BMP2 is required for early heart development during a distinct time period

Mechanisms of Development 91 (2000) 259±270 www.elsevier.com/locate/modo

BMP2 is required for early heart development during a distinct time period Thomas Schlange, Birgit AndreÂe, Hans-Henning Arnold, Thomas Brand* Department of Cell and Molecular Biology, Institute of Biochemistry and Biotechnology, Technical University of Braunschweig, Spielmannstrasse 7, 38106 Braunschweig, Germany Received 23 June 1999; received in revised form 16 November 1999; accepted 17 November 1999

Abstract BMP2, like its Drosophila homologue dpp, is an important signaling molecule for speci®cation of cardiogenic mesoderm in vertebrates. Here, we analyzed the time-course of BMP2-requirement for early heart formation in whole chick embryos and in explants of antero-lateral plate mesoderm. Addition of Noggin to explants isolated at stage 4 and cultured for 24 h resulted in loss of NKX2.5, GATA4, eHAND, Mef2A and vMHC expression. At stages 5±8 the individual genes showed differential sensitivity to Noggin addition. While expression of eHAND, NKX2.5 and Mef2A was clearly reduced by Noggin vMHC was only marginally affected. In contrast, GATA4 expression was enhanced after Noggin treatment. The developmental period during which cardiac mesoderm required the presence of BMP signaling in vivo was assessed by implantation of Noggin expressing cells into stage 4±8 embryos which were then cultured until stage 10±11. Complete loss of NKX2.5 and eHAND expression was observed in embryos implanted at stages 4±6, and expression was still suppressed in stages 7 and 8 implanted embryos. GATA4 expression was also blocked by Noggin at stage 4, however increased at stages 5, 6 and 7. Explants of central mesendoderm, that normally do not form heart tissue were employed to study the time-course of BMP2-induced cardiac gene expression. The induction of cardiac lineage markers in central mesendoderm of stage 5 embryos was distinct for different genes. While GATA4, -5, -6 and MEF2A were induced to maximal levels within 6 h after BMP2 addition, eHAND and dHAND required 12 h to reach maximum levels of expression. NKX2.5 was induced by 6 h and accumulated over 48 h. vMHC and titin were induced at signi®cant levels only after 48 h of BMP2 addition. These results indicate that cardiac marker genes display distinct expression kinetics after BMP2 addition and differential response to Noggin treatment suggesting complex regulation of myocardial gene expression in the early tubular heart. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Heart development; Noggin; BMP2; NKX2.5; GATA genes; Chicken embryo; Gene expression

1. Introduction In vertebrate embryos the heart is one of the ®rst organs to be formed. In early gastrula of the chick embryo (stage 3; (Hamburger and Hamilton, 1951), myocardial precursor cells are located in the mid-primitive streak from which they enter the mesoderm and spread antero-laterally at stage 3±4 (Rosenquist and DeHaan, 1966; Rosenquist, 1970; Garcia-Martinez and Schoenwolf, 1993). These cells form two heart-forming regions (HFR) on each side of Hensen's node at stage 5 (Rawles, 1943; DeHaan, 1965). Around stage 6, the bilateral myocardial precursors migrate anteriorly in association with the underlying endoderm and fuse between stages 7 and 9 into the primitive heart tube along the ventral midline at the level of the anterior intestinal portal (DeHaan, 1965). Cardiac speci®cation takes place during stages 4±5 and cells are committed to the * Corresponding author. Tel.: 149-531-391-5733; fax: 149-531-3918178. E-mail address: [email protected] (T. Brand)

cardiac lineage after midgastrulation, while cardiac differentiation markers are only detectable by stage 7 (Bisaha and Bader, 1991; Han et al., 1992; Antin et al., 1994; Sugi and Lough, 1994; Gannon and Bader, 1995). It is believed that in vertebrate embryos anterior endoderm is the source of inducing signals which recruit mesoderm to the cardiac cell lineage (Sugi and Lough, 1994; Tonegawa et al., 1996; Arai et al., 1997). Experiments with explant cultures suggest that anterior endoderm has instructive potential to reprogram posterior mesoderm to the cardiogenic fate (Schultheiss et al., 1997). Therefore, migration of myocardial progenitor cells into the antero-lateral position is required to bring mesoderm in the correct position in order to receive instructive signals from underlying anterior endoderm (Tam et al., 1997). Anterior lateral endoderm secretes a variety of signaling molecules including various FGFs, activin, insulin-like growth factor II which all appear to promote cell survival and proliferation of cardiogenic cells and differentiating myocytes (Kokan-Moore et al., 1991; Parlow et al., 1991; Sugi and Lough, 1995; Antin et al., 1996; Zhu et al., 1999).

0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(99)00311-1

260

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

We and others have demonstrated that BMP2 is part of the heart-inducing activity that is derived from the pharyngeal endoderm. BMP2 is expressed at stages 4 and 5 of chick embryogenesis in lateral mesendoderm adjacent to the heart forming region and during stages 6±9 in pharyngeal endoderm, underlying cardiogenic mesoderm. Subsequently, during stages 10 and 11, BMP2 is present in the sino-atrial region, but absent from the tubular heart itself (AndreÂe et al., 1998). BMP2 is able to induce the expression of myocardial lineage markers in ectopic locations (Lough et al., 1996; Schultheiss et al., 1997; AndreÂe et al., 1998; Ladd et al., 1998). Moreover, Noggin which binds BMPs with high af®nity prevents myocardial differentiation of lateral mesendoderm cultured in vitro (Schultheiss and Lassar, 1997; Schultheiss et al., 1997; Ladd et al., 1998). The capacity of BMP2 to induce heart-speci®c markers seems to be conserved during evolution, because dpp mutant embryos in Drosophila fail to express tinman and form no heart (Frasch, 1995). Similarly, in zebra®sh the BMP2 mutant swirl lacks cardiogenic mesoderm (Kishimoto et al., 1997). Murine BMP2 mutant embryos display abnormal heart development, but cardiac mesoderm is formed (Zhang and Bradley, 1996). This mild phenotype in mice may be due to the fact that multiple members of the BMP family with overlapping functions are expressed in cardiogenic mesoderm and the essential function in heart formation might only be uncovered in double or triple null embryos (Dudley and Robertson, 1997; Solloway and Robertson, 1999). However, it has been demonstrated recently that myocardial differentiation in murine P19 embryonic teratocarcinoma cells is dependent on a functional BMP signal transduction pathway (Monzen et al., 1999). In recent years several classes of transcription factors have been isolated from the heart. The homeobox gene NKX2.5 is one of the ®rst transcription factors to be expressed in speci®ed myocardial progenitor cells (Schultheiss et al., 1995; Ehrman and Yutzey, 1999). In the chick NKX2.5 expression is ®rst seen at stage 5 in antero-lateral mesoderm. Another NK homeobox gene expressed in the chicken heart is NKX2.8. Between stage 6 and 8 NKX2.8 is expressed in the pharyngeal endoderm underlying the forming heart ®elds (Brand et al., 1997) and between stage 8 and 11 in the forming heart tube. There are three GATA genes, GATA4±6 that are expressed in cardiac mesoderm and endoderm of the gut (Laverriere et al., 1994). The GATA genes in the avian embryo are ®rst transcribed in antero-lateral mesoderm emerging from the primitive streak during gastrulation (Jiang et al., 1998). Subsequently, GATA expression encompasses the cardiac ®eld including both, cardiac progenitor cells and the associated endoderm. At stages 5±8 the expression domains of GATA genes closely overlap with that of BMP2 (Schultheiss et al., 1997; AndreÂe et al., 1998). Transcripts are then localized in the primitive heart tube and in adjacent lateral plate mesoderm. By stage 9 the transcript levels of all three GATA genes, particularly that of GATA4 are low in the anterior heart tube, the presumptive

