Anterior expression of the caudal homologue cCdx-B activates a posterior genetic program in avian embryos

Authors

  • Lisa A. Ehrman,

    1. Division of Molecular Cardiovascular Biology, The Children's Hospital Research Foundation, Cincinnati, Ohio
    Search for more papers by this author
  • Katherine E. Yutzey

    Corresponding author
    1. Division of Molecular Cardiovascular Biology, The Children's Hospital Research Foundation, Cincinnati, Ohio
    • Division of Molecular Cardiovascular Biology, The Children's Hospital Research Foundation, 3333 Burnet Avenue, Cincinnati, OH 45229
    Search for more papers by this author

Abstract

Several families of regulatory genes have been implicated in anteroposterior patterning of gastrulation-stage vertebrate embryos. Members of the Drosophila caudal family of homeobox genes (Cdx) are among the earliest regulators of posterior cell fates. The regulatory cascade initiated by the caudal homologue, cCdx-B, was examined in avian embryos. During gastrulation, cCdx-B is expressed with other posterior patterning genes. In the posterior primitive streak, cCdx-B expression coincides with posteriorly expressed Hox cluster genes and Wnt family members such as Wnt-8c. The hierarchical relationship between these patterning genes was examined after anterior ectopic expression of cCdx-B. cCdx-B expression in anterior cardiogenic cells by means of adenoviral infection leads to the induction of Wnt-8c and the posterior Hox genes, Hoxa-7, Hoxc-6, and Hoxc-8. Cardiogenesis is not inhibited in cCdx-B expressing anterior lateral mesoderm, indicating that anterior cell fates are not respecified with the activation of posterior patterning genes after gastrulation. These results support an important role for cCdx-B in initiating a posterior program of gene expression that includes Wnt signaling molecules and the Hox cluster genes. © 2001 Wiley-Liss, Inc.

INTRODUCTION

The initiation and maintenance of anteroposterior patterning of gastrulation stage vertebrate embryos is a dynamic process involving several families of transcription factors and signaling molecules. In avian embryos, the first indicator of asymmetry in the radial blastoderm is the formation of Koller's sickle in the posterior marginal zone (Eyal-Giladi and Kochav, 1976). This structure represents the initiation of gastrulation and primitive-streak formation along the anteroposterior axis. Extensive fate mapping studies of chicken embryos demonstrate that the progenitors of anterior structures, such as the head and heart, gastrulate before posterior extraembryonic and blood cell lineages (Garcia-Martinez and Schoenwolf, 1992, 1993). Distinct anterior and posterior cell lineages are present in the lateral plate mesoderm in late primitive-streak–stage embryos (Gonzalez-Sanchez and Bader, 1990; Schultheiss et al., 1995). However, positional information in the lateral plate can be altered with retinoid treatment or transplantation (Sundin and Eichele, 1992; Inagaki et al., 1993). Thus, there is plasticity in the anteroposterior patterning of newly gastrulated cells long after the axes are initially established. Although several families of transcription factors and secreted proteins have been implicated in the regulation of anteroposterior patterning during gastrulation, the relationships between many of these factors have not been clearly established (Bachvarova, 1999).

Caudal-related homeobox genes (Cdx) are among the critical regulators of anteroposterior patterning in a diversity of vertebrate and invertebrate embryos (Mlodzik et al., 1985; Frumkin et al., 1991; Duprey et al., 1988; Northrop and Kimelman, 1994; Hunter and Kenyon, 1996). Cdx genes are present in evolutionarily related ParaHox clusters in amphioxus and human genomes (Brooke et al., 1998; Pollard and Holland, 2000). The three vertebrate Cdx paralogs are expressed in overlapping posterior regions in gastrulating mouse, chicken, and Xenopus laevis embryos (Meyer and Gruss, 1993; Gamer and Wright, 1993; Marom and Fainsod, 1997; Pillemer et al., 1998). Later, during organogenesis, Cdx proteins regulate gut development and intestinal-specific gene expression (Frumkin et al., 1991; Suh et al., 1994; Wu and Lengyel, 1998). Functional studies in Drosophila and Xenopus reveal an early and important role for Cdx proteins in anteroposterior patterning. In flies, ectopic anterior expression of Caudal leads to the disruption of head development and segmentation. Conversely, loss of caudal expression results in the absence of abdominal segments (Macdonald and Struhl, 1986; Mlodzik et al., 1990; Wu and Lengyel, 1998). In Xenopus embryos, overexpression of Xcad-3 results in loss of head structures, and expression of an Xcad-3 repressor protein inhibits tail development (Pownall et al., 1996; Isaacs et al., 1998). These studies demonstrate an important role for Caudal homologues in regulating anteroposterior patterning.

