Tbx5 and Tbx4 transcription factors interact with a new chicken PDZ-LIM protein in limb and heart development

Developmental Biology 273 (2004) 106 – 120 www.elsevier.com/locate/ydbio

Tbx5 and Tbx4 transcription factors interact with a new chicken PDZ-LIM protein in limb and heart development Ange Krause, a William Zacharias, a Troy Camarata, a Barbara Linkhart, b Evelyn Law, a Antje Lischke, a Erik Miljan, c and Hans-Georg Simon a,* a

Department of Pediatrics, The Feinberg School of Medicine, Children’s Memorial Institute for Education and Research, Northwestern University, Chicago, IL 60614, USA b Department of Anatomy, University of Wisconsin-Madison, Madison, WI 53706, USA c Human Cell Group Department, ReNeuron Limited, Surrey GU2 7AF England, UK Received for publication 5 December 2003, revised 26 March 2004, accepted 3 May 2004

Abstract The T-domain transcription factors, Tbx5 and Tbx4, play important roles in vertebrate limb and heart development. To identify interacting and potential Tbx-regulating proteins, we performed a yeast two-hybrid screen with the C-terminal domain of Tbx5 as bait. We identified a new PDZ-LIM protein composed of one N-terminal PDZ and three C-terminal LIM domains, which we named chicken LMP-4. Among the Tbx2, 3, 4, 5 subfamily, we observed exclusive interaction with Tbx5 and Tbx4 proteins. Tbx3 nor Tbx2 can substitute for LMP-4 binding. While chicken LMP-4 associates with Tbx5 or Tbx4, it uses distinct LIM domains to bind to the individual proteins. Subcellular colocalization of LMP-4 and Tbx proteins supports the protein interaction and reveals interference of LMP-4 with Tbx protein distribution, tethering the transcription factors to the cytoskeleton. The protein – protein interaction indicates regulation of Tbx function at the level of transcription factor nuclear localization. During chicken limb and heart development, Tbx5/LMP-4 and Tbx4/LMP-4 are tightly co-expressed in a temporal and spatial manner, suggesting that they operate in the same pathway. Surprisingly, chicken LMP-4 expression domains outside those of Tbx5 in the heart led to the discovery of Tbx4 expression in the outflow tract and the right ventricle of this organ. The Tbx4expressing cells coincide with those of the recently discovered secondary anterior heart-forming field. The discrete posterior or anterior expression domains in the heart and the exclusive fore- or hindlimb expression of Tbx5 and Tbx4, respectively, suggest common pathways in the heart and limbs. The identification of a new Tbx5/4-specific binding factor further suggests a novel mechanism for Tbx transcription factor regulation in development and disease. D 2004 Elsevier Inc. All rights reserved. Keywords: Limb development; Cardiac development; Pattern formation; Protein interaction; Tbx4; Tbx5; LMP-4; Yeast two-hybrid; GST co-precipitation; In situ hybridization

Introduction Until recently, little was known about the molecular basis for the morphological differences between forelimbs (arms) and hindlimbs (legs). The identification of the forelimb-specific Tbx5 and the hindlimb-specific Tbx4, however, marked a turning point by providing candidate genes for the control of limb-type identity (Basson et al., * Corresponding author. Department of Pediatrics, The Feinberg School of Medicine, Children’s Memorial Institute for Education and Research (CMIER), Northwestern University, 2300 Children’s Plaza M/C 204, Chicago, IL 60614, USA. Fax: +1-773-880-8266. E-mail address: [email protected] (H.-G. Simon). 0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2004.05.024

1997; Gibson-Brown et al., 1996; Khan et al., 2002; Li et al., 1997; Simon et al., 1997). The encoded T-domain proteins are members of a rapidly growing and highly conserved family of transcription factors that share a region of homology with the DNA-binding domain (Tdomain) of the mouse brachyury (or T) gene product (Herrmann et al., 1990; Kispert and Herrmann, 1993). To understand the function of Tbx5 and Tbx4, several experiments, such as transplantations and the induction of ectopic limbs in the chick, have been conducted to show that Tbx5 and Tbx4 gene expression relates directly to the identity of the limb-type (Gibson-Brown et al., 1998; Isaac et al., 1998; Logan et al., 1998; Ohuchi et al., 1998). Experiments in chicken embryos suggested that forced

A. Krause et al. / Developmental Biology 273 (2004) 106–120

misexpression of Tbx5 in the hindlimb and Tbx4 or Pitx1 in the forelimb can at least in part reprogram skeletal and muscle elements into the opposite limb type (Logan and Tabin, 1999; Rodriguez-Esteban et al., 1999; Takeuchi et al., 1999). Based on this experimental work, Tbx4 and Tbx5 are considered to play a role in the specification of limb-type identity. Recent data from gain- and loss-of-function approaches in zebrafish and chick embryos (Ahn et al., 2002; Ng et al., 2002), as well as mouse embryos lacking Tbx5 (Agarwal et al., 2003; Rallis et al., 2003) suggest a complex interplay of Tbx5, Wnt2b, and Fgf10. While there are some discrepancies regarding the type of Wnt molecule and/or the hierarchy of these molecules in the animal models used, it is clear that Tbx5 also has a very early function in forelimb initiation. In the prospective wing territory of the chicken, Tbx5 transcription factors stimulate expression of the signaling molecules Wnt2b and Fgf10, both of which in turn establish a positive feedback loop and maintain Tbx5 expression. This signaling network is apparently critical for continued limb outgrowth, and in the absence of either factor, no limbs form (Agarwal et al., 2003; Galceran et al., 1999; Kawakami et al., 2001; Rallis et al., 2003; Sekine et al., 1999; Xu et al., 1998). Related experiments using dominant negative constructs in the chicken suggest similar functional roles for Tbx4, Wnt8c, and Fgf10 in the legs (Takeuchi et al., 2003). Thus, the specification of limb-type identity and limb initiation appear to be linked functions, with Tbx5/4 as major but probably not sole players. As in the limbs, Tbx5 appears to be the earliest determinant of vertebrate heart growth and has been implicated in the regulation of cardiomyocyte proliferation (Hatcher et al., 2001). In the mouse, Tbx5 is initially expressed in a domain that corresponds to and is contiguous with the cardiac crescent and the forelimb field (Bruneau et al., 1999). In zebrafish, Tbx5-expressing cells are also initially arranged in two bilateral stripes, which break up into an anterior group of cells contributing to the heart primordium and a posterior group of cells migrating towards the future pectoral fin bud (Ahn et al., 2002). Evidence for coordinated development of heart and limbs and the existence of a cardiomelic field also comes from correlated heart/limb anomalies in Mendelian syndromes in humans (Wilson, 1998). Limb and heart patterning are multistep processes that involve specification of the limb and heart fields, respectively, establishment of defined signaling centers that globally inform cells of their position, interpretation of positional signals, and regulated growth and differentiation. The Tbx proteins are implicated in a pathway that controls and regulates a cascade of molecular events in the cell; however, very little is known about the steps involved. In addition, it is not yet understood how the encoded Tbx proteins interact in a molecular and cellular network to execute their functions.

107

Here, we describe the isolation of a novel chicken PDZLIM domain protein in a two-hybrid screen using the carboxyterminal domain of Tbx5 as bait. The exclusive binding specificity, cellular co-localization and interference with Tbx protein nuclear distribution, as well as co-expression with Tbx5 and Tbx4 in developing limbs and heart suggest a role for this interacting protein in the regulation of Tbx transcription factor activity.

Materials and methods Unless otherwise noted, all standard cloning techniques were performed according to Sambrook et al. (1989). All enzymes and molecular biology reagents were obtained from Roche Molecular Biochemicals (Indianapolis, IN), or as indicated. Yeast two-hybrid library screen and liquid binding assay To identify Tbx5-interacting proteins, we employed the yeast two-hybrid system essentially as described by Golemis et al. (1997). PCR fragments encompassing the carboxyterminal domains of chicken Tbx5, Tbx4, Tbx3, or Tbx2 were subcloned into the LexA DNA-binding-domain fusion vector pMW103, thereby creating our ‘bait’ constructs. The Tbx baits were transformed into Saccharomyces cerevisiae EGY191 (two lexA operators) by the standard Lithium acetate method to give His, Leu prototrophs. Transformants were tested for expression of the LexA – Tbx fusion protein by immunoblotting using anti-Tbx4/5 or anti-lexA antibodies. As ‘prey’, we employed a chicken limb bud (st. 22 – 24) cDNA library fused to the B42 activation domain in the pJG4-5 vector. The EGY191 yeast cells transformed with the Tbx5 bait were cotransformed with the prey cDNA library, and the pRB1840 lacZ reporter plasmid (one lexA operator). Approximately 2  106 yeast transformants were selected for 5 days on glucose plates lacking His, Trp, Leu, and Ura. Colonies were re-screened for interaction of bait and prey using a dual selection based on growth on Leu plates and h-galactosidase (h-gal) activity on X-GAL plates. Colonies that were Leu+ and blue were selected for quantitative liquid h-gal assays according to Serebriiskii and Golemis (2000). Potential positive clones were sequenced from both 5V- and 3V-ends, and sequences were compared with the NCBI database. Preparation of recombinant fusion proteins in Escherichia coli and GST pull-downs Chicken Tbx3, Tbx4 and Tbx5 carboxyterminal protein domains were expressed as T7/His6-tagged fusions in the pET-21b vector (Novagen, Madison, WI), and the various chicken LMP-4 constructs were expressed as GST fusion proteins in the pGEX-4T-1 vector (Pharmacia, Piscataway, NJ). BL21(DE3) E. coli cells at an OD600 of about 0.5 were

