized strains containing two copies of these transgenes.

UAS.Mad and UAS.hSmad1 induced similar wing and leg phenotypes. For example, UAS.Mad/ptc.Gal4 genotypes have ectopic vein tissue between L3 and L5 (Fig 3B). This vein phenotype is consistent with two previous results: clonal analysis showed a role for Mad in vein formation (Fig 1B) and ptc.Gal4 expression in wing disk cells that eventually reside between L3 and L4 (Fig 2H'). In UAS.Mad/ptc.Gal4 wings the distance between L3 and L5 appears reduced. This may be due to the smaller size of vein cells vs. intervein cells. A comparison of wing surface areas showed that UAS.Mad/ptc.Gal4 wings are 22% smaller than wild type. UAS.hSmad1/ptc.Gal4 wings have ectopic vein tissue in roughly the same region (Fig 3C). UAS.hSmad1/ptc.Gal4 wings are 15% smaller than wild type. UAS.Mad and UAS.hSmad1 expression appears to mimic dpp's role in vein formation. UAS.Mad and UAS.hSmad1 expression does not appear to mimic dpp's other roles in wing development (cell proliferation and/or cell survival and anterior/posterior patterning).

View larger version (112K): In this window In a new window Download PPT slide   Figure 3. Phenotypes generated by UAS.Mad, UAS.hSmad1, UAS.Med, and UAS.hSmad4. (A) Wild-type wing. (B) UAS.Mad/ptc.Gal4 small wing with additional vein tissue between L3 and L4 over their full length and between the proximal portions of L4 and L5. (C) UAS.hSmad1/ptc.Gal4 small wing with additional vein tissue between the proximal portions of L3 and L4. (D) UAS.Med/ptc.Gal4 large wing with additional vein tissue between the proximal portions of L3 and L4. (E) UAS.hSmad4/ptc.Gal4 large wing with additional vein tissue between the proximal portions of L3 and L4. (F) Wild-type leg. (G) UAS.Mad/ptc.Gal4 femur has a small ectopic leg on its ventral side. (H) UAS.hSmad1/ptc.Gal4 tibia has a small ectopic leg on its ventral side. (I) UAS.Med/ptc.Gal4 abnormally shaped tibia with bristle overgrowth and duplicated tarsi. (J) UAS.hSmad4/ptc.Gal4 abnormally shaped tibia with normal bristle pattern and duplicated tarsi. One of the duplicated tarsi lies atop the tibia and is indicated by an arrow.

UAS.Med/ptc.Gal4 (Fig 3D) and UAS.hSmad4/ptc.Gal4 (Fig 3E) wings have ectopic vein tissue in the same region as UAS.Mad/ptc.Gal4 (Fig 3B) and UAS.hSmad1/ptc.Gal4 (Fig 3C) wings. This suggests that Med forms complexes with Mad during vein formation. UAS.Med/ptc.Gal4 wings are 11% larger and UAS.hSmad4/ptc.Gal4 wings are 6% larger than wild type. The larger, rather than smaller, size of UAS.Med/ptc.Gal4 and UAS.hSmad4/ptc.Gal4 wings suggests that multi-subfamily signaling Smads can influence wing size and vein formation.

UAS.Mad/ptc.Gal4 flies have an ectopic leg on the ventral side of a normal limb (Fig 3G). The ectopic leg has several segments and terminates in a set of tarsal claws. During leg development, wingless normally represses dpp expression on the ventral side of the limb (THEISEN et al. 1996 ). In limbs expressing ectopic dpp an additional leg develops on the ventral side of the limb (BASLER and STRUHL 1994 ). The similarity between phenotypes that result from ectopic dpp expression and UAS.Mad expression suggests that UAS.Mad is capable of simulating Dpp signals in leg patterning. UAS.hSmad1/ptc.Gal4 flies also have an ectopic leg on the ventral side of a normal limb (Fig 3H). The ectopic leg consists of a single segment. UAS.Med/ptc.Gal4 and UAS.hSmad4/ptc.Gal4 flies have ectopic legs of a different type. Legs from these genotypes have short, abnormally wide tibia that lead to duplicated tarsi of one or more segments (Fig 3I and Fig J). The abnormal tibia of UAS.Med/ptc.Gal4 flies has an additional patterning defect, severe bristle overgrowth.

Overall, these four Smads generated comparable vein and leg phenotypes. The phenotypes of phylogenetically related Smads showed the greatest similarity. The phenotypes suggest that UAS.hSmad1 and UAS.hSmad4 can also simulate Dpp signaling in Drosophila limb development. These findings are consistent with cell culture studies noted above and further support the view that these hSmads transduce BMP signals during human development. The size of UAS.Med and UAS.hSmad4 wings suggests a role for these multi-subfamily signaling Smads not shared with the Dpp/BMP signaling Smads UAS.Mad and UAS.hSmad1.