Wnt signaling proteins also have been implicated in the early stages of posterior patterning. Several Wnts, including Wnt-8c and Wnt-3a, are expressed in the posterior region of chicken and mouse embryos at primitive-streak stage 5 (Hume and Dodd, 1993; Bouillet et al., 1996; Marvin et al., 2001). In mouse and Xenopus embryos, altered expression of Wnt-8 disrupts anterior development (Christian and Moon, 1993; Popperl et al., 1997). There is conflicting evidence for the relationship between caudal/Cdx gene regulation and the Wingless/Wnt signaling pathway. In Drosophila caudal mutant embryos, the posterior domain of expression of the wingless (wg) gene is absent, indicating an upstream regulatory role for caudal in regulating wg gene expression (Wu and Lengyel, 1998). In embryonic mouse intestine, Cdx-1 expression is regulated directly by the Wnt/β-catenin signaling pathway (Lickert et al., 2000). However, the relationship between Cdx and Wnt genes in anteroposterior patterning of gastrulation-stage vertebrate embryos has not been established.

One of the mechanisms by which Cdx homeoproteins regulate early embryonic anteroposterior patterning is by direct regulation of Antennapedia-like Hox cluster gene expression (Subramanian et al., 1995; Charite et al., 1998). Temporal and spatial restriction of Hox gene expression is associated with anteroposterior patterning of the embryo as a whole as well as in patterning a variety of organs and structures in the developing embryo (Krumlauf, 1994). A principal of colinearity in the temporal and spatial regulation of Hox cluster gene expression has been described as a mechanism for establishing positional identity. Activation of posteriorly expressed Hox genes in the lateral plate region occurs relatively late in gastrulation. In avian embryos of primitive-streak stage, restricted patterns of Hox gene expression are present in the lateral plate region (Sundin and Eichele, 1992; Gaunt and Strachan, 1996; Lemaire and Kessel, 1997; Searcy and Yutzey, 1998). At these stages, Cdx expression also is evident in the posterior segments of the embryo (Marom and Fainsod, 1997). Functional Cdx binding sites within the regulatory regions of posteriorly expressed Hox genes have been previously identified (Subramanian et al., 1995; Charite et al., 1998). Global regulation of posterior Hox genes by Cdx homologues is supported by homeotic transformations of multiple vertebrae in mice with targeted mutation of Cdx-1 or Cdx-2 (Subramanian et al., 1995; Chawengsaksophak et al., 1997). These studies provide strong evidence for the regulation of Hox gene expression by Cdx. However, the temporal and spatial limits of this regulatory interaction relative to specification of positionally restricted cell fates has not been thoroughly investigated.

In this study, the functional role of the Caudal homologue cCdx-B in regulating the expression of anteroposterior patterning genes in the chicken embryo was examined. The effect of programmatic changes in anteroposterior patterning genes on cardiomyogenic lineage determination also was assessed. The full-length coding sequence of cCdx-B was isolated and used for functional analysis of altered Cdx expression in the anterior lateral mesoendoderm of chicken embryos in late primitive-streak stage. Anterior expression of cCdx-B induces expression of Wnt-8c and posterior Hox cluster genes. However, cardiac lineage determination is not inhibited by activation of posterior patterning genes in the anterior lateral mesoderm. These studies place cCdx-B high in a regulatory hierarchy upstream of posterior-restricted Wnt gene expression, in addition to regulating posterior Hox cluster genes.