108

A. Krause et al. / Developmental Biology 273 (2004) 106–120

induced with 1 mM isopropyl-h-D-thiogalactopyranoside (IPTG) at 30jC for 3 h to express the respective constructs. Cells were lysed by sonication in phosphate-buffered saline (PBS) containing 1 mM DTT, 5 mM EGTA, 20 mM MgCl2, 1 EDTA-free protease inhibitor mix, and 0.1% Tween-20. The lysate was centrifuged at 35,000 rpm for 45 min at 4jC. Tbx lysates were precleared with glutathione-Sepharose 4B beads (Pharmacia) to remove endogenous GST proteins. LMP-4 lysates were rocked for 1 h at 4jC with beads blocked with 50% calf serum. The LMP-4-bound beads were washed twice in wash buffer (PBS with 5 mM EGTA and 20 mM MgCl2), and then rocked with Tbx lysates (diluted in wash buffer with 10% glycerol) for 1 h at 4jC. Following two washes with wash buffer, the GST complexes were eluted from the beads with 10 mM glutathione at pH 7.4 or by boiling at 100jC in protein gel loading buffer. Bound Tbx proteins were visualized by Western blot via ECL (Pierce Biotechnology Inc., Rockford, IL) with rabbit polyclonal anti-Tbx4, anti-Tbx5 (Khan et al., 2002), anti-His6 (Pharmacia) and anti-T7 antibodies (Novagen). Blots were stripped 30 min at 50jC in Tris buffer with 2% SDS and 100 mM h-mercaptoethanol, and GST proteins were detected with anti-GST antibodies (Pharmacia).

Whole mount in situ hybridization Whole mount in situ hybridizations were performed in an INTAVIS InSitu Pro (San Marcos, CA) essentially as described (Logan et al., 1998; Nelson et al., 1996). Avoiding the conserved T-box, the Tbx4 and Tbx5 antisense probes contain the 3V untranslated regions and most of the transactivation domains. The chicken LMP-4 antisense probe spans the entire open reading frame of the gene. Specific probe hybridization was visualized with the NBT/BCIP staining protocol, which produces a blue/purple precipitation product. Sense RNA probes were used as controls to determine specific hybridization. Post whole mount in situ sectioning After whole mount in situ hybridization, stained embryos were washed into PBS, then incubated in 5% sucrose in PBS followed by an overnight incubation in 15% sucrose at 4jC. The embryos were then incubated overnight at 37jC in gelatin. Embryos were embedded in fresh gelatin, frozen in a methanol/dry ice bath, and stored at 80jC. Ten-micron cryosections were prepared and imaged at 20 magnification.

Cell culture and transfection RNA preparation and RT-PCR expression analysis COS-7 cells were grown to 80% confluency in DMEM supplemented with 10% fetal calf serum, 1% L-glutamine, and penicillin-streptomycin. Cells were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) with Nterminally fused EGFP-Tbx5 or EGFP-Tbx4 and HcRedLMP-4 for direct fluorescence analysis. Transfections were also performed using C-terminally tagged Tbx5-FLAG or Tbx4-FLAG and Myc-LMP-4 for indirect fluorescence observation. FLAG and Myc epitopes were detected using anti-FLAG M2 (Sigma, St. Louis, MO) and anti-c-Myc 9E10 (Sigma) antibodies, respectively. F-actin was stained with Alexa 488-phalloidin (Molecular Probes, Eugene, OR). Cells were fixed in 3% PFA and visualized by confocal microscopy. The 1% Triton X-100 treatment was performed as described (Vallenius and Makela, 2002). Vector-transfected cells were used as negative controls in image analyses. Chicken embryos All chicken embryos used in the experiments were White Leghorn chickens obtained from Charles River SPAFAS (Boston, MA). Eggs were incubated at 37.5jC for 2 to 7 days, and embryos were dissected out and cleared of membranes in PBS. All embryos were staged according to Hamburger and Hamilton (1951) (HH). Embryos to be used for in situ hybridization were fixed overnight in 4% Paraformaldehyde in PBS (PFA) at 4jC. For RT-PCR, limb regions were excised from embryos, frozen on dry ice, and stored at 80jC.

Chicken wing and leg regions at HH stages 9, 11, 13, 15, and 17 were collected as described above. Total RNA was prepared from 30 mg tissue using the RNAeasy miniprep protocol including DNase treatment (QIAGEN, Valencia, CA), and stored at 80jC. Extracted RNA was quantified at an optical density of 260 nm and tested for integrity in a denaturing agarose gel. Probe and primer sets were designed for b-actin, Tbx4, Tbx5 and LMP-4 using Primer Express software (Applied Biosystems, Foster City, CA). The Tbx4 and Tbx5 probes and primers were designed for their transactivation domains, and the LMP-4 probe and primers for the proline-rich region between the PDZ and LIM domains. Employed oligonucleotides specific for the individual genes were: b-actin: (FWD) 5 V- G ATAT T G C T G C G C T C G T T G T T, ( R E V ) 5 VC T G T C T T T C T G G C C C ATA C C A ; ( P R O B E ) 5 VTCCGGTATGTGCAAGGCCGGTTT; Tbx4: (FWD) 5VCAGCCACTTCAGCGTCTACAAC, (REV) 5V-ACGGACGGATGCACTCTTTC, (PROBE) 5V-CAGTCTC A G G T T C G G G A G C G T G T C ; T b x 5 : ( F W D ) 5 VAACAGGGACTGAACACTTCGTACA, (REV) 5V-GGAGC A G A G C T G G C G TA C AT, ( P R O B E ) 5V- C A G C C C A G C G C C A G G C AT G ; L M P - 4 : ( F W D ) 5 VCCTGGGCTGGTGACGAAAC, (REV) 5V-GGCCAATC C T G G T T C T G T C T T, ( P R O B E ) 5 V- A G C A C AATGGCCAGGTCTCCGTAC. RT-PCR was performed in an ABI Prism 7700 sequence detector using Taqman reagents (both from Applied Biosystems, Foster City, CA) essentially as described

A. Krause et al. / Developmental Biology 273 (2004) 106–120

(Hoffman et al., 2002; Mayanil et al., 2001). 25 ng of RNA were used in a 50 Al reaction as follows: Reverse transcription (RT): 48jC for 30 min, 95jC for 15 min; PCR: 40 cycles of 95jC for 15 s and 59jC for 1 min. During linear amplification, aliquots were removed for Tbx4, Tbx5, LMP-4 (after 28 cycles) and b-actin (after 24 cycles), and analyzed on 4% agarose gels to confirm that a single product of expected size was amplified. Band intensities were measured with an Alpha Imager (Alpha Innotech., San Leandro, CA). Reactions were run in triplicate from two different RNA preparations, including no RT and no template controls.

Results Isolation and characterization of Tbx binding proteins In alignments of Tbx proteins in different species, we have identified distinct and highly conserved Tbx4- and Tbx5-specific amino acid residues, particularly in the Cterminal transactivation domains (Khan et al., 2002; Simon, 1999). Guided by the hypothesis that these distinct motifs may have functional significance in binding accessory proteins, we performed a yeast two-hybrid screen (Golemis et al., 1997) with the chicken Tbx5 C-terminal domain as bait. Using h-gal liquid assays (Serebriiskii and Golemis, 2000), 19 potential Tbx5-binding partners were re-screened with Tbx5 and in addition as controls chicken Tbx4, Tbx3, and Tbx2, which belong to the same Tbx subfamily (Simon, 1999). For further studies, we selected clone II-3.48, which interacted specifically with Tbx5 and surprisingly with Tbx4 as well. In contrast, the less structural conserved Tbx2 and Tbx3 did not demonstrate binding, or their interaction was significantly weaker and below the detection level of the assay (data not shown). Chicken LMP-4 is a novel member of the PDZ-LIM family of proteins Clone II-3.48 showed striking sequence homology to the LIM domain, a well-documented protein –protein binding motif (Bach, 2000). We performed 5VRACE to complete the cDNA and acquire a full-length protein-coding region, which spans 416 amino acids. The protein contains an Nterminal PDZ domain, a large proline-rich intervening sequence, and three C-terminal LIM domains (Figs. 1A, C). BLAST searches and phylogenetic tree analyses characterized the chicken isolate as a close relative of human Enigma/LMP-1, LMP-2, LMP-3, and rat LMP-1 proteins, therefore placing it in the LIM mineralizing protein (LMP) subfamily of PDZ-LIM proteins (Fig. 1B). The PDZ and LIM domains of our isolate are highly homologous to those of human LMP-1 and LMP-2. However, considerable sequence differences among the LMP proteins occur within the proline-rich region. Because our