primitive ventricle, but high in the posterior end of the tube, including both the presumptive atrium and the sinus venosus (Kostetskii et al., 1999). In the mouse all four MEF2 genes are expressed in the heart. MEF2B and MEF2C are expressed in the heart primordia beginning at day 7.5 p.c., while MEF2A and MEF2D are detected at day 8.0 (Edmondson et al., 1994). In the chick only expression of MEF2A has been reported and found to be ®rst expressed in the precardiac mesoderm by stage 8 (Buchberger and Arnold, 1999). The basic helix-loop-helix transcription factors eHAND and dHAND are also expressed in the early heart. By wholemount in situ hybridization dHAND and eHAND are coexpressed in the heart forming region by stage 8 and throughout the entire tubular heart by stage 10 in chicken embryos (Srivastava et al., 1995). Thus, multiple transcription factors are expressed in the heart forming region and in the primitive tubular heart. In the present study we examined the temporal requirement of endodermal BMP2 expression during the early stages of heart development. BMP2 was able to rapidly induce multiple cardiac lineage-restricted genes in explanted ventral mesendoderm of stage 5 embryos. The BMP-binding protein Noggin prevented expression of myocardial marker genes when added to explant cultures of antero-lateral mesendoderm. At stage 4 Noggin inhibited expression of NKX2.5, GATA4, eHAND, Mef2A and vMHC. At stages 5±8 individual genes showed differential sensitivity to Noggin addition. The requirement of BMP signals for cardiac mesoderm development in vivo was assessed by implantation of Noggin expressing cells into stage 4±8 embryos. Complete loss of NKX2.5 and eHAND expression was observed at stage 4±6, while at stages 7 and 8 myocardial gene expression was suppressed only in the vicinity of the Noggin source. Interestingly, GATA4 showed a different response to Noggin implantation. While complete loss of GATA4 expression was observed at stage 4, at stages 5±7 GATA4 expression was enhanced in the vicinity of the Noggin source. Thus, we suggest that myocardial cells shortly after gastrulation are still plastic and require the continuous presence of BMP signals until myocardial determination becomes independent of BMP2 at stage 8. Individual myocardial transcription factors display distinct kinetics of induction after BMP2 addition and also differ in their response to Noggin suggesting complex regulatory pathways governing myocardial gene expression in the early tubular heart.

2. Results 2.1. BMP2 programs mesendoderm to express early and late cardiac lineage marker genes We and others have previously shown that BMP2 is involved in early heart formation (Schultheiss et al., 1997; AndreÂe et al., 1998). Here, we investigated the time course

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

of gene induction for an array of cardiac-speci®c genes in cultivated anterior central mesendoderm in the presence of BMP2. Anterior central mesendoderm from stage 5 which is normally not fated to form cardiogenic mesoderm will do so after BMP2 addition (Schultheiss et al., 1997). Central mesendoderm was explanted and cultured in the presence of BMP2 for various length of time (Fig. 1A). RNA was isolated from explant cultures and subjected to RT-PCR with speci®c primers to detect various cardiac marker transcripts. The transcription factors GATA4, -5, and -6 as well as MEF2A were maximally induced by BMP2 within 6 h and high level expression was maintained for at least 48 h (Fig. 2). Transcripts for eHAND and dHAND reached the highest level 12 h after BMP2 addition and NKX2.5 steadily increased between 6 and 48 h. NKX2.8 was not signi®cantly affected by BMP2 addition during the ®rst 12 h, however strongly induced after 24 and 48 h. Induction of vMHC and

261

titin by BMP2 only started at 12 h with continuing increase for 48 h, re¯ecting that these genes constitute late differentiation markers for myocardial cells. Taken together these results indicate the different temporal response of cardiac marker genes to BMP2. The transcription factors GATA4, 5, -6, MEF2A, NKX2.5, are induced rapidly, while eHAND and dHAND respond to BMP2 more slowly followed by NKX2.8 and the late differentiation markers vMHC and titin that are only induced signi®cantly 24 h after addition of BMP2. One may speculate that the different time courses of induction by BMP2 re¯ect the epistasis of genes involved in setting up the early tubular heart. 2.2. GATA genes can be activated by BMP2 outside the heart-forming area We have previously shown the induction of NKX2.5 and

Fig. 1. Graphic representation of the experimental protocols used in this study. (a) In order to study the time course of cardiac marker gene induction in explant cultures, central mesendoderm, which normally does not form cardiac tissue was explanted and cultured in the presence or absence of BMP2 for indicated length of time. (b) In order to de®ne the period of BMP-dependence of cardiac mesoderm, lateral mesendoderm was isolated from the heart forming region (HFR) of stage 4±8 embryos and cultured for 24 h in serum-free medium in the presence, or absence of Noggin. (c) In order to assess the period of Noggin sensitivity of the HFR in vivo, stage 4±8 embryos were placed into New culture and Noggin-expressing, or control cells were implanted into the HFR. The embryos were incubated until they reached stage 8 and the anterior intestinal portal (AIP) was split using a tungsten needle in order to prevent the fusion of the two HFRs. Thus, the effects of Noggin cell implants vs. control cell implants could be studied in the same embryo. Cultures were terminated when the embryos had reached stage 11.

262

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

locations only GATA genes can be ectopically activated, while NKX2.5 can not (AndreÂe et al., 1998). This different regulation might re¯ect that GATA genes are also expressed in the gut, while NKX2.5 is con®ned to the heart, at least during early stages of development. 2.3. Noggin inhibits myocardial marker gene expression in explant cultures

Fig. 2. Expression of cardiogenic marker genes in explant cultures of central mesendoderm of stage 5 embryos determined by RT-PCR analysis. Explants were cultured in the presence (1) or absence (2) of BMP2 for the indicated length of time before total mRNA was isolated from two explants per time point. Transcripts were detected by RT-PCR using gene-speci®c primers as described in Section 4. After gel electrophoresis PCR products were visualized on Southern blot with speci®c cDNA probes or internal oligonucleotides.

