Pannier Is a Transcriptional Target and Partner of Tinman During

Developmental Biology 233, 425–436 (2001) doi:10.1006/dbio.2001.0220, available online at http://www.idealibrary.com on View metadata, citation and similar papers at core.ac.uk brought to you by CORE Pannier is a Transcriptional Target and Partner provided by Elsevier - Publisher Connector of Tinman during Drosophila Cardiogenesis

Kathleen Gajewski,*,1 Qian Zhang,*,1 Cheol Yong Choi,†,1 Nancy Fossett,* Anh Dang,* Young Ho Kim,† Yongsok Kim,† and Robert A. Schulz*,2 *Department of Biochemistry and Molecular Biology, Graduate Program in Genes & Development, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030; and †Laboratory of Molecular Cardiology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892

During Drosophila embryogenesis, the homeobox gene tinman is expressed in the dorsal mesoderm where it functions in the speciﬁcation of precursor cells of the heart, visceral, and dorsal body wall muscles. The GATA factor gene pannier is similarly expressed in the dorsal-most part of the mesoderm where it is required for the formation of the cardial cell lineage. Despite these overlapping expression and functional properties, potential genetic and molecular interactions between the two genes remain largely unexplored. Here, we show that pannier is a direct transcriptional target of Tinman in the heart-forming region. The resulting coexpression of the two factors allows them to function combinatorially in the regulation of cardiac gene expression, and a physical interaction of the proteins has been demonstrated in cultured cells. We also deﬁne functional domains of Tinman and Pannier that are required for their synergistic activation of the D-mef2 differentiation gene in vivo. Together, these results provide important insights into the genetic mechanisms controlling heart formation in the Drosophila model system. © 2001 Academic Press Key Words: cardiogenic factors; D-MEF2; heart; Pannier; protein interactions; Tinman; transcriptional enhancer.

INTRODUCTION (Azpiazu and Frasch, 1993; Bodmer, 1993). Based on these properties, tin-related genes were cloned and characterized The study of heart development in Drosophila has been in vertebrate species, with several shown to be expressed in of signiﬁcant value, not only to our fundamental under- the cardiac lineage (reviewed in Harvey, 1996). Comparable standing of organogenesis in this model system, but also to the role of Dpp in regulating tin and promoting cardio- through the application of genetic insights to the investi- genesis, bone morphogenetic protein signaling induces gation of cardiogenesis and disease in evolutionarily ad- Nkx-2.5 expression and cardiac myogenesis in the chick vanced organisms. A prime example can be found in the embryo (Schultheiss et al., 1997). Likewise, genetic studies homeobox gene tinman (tin) and its vertebrate NK-2 class in the mouse and Xenopus systems have demonstrated counterparts. tin expression in the dorsal mesoderm is essential functions for the Nkx-2.3 and Nkx-2.5 genes in regulated by intercellular signaling mediated by the growth vertebrate heart development (Lyons et al., 1995; Fu et al., factor Decapentaplegic (Dpp) produced by adjacent ectoder- 1998; Grow and Krieg, 1998; Tanaka et al., 1999). Of mal cells (Frasch, 1995; Xu et al., 1998). tin function is medical importance, certain patients with congenital heart required in the dorsal mesoderm for the formation of disease have been shown to carry dominant mutations in progenitors of the heart, visceral, and dorsal somatic the human NKX2–5 gene, resulting in diverse cardiac de- muscles, and the mutation of this transcription factor velopmental abnormalities (Schott et al., 1998; Benson et results in an absence of these mesodermal derivatives al., 1999). tin is expressed in the dorsal mesoderm in a broad pattern 1 K.G., Q.Z., and C.Y.C. contributed equally to this work. that exceeds the limits of the heart-forming region (Yin and 2 To whom correspondence should be addressed. Fax: (713) 790- Frasch, 1998; Gajewski et al., 1999). This observation, 0329. E-mail: [email protected]. coupled with the inability of pan-mesodermal tin expres-