RESULTS

Isolation of cCdx-B Sequence

The full-length sequence of cCdx-B was isolated to initiate functional analyses of its role in anteroposterior patterning in gastrulation-stage chicken embryos. Expression of cCdx-B was initially detected in a degenerate screen for homeobox genes present during heart formation (Searcy and Yutzey, 1998). The full-length coding sequence of cCdx-B was subsequently isolated and encodes for a 264-amino acid protein with high homology to other Cdx proteins throughout its length (Fig. 1A). A partial cDNA sequence for cCdx-B with an incomplete open reading frame has been published previously (Morales et al., 1996). Comparison of the deduced amino acid sequence of the cCdx-B homeodomain reveals an 80% homology with Drosophila caudal and an 87–97% homology with other vertebrate caudal family members (Fig. 1B). The homeodomain sequence of cCdx-B is most identical to mouse Cdx-4 with a homology of 97%. The cCdx-B homeodomain has 87% and 88% identity with cCdx-A and cCdx-C, respectively, indicating lower levels of identity with other Cdx family members (Frumkin et al., 1991; Serrano et al., 1993). Thus, cCdx-B represents the chicken homologue of the Cdx-4 subfamily (Gamer and Wright, 1993).

Figure 1.

A: Primary nucleotide and deduced amino acid sequence of cCdx-B. The homeodomain and the polyadenylation signal are underlined. B: Homology of the cCdx-B homeodomain to other members of the caudal family arranged into the three different subfamilies. The homeodomain for Drosophila caudal is included. cCdx-B was identified as a member of the Cdx-4 subfamily of caudal-related genes based on the homeodomain sequence and conserved regions outside the homeodomain. The dots indicate the amino acid residues that are identical to the cCdx-B homeodomain sequence. The homeodomain sequences used are mouse Cdx-4 (Gamer and Wright, 1993), human Cdx-4 (Horn and Ashworth, 1995), Xenopus Xcad-3 (Northrop and Kimelman, 1994), chick Cdx-A (Frumkin et al., 1991), mouse Cdx-1 (Duprey et al., 1988), human Cdx-1 (Bonner et al., 1995), Xenopus Xcad-2 (Blumberg et al., 1991), chick Cdx-C (Marom and Fainsod, 1997), mouse Cdx-2 (James and Kazenwadel, 1991), human Cdx-2 (Mallo et al., 1997), Xenopus Xcad-1 (Blumberg et al., 1991), Drosophila cad (Mlodzik et al., 1985).

cCdx-B Is Expressed With Posterior Patterning Genes in Gastrulation-Stage Chicken Embryos

Expression of cCdx-B was examined in avian embryos during the critical stages of gastrulation and neurulation (Hamburger and Hamilton, stages 5–8). By late primitive-streak stages, differences in developmental potential of anterior vs. posterior lateral mesoderm are already established (Gonzalez-Sanchez and Bader, 1990; Schultheiss et al., 1995). Between stages 5 and 8, progenitors of the blood and hindgut gastrulate and activate posterior patterning genes (Garcia-Martinez and Schoenwolf, 1992; Marom and Fainsod, 1997). At stage 5, cCdx-B is expressed in the posterior lateral plate but not in the anterior lateral regions of the embryo (Fig. 2A). cCdx-B expression also is evident in the extraembryonic tissue surrounding the embryo. At stages 7 and 8, cCdx-B expression in the lateral plate regresses with Hensen's node but is maintained in the extraembryonic membranes and yolk sac tissue (Fig. 2D,G). Later in development, cCdx-B expression was not detected in the developing hindgut at stages 12–15 (data not shown). This restriction of cCdx-B to gastrulation stage embryos is in contrast to Cdx-1/cCdx-A and Cdx-2/cCdx-C, which are expressed in the hindgut (Duprey et al., 1988; James and Kazenwadel, 1991; Frumkin et al., 1991; Marom and Fainsod, 1997). However, the mouse homologue Cdx-4 also is expressed only during gastrulation and neurulation stages (Gamer and Wright, 1993). Thus, cCdx-B (Cdx-4) seems to function exclusively in early embryonic patterning and is likely not involved in gut patterning and differentiation.