GATA4 by BMP2 in the heart forming region (HFR) of embryos cultured according to New (AndreÂe et al., 1998). NKX2.5 and GATA4 were both induced in the vicinity of the HFR but GATA4 was in addition signi®cantly induced in posterior mesoderm at considerable distance from the HFR. Here we extended this analysis for GATA5 and GATA6 which both are closely related to GATA4. As shown in Fig. 3, BMP2 expressing cells implanted into stage 4 embryos were able to induce ectopic GATA5 and GATA6 expression. While GATA5 was strongly induced in anterior mesendoderm, even adjacent to the notochord (Fig. 3e), induction of GATA6 was clearly weaker and limited to the immediate vicinity of the implanted cells (Fig. 3c,f). The notochord is believed to contain inhibitory signals for heart formation (Schultheiss et al., 1997), which apparently can be overcome by BMP2. This may suggest that inhibition of cardiac formation by the notochord involves BMP binding proteins in this territory (Schultheiss et al., 1997; Capdevila and Johnson, 1998; Streit et al., 1998). As for GATA4, BMP2 also induced GATA5 far posterior to the actual HFR suggesting common regulatory pathways for both genes. In summary, implantation of BMP2 into an anterior domain results in upregulation of the cardiac marker genes NKX2.5, GATA4, (AndreÂe et al., 1998) -5, and -6 (this study) suggesting that all necessary signals are present for the induction of these genes. In contrast, in more posterior

In order to assess the dependence of cardiac marker genes on the presence of endogenous BMP2, we cultured anterolateral mesendoderm explanted from stages 4±8 embryos in the presence or absence of Noggin (Fig. 1B). RNA was isolated from explants 24 h following Noggin addition and subjected to RT-PCR analysis using gene-speci®c probes (Fig. 4). At stage 4 all transcripts analyzed were completely lost after Noggin addition, while from stage 5 onwards distinct sensitivities towards Noggin were observed. eHAND was strongly diminished by Noggin treatment and this effect remained during the entire observation period. NKX2.5, MEF2A, and vMHC were suppressed by Noggin during stages 5±8 but to a weaker extent than eHAND. By stage 8 transcripts for all genes, with the notable exception of GATA4, have almost reached the level observed in control cultures suggesting that Noggin does no longer have an effect on myocardial marker genes by this stage. In contrast, GATA4 expression was enhanced in Noggin treated cultures. This result is in good agreement with in vivo results (see below). Taken together these observations con®rm that BMP2 is essential for myocardial differentiation and expression marker genes. 2.4. Cardiac myogenesis in the embryo is dependent on the continuous presence of BMP2 during stages 4±8 We also compared the effect of Noggin in explant cultures with the in vivo situation. For this purpose Noggin expressing CHO cells were implanted in the HFR of stage 4 embryos in New culture. The cultures were terminated when the embryos reached stages 7±8. As shown in Fig. 5, expression of the three GATA genes and NKX2.5 expression was blocked in the HFR containing the Noggin implant indicating the absolute requirement of BMP signaling for myocardial marker gene expression at stage 4. In order to de®ne the period of dependence of the precardiac mesoderm for BMP signaling, Noggin cells were implanted in the left HFR and control cells at the contra-lateral side of embryos at stages 4±8. In several independent experiments we determined that sidedness of Noggin implantation did not affect the outcome of the experiment (data not shown). In all embryos the anterior intestinal portal (AIP) was split when the embryos had reached stage 8 in order to prevent fusion of the heart ®elds and enable us to compare the effects of Noggin and control cell implantations in the same embryo (Fig. 1c). Subsequently, embryos were cultured overnight until they reached stages 10±11 when the expression of NKX2.5, eHAND and GATA4 was analyzed. NKX2.5 and eHAND expression

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

263

Fig. 3. BMP2 induces ectopic GATA5 and GATA6 expression in the embryo. All embryos are shown in a ventral view with rostral end at the top. (a,b) Wholemount in situ hybridizations for GATA5 expression of embryos subjected to cell implantation and overnight New culture. Red arrow in (a) depicts ectopic GATA5 expression throughout the mesendoderm induced by a cell aggregate which secrets BMP2 (*). (b) Ectopic expression in posterior mesendoderm. (e) Transverse section of the embryo depicted in panel a showing ectopic expression of GATA5 in mesendoderm surrounding the BMP2 cell implant (*), while control cell implants (1) display no ectopic expression. (c,d) Whole-mount in situ hybridizations for GATA6 expression of embryos subjected to cell implantation and overnight New culture. Red arrow points to a BMP2 cell implant (*) surrounded by cells expressing ectopic GATA6, while on the contra-lateral side control cells (1) elicits no expression. (f) Transverse section through the embryo shown in (c).

was signi®cantly blocked by Noggin expressing cells independent of the stage of implantation. While, NKX2.5 and eHAND positive heart tissue was readily formed at the contra-lateral side (Fig. 6a±e,k±o). The differentiation block by Noggin, most prominently in stage 4 and 5 embryos, became clearly weaker when older embryos were implanted (compare Fig. 6a,b with c±e). Implantation

at stage 6 resulted in a complete loss of NKX2.5 and eHAND expression only in a few embryos (Fig. 6c) and at stage 7 and 8 embryos showed reduction but never complete loss of NKX2.5 and eHAND expression (Fig. 6d,e). eHAND expression was not only affected in the precardiac mesoderm but also in the overlying ectoderm (Fig. 6k,k 0 ). eHAND expression in the ectoderm might be under the

Fig. 4. Time course of Noggin inhibition of myocardial marker gene expression in explant cultures. Explants of the heart forming region (HFR) were dissected at the indicated stages and cultured in presence (1) or absence (2) of medium conditioned by Noggin-secreting CHO cells. Total mRNA was isolated from two explants per time point, and subjected to RT-PCR using gene-speci®c primers. After gel electrophoresis PCR products were visualized on Southern blot using speci®c cDNA probes or internal oligonucleotides.

264

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

3. Discussion

Fig. 5. Noggin cell implants prevent myocardial marker gene expression. All embryos are shown in a ventral view with rostral end at the top. Embryos were implanted with Noggin cells at stage 4. (a) Stage 8 embryo subjected to whole-mount in situ hybridization with GATA4 probe. (b) Stage 9 embryo hybridized with GATA5 probe. (c) Stage 7 embryo stained for GATA6. (d) Stage 8 embryo subjected to whole-mount in situ hybridization with NKX2.5 probe. Red arrows point to the loss of marker gene expression as a result of Noggin cell implantation, while green arrow points to the control on the contra-lateral side. (1) marks control cell implant; (*) marks Noggin cell implant.

control of BMP4 and BMP7 which both are expressed in the ectoderm during this time and may be affected by the Noggin implant (Schultheiss et al., 1997). At stage 4, GATA4 expression was also strongly suppressed by Noggin. However, embryos implanted at stages 5±7 paradoxically showed stronger expression in an enlarged area (g±i). In normal embryos GATA4 expression is downregulated in the anterior ventricular segment at stage 9 and is con®ned to the posterior in¯ow segment (Kostetskii et al., 1999). In Noggin implanted embryos, however, the expression domain of GATA 4 was extended anteriorly (dotted line demarcating the anterior limit of GATA4 expression on the control side). The stronger expression of GATA4 on the Noggin implanted side is illustrated on sections (Fig. 6g 0 -i 0 ). This effect became weaker in older implanted embryos Fig. 6i,j). Suppression of tubular heart formation by Noggin was observed in many embryos (Fig. 6a 0 ±o 0 ). In embryos implanted at stages 4±6 a simple epithelium instead of a tubular heart was present on the implanted side and no separation of endocard and myocard was visible (Fig. 6a 0 ±c 0 ,f 0 -h 0 ,k 0 -m 0 ). Coelom formation and separation of splanchnic and somatic mesoderm still occurred normally in the presence of Noggin. At stages 7 and 8 tube-like structures were present on the Noggin implanted side and their dimension were increased compared to embryos implanted at stages 4±6 (Fig. 6e 0 ,j 0 ,o 0 ).