109

isolate does not closely identify with any of the LMP variants, we have designated it as chicken LMP-4. LMP-4 interacts specifically with the transactivation domains of Tbx5 and Tbx4, but not Tbx3 To further demonstrate interaction of chicken LMP-4 with Tbx proteins, we performed in vitro co-precipitation experiments. In Fig. 2, a constant input of GST-LIM2/3 (the original yeast two-hybrid isolate II-3.48, which contains only LIM domains 2 and 3) was bound to glutathione Sepharose beads and incubated with increasing amounts of E. coli lysates containing T7/His6-tagged Tbx3, Tbx4, or Tbx5 proteins. After several washes, the proteins were eluted from the beads and analyzed with various antibodies by Western blot. As shown in Figs. 2A and B, Tbx4 and Tbx5 co-precipitated with GST-LIM2/3, demonstrating interactions of the two transcription factors with LMP-4. In addition, increasing the amount of input Tbx proteins yielded concurrent higher levels of Tbx coprecipitations. The Tbx proteins alone did not co-precipitate with the glutathione beads, and reprobing the blots with anti-GST antibodies verified the presence of GSTLIM2/3 in the reaction. While anti-GST antibodies confirmed the presence of GST-LIM2/3 in the reaction, an interaction between GST-LIM2/3 and Tbx3 could not be observed (Fig. 2C). The exclusive interaction of LMP-4 with Tbx5 and Tbx4 agrees with and confirms our yeast two-hybrid data. To identify the domains of LMP-4 involved in the interaction with Tbx5 and Tbx4, various constructs containing the PDZ domain and proline-rich region, individual LIM domains, or combinations of the LIM domains were generated (Fig. 3A). Saturating amounts of the various GSTtagged LMP-4 protein domains were bound to the beads and subsequently incubated with constant concentrations of Tbx5 and Tbx4 lysates. Fig. 3B demonstrates that Tbx5 binds to GST-tagged LIM2/3, LIM1/2/3, and LIM3. However, no co-precipitations of Tbx5 with GST-tagged LIM1, LIM2, LIM1/2, or the PDZ/proline rich domains were detected, indicating that LIM domain 3 is necessary and sufficient for a Tbx5/LMP-4 interaction. Tbx4, however, demonstrates a different binding profile. In Fig. 3C, Tbx4 co-precipitates with the LIM1/2/3, LIM2/3, LIM1/2, and LIM2 GST fusions, but not with LIM1, LIM3, or the PDZ/ proline-rich domains, which suggests that LIM domain 2 is responsible for the interaction between Tbx4 and LMP-4. In some cases, we observed that the recombinant Tbx4 proteins ran as lower molecular-weight bands (Fig. 3C, lanes 4, 5, and 7). The differences may be due to conformational changes or partial degradations, although the alterations did not affect the performance of the co-precipitations. Both Tbx5 and Tbx4 did not interact with GST alone, and neither protein bound to GST-tagged Zyxin, another member of the LIM-domain protein family (Schmeichel and Beckerle, 1994). Collectively, our co-precipitation results are in agree-

110

A. Krause et al. / Developmental Biology 273 (2004) 106–120

Fig. 1. Chicken LMP-4 is a novel PDZ-LIM protein with two separate functional domains. (A) General architecture and domain arrangement of chicken LMP-4. Protein-coding regions are indicated by a bar, and functional domains are shaded. The 5’ and 3’ untranslated regions are represented by a solid line. (B) Phylogenetic tree of PDZ-LIM proteins. Amino acid sequences of various PDZ-LIM proteins in several species were compared in a phylogenetic tree using MacVector 7.0 software (Oxford Molecular Company, Madison, WI). Chick LMP-4 is boxed; note its location on a unique branch within the cluster of LMP proteins. Gg = Gallus gallus; Hs = Homo sapiens; Mm = mus musculus; Rn = Rattus norvegicus. Genbank accession numbers for all sequences are as follows: Gg Zyxin, A44358; Hs LMP-1, NP_005442; Hs LMP-2, AAK30568; Hs LMP-3, AAK30569; Gg LMP-4, AY376690; Mm Cypher, NP_036048; Mm ENH1, NP_062783; Mm ENH2, NP_062783; Mm ENH3, NP_072048; Mm Oracle-1, AAF33847; Mm Oracle-2, AAF33848; Rn ENH, NP_445778; Rn LMP-1, AAD13197. (C) Amino acid sequence alignment of LIM domains from chick LMP-4, closely related human LMP-2, and more distant relative chicken Zyxin. Conserved residues between LIM domains are indicated by shading. The general LIM domain consensus (CX2CX17 – 19HX2CX2CX2CX16 – 20 CX2C/D/H, from Dawid et al., 1995) verifies the sequences as LIM domains. The (*) symbol indicates residues that may consist of either C, D, or H.

ment with our yeast two-hybrid binding data and demonstrate that LMP-4 interacts selectively with Tbx5 and Tbx4. LMP-4 localizes Tbx5 and Tbx4 to actin filaments The results of our in vitro interaction studies require that both LMP-4 and Tbx5/4 localize to the same cellular compartment. To show this, COS-7 cells were transiently co-transfected or transfected individually with full-length constructs of EGFP-Tbx5 or EGFP-Tbx4 and HcRed-LMP-

4, and processed for direct fluorescent confocal microscopy. Control transfections of the fluorescent protein-expressing vectors did not show specific localization of either EGFP or HcRed (data not shown). As expected, the Tbx5 transcription factor accumulates in the nucleus of the cell (Fig. 4A inset). However, in double transfected cells Tbx5 and LMP4 demonstrate clear co-localization to cytoplasmic sites (Figs. 4A – C). Tbx4 behaved in a similar fashion as Tbx5 and in combination with LMP-4 it was localized to cytoplasmic sites (Figs. 4D – F). These cellular observations

A. Krause et al. / Developmental Biology 273 (2004) 106–120

111

Fig. 3. Identification of the domains of chicken LMP-4 that associate with Tbx5 and Tbx4. E. coli lysates containing T7/His6-taggedTbx and GSTLMP-4 peptides were combined to display an interaction in GST ‘pulldown’ assays. In (B) and (C), lanes 1 and 2 control for Tbx antibody specificity and nonspecific protein binding, respectively (LMP-4 controls not shown). Lanes 3 – 11 display co-precipitations of Tbx5 or Tbx4 with the various LMP-4 peptides, as well as the GST fusion protein and the distantly related LIM protein Zyxin. (A) LMP-4 constructs generated as GST fusions to detect the domains that interact with Tbx4 and Tbx5. (B) Tbx5 coprecipitates with LIM2/3, LIM1/2/3, and LIM3 and is detected with antiTbx5 antibodies. (B) Tbx4 associates with LIM2/3, LIM1/2, LIM1/2/3, and LIM2 and is detected with anti-Tbx4 antibodies. Blots in (B) and (C) were reprobed with anti-GST antibodies to verify the presence of the various input LMP-4 peptides (data not shown).

Fig. 2. Chicken LMP-4 co-precipitates specifically with Tbx5 and Tbx4 proteins. E.coli lysates containing the T7/His6-tagged Tbx and GST-LIM2/3 proteins were combined to display an interaction in GST ‘pull-down’ assays. Lanes 1 and 2 are Tbx and GST-LIM2/3 positive controls, respectively, which demonstrate antibody specificity. Lanes 3 and 7 contain Tbx or GST-LIM2/3 incubated alone with the beads to control for nonspecific protein binding. Lanes 4 – 6 are GST-LIM2/3/Tbx incubated in 1:1, 1:2, and 1:4 ratios, respectively. (A) Tbx5 binds to GST-LIM2/3. Coprecipitated Tbx5 is detected on the Western blot with anti-Tbx5 or antiHis6 antibodies. Anti-GST antibodies demonstrate constant amounts of input GST-LIM2/3 protein. (B) Tbx4 associates with GST-LIM2/3. AntiTbx4 antibodies detect co-precipitated Tbx4. Constant amounts of input GST-LIM2/3 protein were confirmed with anti-GST antibodies. (C) Tbx3 and GST-LIM2/3 do not interact. Because anti-Tbx3 antibodies were unavailable, the blot was processed with anti-His6 antibodies to detect coprecipitated Tbx3. Anti-GST antibodies verify the presence of GST-LIM2/3 protein in the reaction. All blots in (A) – (C) were stripped between the various antibody detection reactions.

correspond with our biochemical protein interactions. We next wished to determine if LMP-4 was associated with the actin cytoskeleton since other PDZ-LIM family members have been found to interact with F-actin-associated proteins (Guy et al., 1999; Nakagawa et al., 2000; Vallenius et al., 2000). In HcRed-LMP-4 transfected cells, we detected LMP-4 fusion proteins in the cytoplasm, associated with phalloidin-labeled actin stress fibers (Figs. 4G –I). Thus, presumably through its interaction with LMP-4, Tbx5 and Tbx4 were co-localized to actin filaments. To test this directly, cells were co-transfected with epitope-tagged Tbx5-FLAG or Tbx4-FLAG and LMP-4-Myc, and processed for indirect fluorescent confocal analysis. In addition, the cells were subjected to detergent extraction to enhance visualization of actin stress fibers (Vallenius and Makela, 2002). Tbx5 was observed in the nucleus and also associated with actin filaments (Figs. 4J– L). Likewise, Tbx4 was