UAS.dSmad2, UAS.hSmad2, UAS.hSmad3, UAS.Med, and UAS.hSmad4: dSmad2, hSmad2, and hSmad3 can transduce TGF-ß/Activin signals. hSmad4 and possibly Med can transduce signals for both TGF-ß subfamilies (NEWFELD et al. 1999 ). hSmad4 forms complexes with hSmad2 and with hSmad3 (WRANA 2000 ). These relationships suggest that these Smads will produce similar phenotypes.

UAS.dSmad2, UAS.hSmad2, UAS.Med, and UAS.hSmad4 induced similar wing phenotypes. When expressed with A9.Gal4 each produced moderately large wings (Fig 4). A9.Gal4 is expressed throughout the wing disk (Fig 2D'). A comparison of wing surface areas reveals that UAS.dSmad2, UAS.hSmad2, and UAS.Med wings are 22% larger than wild type. UAS.hSmad4 wings are 16% larger than wild type. It has been shown that dSmad2 transduces dActivin signals that modestly stimulate cell proliferation in wing development. The influence of dActivin signals on wing cell proliferation is much smaller than that of Dpp signals. Ectopic dActivin signaling results in an 30% increase in wing size (BRUMMEL et al. 1999 ). The moderately large size of the common wing phenotype suggests that UAS.hSmad2, UAS.Med, and UAS.hSmad4 simulate dActivin-, rather than Dpp-, mediated cell proliferation.

View larger version (59K): In this window In a new window Download PPT slide   Figure 4. Phenotypes generated by UAS.dSmad2, UAS.hSmad2, UAS.hSmad3, UAS.Med, and UAS.hSmad4. (A) Wild-type wing. (B) UAS.dSmad2/A9.Gal4 large wing with normal venation. (C) UAS.hSmad2/A9.Gal4 large wing with normal venation and a margin notch. (D) UAS.hSmad3/A9.Gal4 small wing with normal venation except for a small loop in L3. (E) UAS.Med/A9 large wing with ectopic veins near the margin. (F) UAS.hSmad4/A9.Gal4 large wing with ectopic veins near the margin.

Consistent with their ptc.Gal4 phenotypes, UAS.Med/A9.Gal4 and UAS.hSmad4/A9.Gal4 wings also show ectopic veins. These wings have a distal crossvein between L2 and L3 and duplications of L2, L3, and L4 at the margin (Fig 4E and Fig F). The presence of wing size and vein phenotypes provides the first evidence that Med, like its counterpart hSmad4, can signal for both TGF-ß subfamilies. UAS.Med and UAS.hSmad4 influenced vein formation like Dpp/BMP signaling Smads (UAS.Mad and UAS.hSmad1) and moderately increased wing size like TGF-ß/Activin signaling Smads (UAS.dSmad2 and UAS.hSmad2).

In contrast, UAS.hSmad3/A9.Gal4 wings are smaller than wild type (32%). These wings have essentially wildtype venation although loops are occasionally present (Fig 4D). The dramatic difference in wing size suggests that UAS.hSmad3 cannot participate in a dActivin pathway that stimulates cell proliferation. The UAS.hSmad3 wing phenotype suggests that UAS.hSmad3 can inhibit cell proliferation or stimulate apoptosis during wing development.

UAS.Dad, UAS.hSmad6, UAS.hSmad7, UAS.Med, and UAS.hSmad4: hSmad6, hSmad7, and possibly Dad can antagonize signals of both TGF-ß subfamilies (NEWFELD et al. 1999 ). This relationship suggested that these Smads will produce similar phenotypes.

The phenotypes generated by UAS.hSmad6, UAS.hSmad7, and UAS.Dad were comparable. UAS.hSmad6/ptc.Gal4 (Fig 5B) and UAS.hSmad7/ptc.Gal4 (Fig 5C) genotypes have truncated legs. We could not directly compare these phenotypes with those of UAS.Dad because UAS.Dad did not generate leg phenotypes. UAS.Dad was lethal with all but two Gal4 lines (A9.Gal4 and MS1096.Gal4). These lines have virtually no embryonic or leg disk expression (Fig 2D and Fig F). Alternatively, UAS.hSmad6 and UAS.hSmad7 did not generate wing phenotypes with A9.Gal4 or MS1096.Gal4. However, the tiny wings of UAS.Dad/MS1096.Gal4 flies (Fig 5G) and the truncated legs of UAS.hSmad6/ptc.Gal4 (Fig 5B) and UAS.hSmad7/ptc.Gal4 (Fig 5C) flies suggest that these Smads have similar abilities to inhibit limb growth. Perhaps the UAS.Dad/MS1096.Gal4 tiny wing phenotype results from Dad antagonizing both the Dpp and the dActivin cell proliferation pathways.