In this paper we have analyzed the function of BMP2 in early chick cardiac development using both gain-of-function and loss-of-function approaches in vitro and in vivo. It has been previously shown that central mesendoderm while normally not fated to become cardiac mesoderm has the ability to express cardiac marker genes after addition of BMP2, BMP4, or BMP7 making it a useful model for cardiac induction (Schultheiss et al., 1997). Here, we determined the time course of induction of several cardiac marker genes by BMP2 in explanted central mesendoderm. We ®nd that the onset of expression and the time period necessary to reach maximal transcript levels differed for various genes. It should be mentioned that transcripts for some cardiogenic marker genes, such as the GATAs, MEF2A, and NKX2.8 but not NKX2.5 were already detectable by RT-PCR at the time zero of explantation (data not shown). However, these transcripts disappeared during the early hours of culture, unless BMP2 was present. The significance of this observation is presently unclear. 3.1. BMP2 induces cardiac marker genes in central mesendoderm in similar temporal pro®les as in vivo In explant cultures taken from stage 4/5 embryos expression of GATA 4, -5, and -6 and MEF2A is rapidly induced to reach maximal transcript levels within 6 h after BMP2 addition. This is in good agreement with GATA mRNA accumulation in the embryo and re¯ects the overlapping spatiotemporal expression pattern of GATAs and BMP2 (Schultheiss et al., 1997; Jiang et al., 1998). We recently showed that MEF2A in chick is ®rst expressed in cardiac mesoderm of stage 8 embryos (Buchberger and Arnold, 1999). In the explant culture model therefore MEF2A expression seems to be precocious compared to the in vivo situation but the difference might also be the result of the enhanced sensitivity of transcript detection by RT-PCR compared to in situ hybridization. The HAND transcription factors, eHAND and dHAND were induced substantially only after 12 h. In the embryo low level of eHAND accumulates in posterior mesoderm and extraembryonic tissue at stage 5 but signi®cant levels of HAND transcripts within the HFR do not appear before stage 8 (Schultheiss et al., 1995). Thus, GATA and HAND genes respond to BMP2 in explant cultures similar to their appearance in the embryo. Moreover, eHAND was also shown to be activated by BMP2 in animal caps of Xenopus embryos suggesting that this signaling pathway has been conserved during evolution (Sparrow et al., 1998). The onset of NKX2.5 expression in explants also starts 6 hours after BMP2 addition but it continues to rise over at least 48 h. In the mouse eHAND transcription seems to be dependent on NKX2.5 (Biben and Harvey, 1997). This could also be the case in the chick embryo based on the order of activation with HAND genes following the induction of NKX2.5 by BMP2, however, epistatic relationships might

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

265

Fig. 6. Inhibition of cardiac marker gene expression and heart formation by Noggin is age-dependent. All embryos are in a ventral view with the rostral end at the top. Chicken embryos in New culture were implanted with Noggin-expressing cells at stage 4, (panels a,f,k), stage 5 (panels b,g,l), stage 6 (panels c,h,m), stage 7 (panels d,i,n), and stage 8 (panels e,j,o), cultured overnight and subjected to whole-mount in situ hybridization with NKX2.5 (panels a±e), GATA4 (panels f±j), and eHAND (panels k±o). Panels a 0 ±o 0 show transverse sections of the embryos depicted in panels a±o. Bar shows approximate plane of sectioning. Red arrow depicts loss of cardiac gene expression by the Noggin implant; green arrows depict marker gene expression on the contra-lateral side. Where visible the implant of Noggin-expressing cells or control cells is marked. (1) marks control cell implant; (*) marks Noggin cell implant.

not be conserved in every case. NKX2.8 was expressed at baseline in the explanted tissue and BMP2 affected expression only after 24 and 48 h suggesting that NKX2.8 expression may be regulated indirectly by BMP2. NKX2.8 expression is ®rst seen between stage 6 and 8 in the pharyngeal endoderm and only between stage 8 and 11 it is present in the cardiac mesoderm (Brand et al., 1997). Thus, the late response to BMP2 addition probably re¯ects the expression pattern in vivo. Both examined structural genes, titin and vMHC were only signi®cantly induced by 24 h after BMP2 addition, in good correlation to their late activation in the embryo at stage 9 (Tokuyasu and Maher, 1987; Wei et al., 1996). One may speculate that the different time courses of induction by BMP2 re¯ect the epistasis of genes involved in setting up the early tubular heart. Despite the fact that we employed serum-free medium in the explant cultures several unde®ned sources of growth factor activity may be important in addition to BMP2 for cardiac induction in central mesendoderm. The ®bronectin used for coating of the culture dishes was puri®ed from

human plasma and could contain unde®ned growth factor activity. Furthermore, it is unclear to what extent anterocentral endoderm may release growth factors potentially acting as co-factors for cardiac mesoderm induction. In addition, we used conditioned media of BMP2 secreting Q2bn cells that potentially may release co-factors important for the ability of BMP2 to induce heart formation. It has been recently shown that posterior mesendodermal explants do not survive in the presence of BMP2 alone but do so when FGF4 is added. While FGF is probably not important as an instructive inducer of heart formation, it is clear from work in several species that FGFs are important for heart formation (Beimann et al., 1996; Gisselbrecht et al., 1996; Zhu et al., 1999). 3.2. Noggin blocks cardiac marker gene expression in cultures of anterior lateral plate mesendoderm and in the embryo The induction of cardiac-speci®c gene expression im-