sion to produce ectopic heart cells (Yin and Frasch, 1998), Tin. Our results show that pnr is a direct target of Tin implicated the involvement of additional factors in the transcriptional activation during heart formation. Addition- genesis of cardiac cells within the dorsal-most mesoderm. ally, we identify regions of the two proteins that are Based on the requirement of a GATA element for the essential for their combinatorial regulation of cardiac gene function of a heart enhancer of the D-mef2 differentiation expression. These findings contribute needed information gene (Gajewski et al., 1998) and the demonstrated interac- to our understanding of the genetic pathways controlling tion of NKX2–5 with GATA4 in the regulation of cardiac heart development in Drosophila. gene expression in vertebrates (Durocher et al., 1997; Lee et al., 1998; Sepulveda et al., 1998), a role was predicted for a GATA factor in fly heart development. Subsequently, the MATERIALS AND METHODS GATA gene pannier (pnr) was shown to be expressed in the heart forming region and required for cardial cell formation In Vitro DNA Binding Assays (Gajewski et al., 1999). The coexpression of Tin and Pnr Electrophoretic mobility shift and DNase I protection assays results in a synergistic activation of cardiac gene expression were performed as described in Lee et al. (1997). To prepare probes, and the induction of an apparent cardioblast fate in both equimolar amounts of single-strand oligonucleotides were an- mesodermal and nonmesodermal cell types. Together, nealed and end-labeled with T4 kinase. The sequence of the these results provide evidence that the two cardiogenic oligonucleotides used for the gel shift assays were as follows. Ј factors work together to program cells into the cardiac Wild-type probes: pnrA, 5 -CCTCGGTTCTCAAGTGCGGAC- Ј Ј Ј lineage. Additionally, the forced expression of either Pnr or CTG-3 ; pnrB, 5 -CAGGTCCGCACTTGAGAACCGAGG-3 . Mu- tant probes: pnrAm, 5Ј-CCTCGGTTCTgAAGTcCGGACCTG-3Ј; mouse GATA4 throughout the Drosophila mesoderm re- pnrBm, 5Ј-CAGGTCCGgACTTcAGAACCGAGG-3Ј. The mu- sults in supernumerary cardial cells, demonstrating an tated nucleotides are in lower case. For the competitive gel-shift evolutionarily conserved function between the two GATA assays, a 100-fold molar excess of unlabeled DNA was used. For the genes in heart development. footprinting assay, the pnr enhancer DNA was digested with A further complexity in the genetic control of Drosophila HindIII (plus strand labeling) or XbaI (minus strand labeling), and cardiogenesis was found in the analysis of the gene encod- the linearized DNAs were dephosphorylated with alkaline phos- ing the multitype zinc finger protein U-shaped (Ush). This phatase. These DNAs were then digested with XbaI (plus strand) or regulator was previously shown to be an antagonist of Pnr HindIII (minus strand), eluted from a gel, and end-labeled with T4 function in neuronal precursor cell determination, altering kinase. Binding reactions were undertaken with increasing ␮ the properties of the GATA factor from a transcriptional amounts (0.2, 0.4, and 0.6 g) of GST-NK4 protein and labeled DNA (12.5 fmol) in 20 ␮l of binding buffer. After incubation for 15 activator to a repressor in cells where the two are coex- min at room temperature, CaCl2 (final 0.5 mM) and DNase I (0.1 pressed (Haenlin et al., 1997; Garcia-Garcia et al., 1999). unit) were added to the binding mixture. The DNase I digestions Similarly, in the dorsal mesoderm, Ush controls cardial cell were stopped after 1 min incubation by adding stop buffer followed specification through its modulation of Pnr transcriptional by analysis via denaturing polyacrylamide gel electrophoresis. activity (Fossett et al., 2000). A loss of ush function results in an overproduction of cardial cells, while a gain of Expression Analyses of pnr Enhancer–lacZ Fusion function leads to a reduction of cardioblasts, phenotypes Genes that are directly opposite to those seen with pnr. The structurally related Friend of GATA-2 (FOG-2) protein is Transformant strains expressing the 457-bp pnr dorsal meso- expressed during heart formation in the mouse and likewise derm enhancer–lacZ fusion gene have been described (Gajewski functions as a modulator of GATA4 transcriptional activity et al., 1999). To monitor enhancer activity in a tin gain of in cultured cells (Lu et al., 1999; Svensson et al., 1999; function background, pnr-lacZ males were crossed to UAS-tin virgins (Yin and Frasch, 1998), and the resulting heterozygous Tevosian et al., 1999). The recent analysis of targeted gene males were mated to twi-Gal4 virgins (Baylies and Bate, 1996). mutations have demonstrated a vital function for FOG-2 in To assay pnr-lacZ expression in tin loss of function embryos, heart development as mutants display morphogenetic ab- males were crossed to tin346/TM3, Sb virgins, and animals lack- normalities and an absence of the coronary vasculature ing the balancer chromosome were obtained. Embryos were col- (Svensson et al., 2000b; Tevosian et al., 2000). Additional lected from the mating of these flies and immunostained for evidence that Ush and FOG-2 share functional properties is both ␤-galactosidase and Tin, with homozygous embryos iden- provided by the depletion of heart cells in Drosophila tified by the absence of Tin protein. tin346 has been classified as embryos expressing FOG-2 in the mesoderm (N. Fossett, K. a null allele of the gene (Azpiazu and Frasch, 1993). Mutations Gajewski, S. Tevosian, S. Orkin, and R. Schulz, unpub- were introduced into the Tin-binding sites in the pnr enhancer lished results). using the QuickChange Site-Directed Mutagenesis Kit from Stratagene (La Jolla, CA). Primers used were as follows: mutTin1, While substantial data exist on the functions of indi- 5Ј-CCTACCGTTCTCCGGTCTgtAGacCGGACCTGCCTCGAT- vidual genes in the specification of the cardiac lineage in TTCCC-3Ј and 5Ј-GGGAAATCGAGGCAGGTCCGgtCTacA- Drosophila, less is known about genetic and molecular GACCGGAGAACGGTAGG-3Ј; mutTin2, 5Ј-CCTGCCTCGA- interactions between these early-acting cardiogenic factors. TTTCCCGGCTgtAGacCGCGGGACATCATTTTTC-3Ј and In this report, we investigate the regulation of the pnr gene 5Ј-GAAAAATGATGTCCCGCGgtCTacAGCCGGGAAATCGA- and the functional interactions of the GATA protein with GGCAGG-3Ј. Lower case letters correspond to the introduced base