View larger version (39K): In this window In a new window Download PPT slide   Figure 5. Phenotypes generated by UAS.hSmad6, UAS.hSmad7, UAS.hSmad4, UAS.Med, and UAS.Dad. (A) Wild-type leg. (B) UAS.hSmad6/ptc.Gal4 leg truncated at the second tarsal segment. (C) UAS.hSmad7/ptc.Gal4 leg truncated at the second tarsal segment. (D) UAS.hSmad4/ptc.Gal4 leg truncated at the first tarsal segment. (E) UAS.Med/ptc.Gal4 leg truncated at the first tarsal segment. (F) Wild-type wing. (G) UAS.Dad/MS1096 Gal4 tiny wing with no veins and ectopic bristles on the margin.

The UAS.Dad/MS1096.Gal4 (Fig 5G) and UAS.Dad/A9.Gal4 (data not shown) tiny wings are also veinless. This phenotype is likely due to UAS.Dad antagonizing Dpp signals that promote vein formation via Mad and Med. Like hSmad6 and hSmad7 in cell culture, Dad may be capable of antagonizing signals from both TGF-ß subfamilies.

The multi-subfamily signaling Smads, UAS.hSmad4 and UAS.Med, also generated truncated legs with several Gal4 lines. In fact, we noted truncated legs on UAS.hSmad4/ptc.Gal4 (Fig 5D) and UAS.Med/ptc.Gal4 (Fig 5E) flies more frequently than duplicated legs (Fig 3I and Fig J). The truncated leg phenotypes of UAS.hSmad4/ptc.Gal4 (Fig 5D) and UAS.Med/ptc.Gal4 (Fig 5E) flies are similar to those of UAS.hSmad6/ptc.Gal4 (Fig 5B) and UAS.hSmad7/ptc.Gal4 (Fig 5C) flies. The common leg phenotype suggests that antagonist Smads (e.g., hSmad6) may interact with multi-subfamily signaling Smads (e.g., hSmad4) when expressed in Drosophila. Interactions between antagonist and multi-subfamily signaling Smads have been shown in Xenopus injection assays (HATA et al. 1998 ).

UAS.hSmad3 induces cell death:
UAS.hSmad3 consistently generated distinct phenotypes from the phylogenetically related Smads UAS.dSmad2 and UAS.hSmad2. UAS.dSmad2 and UAS.hSmad2 were viable with all Gal4 lines while UAS.hSmad3 was lethal with any Gal4 line with significant embryonic expression. UAS.hSmad3 lethal phenotypes often included small size, brown patches of necrotic tissue, and in pharate adults the absence of entire limbs (data not shown). In cell culture, ectopic expression of hSmad3 induced apoptosis in lung epithelial cells (YANAGISAWA et al. 1998 ). The UAS.hSmad3 phenotypes we observed suggest the hypothesis that UAS.hSmad3 expressed in flies also stimulates apoptosis.

To test this hypothesis, we stained UAS.hSmad3/ptc.Gal4 wing disks (Fig 6B) with acridine orange, an indicator of cell death (ABRAMS et al. 1993 ). UAS.hSmad3/ptc.Gal4 disks showed significantly larger numbers of dead cells than control disks containing only ptc.Gal4 (Fig 6A) or only UAS.hSmad3 (data not shown). The UAS.hSmad3/ptc.Gal4 disks (Fig 6B) show clusters of dead cells in the anterior and posterior compartments. Surprisingly, the distribution of dead cells does not correspond to the expression pattern of ptc.Gal4. However, MILAN et al. 1997 showed that wing disks compensate for cell death in one compartment by inducing cell death in the other compartment. This compensation mechanism appears to be operating in UAS.hSmad3/ptc.Gal4 disks, leading to the observed distribution of dead cells.

View larger version (75K): In this window In a new window Download PPT slide   Figure 6. Phenotypes generated by UAS.hSmad2 and UAS.hSmad3 in acridine orange-stained wing disks. (A) ptc.Gal4 control disk with few brightly stained cells indicating little cell death. (B) UAS.hSmad3/ptc.Gal4 small disk with numerous patches of brightly stained cells in the anterior (arrowhead) and posterior (arrow) compartments. (C) UAS.hSmad3/en.Gal4 very small dysmorphic disk with many patches of brightly stained cells. (D) UAS.Smad2/ptc.Gal4 disk similar to the ptc.Gal4 control disk.