112

A. Krause et al. / Developmental Biology 273 (2004) 106–120

Fig. 4. Chicken LMP-4 localizes Tbx5 to actin filaments. COS-7 cells co-transfected with HcRed-LMP-4 and EGFP-Tbx5 (A – C) or EGFP-Tbx4 (D – E). EGFP-Tbx5 revealed nuclear and cytoplasmic distribution (A), whereas as HcRed-LMP-4 localized only to cytoplasmic sites (B). The merged image displays co-localization of Tbx5 and LMP-4 proteins outside the nucleus (C). Cells transfected with EGFP-Tbx5 alone show exclusive nuclear distribution of Tbx5 (arrowhead and inset in panel A). EGFP-Tbx4 revealed similar localization in the nucleus and cytoplasm (D) while HcRed-LMP-4 again was localized at cytoplasmic sites (E); and both proteins co-localized to cytoplasmic sites (F). COS-7 cells transfected with HcRed-LMP-4 (G – I). Cells stained for actin as visualized with Alexa 488 phalloidin (G) and HcRed-LMP-4 (H). Merged image shows that localization of LMP-4 overlaps with actin cytoskeleton (I). COS-7 cells co-transfected with Tbx5-FLAG and LMP-4-Myc (J – L). Cells were stained for actin (J) and for Tbx5-FLAG using monoclonal anti-FLAG (K). The merged image demonstrates Tbx5 co-localization with actin (L). COS-7 cells co-transfected with Tbx4-FLAG and LMP-4-Myc (M – O). Cells were stained for actin (M) and for Tbx4-FLAG (N). As with Tbx5, the merged image demonstrates Tbx4 co-localization with actin (O). Scale bar = 20 Am.

found to be associated with actin stress fibers as well as being localized to the nucleus (Figs. 4M –O). The Tbx/actin co-localization was only observed in the presence of LMP4, while Tbx5 and Tbx4 alone localize to the nucleus (Fig. 4 insets A and D; arrowhead A). While Tbx5 and Tbx4 can relocate from nuclear to cytoplasmic sites, we have not

detected signs of LMP-4 in the nucleus. Cells transiently transfected with a vector control did not reveal significant fluorescence (data not shown). We conclude that the precise co-localization of LMP-4 and Tbx5/4 and the Tbx protein distribution change from predominantly nuclear to cytoplasmic actin filaments is the result of a physical interaction.

A. Krause et al. / Developmental Biology 273 (2004) 106–120

113

LMP-4 mRNA in the developing limbs co-expresses with Tbx5 and Tbx4 Tbx proteins and LMP-4 must be expressed at the same time in the same tissues and cells to interact during development. We therefore analyzed the temporal and spatial Tbx4, Tbx5, and LMP-4 expression patterns by whole mount in situ hybridization (Wilkinson and Nieto, 1993) in chicken embryos between stages 15 and 32 (Hamburger and Hamilton, 1951). In examinations of several limb bud stages, it was evident that LMP-4 coexpresses with Tbx4 and Tbx5 (Fig. 5). Like Tbx4 and Tbx5, which at early developmental stages mark the cells of the lateral plate mesoderm that contribute to the limb buds (Gibson-Brown et al., 1998; Isaac et al., 1998; Logan et al., 1998; Ohuchi et al., 1998), LMP-4 is expressed in the presumptive fore- and hindlimb fields as visualized by careful microscopic inspection (Figs. 5A, F, K), and RTPCR (Fig. 6). LMP-4 continues to be expressed in the limb mesenchyme throughout forelimb and hindlimb bud outgrowth (inset in Figs. 5C, H, M, and data not shown). The overlap of expression persists into later development, when LMP-4 expression mirrors the Tbx4 and Tbx5 domains in every detail. We also observed a reduced expression of all three genes at late stages in the prechondrogenic condensations without a change in the interdigital mesodermal ‘‘web’’ regions (Figs. 5E, J, O). Sense probe hybridization

Fig. 6. RT-PCR analysis of Tbx4, Tbx5, and LMP-4 expression in earlystage chicken embryos. RT-PCR analysis of Tbx4, Tbx5, LMP-4, and control b-actin expression in dissected presumptive fore- and hindlimbs at developmental stages 9, 11, 13, 15, and 17. The number of PCR cycles was optimized for each gene so that no products were saturated. An average of at least three independent experiments is presented. In the agarose gel system the DNA bands have a slightly different appearance due to different PCR product sizes. Note that Tbx4 and Tbx5 expression is specifically detected in the hindlimb or forelimb fields, respectively, while LMP-4 is expressed in both limb fields at all stages examined.

controls did not reveal staining above background (data not shown). Whole mount in situ hybridization experiments have shown that the onset of Tbx4 and Tbx5 expression occurs between stages 14 and 15 (Gibson-Brown et al., 1998; Isaac et al., 1998; Logan et al., 1998; Ohuchi et al., 1998), and recent work with quantitative RT-PCR has more accurately

Fig. 5. Expression patterns of Tbx4, Tbx5, and LMP-4 in developing chicken wing and leg buds. mRNA expression visualized by whole mount in situ hybridization in stage 15, 18, 25, 29, and 32 chick embryos. Embryos are lying on left side with head at top. Expression domains of Tbx5 (A – E), LMP-4 (F – J), or Tbx4 (K – O) are indicated by the purple alkaline phosphatase reaction product. White arrows indicate locations taken for insets with higher magnification in A, F, and K. Insets in C, H, and M show higher magnifications of transverse sections of limb buds. Black arrowheads indicate the unstained epidermis. Note LMP-4 expression in the heart and eye as well (G). In the heads of some samples, non-specific background staining was observed (A, B, H, K).

114

A. Krause et al. / Developmental Biology 273 (2004) 106–120

estimated the induction of Tbx5/Tbx4 to occur in the wing/leg fields during stage 13 (Saito et al., 2002). To address whether LMP-4 co-expresses with Tbx4 and Tbx5 at these early developmental stages, we employed semi-quantitative RTPCR with dissected wing/leg regions of chick embryos. Since Tbx and LMP-4 genes belong to gene families, we confirmed in BLAST searches the specificity of the employed oligonucleotides for their respective target gene and, in addition to the band intensities, verified the expected amplification product sizes after gel electrophoresis (Fig. 6). Signals for no RT and no template controls were at or below background (data not shown). In accordance with the earlier report from Saito et al. (2002), we detected the onset of expression of Tbx5 in presumptive forelimb and Tbx4 in presumptive hindlimb tissue at stage 13. In addition, we found LMP-4 expression in both presumptive limb tissues at stage 13, verifying that LMP-4 overlaps with Tbx4 and Tbx5 in all stages of detectable expression. RT-PCR also confirmed our in situ hybridization results, demonstrating expression of LMP-4 in both limbs at stages 15 and 17 (Figs. 5A, F, J). Tbx5 and Tbx4 expression domains in the developing heart co-localize with LMP-4 expression In addition to the limbs, LMP-4 is also expressed in the heart, which has previously been reported to express Tbx5 (Bruneau et al., 1999). To explore the regulation of LMP-4 in the developing heart, we performed in situ hybridizations. In early chicken embryos, the developing heart is located outside the body wall, which allows for the straightforward detection of gene expression. However, at later developmental stages, the heart becomes enclosed within the chicken body, and reproducible detection of mRNA expression becomes difficult. We therefore dissected hearts out of later-stage embryos and obtained in situ results identical to those reported (Bruneau et al., 1999), suggesting that performing in situ hybridizations on free hearts is a robust method for generating heart gene expression data. In the developing chicken heart, Tbx5 is initially expressed in the entire bilateral cardiac primordium (Bruneau et al., 1999; Srivastava and Olson, 2000), and we detected expression of Tbx5 and LMP-4 in the early heart (Figs. 5B, G). However, by approximately stage 20, Tbx5 expression is restricted to the left ventricle and both atria (Bruneau et al., 1999; and Figs. 7A – C). This specific expression pattern remains throughout further heart development (Figs. 7G – Fig. 7. Tbx4, Tbx5, and LMP-4 mRNA expression in developing chicken hearts. mRNA expression visualized by whole mount in situ hybridization in stage 22, 25, and 29 hearts shown from several angles. Hearts are positioned with anterior to the top. Right ventricle (RV), left ventricle (LV), right atrium (RA), left atrium (LA), and outflow tract (OT) are indicated. Cells with purple alkaline phosphatase reaction product are positive for Tbx5 (A – C, G – I, M – O), LMP-4 (D – F, J – L, P – R), or Tbx4 (S – U). The strong atrial and left ventricle expression of Tbx5, as well as its exclusion from the right ventricle and outflow tract are consistent with previous reports (Bruneau et al., 1999).