changes. The processing and antibody staining of whole-mount CAATTG-3Ј and 5Ј-CAATTGACGCACTTGCGTCCCTCGC-3Ј; embryos, and the generation of embryo sections, was as described Pnr(G184R), 5Ј-GTGGCGACGGGATAGAACCGGACACTATC-3Ј in Gajewski et al. (1999). and 5Ј-GATAGTGTCCGGTTCTATCCCGTCG CCAC-3Ј; Pnr(C190S), 5Ј-CACTATCTGTCCAACGCCTGCGGAC-3Ј and 5Ј-GTCCGCAGGCGTTGGACAGATAGTG-3Ј; Pnr(D242S), 5Ј-GG- Colocalization of Tin and Pnr in Cultured Cells AGGCGCAACAACAGTGGCGAACCCGTG-3Ј and 5Ј- Ј Ј CV-1 cells were cotransfected with GFP- and Myc-tagged Tin CACGGGTTCGCCACTGTTGTTGCGCCTCC-3 ; Pnr(C247S), 5 - Ј Ј and Pnr expression vectors (1 ␮g per transfection). For the colocal- GGCGAACCCGTGTCCAATGCCTGCG-3 and 5 -CGCA- Ј ization analysis, cells were stained with DAPI, fixed 36 h after GGCATTGGACACGGGTTCGCC-3 ; Pnr (G457stop), Ј Ј Ј transfection, and subjected to immunohistochemistry using an 5 -CGGAACCCATTGACACCATGTGACG-3 and 5 -GT- Ј anti-Myc antibody (Invitrogen, San Diego, CA). Signals were de- CACATGGTGTCAATGGGTTCCG-3 . The mutations were veri- tected by confocal microscopy and analyzed as described in Kim et fied by DNA sequencing, followed by cloning into the XhoI and XbaI al. (1999). For coimmunoprecipitation experiments, nuclear ex- sites of pUAST. tracts were prepared and subjected to immunoprecipitation with anti-Myc antibody followed by Western blot with anti-GFP anti- GST Pull-Down Experiments body (Clontech, Palo Alto, CA) as described previously (Choi et al., 1999a). GST-Tin fusion proteins and the Pnr in vitro translation (IVT) product were generated by using methods described previously (Choi et al., 1999b). Fusion proteins were recovered from the Generation of tin and pnr Gene Mutants supernatants of sonicated BL21DE3 cells (Invitrogen) and coupled to GSH beads (Amersham Pharmacia, Piscataway, NJ) at room For the production of the various UAS-Tin and UAS-Pnr con- temperature. Beads were washed thoroughly to remove unbound structs used in this study, cDNAs encoding either full-length or protein and stored at 4°C until use. The Pnr IVT product was either mutant forms of Tin (Choi et al., 1999b) or Pnr (Winick et al., 1993) used immediately upon synthesis or stored at Ϫ70°C for up to 3 were introduced into the P element vector pUAST (Brand and days. For the pull-down reactions, the relative protein concentra- Perrimon, 1993). Transgenic lines harboring the UAS-cDNAs were tion of the GSH-GST-Tin constructs or GSH-GST beads were established by using standard procedures as described in Gajewski determined by using protein separation by SDS-PAGE and staining et al. (1997). with Coomassie blue. The volume of beads was adjusted by using To generate Tin(N351Q), a NotI–ApaI DNA fragment from pCMV- unbound beads so that all preparations had the same relative NK4 N351Q was inserted into the NotI and ApaI sites of pUAST. The amount of protein in a 60-␮l volume. As a control, the aliquots of construction of Tin(1–35, 46–416) involved a multistep procedure. A adjusted bead volumes were assayed again for the same relative Ј 203-bp DNA fragment corresponding to the 5 -nontranslated region protein concentration, using SDS-PAGE and Coomassie blue stain- through amino acid 35 was amplified by PCR with primer NK453 and ing. A total of 60 ␮l of GSH-GST-Tin beads or GSH-GST beads Ј Ј primer LEE, 5 -CTGGAATTCATCCGCGTTCGTCAGGCT-3 . The were incubated with 10 ␮lof35S-methionine-labeled Pnr in binding DNA was codigested with EcoRI and SalI and subcloned into the buffer (20 mM Tris–HCl, pH 7.6, 2.5 mM MgCl2, 1 mM EDTA, 0.5 corresponding sites of pKSII-A12, which contains the sequence encod- mM DTT, 0.15% NP-40, and 25 mg/ml BSA) for2hatroom ing Tin amino acids 46 through 416. The resulting internally deleted temperature with nutation. The beads were then washed six times DNA was then cloned into the NotI site of pUAST. For the construc- with buffer (20 mM Tris–HCl, pH 8.0, and 1% NP-40) to remove tion of Tin(111–416), Tin(152–416), Tin(192–416), Tin(232–416), unbound IVT product. The GST-Tin/Pnr complexes were eluted Tin(272–416), and Tin(1–109, 192–416), tin cDNAs were generated by from the beads by resuspension in 20 ␮l of SDS reducing buffer and PCR amplification with specific primers corresponding to the indi- heating for 5 min at 90°C. The entire volume of each reaction was Ј cated amino acid residues. The primers included NK4–110 3–5, 5 - loaded onto 7.5% SDS-PAGE gels. Gels were run for 40 min at 150 Ј GATCCATGGATGTGGTGGCGCCATCGTC-3 and NK4–110 V, stained and dried, and exposed to a phosphor screen overnight Ј Ј 5–3, 5 -GATCCATGGTCGTCGTCCCTATCGCCACTC-3 ; NK4– followed by image processing with ImageQuant 5.0. 151, 5Ј-GATCCATGGTCCAGCACTCCGCCTCGGCCTAC-3Ј; NK4–191, 5Ј-GATCCATGGGTGGAGCCACGCCCTCGGCG-3Ј; NK4–231, 5Ј-GATCCATGGCGACGCAGAAGCTGTGCGT- RESULTS CAAT-3Ј; NK4–271, 5Ј-GATCCATGGTGACCTCCTCGCGTTC- CGAG-3Ј; NK4-term, 5Ј-GATTCTAGACTACATGTG- Ј Ј A pnr Dorsal Mesoderm Enhancer Is Regulated by CTGCATCTGTTG-3 . The 5 -nontranslated region of Tin was also the Tin Homeodomain Protein amplified with primers NK453NotI, 5Ј-GATGCGGCCGCA- ATTCAAGTGTCGCAA-3Ј and NK4 ATG, 5Ј-GATCA- Around the time of heart precursor cell specification, pnr CCATGGCTCGCTGCGTGATCACTTG-3Ј. The coding region RNA is detected in the dorsal mesoderm in cells that also DNAs were digested with NcoI and XbaI, while the 5Ј noncoding express the tin gene (Gajewski et al., 1999). We have DNA was digested with NotI and NcoI, followed by the cloning of the previously identified a pnr enhancer that is active in this resulting DNAs into the NotI and XbaI sites of the pUAST vector. heart-forming region and maps to a 457-bp DNA immedi- Immunostaining of embryos expressing the various UAS-tin con- ately upstream of the gene (Fig. 1A). Because the enhancer structs with a Tin antibody (Jagla et al., 1997) showed that all mutant proteins were stably expressed and detectable in the transgenic lines contains two putative Tin recognition sites, we investi- used. gated the binding of the factor to this DNA. A GST-Tin To generate the UAS-Pnr constructs, mutant cDNAs were gener- fusion protein was used in a gel-shift assay to test its ability ated using the QuickChange Site-Directed Mutagenesis Kit. Primers to bind to the Tin1 sequence TCAAGTG, a known recog- used were as follows: Pnr(E168K), 5Ј-CGAGGGACGCAAGTGCGT- nition element of the homeodomain protein in mesodermal

FIG. 1. Identiﬁcation of Tin binding sites in a regulatory sequence upstream of the pnr gene. (A) Schematic of the pnr gene and its adjacent 5Ј-ﬂanking sequence. The location of a 457-bp dorsal mesoderm enhancer is shown on the left; Tin1 and Tin2 indicate consensus binding sites (TCAAGTG) for the protein. (B) Electrophoretic mobility shift assay using a Tin fusion protein (GST-NK4) and 32P-labeled wild-type (WT) or mutant (MT) oligonucleotides corresponding to the Tin1 binding site. (C) DNase I protection assay performed with either a 32P-labeled (Ϫ)or(ϩ) strand of the pnr dorsal mesoderm enhancer and increasing amounts of the Tin fusion protein (GST-NK4). Lanes: G ϩ A, chemical sequencing reaction of the pnr DNA; BSA, reaction with bovine serum albumin control protein added. Boxes 1 and 2 indicate protected sequences corresponding to the Tin1 and Tin2 sites.