To determine if UAS.hSmad3 induces cell death only with ptc.Gal4, we stained UAS.hSmad3/en.Gal4 wing disks. There are many more dead cells in UAS.hSmad3/en.Gal4 wing disks (Fig 6C) than in UAS.hSmad3/ptc.Gal4 wing disks (Fig 6B). This is likely due to the larger domain of expression for en.Gal4 (throughout the posterior compartment; TABATA et al. 1995 ) vs. ptc.Gal4 (a central stripe; Fig 2H'). UAS.hSmad3/en.Gal4 disks are small and dysmorphic possibly due to the large amount of UAS.hSmad3-induced cell death.

A comparison of the amino acid sequences of hSmad2 and hSmad3 using the alignment of NEWFELD et al. 1999 revealed 91% overall identity and 97% identity in the MH2 domain for these Smads (Table 1). Given this extensive identity, we stained UAS.hSmad2/ptc.Gal4 wing disks with acridine orange. We reasoned that perhaps the large wings seen in UAS.hSmad2/ptc.Gal4 flies (Fig 4C) were the result of overproliferation in response to UAS.hSmad2-induced cell death. However, the level of cell death in UAS.hSmad2/ptc.Gal4 (Fig 6D) and UAS.hSmad2/en.Gal4 (data not shown) wing disks was similar to the ptc.Gal4 control (Fig 6A) and the UAS.hSmad2 control disks (data not shown). These results suggest that hSmad2 and hSmad3 can have distinct developmental roles.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The UAS.Smad phenotypic data showed that phylogenetically related Smad family members (Mad/hSmad1, dSmad2/hSmad2, Med/hSmad4, and Dad/hSmad6/hSmad7) induced similar phenotypes. This result supports our hypothesis that an hSmad performs roles in human development analogous to the ones their dSmad counterpart plays in Drosophila development. We suggest that the developmental roles of hSmads can now be more profitably investigated using clues from dSmads. For example, tinman is a Mad/Med target gene for Dpp signals during the subdivision of the embryonic mesoderm (XU et al. 1998 ). On the basis of our results, the highly conserved human homologs of tinman (PARK et al. 1998 ) are candidate targets of hSmad1 and hSmad4 in human mesodermal cells.

Many of the phenotypes we observed reinforce known roles for dSmads. For example, the moderately large wing phenotype seen with UAS.dSmad2 is consistent with a role in a dActivin pathway that stimulates cell proliferation in wing development. However, other phenotypes suggest new roles for dSmads. For example, moderately large wings generated with several Gal4 lines suggest that Med participates in dActivin signaling. The tiny wings generated with MS1096.Gal4 suggest that Dad may have the ability to antagonize both Dpp and dActivin signaling. In addition, the common truncated leg phenotype generated by Med, hSmad6, and hSmad7 suggests that Med may interact with antagonist Smads such as Dad. These potential roles for Med and Dad are consistent with activities already shown for their human counterparts. For example, hSmad4, hSmad6, and hSmad7 can influence signals from both TGF-ß subfamilies in cell culture (MASSAGUE and CHEN 2000 ) and hSmad4 can interact with hSmad6 in Xenopus injection assays (HATA et al. 1998 ).

Our demonstration that UAS.hSmad3 can induce cell death in wing disks is the first instance where cell death is associated with TGF-ß/Activin signaling in Drosophila. Previous studies connected cell death in imaginal disks with the loss of Dpp signals. Increased cell death is seen in eye and wing disks in certain dpp mutants (BRYANT 1988 ; MASUCCI et al. 1990 ; ADACHI-YAMADA et al. 1999 ). The ability of hSmad3 to induce cell death is consistent with the colon tumor phenotype of Smad3 knockout mice (ZHU et al. 1998 ). Tumors may result from the overgrowth of colon cells not properly targeted for apoptosis in the absence of Smad3.

The distinct phenotypes generated by UAS.hSmad2 and UAS.hSmad3 suggest that Smads with similar sequences induce similar phenotypes only when comparing Smads from different species. Expression of the nearly identical Smads, hSmad2 and hSmad3, always generated distinct phenotypes. Consistent with our results, functional differences between these Smads have been reported in the regulation of Activin-inducible genes in cell culture (e.g., LABBE et al. 1998 ).