Figure 2.

Whole-mount in situ hybridization of early avian embryos with cCdx-B (A,D,G), cWnt-8c (B,E,H), and Hoxa-7 (C,F,I) digoxigenin labeled RNA probes. At stage 5, cCdx-B (A) and cWnt-8c (B) are expressed in the region posterior to Hensen's node. cCdx-B is also expressed in the extraembryonic tissue anterior and lateral to the heart forming region. At stage 7, the posterior expression of cCdx-B (D) and cWnt-8c (E) regresses posteriorly with Hensen's node. cWnt-8c expression is retained rostral to the first somite. Expression of Hoxa-7 is first detected by in situ hybridization at stage 7 in the posterior tail bud region (F). At stage 8, overlapping expression of cCdx-B (G), cWnt-8c (H), and Hoxa-7 (I) is detected in the posterior primitive streak.

Comparison of the expression of cCdx-B with other potential regulators of the posterior genetic program, cWnt-8c and Hoxa-7, was initiated to determine whether these genes have overlapping expression in the posterior cells. cWnt-8c is expressed in an overlapping domain with cCdx-B throughout the posterior region of the embryo below the level of Hensen's node at stage 5 (Fig. 2A,B). At stages 7 and 8, the posterior expression of both cWnt-8c and cCdx-B regresses with Hensen's node (Fig. 2D,E,G,H). Unlike cCdx-B, cWnt-8c transcripts are not detected in extra-embryonic cell lineages but are present in the neural plate rostral to the first somite (Fig. 2E,H). The posterior domain of cWnt-8c expression seems to have a graded distribution of transcripts along the medial-lateral axis with a higher level of expression in the axial region of the embryo. Expression of Hoxa-7 was examined as a representative of the posteriorly expressed Hox cluster genes. Unlike cCdx-B or cWnt-8c, Hoxa-7 expression is not apparent in newly gastrulated posterior cell progenitors at stage 5 (Fig. 2C). Hoxa-7 expression is first detected at stage 7 and is more restricted to the posterior axial region than cCdx-B or cWnt-8c (Fig. 2F,I). Thus, the overlapping expression domain of cCdx-B, cWnt-8c, and Hoxa-7 encompasses the posterior primitive streak.

Regulation of Posterior Patterning Genes by cCdx-B

Functional studies were initiated to examine hierarchical relationships between cCdx-B and other posterior patterning genes. A replication-deficient recombinant adenovirus (AdCdx-B) expressing cCdx-B from a cytomegalovirus (CMV) promoter was generated. Viral expression of full-length coding cCdx-B transcripts was verified by Northern blot analysis of RNA isolated from infected 10T1/2 cells (Fig. 3A). To confirm that the cCdx-B transcripts generated from the recombinant adenovirus were translated into functional protein, a transcriptional response assay was performed. Functional Cdx binding sites are present in the promoter region of human guanylyl cyclase C (Park et al., 2000). In HepG2 liver cells, a luciferase reporter plasmid containing 5′ sequences from human guanylyl cyclase C (-257 GCC/luc) is trans-activated by cotransfected Cdx-2. A similar trans-activation assay was used to confirm cCdx-B expression from the recombinant adenovirus. HepG2 cells transfected with -257 GCC/luc were infected with either AdCdx-B or Adβgal control adenovirus. AdCdx-B infection resulted in a fivefold induction of the -257 GCC/luc reporter plasmid expression in comparison to pGL3 basic backbone plasmid (Fig. 3B). Infection with Adβgal did not increase -257 GCC/luc expression, confirming the trans-activation by the cCdx-B adenovirus was a result of the cCdx-B protein. These experiments demonstrate that AdCdx-B produces functional protein capable of activating transcription of Cdx target genes.