266

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

plies but does not prove that BMP2 is an essential signal in the process of cardiogenesis. We attempted to demonstrate the role of endogenous BMP2 by application of one its natural inhibitors, Noggin. Antero-lateral explants were isolated from the HFR of stages 4±8 embryos and placed into culture using serum-free medium. In the presence of conditioned medium of Noggin-expressing CHO cells all three cardiac GATA genes, eHAND, NKX2.5, NKX2.8, MEF2A and vMHC were inhibited in explants taken from stages 4 embryos. This suggests that these genes are indeed absolutely dependent on BMP2 during a short period of cellular competence. Subsequently, in explants taken from stages 5±7 embryos expression of all genes but GATA4 were signi®cantly suppressed by Noggin but some expression remained in each case. This effect was especially pronounced for eHAND and less for other genes. vMHC was least affected by Noggin. Similar results were recently reported by Schultheiss and Lassar using a related in vitro culture system (Schultheiss and Lassar, 1997). However, they reported that Noggin interferes with NKX2.5 and vMHC expression during stages 4 and 5 but not later. The slightly different culture protocols may explain the variable results. We used serum-free medium, as we and others observed that serum contains factors which can promote cardiac differentiation independent of BMP2 (Lough et al., 1996). Thus, the faster independence of BMP signals observed by Schultheiss and Lassar may be caused by serum factors. Our results in vitro and in vivo are in good agreement. A similar time period of sensitivity of cardiac mesoderm formation for inhibition by TPA was described previously (Gonzalez-Sanchez and Bader, 1990). These authors found sensitivity to TPA during stages 4±7, while at stage 8 cellular differentiation was insensitive to TPA treatment. BrdU is another drug shown to interfere with early cardiac differentiation (Chacko and Joseph, 1974). Whether Noggin, TPA and BrdU block cardiac differentiation in related or different signal transduction pathways remains to be seen. Most interestingly, GATA4 showed a different response to Noggin then the other genes. When implanted at stage 4 and stained for GATA gene expression at stage 8, or 11 loss of GATA4 expression was noted similar to NKX2.5 and eHAND. However, in embryos implanted between stages 5 and 7 GATA4 expression was enhanced. Initially at stages 5±8, GATA genes are expressed in the cardiac ®eld, including both cardiac progenitor cells and the associated endoderm. Transcripts are then localized to the primitive heart tube and the adjacent lateral plate mesoderm. By stage 9, transcript levels of all three GATA genes and particularly GATA4 are down-regulated in the presumptive ventricle, remaining higher in the posterior end of the tube including both the presumptive atrium and the sinus venosus (Kostetskii et al., 1999). The posterior expression domain of GATA4 is under the control of retinoic acid and is lost in vitamin A de®cient quail embryos (Kostetskii et al., 1999). The enhanced expression of GATA4 by Noggin could be due

to alterations of retinoic acid synthesis. However, a more likely explanation is that down-regulation of GATA4 in the presumptive ventricular segment is dependent on the extent of myocardial differentiation. Due to the inhibitory in¯uence of Noggin on myocardial differentiation, downregulation of GATA4 might not take place. There is a close inverse correlation in the Noggin-implanted embryos between the level of GATA4 expression in the presumptive ventricle and the extent of tube formation and cardiac histogenesis. It has been recently shown that BMP2 induces the inhibitory Smad molecule Smad6 (Yamada et al., 1999). Beginning at stage 5 Smad6 is expressed in the cardiogenic region overlapping with BMP2 and NKX2.5. Potentially, Smad6 is involved in the downregulation of GATA4 in differentiated myocardial tissue. Consistent with this notion is the fact that Noggin implantation results in down-regulation of Smad6 (Yamada et al., 1999). When Noggin was implanted at stages 4±6 the splanchnopleura remained organized as a simple epithelium and no thickening of the premyocardial mesoderm was observed. This phenotype is reminiscent of Xenopus embryos which were injected with dominant-negative NKX2.3 and NKX2.5 constructs which results in lack of heart formation (Grow and Krieg, 1998). Interference of Noggin with myocardial differentiation was gradually lost in older embryos suggesting that cardiac speci®cation is complete at stage 8. Noggin also lost most of its inhibitory effect on cardiac gene expression when added to lateral mesendoderm explants taken from stage 8 embryo. Thus, the time period of Noggin sensitivity in vivo and in vitro appeared quite similar, although different methods of transcript detection were used. The fact that cardiogenic mesoderm loses sensitivity to Noggin and presumably the requirement for BMP2 around stage 8 suggests that by this time of development cardiogenic cells have acquired stable molecular identity. Apparently, the expression of NKX2.5 or GATA genes is not suf®cient for stable commitment of mesoderm to the cardiac cell lineage. Several other myocardial transcription factors have been identi®ed recently and may play a role at the time when cardiac mesoderm becomes BMP2-independent. These include serum response factor (SRF), pCMF1 and TBX5 (Croissant et al., 1996; Wei et al., 1996; GibsonBrown et al., 1998). We also observed that BMP7 which has a much lower af®nity to Noggin than BMP2 (Croissant et al., 1996; Wei et al., 1996; Gibson-Brown et al., 1998) is expressed in the myocardium of stage 9 chick embryos. Thus, BMP7 and several other BMPs present in the early heart (T. Schlange, unpublished observation) may substitute for the function of BMP2 at later stages. Alternatively, an additional signaling pathway may be required for cellular differentiation at this developmental stage. Interestingly the mouse null mutant of the signaling molecule CRIPTO fails to form functional cardiac mesoderm despite normal accumulation of GATA4, MEF2C, eHAND and dHAND suggesting that it may function in addition to or downstream of BMP2 (Xu et al., 1999).

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

3.3. BMP2 signaling in cardiogenesis is evolutionary conserved and probably delimits the heart forming area Primary heart formation involves several steps, such as cell type speci®cation, differentiation and heart morphogenesis. Based on our observations we would like to propose that the ®rst step involves antero-lateral mesoderm recruitment to the cardiogenic lineage in response to BMP2 in the most lateral mesendoderm at stage 4. Subsequently, cells remaining in contact with the underlying pharyngeal endoderm which also expresses BMP2 retain the cardiogenic fate, while cells outside this area probably lose the potential to form heart. This still plastic state of speci®cation is supported by our observation that early cardiogenic cells in the embryo loose NKX2.5, GATA, and eHAND expression when challenged with Noggin, while at stage 8 marker gene expression appears to be stably ®xed. In addition there seems to be heterochrony in the state of differentiation in different segments of the primitive tubular heart. At the posterior boundaries of the heart ®eld there is probably continuous recruitment of new mesodermal cells to the cardiogenic lineage, while more differentiated cells move anteriorly (Markwald et al., 1999). Different segments of the heart probably arise at different time-points during development. Consistent with this view is the fact that BMP2 is expressed much longer in the posterior pharyngeal endoderm which is in contact with the sinoatrial segments of the heart (AndreÂe et al., 1998). Thus, Noggin interference with cardiac marker gene expression later than stage 6 is probably affecting prospective cardiac mesoderm fated to form posterior segments of the tubular heart cells which at this time are still plastic in their state of speci®cation. Further work is required to identify additional factors important in speci®cation and morphogenesis of the early tubular heart.

4. Experimental procedures 4.1. Cell culture and explant culture CHO.B3.A4 cells expressing Xenopus Noggin are a generous gift of Dr. Richard Harland, UC Berkley (Lamb et al., 1993). Parental CHO dhfr 2 cells which were used as controls, were cultured in aMEM with nucleosides (GibcoLife Science, Heidelberg) containing 10% FCS. Noggin producing CHO cells were cultured in aMEM without nucleosides containing 10% dialyzed FCS (both Gibco), 80 mM methotrexate (Sigma), 1% sodium pyruvate and 1% non-essential amino acids (Lamb et al., 1993). To produce spherical cell aggregates for implantation CHO.B3.A4 cells and control cells were grown to approximately 90% con¯uence in 90 mm f culture dishes, mildly trypsinized, seeded into bacteriological Petri dishes, and incubated in aMEM with nucleosides, 10% FCS for 2 days. Under these conditions the cells formed aggregates