What then are the critical amino acid differences between these nearly identical hSmads? Since hSmad2 and dSmad2 are more similar in the MH1 domain (Table 1) we believe that amino acid differences between hSmad2 and hSmad3 in the MH1 domain are likely responsible for their distinct phenotypes. One obvious sequence difference between hSmad2 and hSmad3 is a 30-amino-acid insertion in the MH1 domain of hSmad2 (NEWFELD et al. 1999 ). Two cell culture studies suggest that this insertion is the source of functional differences between hSmad2 and hSmad3 (DENNLER et al. 1999 ; YAGI et al. 1999 ). However, this is unlikely to be the source of the different ptc.Gal4 phenotypes of UAS.hSmad2 and UAS.hSmad3 that we observed for the following reason. Neither hSmad3 nor dSmad2 have the insertion, yet it is dSmad2 and hSmad2 that display similar phenotypes (Fig 3C and Fig F). We believe that amino acid differences between hSmad2 and hSmad3 in the MH1 domain outside the insertion are responsible for their distinct phenotypes. Our view is supported by another cell culture study. NAGARAJAN and CHEN 2000 claim that the amino-terminal region of the MH1 domain, not the insertion, is responsible for observed functional differences between hSmad2 and hSmad3.

A detailed sequence comparison of hSmad2 with hSmad3 reveals that there are just 8 amino acid differences in the MH1 domain (Table 2). Of these, the first three differences are in the region identified by NAGARAJAN and CHEN 2000 and the remaining five differences are clustered within an 11-amino-acid stretch corresponding to the Helix4/Loop5 region in the crystal structure of hSmad3 (SHI et al. 1998 ). When the MH1 domain of dSmad2 is included in the comparison (Table 2), it is identical to hSmad2 at five of the eight substituted positions and to hSmad3 at just one position. We are attempting to identify the amino acid differences that underlie the phenotypic differences between hSmad2 and hSmad3 using systematically mutagenized transgenes.

In summary, our analysis of hSmad and dSmad transgenes supports the hypothesis that phylogenetically related Smads fulfill developmental roles that are conserved between humans and Drosophila. The results also suggest a number of new hypotheses regarding roles for human and Drosophila Smads in pattern formation, cell proliferation, and cell death. The data suggest that a small number of amino acid differences between two very similar Smads in the same species can confer distinct activities. Overall, our study demonstrates that transgenic methods in Drosophila can provide new information about mammalian members of developmentally important multigene families.

	FOOTNOTES

¹ Present address: Exelixis, Inc., South San Francisco, CA 94083.
² Present address: Department of Molecular and Cellular Biology, University of Arizona, Tucson, AZ 85721.

	ACKNOWLEDGMENTS

W. Gelbart, N. Perrimon, G. Struhl, T. Tabata, and the Bloomington Stock Center provided flies. J. Wrana, D. Falb, G. Riggins, and M. Schutte provided hSmad cDNAs. F. Cifuentes provided technical assistance. R. Wisotzkey and M. Stapleton shared unpublished data. This study was initiated in W. Gelbart's lab. S.J.N. is supported by a Basil O'Connor Starter Scholar Research Award from the March of Dimes and a Research Incentive Award from Arizona State University.

Manuscript received November 3, 2000; Accepted for publication December 15, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABRAMS, J., K. WHITE, L. FESSLER, and H. STELLER, 1993  Programmed cell death during Drosophila embryogenesis. Development 117:29-43[Abstract].

ADACHI-YAMADA, T., K. FUJIMURA-KAMADA, Y. NISHIDA, and K. MATSUMOTO, 1999  Distortion of proximodistal information causes JNK-dependent apoptosis in Drosophila wings. Nature 400:166-169[Medline].

BASLER, K. and G. STRUHL, 1994  Compartment boundaries and the control of Drosophila limb pattern by Hedgehog protein. Nature 368:208-214[Medline].

BRAND, A. and N. PERRIMON, 1993  Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118:401-415[Abstract].

BRUMMEL, T., S. ABDOLLAH, T. HAERRY, M. SHIMELL, and J. MERRIAM et al., 1999  The Drosophila Activin receptor Baboon signals through dSmad2 and controls cell proliferation but not patterning during larval development. Genes Dev. 13:98-111[Abstract/Free Full Text].

BRYANT, P., 1988  Localized cell death caused by mutations in the Drosophila gene coding for a TGF-ß homolog. Dev. Biol. 128:386-395[Medline].

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

CALLEJA, M., E. MORENA, S. PELAZ, and G. MORATA, 1996  Visualization of gene expression in living adult Drosophila. Science 274:252-255[Abstract/Free Full Text].

CAPDEVILA, J. and I. GUERRERO, 1994  Targeted expression of the signaling molecule Dpp induces pattern duplications and growth alterations in Drosophila wings. EMBO J. 13:4459-4468[Medline].