Figure 3.

Recombinant adenovirus AdCdx-B generates functional cCdx-B protein. A: Expression of cCdx-B transcripts from recombinant adenovirus was detected by Northern blot analysis of 10T1/2 cells infected with two preparations of AdCdx-B adenovirus. No endogenous cCdx-B expression was detected in uninfected 10T1/2 cells. The two recombinant adenoviral preparations generate cCdx-B transcripts of the predicted size. B: Expression of a luciferase reporter plasmid containing 5′ sequences from human guanylyl cyclase C (-257 GCC/luc) was determined after infection with either AdCdx-B or Adβgal in HepG2 cells. The backbone plasmid pGL3 was included as a control for background luciferase activity. Relative luciferase activity was normalized to protein concentration and represents the average of three experiments (± SEM).

The cCdx-B recombinant adenovirus was used to determine whether cCdx-B expression is sufficient to activate a posterior profile of gene expression. At stage 5, anterior lateral mesoderm includes cardiomyogenic progenitors that express cNkx-2.5 and anterior Hox cluster genes (Schultheiss et al., 1995; Searcy and Yutzey, 1998; Ehrman and Yutzey, 1999). The posterior patterning genes cCdx-B, cWnt-8c, and Hoxa-7 are not expressed in these cells (Fig. 2A–C). The ability of cCdx-B to activate expression of cWnt-8c and posterior Hox cluster genes in anterior segments of the embryo was examined. Stage 5 anterior lateral mesoendoderm was explanted and infected with either AdCdx-B or Adβgal control virus. Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of cultured explants was performed by using specific primers designed to amplify cCdx-B, cWnt-8c, or Hoxa-7. Equivalent RNA input was confirmed by amplification of gapd. The expression of cNkx-2.5, an early cardiac lineage marker was also analyzed (Schultheiss et al., 1995).

The introduction of cCdx-B into the anterior lateral heart cells by adenoviral infection resulted in the induction of the posterior patterning genes, cWnt-8c and Hoxa-7 (Fig. 4, lane 1). In contrast, cWnt-8c and Hoxa-7 expression was not detected in uninfected or Adβgal infected anterior lateral mesoendoderm. Activation of additional posterior Hox cluster genes was examined with a degenerate PCR screen of Hox genes expressed in AdCdx-B infected explants. Hoxc-6 and Hoxc-8 expression were identified in anterior explants with ectopic expression of cCdx-B (data not shown). These posterior Hox genes were not detected in a previous screen of Hox genes expressed in anterior mesoendoderm of the cardiogenic region (Searcy and Yutzey, 1998). Cardiogenesis seemed to be unaffected by activation of posterior patterning genes in anterior mesoendoderm infected with cCdx-B adenovirus. Expression of cNkx-2.5 was consistent between AdCdx-B– and Adβgal–infected explants. AdCdx–B infected anterior cells also expressed myosin heavy chains as indicated by MF-20 reactivity and were observed to beat (data not shown). Thus, the anterior cardiogenic mesoderm cannot be respecified to another cell fate with the activation of posterior patterning genes at these stages. These data provide novel evidence in the chicken embryo that cCdx-B regulates Wnt and posterior Hox cluster gene expression.

Figure 4.

Posterior gene expression in anterior mesoendoderm explants is induced by ectopic cCdx-B expression. Expression of cCdx-B, cWnt-8c, Hoxa-7, cNkx-2.5, and gapd were monitored by reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of RNA collected from anterior lateral explants infected with AdCdx-B, control Adβgal, or no virus. Ectopic expression of cCdx-B leads to the induction of the posterior genes cWnt-8c and Hoxa-7 in anterior cardiogenic mesoendoderm. Transcript levels for the anterior cardiogenic marker cNkx-2.5 are unchanged. This panel is a representative result of four independent experiments.