267

of different sizes which were selected for implantation under the stereo microscope. Q2bn cells producing BMP2 and control cells were cultured as described previously (AndreÂe et al., 1998). In order to generate BMP2 and Noggin conditioned media Q2bn cells expressing human BMP2 and Noggin producing CHO cells were cultured until they reached con¯uency. Cells were washed twice with PBS in order to remove serum and were cultured for 2 days in serum-free M199 medium. The medium was sterile-®ltered and used directly for the induction experiments. Control medium was generated in the same manner using control Q2bn cells (AndreÂe et al., 1998) and control CHO dhfr 2 cells. For explant cultures chicken embryos of stage 5 were stretched out on a glass ring in Pannett-Compton saline. The embryo was digested for 30 s using 1 units/ml Dispase (Boehringer Mannheim). The enzyme was blocked subsequently by adding 0.02% EDTA in PBS. Central or lateral anterior mesendodermal fragments were isolated with tungsten needles. Culture plates were coated for 30 min at 378C with 1 mg/ml ®bronectin (isolated from human plasma, Gibco-Life Science, Heidelberg) and than washed twice with M199 medium. Explants to be cultured were transferred to the ®bronectin-coated culture plates and were incubated for 30 min in 100 ml serum-free M199 medium containing antibiotics. Subsequently, conditioned medium containing either Noggin or BMP2 was added to the explants. Induction experiments with BMP2 and central mesendodermal explants were terminated after 6, 12, 24 and 48 h of culture (Fig. 1A). Inhibition experiments were performed as follows: explants of antero-lateral mesendoderm were isolated from stage 4±8 embryos and placed into 200 ml of serum-free M199. Subsequently, 200 ml of conditioned M199 medium of Noggin-producing, or control CHO cells was added (Fig. 1C). Explants were cultured for additional 24 h. For termination of explant cultures, cells were washed once with PBS and subsequently were lysed in guanidinium containing lysis buffer and were frozen at 2808C. Total RNA was isolated as described (Schultheiss et al., 1997). A total of three independent experiments were run for each individual marker gene analyzed. 4.2. New culture and cell implantation White Leghorn eggs (Lohman, Cuxhafen) were incubated until they reached stage 4 unless otherwise stated. Embryos were placed in New culture (New, 1955; AndreÂe et al., 1998) and CHO cell aggregates were implanted into the embryo by making a small slit into the endoderm with tungsten needles. In some embryos either CHO.B3.A4 cells, or Q2bn cells producing BMP2 and the respective control cells were implanted at contra-lateral positions (Fig. 1B,D). Embryos were cultured until they reached stages 8±9. In order to test for the time-dependent effects of Noggin on myocardial differentiation, eggs were cultured until they reached, stages 4, 5, 6, 7, or 8. Embryos were implanted

268

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

with CHO.B3.A4 cell aggregates and cultured until they reached stage 8. The AIP was split using tungsten needles in order to prevent fusion of both cardiac anlagen and thereby a comparison of the HFR receiving the Noggin implant and the control HFR was possible (Fig. 1E). After overnight culture embryos were ®xed in buffered 4% paraformaldehyde and subjected to whole-mount in situ hybridization as described (AndreÂe et al., 1998). 4.3. Whole-mount in situ hybridization Whole-mount in situ hybridization was performed as previously described (Brand et al., 1997). Chicken embryos were ®xed overnight in buffered 4% paraformaldehyde. Subsequently, embryos were dehydrated by incubation into increasing concentrations of methanol. Antisense probes were synthesized using linearized plasmids in the presence of digoxigenin-labeled UTP with either T3, or T7 polymerase and detected with anti-digoxigenin alkaline phosphatase antibody (Boehringer Mannheim). Color reaction was performed with 4-nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl-phosphate (Boehringer Mannheim). After the color reaction was complete, embryos were washed in PBT and re®xed in buffered 4% paraformaldehyde. The chicken NKX2.5 probe was a 550 bp fragment encoding homeobox and the NK2 domain (Buchberger et al., 1996). The chicken GATA4 probe was a 750 bp cDNA fragment of the zinc ®nger region (Laverriere et al., 1994). The chicken GATA5 probe consisted of a 500 bp HincII-EcoRI fragment from the 3 0 end (Laverriere et al., 1994). The chicken GATA6 probe consisted of a 500 bp HindIII-XhoI fragment from the 3 0 end (Laverriere et al.,

1994). The eHAND was 521 bp probe consisting of bp 104± 625 of the coding sequence (Schultheiss et al., 1995). Embryos to be sectioned were in®ltrated overnight in 15% sucrose in PBS at 48C and then embedded in 15% sucrose/7.5% gelatin in PBS. After the gelatin had set, blocks were trimmed, frozen in isopentane cooled on dry ice, and sectioned using a cryostat. Whole-mount embryos were photographed through a Leica M10 photomicroscope on 1% agar plates and on a Zeiss Axiomat with Nomarski optics for sectioned material using Kodak T64 ®lms. 4.4. RT-PCR Cultured explants were solubilized in lysis buffer and stored at 2808C. Subsequently total RNA was isolated as previously described (Schultheiss et al., 1997). cDNA was synthesized from DNase treated total RNA using AMV reverse transcriptase. PCR was performed using the primer pairs shown in Table 1. The PCR products were size separated on 2% agarose gels and blotted on Biodyne nylon membrane (Pall Gelman Sciences, Rossbach). In addition, PCR products were sequenced to prove identity. Southern blot hybridizations were performed using either internal oligonucleotides, or cDNA probes. The following internal oligonucleotide primer were used to probe the ampli®ed products by Southern blot hybridization: b-actin: 5 0 -CTCTTCCAGCCATCTTTCTTG-3 0 titin: 5 0 -CTGCTGCAGAAGCGAGTTCC-3 0 VMHC: 5 0 -GAAATAGCTGAATCTCAAGT-3 0 GATA5: 5 0 -TCAACCTTCGAGTATTTGGA-3 0 GATA6:5 0 -TGCGTCAACTGCGGCTCCAT-3 0

Table 1 Primer pairs Name

Primer sequence

Product size (bp)

Ann. temp. (8C)

Cycles

NKX2.5

5 0 -CCTTCCCCGGCCCCTACTAC-3 0 5 0 -CTGCTGCTTGAACCTTCTCT-3 0 5 0 -ACCGATCCCCCCCCGGAGGAT-3 0 5 0 -AGTTTGGGAGAACAAAACG-3 0 5 0 -CTCCTACTCCAGCCCTTACC-3 0 5 0 -GCCCTGTGCCATCTCTCCTC-3 0 5 0 -TAACGGGCTCTACTCCAGC-3 0 5 0 -ATGAAGACAGCCTCTTCTGG-3 0 5 0 -AGAACTCCGTGCTGCACTGCC-3 0 5 0 -TGGTGTGGCAGTTGGCACAGG-3 0 5 0 -TGCCACCAGGAGCGCGCTTAT-3 0 5 0 -CGGGGCTCAGGGGTTCAGTTC-3 0 5 0 -CGCGCTCACTGCTTGAGCTCC-3 0 5 0 -CACGGCGATGAGTCTTGTGGG-3 0 5 0 -GCGTGAGCACATATGAAGTG-3 0 5 0 -CAGTATGCCAGCACTGATATG-3 0 5 0 -TCTTATATCTGGGAGCCAGG-3 0 5 0 -GCTACAAACACCAAGCAGAG-3 0 5 0 -CTTCATCTGCAATCTGGACC-3 0 5 0 -GCTACAAACACCAAGCAGAG-3 0 5 0 -ATCACAGGGGTGTGGGTGTT-3 0 5 0 -AATGAGAGGTTCAGGTGCCC-3 0