CHANG, H., D. HUYLEBROECK, K. VERSCHUEREN, O. GUO, and M. M. MATZUK et al., 1999  Smad5 knockout mice die at mid-gestation due to multiple embryonic and extraembryonic defects. Development 126:1631-1642[Abstract].

CONLEY, C., R. SILBURN, M. SINGER, A. RALSTON, and D. ROHWER-NUTTER et al., 2000  Crossveinless 2 contains cysteine-rich domains and is required for high levels of BMP-like activity during the formation of the cross veins in Drosophila. Development 127:3947-3959[Abstract].

DENNLER, S., S. HUET, and J. GAUTHIER, 1999  A short amino acid sequence in MH1 domain is responsible for functional differences between Smad2 and Smad3. Oncogene 18:1643-1648[Medline].

DERYNCK, R., W. GELBART, R. HARLAND, C. HELDIN, and S. KERN et al., 1996  Nomenclature: vertebrate mediators of TGF-ß family signals. Cell 87:173[Medline].

EPPERT, K., S. SCHERER, H. OZCELIK, R. PIRONE, and P. HOODLESS et al., 1996  MADR2 maps to 18q21 and encodes a TGF-ß regulated MAD-related protein that is functionally mutated in colorectal carcinoma. Cell 86:543-552[Medline].

GELBART, W., 1989  The decapentaplegic gene: a TGF-ß homolog controlling pattern formation in Drosophila. Development 107(Suppl.):65-74.

HAERRY, T., O. KHALSA, M. O'CONNOR, and K. WHARTON, 1998  Synergystic signaling by two BMP ligands through the SAX and TKV receptors controls wing growth and patterning in Drosophila. Development 125:3977-3987[Abstract].

HAHN, S., M. SCHUTTE, A. HOQUE, C. MOSKALUK, and L. DA COSTA et al., 1996  DPC-4: a candidate tumor suppressor gene at human chromosome 18q21. Science 271:350-353[Abstract].

HATA, A., G. LAGNA, J. MASSAGUÉ, and A. HEMMATI-BRIVANLOU, 1998  Smad6 inhibits BMP/Smad1 signaling by specifically competing with the Smad4 tumor suppressor. Genes Dev. 12:186-197[Abstract/Free Full Text].

HIGGINS, D., A. BLEASBY, and R. FUCHS, 1992  CLUSTAL V: improved software for multiple sequence alignment. Comput. Appl. Biosci. 8:189-191[Abstract/Free Full Text].

HOODLESS, P., T. HAERRY, S. ABDOLLAH, M. STAPLETON, and M. O'CONNOR et al., 1996  MADR1, a Mad-related protein that functions in BMP2 signaling pathways. Cell 85:489-500[Medline].

HOWE, J., S. ROTH, J. RINGOLD, R. W. SUMMERS, and H. J. JARVINEN et al., 1998  Mutations in the Smad4 gene in juvenile polyposis. Science 280:1086-1088[Abstract/Free Full Text].

HUDSON, J., S. PODOS, K. KEITH, S. SIMPSON, and E. FERGUSON, 1998  Drosophila Medea is required downstream of dpp and encodes a functional homolog of human Smad4. Development 125:1407-1420[Abstract].

JIANG, J. and G. STRUHL, 1995  PKA and Hedgehog signaling in Drosophila limb development. Cell 80:563-572[Medline].

LABBE, E., C. SILVESTRI, P. HOODLESS, J. WRANA, and L. ATTISANO, 1998  Smad2 and Smad3 positively and negatively regulate TGF-ß dependent transcription through the forkhead DNA-binding protein FAST2. Mol. Cell 2:109-120[Medline].

LAGNA, G., A. HATA, A. HEMMATI-BRIVANLOU, and J. MASSAGUÉ, 1996  Partnership between DPC4 and SMADs in TGF-ß signaling pathways. Nature 383:832-836[Medline].

MASSAGUÉ, J. and Y. CHEN, 2000  Controlling TGF-ß signaling. Genes Dev. 14:627-644[Free Full Text].

MASUCCI, J., R. MILTENBERGER, and F. HOFFMANN, 1990  Pattern-specific expression of Drosophila dpp in imaginal disks is regulated by 3' elements. Genes Dev. 4:2011-2023[Abstract/Free Full Text].

MILAN, M., S. CAMPUZANO and A. GARCIA-BELLIDO, 1997 Developmental parameters of cell death in the wing disc of Drosophila. Proc. Natl. Acad. Sci. USA 94: 5691–5696.34.

MILAN, M., F. DIAZ-BENJUMEA, and S. COHEN, 1998  Beadex encodes an LMO protein that regulates Apterous LIM-homeodomain activity in Drosophila wing development: a model for LMO oncogene function. Genes Dev. 12:2912-2920[Abstract/Free Full Text].