DISCUSSION

The function of a chicken caudal-related gene, cCdx-B, in regulating a posterior program of gene expression was examined in gastrulation-stage avian embryos. Analysis of the overlapping expression of cCdx-B with posteriorly expressed members of the Hox and Wnt gene families supports a regulatory role for cCdx-B in posterior pattern formation. Ectopic expression of cCdx-B in the anterior mesoendoderm of the chicken embryo results in the activation of the posteriorly expressed genes, Hoxa-7 and cWnt-8c. These data suggest that the chicken homologue cCdx-B may have a role in posterior development through the regulation of posterior Hox patterning genes and Wnt signaling proteins. This hierarchy of patterning genes resembles the sequential expression of segment polarity genes, including wingless and homeotic genes initiated by a gradient of Caudal protein in Drosophila embryos (Mlodzik et al., 1990; Wu and Lengyel, 1998).

In this study, we present novel evidence that Cdx proteins regulate the expression of members of the vertebrate Wnt family of secreted signaling proteins. cWnt-8c is expressed in the posterior region of the chicken embryo in an overlapping domain with cCdx-B and is induced with cCdx-B expression in the anterior mesoendoderm. Wnt-8c has been implicated in posterior patterning of a variety of vertebrate embryos. Misexpression of cWnt-8c by transgenesis in the mice leads to truncations in the anterior region of the mouse embryo (Popperl et al., 1997). Similarly, overexpression of Xwnt-8 after the midblastula transition in Xenopus also results in the disruption of anterior structures (Christian and Moon, 1993). Studies performed in Drosophila embryos also indicate that Drosophila caudal is upstream of wg expression (Wu and Lengyel, 1998). Therefore, regulation of Wnts by Cdx proteins may represent a conserved regulatory interaction important in posterior cell fate specification. It is not known whether the regulation of Wnt expression by Cdx occurs by a direct or indirect mechanism.

Previous studies in mouse and Xenopus embryos have associated caudal-related genes with regulation of Hox posterior patterning genes. Anterior homeotic transformations of the vertebrae occur in mice with targeted mutation of either Cdx-1 or Cdx-2 (Subramanian et al., 1995; Chawengsaksophak et al., 1997). In Cdx-1 mutant mice, the expression of Hoxa-7, Hoxc-5, Hoxc-6, Hoxc-8, and Hoxd-3 was shifted more posterior (Subramanian et al., 1995). Regulation of posterior Hox gene expression by Cdx also was observed in Xenopus embryos. Overexpression of Xcad-3 in embryos results in the dramatic deletion of anterior head structures and in the up-regulation of Hoxa-7, Hoxc-6, Hoxb-7, and Hoxb-9 (Isaacs et al., 1998). In mice, transgenic expression of Cdx-4 in the neural tube leads to anterior induction of Hoxb-8 (Charite et al., 1998). In the present study, ectopic expression of chicken cCdx-B in anterior mesoendoderm results in the activation Hoxa-7, Hoxc-6, and Hoxc-8. In each experimental system, the Hox genes activated by Cdx homologues belong to Ultrabithorax and Abdominal-A paralogous groups (Subramanian et al., 1995; Pownall et al., 1996; Charite et al., 1998).

Hox genes in the Ultrabithorax and Abdominal-A paralogous groups are direct downstream targets of Cdx homeoproteins. Functional Cdx binding sites have been identified in mouse Hoxa-7, Hoxc-8, and Hoxb-8 regulatory sequences (Shashikant et al., 1995; Subramanian et al., 1995; Charite et al., 1998). Thus, activation of Hoxa-7 and Hoxc-8 in anterior mesoendoderm infected with cCdx-B adenovirus likely represents a direct regulatory interaction between cCdx-B protein and regulatory elements of these genes. By extension, the detection of Hoxc-6 expression in cCdx-B–infected tissue may represent an additional target of cCdx-B within the posterior Hox cluster genes. However cCdx-B is not sufficient to activate posterior Hox expression in all tissues where it is expressed, because the expression domain of cCdx-B is much broader than the expression domain of posterior Hox genes and extends into the extraembryonic tissue. Thus, it is likely that cCdx-B acts in conjunction with other regulatory proteins to activate Hox gene expression. These other transcription factors may include other classes of homeoproteins or retinoid responsive genes (Charite et al., 1998; Houle et al., 2000). However, because ectopic cCdx-B is sufficient to activate Hoxa-7, Hoxc-6, and Hoxc-8 expression in anterior lateral mesoendoderm, these accessory factors must be more broadly expressed in the embryo than the posterior patterning genes.