enhancers of the D-mef2, tin, and ␤3 tubulin genes (Gajew- and mesectoderm under the control of the twi-Gal4 driver. ski et al., 1997; Lee et al., 1997; Xu et al., 1998; Kremser et We observed an expanded function of the enhancer within al., 1999). Tin can bind speciﬁcally to the Tin1 consensus the mesoderm, coupled with ectopic activity in midline but not to a mutant version of the sequence (Fig. 1B). DNase cells of the central nervous system (CNS) due to the forced I protection assays were also performed with the fusion expression of Tin (Figs. 2C and 2D). Conversely, in tin null protein on the pnr DNA, and two separate footprints were embryos, a complete absence of ␤-galactosidase expression obtained that correspond to the Tin1 and Tin2 sequences was observed which demonstrated a requirement of Tin (Fig. 1C). These in vitro experiments demonstrate that Tin function for enhancer activity (Fig. 2H). As a control, we can recognize and bind to two sites within the pnr dorsal detected a normal pattern of pnr-lacZ transgene expression mesoderm enhancer, suggesting the regulatory DNA might in tin heterozygous embryos (Fig. 2G). be a direct target of Tin transcriptional activity. The functional importance of the Tin1 and Tin2 sites The deﬁned pnr enhancer functions in the dorsal-most within the pnr control region was assessed by individually cells of the mesoderm and in cells of the amnioserosa (Figs. or simultaneously mutating these sequences and testing 2A and 2B; Gajewski et al., 1999). To determine whether its the effects of the alterations on enhancer function. Trans- activity was regulated by Tin around the time of heart cell formant lines that contained the mutTin1, mutTin2, or formation, we monitored the expression of a pnr enhancer– mutTin1/2 pnr enhancer-lacZ transgenes showed identical lacZ fusion gene in tin gain and loss of function embryos. patterns of a loss of expression from the dorsal mesoderm Initially, we used the Gal4/UAS binary system (Brand and (Figs. 2E and 2F). Thus, the presence of both Tin sites is Perrimon, 1993) to express tin throughout the mesoderm required for the normal activity of the enhancer. The results

FIG. 2. Tin is a transcriptional regulator of the pnr dorsal mesoderm enhancer. (A) Ventral view of a whole-mount wild-type embryo expressing the 457-bp pnr enhancer-lacZ transgene in the dorsal mesoderm. (B) Cross section of an embryo of the same genotype as (A) showing enhancer activity in the dorsal mesoderm. (C) Ventral view of pnr-lacZ expression in a whole-mount embryo of the genotype twi-Gal4; UAS-tin. ␤-galactosidase activity is detected in an expanded pattern in the mesoderm (open arrowheads) and de novo along the ventral midline (closed arrowhead). (D) Cross section of an embryo of the same genotype as (C) showing ectopic enhancer function in the mesoderm (open arrowhead) and derivatives of the mesectoderm (closed arrowhead). (E) Cross section of an embryo expressing the mutTin1 pnr-lacZ transgene. ␤-galactosidase activity is lost from the dorsal mesoderm (open arrowhead) and gained in the dorsal ectoderm. (F) Cross section of an embryo expressing the mutTin2 pnr-lacZ transgene. The arrowhead points to the absence of enhancer activity. (G) Cross section of a tin heterozygous embryo double stained for Tin protein (blue) and ␤-galactosidase activity (brown) produced under the control of the pnr enhancer. (H) Lack of pnr enhancer function (arrowhead) in a tin homozygous mutant embryo, deﬁnitively identiﬁed by the absence of Tin protein. All embryos are at stage 11. Abbreviations: de, dorsal ectoderm; dm, dorsal mesoderm; WT, wild type. FIG. 3. Nuclear colocalization and physical interaction of Pnr and Tin in cultured cells. (A) Confocal images showing colocalization of Tin (green) and Pnr (red) within the nucleus (MERGE, yellow). Cells were cotransfected with GFP-Tin and Myc-Pnr expression vectors and proteins were detected with GFP signal and anti-Myc antibody. Nuclei were also stained with DAPI (blue). (B) Nuclear extracts from cotransfected cells were subjected to immunoprecipitation followed by Western blot analysis of proteins with the indicated antibodies. Tin and Pnr are coimmunoprecipitated, indicating the two proteins interact with each other in the nucleus. Abbreviations: IgG, IgG band; IP, immunoprecipitation; WB, Western blot.

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Functional Interactions of Pannier and Tinman 431 of these mutagenesis experiments coupled with the tin gain of the Tin recognition sequences (Gajewski et al., 1997, and loss of function studies demonstrated the pnr dorsal 1998). Coexpression of the two factors in CNS midline cells mesoderm enhancer is directly regulated by the Tin home- results in the ectopic activation of the D-mef2 enhancer odomain protein. We also note that the transformant em- normally expressed only in cardial cells (Fig. 4B; Gajewski bryos displayed ectopic ␤-galactosidase activity in the dor- et al., 1999). This result is compatible with the nuclear sal ectoderm. The inappropriate ectodermal expression of colocalization and physical interaction of Tin and Pnr in the mutated enhancers may be indicative of the existence of cultured cells and provides an embryological assay for a repressor protein that interacts with the Tin recognition identifying regions of the proteins that are essential for sequences in these cells (Xu et al., 1998). their functional synergism. Nine deleted or point mutant Since a single GATA sequence was also present in the pnr versions of Tin were tested in the synergism assay (Fig. 4A). enhancer at a position distal to the two required Tin sites Tin(N351Q) has a single amino acid change in the home- (Winick et al., 1993), we monitored the expression of the odomain and is unable to bind DNA (Choi et al., 1999b). pnr-lacZ transgene in a twi-GAL4;UAS-pnr genetic back- Coexpression of this mutant with wild-type Pnr fails to ground. No ectopic expression of the reporter sequence was activate the D-mef2 enhancer (Fig. 4C). While a competent observed in these embryos (data not shown), suggesting that homeodomain must be present in Tin for synergism with pnr enhancer activity is not subject to transcriptional Pnr, this region by itself is not sufficient as it fails in the autoregulation. coactivation assay (Fig. 4A; data not shown). The TN domain is a highly conserved 12 amino acid region found in Tin and most other NK-2 class proteins (Harvey, 1996). A Colocalization and Physical Interaction of the Pnr 10-amino acid deletion was made within this domain to and Tin Proteins in Cultured Cells generate the Tin(1–35, 46–416) mutant, but this altered We have previously shown that the coexpression of Pnr protein was still able to function combinatorially with Pnr and Tin in the embryo resulted in a synergistic activation of (Fig. 4D). Thus, the TN domain is dispensable in the cardiac gene expression and the ectopic induction of heart- synergism assay. like cells (Gajewski et al., 1999). To investigate a possible A transcriptional activation domain has been mapped to physical interaction of the two proteins, we cotransfected the N-terminal 114 amino acids of Tin by using a cell mammalian CV-1 cells with Pnr and Tin expression vectors transfection strategy (Choi et al., 1999b). To determine and monitored the cellular distribution of the factors by whether this region was required for functional interaction confocal microscopy and coimmunoprecipitation analysis. with Pnr, the Tin(111–416) deletion was generated and As shown in Fig. 3A, cells expressing GFP-tagged Tin and tested. This truncated protein remained competent to syn- Myc-tagged Pnr exhibited colocalization of the proteins in ergize with Pnr in the activation of the D-mef2 enhancer, the nucleus. Furthermore, analysis of nuclear extracts by showing that the Tin transactivation domain was not immunoprecipitation followed by Western blot revealed required (Fig. 4E). However, larger N-terminal deletions that Tin was coimmunoprecipitated with Pnr (Fig. 3B, lane resulted in Tin proteins that were functionally inactive (Fig. 6). Conversely, Pnr was also detected in the Tin immuno- 4A). Specifically, removal of an additional 41 amino acids in precipitate (Fig. 3B, lane 12). These results indicate that the Tin(152–416) identified residues 111 to 151 as essential for two proteins form a physical complex in the cultured cells, Tin synergism with Pnr (Fig. 4F). The Tin(1–109, 192–416) consistent with their ability to work combinatorially in the variant that contains the transactivation domain and home- regulation of cardiac gene expression. odomain, but lacks internal sequences including the required 41-amino acid region, is likewise nonfunctional in the D-mef2 enhancer coactivation assay (Fig. 4G). There- Identification of Tin Protein Domains Essential for fore, these studies identify two distinct regions of Tin Its Functional Interaction with Pnr needed for its combinatorial function with Pnr, an internal The D-mef2 gene is a direct transcriptional target of Tin segment of 41 amino acids adjacent to the transactivation and Pnr in cardioblasts. A defined heart enhancer for the domain and the conserved homeodomain. gene contains a pair of essential Tin binding sites and a In support of this in vivo data, we delineated regions of required GATA element located in close proximity to one Tin that were able to bind to the Pnr protein. Eight GST-Tin