NAGARAJAN, R. and Y. CHEN, 2000  Structural basis for the functional difference between Smad2 and Smad3 in FAST-2 mediated transcription. Biochem. J. 350:253-259.

NELLEN, D., R. BURKE, G. STRUHL, and K. BASLER, 1996  Direct and long-range action of a DPP gradient. Cell 85:357-368[Medline].

NEWFELD, S., E. CHARTOFF, J. GRAFF, D. MELTON, and W. GELBART, 1996  Mad encodes a conserved cytoplasmic protein required in DPP/TGF-ß responsive cells. Development 122:2099-2108[Abstract].

NEWFELD, S., R. WISOTZKEY, and S. KUMAR, 1999  Molecular evolution of a developmental pathway: phylogenetic analyses of TGF-ß family ligands, receptors and Smad signal transducers. Genetics 152:783-795[Abstract/Free Full Text].

PADGETT, R., J. WOZNEY, and W. GELBART, 1993  Human BMP sequences confer normal dorsal-ventral patterning in the Drosophila embryo. Proc. Natl. Acad. Sci. USA 90:2905-2909[Abstract/Free Full Text].

PARK, M., C. LEWIS, D. TURBAY, A. CHUNG, and J. N. CHEN et al., 1998  Differential rescue of visceral and cardiac defects in Drosophila by vertebrate tinman-related genes. Proc. Natl. Acad. Sci. USA 95:9366-9371[Abstract/Free Full Text].

RAFTERY, L. and D. SUTHERLAND, 1999  TGF-ß family signal transduction in Drosophila development: from Mad to Smads. Dev. Biol. 210:251-268[Medline].

RIGGINS, G., S. THIAGALINGAM, E. ROZENBLUM, C. WEINSTEIN, and S. KERN et al., 1996  Mad-related genes in the human. Nat. Genet. 13:347-349[Medline].

RIGGINS, G., K. KINZLER, B. VOGELSTEIN, and S. THIAGALINGAM, 1997  Frequency of Smad gene mutations in human cancers. Cancer Res. 57:2578-2580[Abstract/Free Full Text].

SAMPATH, T., K. RASHKA, J. DOCTOR, R. TUCKER, and F. HOFFMANN, 1993  Drosophila TGF-ß superfamily proteins induce endochondrial bone formation in mammals. Proc. Natl. Acad. Sci. USA 90:6004-6008[Abstract/Free Full Text].

SEKELSKY, J., S. NEWFELD, L. RAFTERY, E. CHARTOFF, and W. GELBART, 1995  Genetic characterization and cloning of Mad, a gene required for dpp function in Drosophila. Genetics 139:1347-1358[Abstract].

SHI, Y., Y. WANG, L. JAYARAMAN, H. YANG, and J. MASSAGUÉ et al., 1998  Crystal structure of a Smad MH1 domain: DNA binding in TGF-ß signaling. Cell 94:585-594[Medline].

SINGER, M., A. PENTON, V. TWOMBLY, F. HOFFMANN, and W. GELBART, 1997  Signaling through both type I DPP receptors is required for anterior-posterior patterning of the entire Drosophila wing. Development 124:79-89[Abstract].

SPEICHER, S., U. THOMAS, U. HINZ, and E. KNUST, 1994  The Serrate locus of Drosophila and its role in morphogenesis of the wing imaginal discs: control of cell proliferation. Development 120:535-544[Abstract].

SPREIJ, T., 1971  Cell death during the development of the imaginal disks of Calliphora.. Neth. J. Zool. 21:221-264.

STAEHLING-HAMPTON, K., P. JACKSON, M. CLARK, A. BRAND, and F. HOFFMANN, 1994  Specificity of BMP-related factors: cell fate and gene expression changes in Drosophila embryos induced by Dpp but not 60A. Cell Growth Differ. 5:585-593[Abstract].

STURTEVANT, M. and E. BIER, 1995  Analysis of the genetic hierarchy guiding wing vein development in Drosophila. Development 121:785-801[Abstract].

TABATA, T., C. SCHWARTZ, E. GUSTAVSON, Z. ALI, and T. KORNBERG, 1995  Creating a Drosophila wing de novo, the role of engrailed and the compartment border hypothesis. Development 121:3359-3369[Abstract].

TAKAKU, K., M. OSHIMA, H. MIYOSHI, M. MATSUI, and M. F. SELDIN et al., 1998  Intestinal tumorigenesis in compound mutant mice of both Smad4 and Apc genes. Cell 92:645-656[Medline].