221

53

30

243

53

30

224

52

30

319

55

30

281

63

30

521

64

30

666

60

30

294

56

30

211

55

30

490

53

30

409

60

28

NKX2.8 GATA4 GATA5 GATA6 EHAND DHAND MEF2A VMHC Titin b-actin

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

In all other cases cDNA probes were used to probe the Southern blots. For NKX2.5 a partial cDNA probe was used as described (AndreÂe et al., 1998). NKX2.8 was a partial cDNA as described (Brand et al., 1997). MEF2A probe was a 401 bp product encompassing the 5 0 end of the cDNA (Drazen Sosic and Eric Olson unpublished). The dHAND probe was 1100 bp full-length cDNA probe and eHAND was 521 bp probe consisting of bp 104±625 of the coding sequence (Srivastava et al., 1995). For GATA4 a 750 bp cDNA fragment was used (Laverriere et al., 1994). Acknowledgements Expert technical assistance by Kerstin Zander is gratefully acknowledged. The authors thank Charlotte Klaue for secretarial help. We thank Dr. Richard Harland for the generous gift of CHO cells expressing Xenopus Noggin and Delphine Duprez for providing the Q2bn cells expressing BMP2. We are indebted to Todd Evans for providing GATA probes, Depaak Srivastava for providing HAND probes and Dravor Sosic and Eric Olson for providing the MEF2A cDNA. This work was supported by the Deutsche Forschungsgemeinschaft, SFB, Teilprojekt A1 (T.B and H.A.) References AndreÂe, B., Duprez, D., Vorbusch, B., Arnold, H.-H., Brand, T., 1998. BMP-2 induces ectopic expression of cardiac lineage markers and interferes with somite formation in chicken embryos. Mech. Dev. 70, 119± 131. Antin, P., Taylor, R., Yatskievych, T., 1994. Precardiac mesoderm is speci®ed during gastrulation in quail. Dev. Dyn. 200, 144±154. Antin, P.B., Yatskievych, T., Dominguez, J.L., Chief®, P., 1996. Regulation of avian precardiac mesoderm development by insulin and insulin-like growth factors. J Cell Physiol 168, 42±50. Arai, A., Yamamoto, K., Toyama, J., 1997. Murine cardiac progenitor cells require visceral embryonic endoderm and primitive streak for terminal differentiation. Dev Dyn 210, 344±353. Beimann, M., Shilo, B.-Z., Volk, T., 1996. Heartless, a Drosophila FGF receptor homolog, is essential for cell migration and establishment of several mesodermal lineages. Genes Dev 10, 2993±3002. Biben, C., Harvey, R.P., 1997. Homeodomain factor Nkx2-5 controls left/ right asymmetric expression of bHLH gene eHand during murine heart development. Genes Dev 11, 1357±1369. Bisaha, J., Bader, D., 1991. Identi®cation and characterization of a ventricular-speci®c avian myosin heavy chain. VMHC1: expression in differentiating cardiac and skeletal muscle. Dev. Biol. 148, 355±364. Brand, T., Andree, B., Schneider, A., Buchberger, A., Arnold, H.H., 1997. Chicken NKx2-8, a novel homeobox gene expressed during early heart and foregut development. Mech. Dev. 64, 53±59. Buchberger, A., Arnold, H.H., 1999. The MADS domain containing transcription factor cMef2a is expressed in heart and skeletal muscle during embryonic chick development. Dev. Genes Evol. 209, 376±381. Buchberger, A., Pabst, O., Brand, T., Seidl, K., Arnold, H.-H., 1996. Chick Nkx-2.3 represents a novel family member of vertebrate homologues to the Drosophila homeobox gene tinman. Differential expression of cNKx-2.3 and cNKx-2.5 during cardiac and gut development. Mech. Dev. 56, 151±163. Capdevila, J., Johnson, R., 1998. Endogenous and ectopic expression of

269

noggin suggests a conserved mechanism for regulation of BMP function during limb and somite patterning. Dev. Biol. 197, 205±217. Chacko, S., Joseph, X., 1974. The effect of 5-bromodeoxyuridine (BrdU) on cardiac muscle differentiation. Dev. Biol. 40, 340±354. Croissant, J.D., Kim, J.H., Eichele, G., Goering, L., Lough, J., Prywes, R., Schwartz, R.J., 1996. Avian serum response factor expression restricted primarily to muscle cell lineages is required for alpha-actin gene transcription. Dev. Biol. 177, 250±264. DeHaan, R., 1965. Migration patterns of precardiac mesoderm in the early chick embryo. Exp. Cell Res. 29, 544±560. Dudley, A., Robertson, E., 1997. Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 de®cient embryos. Dev. Dyn. 208, 349±362. Edmondson, D., Lyons, G.E., Martin, J.F., Olson, E.N., 1994. Mef2 gene expression marks the cardiac and skeletal muscle lineages during mouse embryogenesis. Development 120, 1251±1263. Ehrman, L.A., Yutzey, K.E., 1999. Lack of regulation in the heart forming region of avian embryos. Dev. Biol. 207, 163±175. Frasch, M., 1995. Induction of visceral and cardiac mesoderm by ectodermal Dpp in the early Drosophila embryo. Nature 374, 464±467. Gannon, M., Bader, D., 1995. Initiation of cardiac differentiation occurs in the absence of anterior endoderm. Dev. Biol. 121, 2439±2450. Garcia-Martinez, V., Schoenwolf, G., 1993. Primitive-streak origin of the cardiovascular system in avian embryos. Dev. Biol. 159, 706±719. Gibson-Brown, J., Agulnik, S., Silver, L., Papaioannou, V., 1998. Expression of T-box genes Tbx2-Tbx5 during chick organogenesis. Mech. Dev. 74, 165±169. Gisselbrecht, S., Skeath, J., Doe, C., Michelson, A., 1996. heartless encodes a ®broblast growth factor receptor (DFR1/DFGF-R2) involved in the directional migration of early mesodermal cells in the Drosophila embryo. Genes Dev. 10, 3003±3017. Gonzalez-Sanchez, A., Bader, D., 1990. In vitro analysis of cardiac progenitor cell differentiation. Dev. Biol. 139, 197±209. Grow, M.W., Krieg, P.A., 1998. Tinman function is essential for vertebrate heart development: elimination of cardiac differentiation by dominant inhibitory mutants of the tinman-related genes, XNkx2-3 and XNkx2-5. Dev. Biol. 204, 187±196. Hamburger, V., Hamilton, H., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49±92. Han, Y., Dennis, J.E., Cohen-Gould, L., Bader, D.M., Fischman, D.A., 1992. Expression of sarcomeric myosin in the presumptive myocardium of chicken embryos occurs within six hours of myocyte commitment. Dev. Dyn. 193, 257±265. Jiang, Y., Tarzami, S., Burch, J.B.E., Evans, T., 1998. Common role for each of the cGATA4/5/6 genes in the regulation of cardiac morphogenesis. Dev. Genet. 22, 263±277. Kishimoto, Y., Lee, K.H., Zon, L., Hammerschmidt, M., Schulte-Merker, S., 1997. The molecular nature of zebra®sh swirl: BMP2 function is essential during early dorsoventral patterning. Development 124, 4457± 4466. Kokan-Moore, N.-P., Bolender, D.L., Lough, J., 1991. Secretion of inhibin bA by endoderm cultured from early embryonic chicken. Dev. Biol. 146, 242±245. Kostetskii, I., Jiang, Y., Kostetskaia, E., Yuan, S., Evans, T., Zile, M., 1999. Retinoid signaling required for normal heart development regulates GATA- 4 in a pathway distinct from cardiomyocyte differentiation. Dev. Biol. 206, 206±218. Ladd, A.N., Yatskievych, T.A., Antin, P.B., 1998. Regulation of avian cardiac myogenesis by Activin/TGFbeta and bone morphogenetic proteins. Dev. Biol. 204, 407±419. Lamb, T.M., Knecht, A.K., Smith, W.C., Stachel, S.E., Economides, A.N., Stahl, N., Yancopolous, G.D., Harland, R.M., 1993. Neural induction by the secreted polypeptide noggin. Science 262, 713±718. Laverriere, A.C., MacNeill, C., Mueller, C., Poelmann, R.E., Burch, J.B., Evans, T., 1994. GATA-4/5/6, a subfamily of three transcription factors transcribed in developing heart and gut. J. Biol. Chem. 269, 23177± 23184.