FIG. 4. Protein domains needed for transcriptional synergism of Tin with Pnr. (A) Full-length, point mutant, truncated, and deleted forms of Tin used in the D-mef2 enhancer-lacZ activation assay. The (ϩ)or(Ϫ) designations indicate the ability or inability of a protein to synergize with Pnr. Abbreviations: TN, conserved TN domain; HD, homeodomain. (B–G) Stage 11 embryos stained for ␤-galactosidase activity due to the expression of the D-mef2-lacZ fusion gene. All embryos are of the genotype twi-Gal4; IIA237 D-mef2-lacZ/UAS-tin or UAS-mutant tin; UAS-pnr. Filled arrowheads point to the de novo expression of D-mef2-lacZ in CNS midline cells while open arrowheads indicate the lack of enhancer activation in these cells. All embryos show a strong overactivation of the fusion gene in the dorsal mesoderm due to the expression of wild type Pnr. This staining is in focus (arrows) in the embryos of (C) and (F).

inability of the truncated protein to synergize with Tin demonstrates an essential requirement of this Pnr sequence (Fig. 6B). Pnr(E168K) and Pnr(C190S) contain single amino acid changes in the N-terminal zinc finger that correspond to mutations found in the dominant alleles pnrD4 and pnrD3 (Ramain et al., 1993). These mutations may affect the formation of the first zinc finger and result in proteins that heterodimerize poorly with the Ush antagonist (Haenlin et al., 1997). However, both Pnr proteins were able to synergize with Tin and direct D-mef2 expression in the CNS (Figs. 6C and 6D). In contrast, the mutation of a conserved cysteine residue in zinc finger 2 in Pnr(C247S) inactivated the protein in the synergism assay (Fig. 6E). This amino acid change is likely to influence the formation of the C-terminal zinc finger and identifies this region as an essential functional domain of Pnr in the coactivation of D-mef2. It is important to note that, although Pnr(1–457) and Pnr(C247S) fail to synergize with Tin, they are competent to bind the homeodomain protein in the GST pull- FIG. 5. Analysis of the Pnr interaction domains of Tin. Various down assay (data not shown). In combination, these results GST-Tin fusion proteins were purified from bacterial cultures, substantiate that intrinsic functional properties of Pannier fused to glutathione-Sepharose beads, and subjected to pull-down 35 are perturbed in the two mutant forms of the GATA factor. assays with in vitro-translated S-methionine-labeled Pnr protein. The designations of the Tin proteins are given as A1 to A34, and Previous studies have shown that mouse GATA4, but not their amino acid compositions are listed in parentheses. The GATA1, could work with Tin in inducing cardial cell relative binding of Pnr to the wild-type or mutant forms of Tin was markers when coexpressed in the midline cells (Gajewski et determined from an average of six assays and indicated by (ϩ)or(Ϫ) al., 1999). One theory for the functional difference between on the right. The inset at the bottom shows the data from one of the the vertebrate factors in the Drosophila cardiogenic assay experiments and includes a sample of input 35S-labeled Pnr and a was that certain key amino acids, present in Pnr and GST protein control. GATA4, were not conserved in GATA1. We generated two Pnr proteins that changed amino acids found exclusively in Pnr and GATA4 to more closely resemble the GATA1 sequence. However, the zinc finger 1 mutant Pnr(G184R) fusion proteins containing all or part of the cardiogenic and the zinc finger 2 mutant Pnr(D242S) were both compe- factor were used in in vitro pull-down assays with 35S- tent in the enhancer coactivation test, demonstrating that labeled full-length Pnr (Fig. 5). The strongest interactions these conserved residues were not essential for Pnr func- were observed with either wild-type Tin or truncated ver- tional interaction with Tin (Figs. 6F and 6G). sions that contained the homeodomain region (Tin A1, A2, A23, and A31). Additionally, weaker but discernible binding of Tin and Pnr was observed with truncated proteins DISCUSSION that contained the functionally defined 111 to 151 region (Tin A4, A5, and A34). These molecular results are consis- Results from the current study contribute significantly to tent with the requirement of this internal domain for the our understanding of the genetic control of cardiogenesis in functional synergism of Tin with Pnr in the activation of Drosophila. Together with recent work from our lab and the D-mef2 cardiac enhancer in the CNS. others, we can elaborate a detailed regulatory hierarchy programming cardial cell specification and diversification (Fig. 7). The selective induction of the Tin homeodomain Identification of Pnr Protein Domains Essential for protein in the dorsal mesoderm is mediated by Dpp signals Its Functional Interaction with Tin emanating from the dorsal ectoderm (Frasch, 1995). This We also used the ectopic activation assay to determine involves the binding of Dpp-activated Smad proteins, and those regions of Pnr that were essential for its functional Tin itself, to a Dpp-responsive enhancer so as to activate tin synergism with Tin. Six deleted or point mutant forms were expression (Yin et al., 1997; Xu et al., 1998). The present tested for enhancer activation in CNS midline cells (Fig. work demonstrates that pnr is a direct transcriptional target 6A). Pnr(1–457) represents a C-terminal truncation of the of Tin in cells of the heart forming region. It is noteworthy GATA factor that maintains zinc fingers 1 and 2, but that two recent reports have likewise demonstrated the deletes two putative amphipathic ␣ helices. This direct activation of cardiac enhancers of the chick and C-terminal region has been shown to contain a transcrip- mouse gata6 genes by Nkx2.5 (Davis et al., 2000; Molken- tional activation domain (Haenlin et al., 1997), and the tin et al., 2000). These similar findings suggest that impor-