I), although by stage 29, Tbx5 expression has been reduced to a lower level (Figs. 7M –O). Since we had identified LMP-4 based on its interaction with Tbx5 and given the strict coexpression of LMP-4 and Tbx5 mRNA in the forelimbs, we expected that the domains of LMP-4 expression would

A. Krause et al. / Developmental Biology 273 (2004) 106–120

115

Fig. 8. Tbx4 expression in cardiac tissue during chicken development. mRNA expression visualized by whole mount in situ hybridization. Chick embryos or dissected hearts are positioned so that anterior is at the top and posterior at the bottom. Right ventricle (RV), left ventricle (LV), right atrium (RA), left atrium (LA), and outflow tract (OT) are indicated. Tbx4-positive cells at developmental stages 21 (A), 25 (B), and 29 (C) are indicated by the purple reaction product. White arrow in (A) indicates the band of Tbx4 expressing cells just anterior of the heart. Black arrow in (B) indicates Tbx4 expressing cells along the lateral side of the embryo towards the dorsal side. Inset in (B) shows a heart dissected from a whole mount in situ to emphasize light expression in OT and RV, and to provide a transition for the same view in (C).

overlap precisely with those of Tbx5 in the heart as well. LMP-4, however, is surprisingly expressed in every cardiac chamber and the outflow tract, despite the absence of Tbx5 expression in some areas of the heart (Fig. 7). The incomplete overlap of LMP-4 and Tbx5 expression domains suggested that there was another binding partner for LMP-4 in the heart. Based on our binding results and earlier expression studies in the mouse heart (Chapman et al., 1996), we suspected Tbx4 as a candidate for co-expression with LMP-4. Surprisingly, we were able to detect a new expression domain for Tbx4 in the outflow tract of the developing heart at stage 29 by in situ hybridization. The strong Tbx4 expression in the outflow tract correlates with LMP-4 expression, and is outside of the domains of Tbx5 expression (Figs. 7S, T, U). Careful examination also revealed that Tbx4 is expressed in the right ventricle, although at lower levels. Having discovered unexpected Tbx4 expression in stage 29 cardiac tissue, we re-examined its expression profile at earlier stages. We did not detect noticeable Tbx4 levels in the developing heart at stages earlier than 25. However, at stage 18 (Fig. 5L, and data not shown) and stage 21 (Fig. 8A), Tbx4 is expressed in a region immediately anterior and lateral to the heart. Tbx4 expression remains in this anterior/ lateral domain at stage 25, and after careful inspection we also detected expression in some cells of the outflow tract and the right ventricle (Fig. 8B). As development proceeds, the domain of Tbx4-expressing cells expands gradually more posteriorly, and expressing cells are observed down into the proximal outflow tract and the right ventricle (Fig. 8C, and data not shown). No Tbx5 expression was detected at any stage in these anterior cardiac cells.

Discussion Chicken LMP-4 is a new member of the PDZ-LIM family of proteins and interacts specifically with Tbx5 and Tbx4 Chicken LMP-4 belongs to a recently characterized family of cytoplasmic PDZ-LIM proteins. Like the family

members Enigma/LMP-1 (Boden et al., 1998; Wu and Gill, 1994), ENH (Kuroda et al., 1996), ZASP/Cypher1 (Faulkner et al., 1999; Zhou et al., 1999), and Oracle (Passier et al., 2000), chicken LMP-4 contains an N-terminal PDZ domain and three C-terminal LIM domains. The PDZ domain is an approximately 85-amino-acid h-barrel structure (Doyle et al., 1996; Morais Cabral et al., 1996), whereas the LIM domain is an approximately 50-amino-acid-long, cysteinerich module that forms a double zinc-finger-like structure (Brown et al., 2001; Perez-Alvarado et al., 1994). PDZ and LIM domains are proposed to function in the assembly and disassembly of protein complexes (Bach, 2000; Fanning and Anderson, 1999). In Enigma, ALP, and the ZASP/Cypher1 proteins, the PDZ domains facilitate interaction with the cytoskeleton. Enigma binds to the actin-binding protein skeletal h-tropomyosin (Guy et al., 1999), whereas ALP, a muscle-specific PDZ-LIM family member, associates with an abundant skeletal muscle actinbinding protein, a-actinin-2 (Xia et al., 1997). LIM domains are known to mediate protein interactions, and the close family member Enigma binds to the insulin receptor, receptor tyrosine kinases, and protein kinase C (Durick et al., 1998; Kuroda et al., 1996; Wu and Gill, 1994). Proteins containing LIM or PDZ domains play important roles in fundamental biological processes including cytoskeleton organization, cell lineage specification, and organ development (Bach, 2000; Dawid et al., 1998; Fanning and Anderson, 1999). PDZ-LIM proteins are scaffolding proteins, and as such the two modular protein interaction motifs, PDZ and LIM, may enable chicken LMP-4 to form protein complexes and potentially link Tbx transcription factors with other cellular proteins. Tbx5 has been shown to interact through its N-terminal T-domain with the cardiac homeodomain protein Nkx2-5 (Bruneau et al., 2001; Hiroi et al., 2001) and the zinc-finger transcription factor Gata4 (Garg et al., 2003). In both cases, the cooperative binding of Tbx5 and Nkx2-5 or Gata4 resulted in synergistic transcriptional activation of heartspecific gene promoters, indicating considerable complexity in Tbx5 regulation of target genes. Here, we show the

116

A. Krause et al. / Developmental Biology 273 (2004) 106–120

interaction of LMP-4 with the C-terminal transactivation domains of Tbx5 and Tbx4, which suggests a very different function, possibly in the regulation of transcription factor activity. The distinct Tbx5- and Tbx4-specific residues in this domain suggest differential binding, and this idea is supported by our conclusive mapping of the interaction of Tbx5 and Tbx4 to distinct subdomains of the LMP-4 protein. However, while LIM domain 2 and LIM domain 3 are necessary and sufficient for Tbx5 and Tbx4 binding, respectively, the additional LIM domains may act cooperatively to ensure stable interactions.

so-called synexpression groups. Member genes are not only expressed in the same temporal and spatial pattern, but also function in a common pathway (Fu¨rthauer et al., 2002; Tsang et al., 2002). Based on this argument, and given the closely related expression patterns of LMP-4/Tbx5 and LMP-4/Tbx4, we speculate that chicken LMP-4 and its binding partners Tbx5 or Tbx4 belong to and operate in the same functional pathway.

Chicken LMP-4 co-localizes with Tbx5 or Tbx4 in cells and is co-expressed in the limbs and heart

The initial events of heart development take place in paired heart-forming fields in the anterior lateral region of the embryo. Cells from this paired heart field contribute to the early heart tube, and while the heart is growing into a four-chambered structure it is thought that this heart-forming field controls the development of such posterior elements as the atria and left ventricle. Transcription factors including Tbx5, Nkx2-5, and Gata4 are expressed in the paired heartforming fields and throughout the patterning of both atria and the left ventricle. In addition, mouse embryos lacking either factor reveal significant cardiac malformations, which suggests essential roles for the transcription factors in these processes (Bruneau et al., 2001; Kelly and Buckingham, 2002; Lyons et al., 1995; Srivastava and Olson, 2000). Recently, a secondary heart-forming field originating around stage 12 chick embryos in the anterior cardiogenic mesoderm has been described, and cells from this region are recruited into the developing outflow tract and right ventricle (Kelly et al., 2001; Mjaatvedt et al., 2001; Waldo et al., 2001). In support of this finding, we observed previously unrecognized Tbx4-specific expression in a small band of cells just anterior and lateral to the heart, beginning from at least stage 18, which appears to be in the region of the aortic sac and the anterior heart field (AHF) at early developmental stages. Like the AHF, the Tbx4 expression domain extends dynamically, and at later stages, Tbx4-expressing cells enter the outflow tract and right ventricle. Future experimentation with cell-type specific marker genes will clarify whether Tbx4 expression overlaps with the AHF or may represent other cell types instead. Tbx4 may function in progenitor cells of the AHF in the same way that Tbx5 functions in progenitor cells of the posterior heart-forming region. Tbx5 has the ability to partner with other cardiac transcription factors and cooperatively regulates cardiac-specific genes (Bruneau et al., 2001; Garg et al., 2003; Hiroi et al., 2001). These Nkx2-5 and Gata4 proteins are also expressed in cells originating from the AHF (Lyons et al., 1995; Tanaka et al., 1999; Waldo et al., 2001)), and given the virtual identity of the Tbx4 and Tbx5 T-domain binding sites, they may interact with Tbx4 to regulate downstream genes that are important in the development of such anterior structures as the right ventricle and the outflow tract. This hypothesis can be now experimentally tested.