THEISEN, H., T. HAERRY, M. O'CONNOR, and J. MARSH, 1996  Developmental territories created by mutual antagonism between Wingless and Decapentaplegic. Development 122:3939-3984[Abstract].

TOPPER, J., J. CAI, Y. QIU, K. ANDERSON, and Y. XU et al., 1997  Vascular Mads: two novel Mad-related genes selectively inducible by flow in human vascular endothelium. Proc. Natl. Acad. Sci. USA 94:9314-9319[Abstract/Free Full Text].

WIERSDORFF, V., T. LECUIT, S. COHEN, and M. MLODZIK, 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[Abstract].

WILDER, E. and N. PERRIMON, 1995  Dual functions of wingless in the Drosophila leg imaginal disc. Development 121:477-488[Abstract].

WRANA, J., 2000  Regulation of Smad activity. Cell 100:189-192[Medline].

XU, X., Z. YIN, J. HUDSON, E. FERGUSON, and M. FRASCH, 1998  Smad proteins act in combination with synergystic and antagonistic regulators to target Dpp responses to the Drosophila mesoderm. Genes Dev. 12:2354-2370[Abstract/Free Full Text].

YAGI, K., D. GOTO, T. HAMAMOTO, S. TAKENOSHITA, and M. KATO et al., 1999  Alternatively spliced variant of Smad2 lacking exon 3: comparison with wild-type Smad2 and Smad3. J. Biol. Chem. 274:703-709[Abstract/Free Full Text].

YANAGISAWA, K., H. OSADA, A. MASUDA, M. SAITO, and Y. YATABE et al., 1998  Induction of apoptosis by Smad3 and down regulation of Smad3 expression in response to TGF-ß in human normal lung epithelial cells. Oncogene 17:1743-1747[Medline].

ZHU, Y., J. RICHARDSON, L. PARADA and J. GRAFF, 1998 Smad3 mutant mice develop metastatic colorectal cancer. Cell 94: 703–714.6.

This article has been cited by other articles:

				J. Batut, B. Schmierer, J. Cao, L. A. Raftery, C. S. Hill, and M. Howell Two highly related regulatory subunits of PP2A exert opposite effects on TGF-{beta}/Activin/Nodal signalling Development, September 1, 2008; 135(17): 2927 - 2937. [Abstract] [Full Text] [PDF]

				Y. A. Zeng, M. Rahnama, S. Wang, W. Sosu-Sedzorme, and E. M. Verheyen Drosophila Nemo antagonizes BMP signaling by phosphorylation of Mad and inhibition of its nuclear accumulation Development, June 1, 2007; 134(11): 2061 - 2071. [Abstract] [Full Text] [PDF]

				N. T. Takaesu, C. Hyman-Walsh, Y. Ye, R. G. Wisotzkey, M. J. Stinchfield, M. B. O'Connor, D. Wotton, and S. J. Newfeld dSno Facilitates Baboon Signaling in the Drosophila Brain by Switching the Affinity of Medea Away From Mad and Toward dSmad2 Genetics, November 1, 2006; 174(3): 1299 - 1313. [Abstract] [Full Text] [PDF]

				C. Molnar, A. Lopez-Varea, R. Hernandez, and J. F. de Celis A Gain-of-Function Screen Identifying Genes Required for Vein Formation in the Drosophila melanogaster Wing Genetics, November 1, 2006; 174(3): 1635 - 1659. [Abstract] [Full Text] [PDF]

				S. E. Bondos, X.-X. Tan, and K. S. Matthews Physical and Genetic Interactions Link Hox Function with Diverse Transcription Factors and Cell Signaling Proteins Mol. Cell. Proteomics, May 1, 2006; 5(5): 824 - 834. [Abstract] [Full Text] [PDF]

				J. Massague, J. Seoane, and D. Wotton Smad transcription factors Genes & Dev., December 1, 2005; 19(23): 2783 - 2810. [Abstract] [Full Text] [PDF]

				N. T. Takaesu, E. Herbig, D. Zhitomersky, M. B. O'Connor, and S. J. Newfeld DNA-binding domain mutations in SMAD genes yield dominant-negative proteins or a neomorphic protein that can activate WG target genes in Drosophila Development, November 1, 2005; 132(21): 4883 - 4894. [Abstract] [Full Text] [PDF]

				M. Muzzopappa and P. Wappner Multiple roles of the F-box protein Slimb in Drosophila egg chamber development Development, June 1, 2005; 132(11): 2561 - 2571. [Abstract] [Full Text] [PDF]

T. Furumatsu, M. Tsuda, N. Taniguchi, Y. Tajima, and H. Asahara Smad3 Induces Chondrogenesis through the Activation of SOX9 via CREB-binding Protein/p300 Re