270

T. Schlange et al. / Mechanisms of Development 91 (2000) 259±270

Lough, J., Barron, M., Brogley, M., Sugi, Y., Bolender, D.L., Zhu, X., 1996. Combined BMP-2 and FGF-4, but neither factor alone, induces cardiogenesis in non-precardiac embryonic mesoderm. Dev. Biol. 178, 198± 202. Markwald, R., Trusk, T., Moreno-Rodriguez, R., 1999. Formation and septation of the tubular heart. Integrating dynamics of morphology with emerging molecular concepts. In: de la Cruz, M., Markwald, R. (Eds.). Living Morphogenesis of the Heart, Cardiovascular Morphogenesis, BirkhaÈuser, Boston, MA, pp. 43±84. Monzen, K., Shiojima, I., Hiroi, Y., Kudoh, S., Oka, T., Takimoto, E., Hayashi, D., Hosoda, T., Habara-Ohkubo, A., Nakaoka, T., Fujita, T., Yazaki, Y., Komuro, I., 1999. Bone morphogenetic proteins induce cardiomyocyte differentiation through the mitogen-activated protein kinase TAK1 and cardiac transcription factors Csx/Nkx-2.5 and GATA-4. Mol. Cell. Biol. 19, 7096±7105. New, D., 1955. A new technique for the culture of the chick embryo in vitro. J. Embryol. Exp. Morphol. 3, 326±331. Parlow, M.H., Bolender, D.L., Kokan-Moore, N.P., Lough, J., 1991. Localization of bFGF-like proteins as punctate inclusions in the pre-septation myocardium of the chicken embryo. Dev. Biol. 146, 139±147. Rawles, M., 1943. The heart-forming regions of the early chick blastoderm. Physiol. Zool. 16, 22±42. Rosenquist, G., 1970. Location and movement of cardiogenic cells in the chick embryo: the heart-forming portion of the primitive streak. Dev. Biol. 22, 461±475. Rosenquist, G., DeHaan, R., 1966. Migration of precardiac cells in the chick embryo: a radioautographic study. Contrib. Embryol. 38, 111± 121. Schultheiss, T., Lassar, A., 1997. Induction of chick cardiac myogenesis by bone morphogenetic proteins. Cold Spring Harb. Symp. Quant. Biol. 62, 413±419. Schultheiss, T.M., Xydas, S., Lassar, A.B., 1995. Induction of avian cardiac myogenesis by anterior endoderm. Development 121, 4203±4214. Schultheiss, T., Burch, J., Lassar, A., 1997. A role for bone morphogenetic proteins in the induction of cardiac myogenesis. Genes Dev. 11, 451± 462. Solloway, M.J., Robertson, E.J., 1999. Early embryonic lethality in Bmp5;Bmp7 double mutant mice suggests functional redundancy within the 60A subgroup. Development 126, 1753±1768. Sparrow, D.B., Kotecha, S., Towers, N., Mohun, T.J., 1998. Xenopus eHAND: a marker for the developing cardiovascular system of the

embryo that is regulated by bone morphogenetic proteins. Mech. Dev. 71, 151±163. Srivastava, D., Cserjesi, P., Olson, E.N., 1995. A subclass of bHLH proteins required for cardiac morphogenesis. Science 270, 1995±1999. Streit, A., Lee, K., Woo, I., Roberts, C., Jessell, T., Stern, C., 1998. Chordin regulates primitive streak development and the stability of induced neural cells, but is not suf®cient for neural induction in the chick embryo. Development 125, 507±519. Sugi, Y., Lough, J., 1994. Anterior endoderm is a speci®c effector of terminal cardiac myocyte differentiation in cells from the embryonic heart forming region. Dev. Dyn. 200, 155±162. Sugi, Y., Lough, J., 1995. Activin-A and FGF-2 mimic the inductive effects of anterior endoderm on terminal cardiac myogenesis. Dev. Biol. 168, 567±574. Tam, P.P., Parameswaran, M., Kinder, S.J., Weinberger, R.P., 1997. The allocation of epiblast cells to the embryonic heart and other mesodermal lineages: the role of ingression and tissue movement during gastrulation. Development 124, 1631±1642. Tokuyasu, K.T., Maher, P.A., 1987. Immunocytochemical studies of cardiac myo®brillogenesis in early chick embryos. I. Presence of immuno¯uorescent titin spots in premyo®bril stages. J. Cell Biol. 105, 2781± 2793. Tonegawa, A., Moriya, M., Tada, M., Nishimatsu, S., Katagiri, C., Ueno, N., 1996. Heart formative factor(s) is localized in the anterior endoderm of early Xenopus neurula. Roux's Arch. Dev. Biol. 205, 282±289. Wei, Y., Bader, D., Litvin, J., 1996. Identi®cation of a novel cardiac-speci®c transcript critical for cardiac myocyte differentiation. Development 122, 2779±2789. Xu, C., Liguori, G., Persico, M.G., Adamson, E.D., 1999. Abrogation of the Cripto gene in mouse leads to failure of postgastrulation morphogenesis and lack of differentiation of cardiomyocytes. Development 126, 483± 494. Yamada, M., Szendro, P.I., Prokscha, A., Schwartz, R.J., Eichele, G., 1999. Evidence for a role of Smad6 in chick cardiac development. Dev. Biol. 215, 48±61. Zhang, H., Bradley, A., 1996. Mice de®cient for BMP2 are nonviable and have defects in amnion/chorion and cardiac development. Development 122, 2977±2986. Zhu, X., Sasse, J., Lough, J., 1999. Evidence that FGF receptor signaling is necessary for endoderm- regulated development of precardiac mesoderm. Mech. Ageing Dev. 108, 77±85.