To provide a conceptual framework for their interaction, we have expressed fluorescent fusion proteins of Tbx5/4 and LMP-4 in cells. Tbx5 and Tbx4 produced comparable results suggesting conserved functions in a complex with LMP-4. In agreement with other studies, Tbx5 transfected into COS-7 cells reveals nuclear localization (Takeuchi et al., 1999). In contrast, our studies also demonstrate Tbx5 colocalization with the actin cytoskeleton when LMP-4 is present. This indicates that LMP-4/Tbx protein interaction interferes with the transcription factor’s default nuclear localization. Thus, LMP-4 may regulate the transcription factor availability in a cell. This hypothesis is strengthened by the recent finding that human TBX5 nuclear localization is not uniform during organogenesis (Collavoli et al., 2003). LMP-4 is expressed in forelimbs and hindlimbs, the heart, and the eye—all organs that have been previously shown to express Tbx5 and/or Tbx4 (Bruneau et al., 1999; Gibson-Brown et al., 1998; Logan et al., 1998). From early pre-limb stages throughout later stages, when limbs clearly differentiate into digits and toes, we observed strict temporal and spatial co-expression of LMP-4 with its binding partners Tbx5 or Tbx4. All three genes revealed reduced expression in limb prechondrogenic condensations, while high levels of gene expression remained in the interdigital mesenchymal ‘‘web’’ regions. The interdigital mesenchyme controls the identity of individual digits by presumably modulating the signaling of bone morphogenetic proteins (BMPs) (Dahn and Fallon, 2000), and it may be that LMP-4/Tbx interactions also have a role in digit specification. This is especially likely since the PDZ-LIM family member rat LMP-1 functions downstream of BMP-6 signaling (Boden et al., 1998). Similarly, during heart development, LMP-4 mirrors the expression patterns of Tbx5 and Tbx4, and it seems reasonable to speculate that balanced levels of the binding partners are important for their function. Indeed, misexpression of Tbx5 in developing chicken limbs or chicken and mouse hearts results in limb truncations and loss of ventricular trabeculations, respectively (Liberatore et al., 2000; Rodriguez-Esteban et al., 1999). Recent discoveries in cell signaling during development have elucidated genetic networks where genes function in

A refined view for Tbx gene activity during heart development

A. Krause et al. / Developmental Biology 273 (2004) 106–120

In addition, chicken LMP-4 is expressed in all regions of the heart where we detect Tbx4 or Tbx5. LMP-4 protein may be required for these cardiomyocytes to be competent to correctly respond to Tbx4 and Tbx5 activity during heart morphogenesis. If LMP-4 acts to regulate the activity of Tbx4 and Tbx5, this may explain the observation that LMP4 is expressed in the outflow tract and right ventricle before Tbx4 in the heart proper, where LMP-4 may help prepare the anterior heart cells to become competent for Tbx4 function. LMP-4 could act similarly with Tbx5 early in posterior heart development. The idea that LMP-4 may be essential for Tbx function is further supported by our data in the limbs, where LMP-4 also precedes Tbx4 and Tbx5 expression. Development and disease in the limbs and heart Recent experimental work suggests a relationship between the limbs and the heart in development. Tbx2, 3, and 5 have been implicated in the development of the vertebrate heart, and all of these are also central players in the specification and patterning of the limbs (Gibson-Brown et al., 1998; Isaac et al., 1998; Logan et al., 1998; Ryan and Chin, 2003; Yamada et al., 2000). Fgfs also play important roles in limb development by operating in the same regulatory pathway with Tbx4 and Tbx5 (Agarwal et al., 2003; Martin, 1998; Rallis et al., 2003; Takeuchi et al., 2003), and several have also been implicated in heart development (Kelly et al., 2001). Interestingly, work in pre-limb stages of the mouse and zebrafish indicates that the heart and the forelimbs or pectoral fins may arise from a common Tbx5expressing population of progenitor cells at the midline (Ahn et al., 2002; Bruneau et al., 1999). Our results further expand the limb/heart relationship by showing chick LMP-4 expression in both the heart and limbs. In addition, the newly identified expression of Tbx4 in the heart now places this transcription factor within the group of heart and limb developmental regulators.

117

Clinical manifestations also support the existence of cells and molecules common to heart and limb development. The expression domains of Tbx5 in the forelimbs and heart, as well as gene disruption experiments in the mouse explain how the major defects of Holt-Oram syndrome (HOS) occur in the upper limbs and left ventricle and both atria of the heart (Bruneau et al., 2001). However, the right ventricle and outflow tract remain unaffected in HOS. Numerous reports on congenital heart disease involving outflow tract defects (Ferencz, 1990; Kelly and Buckingham, 2002) suggest that other factors contribute to the development of the anterior heart structures. Our new Tbx4 expression data in the heart point to Tbx4 as one such factor, particularly since its temporal and spatial dynamic expression mirrors the recruitment of AHF cells into the outflow tract and right ventricle. This new possibility for Tbx4 as a developmental regulator of the heart and the limbs, as well as the expression of chicken LMP-4 in the various chambers of the heart and in both fore- and hindlimbs, may explain many heart/ limb syndromes beyond those seen in HOS (Wilson, 1998). Tbx4 knockout mice have been recently described but surprisingly do not reveal heart phenotypes (Naiche and Papaioannou, 2003). These Tbx4-deficient mice have problems with the allantoic connection to the placenta and die early in embryogenesis, and it is therefore possible that no late heart phenotypes could be observed. A hypothesis for LMP-4 function Current data suggest that members of the PDZ-LIM family of proteins play important roles as adaptors that direct LIM domain-binding proteins to cytoplasmic sites to mediate intracellular signaling. Rat LMP-1, for example, has been implicated as a mediator in the BMP signaling pathway in osteoblast differentiation (Boden et al., 1998). Based on the behavior of its family members and the data presented here, chicken LMP-4 may serve as an adaptor by

Fig. 9. Potential function of chick LMP-4. Like its family members, chick LMP-4 is thought to function as a molecular scaffolding protein, enabling the formation of protein complexes and potentially linking different signaling pathways (Bach, 2000; Fanning and Anderson, 1999). LMP-4 may associate with the actin cytoskeleton and/or membrane-associated proteins, thereby potentially tethering the Tbx transcription factors outside the nucleus. The dynamic formation and dissociation of such protein complexes may regulate Tbx5 or Tbx4 transcription factor subcellular localization and activity in their respective cell environments.

118

A. Krause et al. / Developmental Biology 273 (2004) 106–120

localizing Tbx5 or Tbx4 proteins to actin filaments and/or submembrane sites (Figs. 4 and 9). In response to cell surface receptor or other stimulation, the Tbx proteins might then be released to their unbound ‘active’ form and translocate to the nucleus. The localization of the transcription factors away from the nucleus could have profound consequences for the regulation of transcriptional activity. This hypothesis can be tested, and future gain- and loss-offunction experiments in combination with transcription reporter assays will clarify this point. While Tbx4 and Tbx5 activity would be affected in the heart and limbs, the heart-specific transcription factors Nkx2-5 and Gata4, by virtue of their association with the N-terminal domain of Tbx5 and possibly Tbx4, may be regulated by LMP-4 as well. The identification of the cardiac-specific binding partners—and now LMP-4—indicates complex regulatory networks of Tbx function and provides the first inroads into their elucidation. These emerging data on the molecular, cell, and disease levels establish a basis for parallel events in limb and heart development, and suggest that common regulatory pathways are crucial for proper differentiation and growth of these embryonic structures.

Acknowledgments We are grateful to Drs. C. Tabin for providing the chicken limb bud cDNA library, E. Golemis for providing reagents for and expert advice on the yeast two-hybrid system, and M.C. Beckerle for Zyxin cDNA. We thank L. Sleiter, J. Hsu, E. Wroblewski, and D. Isphording for expert technical assistance. We would also like to thank P. Fey, as well as Drs. B. Dettman, F. Szele, and J. Topczewski for critical reading of the manuscript, and members of the Developmental Systems Biology Core for stimulating discussions. This work is supported by a Schweppe Career Award and NIH grant ES012725-01 (to H.-G.S.).