FIG. 6. Protein domains needed for the transcriptional synergism of Pnr with Tin. (A) Full-length, truncated, and point mutant forms of Pnr used in the Dmef2 enhancer-lacZ activation assay. The (ϩ)or(Ϫ) designations indicate the ability or inability of a protein to synergize with Tin. Abbreviations: zf1 and zf2, zinc ﬁnger domains 1 and 2; h1 and h2, putative amphipathic ␣ helices 1 and 2. (B–G) Stage 11 embryos stained for ␤-galactosidase activity due to the expression of the D-mef2-lacZ fusion gene. All embryos are of the genotype twi-Gal4; IIA237 D-mef2-lacZ/UAS-tin; UAS-pnr or UAS-mutant pnr. Filled arrowheads point to de novo expression of D-mef2-lacZ in CNS midline cells, while open arrowheads indicate the lack of enhancer activation in these cells. All embryos show a background activation of the fusion gene in the cephalic mesoderm due to the expression of wild-type Tin. This staining is in focus (arrows) in the embryos of (B) and (E).

two pairs per segment that express the orphan steroid receptor Seven Up (Svp; Fig. 7). A regulatory interaction exists between these transcription factor genes as one identified function of Svp in the two cell pairs is to repress the activity of tin (Gajewski et al., 2000). The resulting mutually exclusive expression of Tin and Svp leads to the activation of different sets of target genes, including the ␤3 tubulin structural gene in the four cell pair population and independent enhancers for the D-mef2 differentiation gene in either of the cell groups (Kremser et al., 1999; Gajewski et al., 2000). This presumably confers a needed functional diversity to cardial cells within the mature heart. Our survey of Tin and Pnr protein motifs needed in the D-mef2 coactivation assay reveals both factors require competent DNA-binding domains and at least one additional region for their synergistic interaction. The necessity of the former is not surprising as the D-mef2 control region contains clustered Tin and GATA elements that must be present for enhancer activity (Gajewski et al., 1997, 1998). Also, several studies have demonstrated that the physical association of NKX2.5 and GATA4 is mediated through the homeodomain and zinc finger 2 domains (Durocher et al., FIG. 7. Schematic of a genetic regulatory hierarchy controlling cardial cell specification and diversification during heart develop- 1997; Lee et al., 1998; Sepulveda et al., 1998). Given the ment in Drosophila. Arrows between genes indicate a direct or extensive sequence homology between the fly and mouse indirect transcriptional activation of a downstream locus. The factors in these two regions, comparable molecular interac- merging lines between tinman and pannier represent the combi- tions would be predicted for Tin and Pnr. natorial function of the two genes in specifying cardioblast precur- An unexpected finding of our current work is that, while sor cells. The blocked lines indicate a negative regulatory effect of the C-terminal transactivation domain of Pnr is required in one gene upon another. Boxes highlight progenitor or genetically the combinatorial assay, the N-terminal transactivation distinct groups of heart cells. The schematic is a simplification of domain of Tin is not. One could envision a mechanism the total genetic complexity controlling heart formation in Dro- wherein the presence of the single domain provided by Pnr sophila and does not necessarily list all essential components of is sufficient for the activation properties of the het- this developmental process. Abbreviations: 4 cc and 2 cc pairs, four or two pairs of cardial cells per segment of the dorsal vessel that erodimeric complex. Additionally, it can not be ruled out express the Tin or Svp transcription factors and separable enhanc- that a second transactivation domain exists in Tin that was ers for the D-mef2 differentiation gene. not revealed previously in cell transfection studies (Ranga- nayakulu et al., 1998; Choi et al., 1999b). Also of note is the nonrequirement of a proposed cardiogenic domain of Tin that maps to the N-terminus of the protein (Park et al., tant regulatory interactions between tin and gata class 1998; Ranganayakulu et al., 1998). Specifically, Tin(111– genes have been evolutionarily conserved during heart 416) is competent to work with Pnr in the cooperative development. We also show that Pnr and Tin colocalize to activation of the D-mef2 heart enhancer, despite the ab- the nucleus of transfected cells, where they form a physical sence of residues 1 through 42. complex. These observations are in accord with the genetic Instead, an internal 41-amino acid region between the requirement of pnr and tin for heart precursor formation Tin transactivation domain and homeodomain has emerged and the ability of the two factors to work combinatorially in as a vital sequence for functional interaction with Pnr. Choi the activation of cardiac gene expression and the production et al. (1999b) has recently assigned a repressor activity of of ectopic cardioblast-like cells (Azpiazu and Frasch, 1993; Tin to residues 111 through 188, and it is plausible that, Bodmer, 1993; Gajewski et al., 1999). Their functional based on the biological assay being used, multiple func- interplay, along with the negative modulatory effect of Ush tional characteristics may be uncovered within this region. on Pnr, culminates in the precise specification of the cardial In the context of Tin’s synergistic interaction with Pnr in cell lineage within the dorsal mesoderm (Fossett et al., regulating a defined cardiac enhancer, association of the 2000). two through this domain may prevent Pnr from interacting Subsequent to specification, cardiac progenitors undergo with other proteins such as Ush. At the same time, because symmetric or asymmetric cell divisions to generate non- Tin has the potential to act as a transcriptional repressor equivalent populations of cardial cells (Gajewski et al., that recruits Groucho via this domain (Choi et al., 1999b), 2000; Ward and Skeath, 2000). These include four pairs of the interaction of Tin and Pnr through the essential 111 to cells per segment of the dorsal vessel that express Tin and 151 subregion may be beneficial to Tin in its role as a

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Functional Interactions of Pannier and Tinman 435 transcriptional activator by eliminating its possible associa- deacetylase complex to repress transcription. J. Biol. Chem. 274, tion with inhibitory cofactors. Preliminary results suggest 33194–33197. the molecular interaction of Tin and Pnr may be due in part Choi, C. Y., Lee, Y. M., Kim, Y. H., Park, T., Jeon, B. H., Schulz, to the presence of this domain. R. A., and Kim, Y. (1999b). The homeodomain transcription Relative to the functions of Tin and Pnr in programming factor NK-4 acts as either a transcriptional activator or repressor and interacts with the p300 coactivator and the Groucho core- the cardial cell lineage, it will be critical to decipher the pressor. J. Biol. Chem. 274, 31543–31552. complex relationship of these two cooperative factors with Davis, D. L., Wessels, A., and Burch, J. B. (2000). An Nkx-dependent that of the Ush antagonist. How the association of this enhancer regulates cGATA-6 gene expression during early stages multitype zinc finger protein with Pnr changes the GATA of heart development. Dev. Biol. 217, 310–322. factor from a transcriptional activator to a repressor re- Deconinck, A. E., Mead, P. E., Tevosian, S. G., Crispino, J. D., Katz, mains a provocative question. Perhaps Ush recruits a re- S. G., Zon, L. I., and Orkin, S. H. (2000). FOG acts as a repressor pressor protein and histone deacetylase to assemble a core- of red blood cell development in Xenopus. Development 127, pressor complex. All FOG class proteins, including Ush, 2031–2040. contain a binding motif for the transcriptional corepressor Durocher, D., Charron, F., Warren, R., Schwartz, R. J., and Nemer, C-terminal Binding Protein (CtBP; Fox et al., 1999). How- M. (1997). The cardiac transcription factors Nkx2–5 and GATA-4 ever, the importance of this heterologous protein interac- are mutual cofactors. EMBO J. 16, 5687–5696. tion has not been firmly established due to conflicting Fossett, N., Zhang, Q., Gajewski, K., Choi, C. Y., Kim, Y., and Schulz, R. A. (2000). The multitype zinc-finger protein U-shaped reports on the requirement of CtBP for FOG repressor functions in heart cell specification in the Drosophila embryo. function (Fox et al., 1999; Deconinck et al., 2000; Svensson Proc. Natl. Acad. Sci. USA 97, 7348–7353. et al., 2000a). Whether Ush interacts with CtBP or another Fox, A. H., Liew, C., Holmes, M., Kowalski, K., Mackay, J., and Drosophila corepressor protein, the specification of the Crossley, M. (1999). Transcriptional cofactors of the FOG family cardial cell lineage will likely depend on the balance be- interact with GATA proteins by means of multiple zinc fingers. tween a promoting Tin-Pnr activation system and an in- EMBO J. 18, 2812–2822. hibitory Ush-Pnr repressor complex within cells of the Frasch, M. (1995). Induction of visceral and cardiac mesoderm by dorsal mesoderm. ectodermal Dpp in the early Drosophila embryo. Nature 374, 464–467. Fu, Y., Yan, W., Mohun, T. J., and Evans, S. M. (1998). Vertebrate ACKNOWLEDGMENTS tinman homologues XNkx2–3 and XNkx2–5 are required for heart formation in a functionally redundant manner. Develop- We thank M. Frasch for providing the Tin antibody, C. Y. Elliott ment 125, 4439–4449. and T. Vachris for technical help, and R. Grenda for assistance with Gajewski, K., Kim, Y., Lee, Y. M., Olson, E. N., and Schulz, R. A. figures. DNA sequences were determined by the M.D. Anderson (1997). D-mef2 is a target for Tinman activation during Drosoph- DNA Core Sequencing Facility supported by National Institutes of ila heart development. EMBO J. 16, 515–522. Health Grant CA16672. This research was supported by grants Gajewski, K., Kim, Y., Choi, C. Y., and Schulz, R. A. (1998). from the American Heart Association, Muscular Dystrophy Asso- Combinatorial control of Drosophila mef2 gene expression in ciation, and National Institutes of Health (HL59151) (to R.A.S.) and cardiac and somatic muscle cell lineages. Dev. Genes Evol. 208, the National Heart, Lung, and Blood Institute intramural research 382–392. program (to Y.K.). Gajewski, K., Fossett, N., Molkentin, J. D., and Schulz, R. A. (1999). The zinc finger proteins Pannier and GATA4 function as cardiogenic factors in Drosophila. Development 126, 5679–5688. REFERENCES Gajewski, K., Choi, C. Y., Kim, Y., and Schulz, R. A. (2000). Genetically distinct cardial cells within the Drosophila heart. Azpiazu, N., and Frasch, M. (1993). tinman and bagpipe: Two Genesis 28, 36–43. homeobox genes that determine cell fates in the dorsal meso- Garcia-Garcia, M. J., Ramain, P., Simpson, P., and Modolell, J. derm of Drosophila. Genes Dev. 7, 1325–1340. (1999). Different contributions of pannier and wingless to the Baylies, M. K., and Bate, M. (1996). twist: A myogenic switch in patterning of the dorsal mesothorax of Drosophila. Development Drosophila. Science 272, 1481–1484. 126, 3523–3532. Benson, D. W., Silberbach, G. M., Kavanaugh-McHugh, A., Cottrill, Grow, M. W., and Krieg, P. A. (1998). Tinman function is essential C., Zhang, Y., Riggs, S., Smalls, O., Johnson, M. C., Watson, for vertebrate heart development: Elimination of cardiac differ- M. S., Seidman, J. G., et al. (1999). Mutations in the cardiac entiation by dominant inhibitory mutants of the tinman-related transcription factor NKX2.5 affect diverse cardiac developmental genes, XNkx2–3 and XNkx2–5. Dev. Biol. 204, 187–196. pathways. J. Clin. Invest. 104, 1567–1573. Haenlin, M., Cubadda, Y., Blondeau, F., Heitzler, P., Lutz, Y., Bodmer, R. (1993). The gene tinman is required for specification of Simpson, P., and Ramain, P. (1997). Transcriptional activity of the heart and visceral muscles in Drosophila. Development 118, Pannier is regulated by heterodimerization of the GATA DNA- 719–729. binding domain with a cofactor encoded by the u-shaped gene of Brand, A. H., and Perrimon, N. (1993). Targeted gene expression as Drosophila. Genes Dev. 11, 3096–3108. a means of altering cell fates and generating dominant pheno- Harvey, R. P. (1996). NK-2 homeobox genes and heart develop- types. Development 118, 401–415. ment. Dev. Biol. 178, 203–216. Choi, C. Y., Kim, Y. H., Kwon, H. J., and Kim, Y. (1999a). The Jagla, K., Frasch, M., Jagla, T., Dretzen, G., Bellard, F., and Bellard, homeodomain protein NK-3 recruits Groucho and a histone M. (1997). ladybird, a new component of the cardiogenic path-