References Agarwal, P., Wylie, J.N., Galceran, J., Arkhitko, O., Li, C., Deng, C., Grosschedl, R., Bruneau, B.G., 2003. Tbx5 is essential for forelimb bud initiation following patterning of the limb field in the mouse embryo. Development 130, 623 – 633. Ahn, D., Kourakis, M.J., Rohde, L.A., Silver, L.M., Ho, R.K., 2002. T-box gene tbx5 is essential for formation of the pectoral limb bud. Nature 417, 754 – 758. Bach, I., 2000. The LIM domain: regulation by association. Mech. Dev. 91, 5 – 17. Basson, C.T., Bachinsky, D.R., Lin, R.C., Levi, T., Elkins, J.A., Soults, J., Grayzel, D., Kroumpouzou, E., Trail, T.A., Leblanc-Straceski, J., Renault, B., Kucherlapati, R., Seidman, J.G., Seidman, C.E., 1997. Mutations in human TBX5 cause limb and cardiac malformations in Holt-Oram syndrome. Nat. Genet. 15, 30 – 35. Boden, S.D., Liu, Y., Hair, G.A., Helms, J.A., Hu, D., Racine, M., Nanes, M.S., Titus, L., 1998. LMP-1, a LIM-domain protein, mediates BMP-6 effects on bone formation. Endocrinology 139, 5125 – 5134. Brown, S., Coghill, I.D., McGrath, M.J., Robinson, P.A., 2001. Role of

LIM domains in mediating signaling protein interactions. IUBMB Life 51, 359 – 364. Bruneau, B.G., Logan, M., Davis, N., Levi, T., Tabin, C.J., Seidman, J.G., Seidman, C.E., 1999. Chamber-specific cardiac expression of Tbx5 and heart defects in Holt-Oram syndrome. Dev. Biol. 211, 100 – 108. Bruneau, B.G., Nemer, G., Schmitt, J.P., Charron, F., Robitaille, L., Caron, S., Conner, D.A., Gessler, M., Nemer, M., Seidman, C.E., Seidman, J.G., 2001. A murine model of Holt-Oram syndrome defines roles of the T-box transcription factor Tbx5 in cardiogenesis and disease. Cell 106, 709 – 721. Chapman, D.L., Garvey, N., Hancock, S., Alexiou, M., Agulnik, S.I., Gibson-Brown, J.J., Cebra-Thomas, J., Bollag, R.J., Silver, L.M., Papaioannou, V.E., 1996. Expression of the T-box family genes, Tbx1-Tbx5, during early mouse development. Dev. Dyn. 206, 379 – 390. Collavoli, A., Hatcher, C.J., He, J., Okin, D., Deo, R., Basson, C.T., 2003. TBX5 nuclear localization is mediated by dual cooperative intramolecular signals. J. Mol. Cell. Cardiol. 35, 1191 – 1195. Dahn, R.D., Fallon, J.F., 2000. Interdigital regulation of digit identity and homeotic transformation by modulated BMP signaling. Science 289, 438 – 441. Dawid, I.B., Toyama, R., Taira, M., 1995. LIM domain proteins. C.R. Acad. Sci. III 318, 295 – 306. Dawid, I.B., Breen, J.J., Toyama, R., 1998. LIM domains: multiple roles as adapters and functional modifiers in protein interactions. Trends Genet. 14, 156 – 162. Doyle, D.A., Lee, A., Lewis, J., Kim, E., Sheng, M., MacKinnon, R., 1996. Crystal structures of a complexed and peptide-free membrane proteinbinding domain: molecular basis of peptide recognition by PDZ. Cell 85, 1067 – 1076. Durick, K., Gill, G.N., Taylor, S.S., 1998. Shc and Enigma are both required for mitogenic signaling by Ret/ptc2. Mol. Cell. Biol. 18, 2298 – 2308. Fanning, A.S., Anderson, J.M., 1999. PDZ domains: fundamental building blocks in the organization of protein complexes at the plasma membrane. J. Clin. Invest. 103, 767 – 772. Faulkner, G., Pallavicini, A., Formentin, E., Comelli, A., Ievolella, C., Trevisan, S., Bortoletto, G., Scannapieco, P., Salamon, M., Mouly, V., Valle, G., Lanfranchi, G., 1999. ZASP: a new Z-band alternatively spliced PDZ-motif protein. J. Cell Biol. 146, 465 – 475. Ferencz, C., 1990. On the birth prevalence of congenital heart disease. J. Am. Coll. Cardiol. 16, 1701 – 1702. Fu¨rthauer, M., Lin, W., Ang, S.L., Thisse, B., Thisse, C., 2002. Sef is a feedback-induced antagonist of Ras/MAPK-mediated FGF signalling. Nat. Cell Biol. 4, 170 – 174. Galceran, J., Farinas, I., Depew, M.J., Clevers, H., Grosschedl, R., 1999. Wnt3a / like phenotype and limb deficiency in Lef1( / )Tcf1( / ) mice. Genes Dev. 13, 709 – 717. Garg, V., Kathiriya, I.S., Barnes, R., Schluterman, M.K., King, I.N., Butler, C.A., Rothrock, C.R., Eapen, R.S., Hirayama-Yamada, K., Joo, K., Matsuoka, R., Cohen, J.C., Srivastava, D., 2003. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. Nature 424, 443 – 447. Gibson-Brown, J.J., Agulnik, S.I., Champman, D.L., Alexiou, M., Garvey, N., Silver, L.M., Papaioannou, V.E., 1996. Evidence of a role for T-box genes in the evolution of limb morphogenesis and the specification of forelimb/hindlimb identity. Mech. Dev. 56, 93 – 101. Gibson-Brown, J.J., Agulnik, S.I., Silver, L.M., Niswander, L., Papaioannou, V.E., 1998. Involvement of T-box genes Tbx2 – Tbx5 in vertebrate limb specification and development. Dev., Suppl. 125, 2499 – 2509. Golemis, E.A., Serebriiskii, I., Gyuris, J., Brent, R., 1997. Interaction trap/two-hybrid system to identify interacting proteins. In: Ausubel, R., Brent, R., Kingston, R., Moore, D., Seidman, J., Struhl, K. (Eds.), Current Protocols in Molecular Biology, vol. 3. John Wiley and Sons, New York, pp. 20.1.1 – 20.1.35. Guy, P.M., Kenny, D.A., Gill, G.N., 1999. The PDZ domain of the LIM protein enigma binds to beta-tropomyosin. Mol. Biol. Cell 10, 1973 – 1984.

A. Krause et al. / Developmental Biology 273 (2004) 106–120 Hamburger, V., Hamilton, H.L., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49 – 92. Hatcher, C.J., Kim, M.S., Mah, C.S., Goldstein, M.M., Wong, B., Mikawa, T., Basson, C.T., 2001. TBX5 transcription factor regulates cell proliferation during cardiogenesis. Dev. Biol. 230, 177 – 188. Herrmann, B.G., Labeit, S., Poustka, A., King, T.R., Lehrach, H., 1990. Cloning of the T gene required in mesoderm formation in the mouse. Nature 343, 617 – 622. Hiroi, Y., Kudoh, S., Monzen, K., Ikeda, Y., Yazaki, Y., Nagai, R., Komuro, I., 2001. Tbx5 associates with Nkx2-5 and synergistically promotes cardiomyocyte differentiation. Nat. Genet. 28, 276 – 280. Hoffman, M.P., Kidder, B.L., Steinberg, Z.L., Lakhani, S., Ho, S. et al., 2002. Gene expression profiles of mouse submandibular gland development: FGFR1 regulates branching morphogenesis in vitro through BMPand FGF-dependent mechanisms. Development 129, 5767 – 5778. Isaac, A., Rodriguez Esteban, C., Ryan, A., Altabef, M., Tsukui, T., Patel, K., Tickle, C., Izpisua Belmonte, J.C., 1998. Tbx genes and limb identity in chick embryo development. Development 125, 1867 – 1875. Kawakami, Y., Capdevila, J., Buscher, D., Itoh, T., Rodriguez Esteban, C., Izpisua Belmonte, J.C., 2001. WNT signals control FGF-dependent limb initiation and AER induction in the chick embryo. Cell 104, 891 – 900. Kelly, R.G., Buckingham, M.E., 2002. The anterior heart-forming field: voyage to the arterial pole of the heart. Trends Genet. 18, 210 – 216. Kelly, R.G., Brown, N.A., Buckingham, M.E., 2001. The arterial pole of the mouse heart forms from Fgf10-expressing cells in pharyngeal mesoderm. Dev. Cell 1, 435 – 440. Khan, P.A., Linkhart, B., Simon, H.G., 2002. Different regulation of T-box genes Tbx4 and Tbx5 during limb development and limb regeneration. Dev. Biol. 250, 383 – 392. Kispert, A., Herrmann, B.G., 1993. The Brachyury gene encodes a novel DNA binding protein. EMBO J. 12, 4898 – 4899. Kuroda, S., Tokunaga, C., Kiyohara, Y., Higuchi, O., Konishi, H., Mizuno, K., Gill, G.N., Kikkawa, U., 1996. Protein – protein interaction of zinc finger LIM domains with protein kinase C. J. Biol. Chem. 271, 31029 – 31032. Li, Q.Y., Newbury-Ecob, R.A., Terrett, J.A., Wilson, D.I., Curtis, A.R.J., Yi, C.H., Gebuhr, T., Bullen, P.J., Robson, S.C., Strachan, T., Bonnet, D., Lyonnet, S., Young, I.D., Raeburn, J.A., Buckler, A.J., Law, D.J., Brook, J.D., 1997. Holt-Oram syndrome is caused by mutations in TBX5, a member of the Brachyury (T) gene family. Nat. Genet. 15, 21 – 29. Liberatore, C.M., Searcy-Schrick, R.D., Yutzey, K.E., 2000. Ventricular expression of tbx5 inhibits normal heart chamber development. Dev. Biol. 223, 169 – 180. Logan, M., Tabin, C.J., 1999. Role of Pitx1 upstream of Tbx4 in specification of hindlimb identity. Science 283, 1736 – 1739. Logan, M., Simon, H.G., Tabin, C., 1998. Differential regulation of T-box and homeobox transcription factors suggests roles in controlling chick limb-type identity. Dev., Suppl. 125, 2825 – 2835. Lyons, I., Parsons, L.M., Hartley, L., Li, R., Andrews, J.E., Robb, L., Harvey, R.P., 1995. Myogenic and morphogenetic defects in the heart tubes of murine embryos lacking the homeo box gene Nkx2 – 5. Genes Dev. 9, 1654 – 1666. Martin, G.R., 1998. The roles of FGFs in the early development of vertebrate limbs. Genes Dev. 12, 1571 – 1586. Mayanil, C.S., George, D., Freilich, L., Miljan, E.J., Mania-Farnell, B. et al., 2001. Microarray analysis detects novel Pax3 downstream target genes. J. Biol. Chem. 276, 49299 – 49309. Mjaatvedt, C.H., Nakaoka, T., Moreno-Rodriguez, R., Norris, R.A., Kern, M.J., Eisenberg, C.A., Turner, D., Markwald, R.R., 2001. The outflow tract of the heart is recruited from a novel heart-forming field. Dev. Biol. 238, 97 – 109. Morais Cabral, J.H., Petosa, C., Sutcliffe, M.J., Raza, S., Byron, O., Poy, F., Marfatia, S.M., Chishti, A.H., Liddington, R.C., 1996. Crystal structure of a PDZ domain. Nature 382, 649 – 652. Naiche, L.A., Papaioannou, V.E., 2003. Loss of Tbx4 blocks hindlimb