way in Drosophila required for diversification of heart precur- Sepulveda, J. L., Belaguli, N., Nigam, V., Chen, C. Y., Nemer, M., sors. Development 124, 3471–3479. and Schwartz, R. J. (1998). GATA-4 and Nkx-2.5 coactivate Kim, Y. H., Choi, C. Y., and Kim, Y. (1999). Covalent modification Nkx-2 DNA binding targets: Role for regulating early cardiac of the homeodomain-interacting protein kinase 2 (HIPK2) by the gene expression. Mol. Cell. Biol. 18, 3405–3415. ubiquitin-like protein SUMO-1. Proc. Natl. Acad. Sci. USA 96, Svensson, E. C., Tufts, R. L., Polk, C. E., and Leiden, J. M. (1999). 12350–12355. Molecular cloning of FOG-2: A modulator of transcription factor Kremser, T., Gajewski, K., Schulz, R. A., and Renkawitz-Pohl, R. GATA-4 in cardiomyocytes. Proc. Natl. Acad. Sci. USA 96, (1999). Tinman regulates the transcription of the ␤3 tubulin gene 956–961. in the dorsal vessel of Drosophila. Dev. Biol. 216, 327–339. Svensson, E. C., Huggins, G. S., Dardik, F. B., Polk, C. E., and Lee, Y. M., Park, T., Schulz, R. A., and Kim, Y. (1997). Twist- Leiden, J. M. (2000a). A functionally conserved N-terminal mediated activation of the NK-4 homeobox gene in the visceral domain of the Friend of GATA-2 (FOG-2) protein represses mesoderm of Drosophila requires two distinct clusters of E-box GATA4-dependent transcription. J. Biol. Chem. 275, 20762– regulatory elements. J. Biol. Chem. 272, 17531–17541. 20769. Lee, Y., Shioi, T., Kasahara, H., Jobe, S. M., Wiese, R. J., Markham, Svensson, E. C., Huggins, G. S., Lin, H., Clendenin, C., Jiang, F., B. E., and Izumo, S. (1998). The cardiac tissue-restricted ho- Tufts, R., Dardik, F. B., and Leiden, J. M. (2000b). A syndrome of meobox protein Csx/Nkx2.5 physically associates with the zinc tricuspid atresia in mice with a targeted mutation of the gene finger protein GATA4 and cooperatively activates atrial natri- encoding Fog-2. Nat. Genet. 25, 353–356. uretic factor gene expression. Mol. Cell. Biol. 18, 3120–3129. Tanaka, M., Chen, Z., Bartunkova, S., Yamasaki, N., and Izumo, S. Lu, J. R., McKinsey, T. A., Xu, H., Wang, D. Z., Richardson, J. A., (1999). The cardiac homeobox gene Csx/Nkx2.5 lies genetically and Olson, E. N. (1999). FOG-2, a heart- and brain-enriched upstream of multiple genes essential for heart development. cofactor for GATA transcription factors. Mol. Cell. Biol. 19, Development 126, 1269–1280. 4495–4502. Tevosian, S. G., Deconinck, A. E., Cantor, A. B., Rieff, H. I., Lyons, L., Parsons, L. M., Hartley, L., Li, R., Andrews, J. E., Robb, Fujiwara, Y., Corfas, G., and Orkin, S. H. (1999). FOG-2: A novel L., and Harvey, R. P. (1995). Myogenic and morphogenetic GATA-family cofactor related to multitype zinc-finger proteins Friend of GATA-1 and U-shaped. Proc. Natl. Acad. Sci. USA 96, defects in the heart tubes of murine embryos lacking the ho- 950–955. meobox gene Nkx-2.5. Genes Dev. 9, 1654–1666. Tevosian, S. G., Deconinck, A. E., Tanaka, M., Schinke, M., Molkentin, J. D., Antos, C., Mercer, B., Taigen, T., Miano, J. M., Litovsky, S. H., Izumo, S., Fujiwara, Y., and Orkin, S. H. (2000). and Olson, E. N. (2000). Direct activation of a GATA6 cardiac FOG-2, a cofactor for GATA transcription factors, is essential for enhancer by Nkx2.5: Evidence for a reinforcing regulatory net- heart morphogenesis and development of coronary vessels from work of Nkx2.5 and GATA transcription factors in the develop- epicardium. Cell 101, 729–739. ing heart. Dev. Biol. 217, 301–309. Ward, E. J., and Skeath, J. B. (2000). Characterization of a novel Park, M., Lewis, C., Turbay, D., Chung, A., Chen, J. N., Evans, S., subset of cardiac cells and their progenitors in the Drosophila Breitbart, R. E., Fishman, M. C., Izumo, S., and Bodmer, R. (1998). embryo. Development 127, 4959–4969. Differential rescue of visceral and cardiac defects in Drosophila Winick, J., Abel, T., Leonard, M. W., Michelson, A. M., Chardon- by vertebrate tinman-related genes. Proc. Natl. Acad. Sci. USA Loriaux, I., Holmgren, R. A., Maniatis, T., and Engel, J. D. (1993). 95, 9366–9371. A GATA family transcription factor is expressed along the Ramain, P., Heitzler, P., Haenlin, M., and Simpson, P. (1993). embryonic dorsoventral axis in Drosophila melanogaster. Devel- pannier achaete scute Drosophila , a negative regulator of and in , opment 119, 1055–1065. encodes a zinc finger protein with homology to the vertebrate Xu, X., Yin, Z., Hudson, J. B., Ferguson, E. L., and Frasch, M. (1998). transcription factor GATA-1. Development 119, 1277–1291. Smad proteins act in combination with synergistic and antago- Ranganayakulu, G., Elliott, D. A., Harvey, R. P., and Olson, E. N. nistic regulators to target Dpp responses to the Drosophila (1998). Divergent roles for NK-2 class homeobox genes in cardio- mesoderm. Genes Dev. 12, 2354–2370. genesis in flies and mice. Development 125, 3037–3048. Yin, Z., and Frasch, M. (1998). Regulation and function of tinman Schott, J. J., Benson, D. W., Basson, C. T., Pease, W., Silberbach, during dorsal mesoderm induction and heart specification in G. M., Moak, J. P., Maron, B. J., Seidman, C. E., and Seidman, J. G. Drosophila. Dev. Genet. 22, 187–200. (1998). Congenital heart disease caused by mutations in the transcription factor NKX2–5. Science 281, 108–111. Submitted for publication November 8, 2000 Schultheiss, T. M., Burch, J. B., and Lassar, A. B. (1997). A role for Revised January 25, 2001 bone morphogenetic proteins in the induction of cardiac myo- Accepted January 25, 2001 genesis. Genes Dev. 11, 451–462. Published online April 16, 2001