119

development and affects vascularization and fusion of the allantois. Development 130, 2681 – 2693. Nakagawa, N., Hoshijima, M., Oyasu, M., Saito, N., Tanizawa, K., Kuroda, S., 2000. ENH, containing PDZ and LIM domains, heart/skeletal muscle-specific protein, associates with cytoskeletal proteins through the PDZ domain. Biochem. Biophys. Res. Commun. 272, 505 – 512. Nelson, C.E., Morgan, B.A., Burke, A.C., Laufer, E., DiMambro, E., Murtaugh, L.C., Gonzales, E., Tessarollo, L., Parada, L.F., Tabin, C., 1996. Analysis of Hox gene expression in the chick limb bud. Development 122, 1449 – 1466. Ng, J.K., Kawakami, Y., Bu¨scher, D., Raya, A., Itoh, T., Koth, C.M., Rodriguez Esteban, C., Rodriguez-Leon, J., Garrity, D.M., Fishman, M.C., Izpisua Belmonte, J.C., 2002. The limb identity gene Tbx5 promotes limb initiation by interacting with Wnt2b and Fgf10. Development 129, 5161 – 5170. Ohuchi, H., Takeuchi, J., Yoshioka, H., Ishimaru, Y., Ogura, K., Takahashi, H., Ogura, T., Noji, S., 1998. Correlation of wing-leg identity in ectopic FGF-induced chimeric limbs with the differential expression of chick Tbx5 and Tbx4. Development 125, 51 – 60. Passier, R., Richardson, J.A., Olson, E.N., 2000. Oracle, a novel PDZ-LIM domain protein expressed in heart and skeletal muscle. Mech. Dev. 92, 277 – 284. Perez-Alvarado, G.C., Miles, C., Michelsen, J.W., Louis, H.A., Winge, D.R., Beckerle, M.C., Summers, M.F., 1994. Structure of the carboxy-terminal LIM domain from the cysteine rich protein CRP. Nat. Struct. Biol. 1, 388 – 398. Rallis, C., Bruneau, B.G., Buono, J.D., Seidman, C.E., Seidman, J.G., Nissim, S., Tabin, C.J., Logan, M.P.O., 2003. Tbx5 is required for forelimb bud formation and contiued outgrowth. Development 130, 2741 – 2751. Rodriguez-Esteban, C., Tsukui, T., Yonei, S., Magallon, J., Tamura, K., Izpisua-Belmonte, J.C., 1999. The T-box genes Tbx4 and Tbx5 regulate limb outgrowth and identity. Nature 398, 814 – 818. Ryan, K., Chin, A.J., 2003. T-box genes and cardiac development. Birth Defects Res. C, Embryo Today 69, 25 – 37. Saito, D., Yonei-Tamura, S., Kano, K., Ide, H., Tamura, K., 2002. Specification and determination of limb identity: evidence for inhibitory regulation of Tbx gene expression. Development 129, 211 – 220. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Schmeichel, K.L., Beckerle, M.C., 1994. The LIM domain is a modular protein-binding interface. Cell 79, 211 – 219. Sekine, K., Ohuchi, H., Fujiwara, M., Yamasaki, M., Yoshizawa, T., Sato, T., Yagishita, N., Matsui, D., Koga, Y., Itoh, N., Kato, S., 1999. Fgf10 is essential for limb and lung formation. Nat. Genet. 21, 138 – 141. Serebriiskii, I.G., Golemis, E.A., 2000. Uses of lacZ to study gene function: evaluation of beta-galactosidase assays employed in the yeast twohybrid system. Anal. Biochem. 285, 1 – 15. Simon, H.-G., 1999. T-box genes and the formation of vertebrate forelimband hindlimb specific pattern. Cell Tissue Res. 296, 57 – 66. Simon, H.-G., Kittappa, R., Khan, P.A., Tsilfidis, C., Liversage, R.A., Oppenheimer, S., 1997. A novel family of T-box genes in urodele amphibian limb development and regeneration: candidate genes involved in vertebrate forelimb/hindlimb patterning. Development 124, 1355 – 1366. Srivastava, D., Olson, E.N., 2000. A genetic blueprint for cardiac development. Nature 407, 221 – 226. Takeuchi, J.K., Koshiba-Takeuchi, K., Matsumoto, K., Vogel-Hopker, A., Naitoh-Matsuo, M., Ogura, K., Takahashi, N., Yasuda, K., Ogura, T., 1999. Tbx5 and Tbx4 genes determine the wing/leg identity of limb buds. Nature 398, 810 – 814. Takeuchi, J.K., Koshiba-Takeuchi, K., Suzuki, T., Kamimura, M., Ogura, K., Ogura, T., 2003. Tbx5 and Tbx4 trigger limb initiation through activation of the Wnt/Fgf signaling cascade. Development 130, 2729 – 2739. Tanaka, M., Chen, Z., Bartunkova, S., Yamasaki, N., Izumo, S., 1999. The

120

A. Krause et al. / Developmental Biology 273 (2004) 106–120

cardiac homeobox gene Csx/Nkx2.5 lies genetically upstream of multiple genes essential for heart development. Development 126, 1269 – 1280. Tsang, M., Friesel, R., Kudoh, T., Dawid, I.B., 2002. Identification of Sef, a novel modulator of FGF signalling. Nat. Cell Biol. 4, 165 – 169. Vallenius, T., Makela, T.P., 2002. Clik1: a novel kinase targeted to actin stress fibers by the CLP-36 PDZ-LIM protein. J. Cell Sci. 115, 2067 – 2073. Vallenius, T., Luukko, K., Makela, T.P., 2000. CLP-36 PDZ-LIM protein associates with nonmuscle alpha-actinin-1 and alpha-actinin-4. J. Biol. Chem. 275, 11100 – 11115. Waldo, K.L., Kumiski, D.H., Wallis, K.T., Stadt, H.A., Hutson, M.R., Platt, D.H., Kirby, M.L., 2001. Conotruncal myocardium arises from a secondary heart field. Development 128, 3179 – 3188. Wilkinson, D.G., Nieto, M.A., 1993. Detection of messenger RNA by in situ hybridization to tissue sections and whole mounts. Methods Enzymol. 225, 361 – 373. Wilson, G.N., 1998. Correlated heart/limb anomalies in Mendelian syndromes provide evidence for a cardiomelic developmental field. Am. J. Med. Genet. 76, 297 – 305.

Wu, R.Y., Gill, G.N., 1994. LIM domain recognition of a tyrosine-containing tight turn. J. Biol. Chem. 269, 25085 – 25090. Xia, H., Winokur, S.T., Kuo, W.L., Altherr, M.R., Bredt, D.S., 1997. Actinin-associated LIM protein: identification of a domain interaction between PDZ and spectrin-like repeat motifs. J. Cell Biol. 139, 507 – 515. Xu, X., Weinstein, M., Li, C., Naski, M., Cohen, R.I., Ornitz, D.M., Leder, P., Deng, C., 1998. Fibroblast growth factor receptor 2 (FGFR2)-mediated reciprocal regulation loop between FGF8 and FGF10 is essential for limb induction. Development 125, 753 – 765. Yamada, M., Revelli, J.P., Eichele, G., Barron, M., Schwartz, R.J., 2000. Expression of chick Tbx-2, Tbx-3, and Tbx-5 genes in early heart development: evidence for BMP2 induction of Tbx2. Dev. Biol. 228, 95 – 105. Zhou, Q., Ruiz-Lozano, P., Martone, M.E., Chen, J., 1999. Cypher, a striated muscle-restricted PDZ and LIM domain-containing protein, binds to alpha-actinin-2 and protein kinase C. J. Biol. Chem. 274, 19807 – 19813.