Interestingly, the cardiogenic marker cNkx-2.5 seemed to be unaffected by the ectopic expression of cCdx-B and the subsequent activation of posterior patterning genes in the cardiac progenitors. Recently Wnt-8c has been shown to inhibit cardiogenesis in Xenopus and chicken embryos (Schneider and Mercola, 2001; Marvin et al., 2001). However, the presence of the Wnt antagonist Crescent is sufficient to induce cardiogenesis in the presence of high Wnt activity. Therefore, Crescent likely prevents cWnt-8c from inhibiting cardiogenesis in AdCdx-B infected anterior explants. In Xenopus and Drosophila embryos, ectopic expression of Cdx proteins from the earliest stages of development leads to global changes in patterning and deletion of anterior structures (Mlodzik et al., 1990; Isaacs et al., 1998). In the chick, manipulation of cCdx-B expression in anterior lateral mesoendoderm did not result in inhibition of cardiogenesis or altered anteroposterior cell fate determination. This finding is likely due to the introduction of cCdx-B expression in cells that have already gastrulated and been influenced by patterning genes. In addition, detectable levels of protein expression from the recombinant adenovirus are not observed for several hours after infection, which further delays the onset of cCdx-B activity. At this relatively late stage after axis determination, the anterior mesoendoderm still has the potential to activate a spectrum of posterior patterning genes. However, cardiogenic cell fates cannot be respecified in the anterior lateral plate mesoderm, even in the presence of the posterior regulatory cascade initiated by cCdx-B.

Caudal/Cdx homologues are among the critical early determinants of anteroposterior axis determination and patterning in a broad spectrum of invertebrate and vertebrate embryos. In Drosophila, maternally expressed caudal is present as a posterior to anterior gradient in the syncytial blastoderm (Mlodzik et al., 1985). Altered caudal expression leads to misexpression of segment polarity genes such as fushi tarazu and wingless (Mlodzik et al., 1990; Wu and Lengyel, 1998). However, Caudal has not been shown to directly regulate homeotic gene expression in flies. Cdx homologues in vertebrate embryos seem to function somewhat later in development. In Xenopus, Xcad-3 is an immediate early target of FGF signaling during gastrulation (Isaacs et al., 1998). Among the downstream targets of Cdx in Xenopus, mouse and chicken embryos are the posterior Hox cluster genes (Subramanian et al., 1995; Pownall et al., 1996; Charite et al., 1998; this study). In addition, we have found that ectopic anterior expression of cCdx-B induces cWnt-8c expression. The activation of Wnt expression by Cdx represents a conservation of wg regulation by Caudal observed in Drosophila embryos. It remains to be seen whether this regulatory mechanism is widespread in vertebrate embryos and whether it represents a direct interaction between Cdx and Wnt gene regulatory elements. Together, these studies support an important role for Cdx proteins as master regulators of early embryonic patterning through the activation of a posterior genetic program that includes Wnt-8c and posterior Hox cluster genes.

EXPERIMENTAL PROCEDURES

Chicken cCdx-B cDNA Isolation

The homeodomain of cCdx-B was isolated from a degenerate PCR screen of a stage 8–11 chick heart cDNA pScreen library by using primers specific to the highly conserved first and third helices of the homeobox DNA binding domain (Searcy and Yutzey, 1998). A 1.4-kb clone for cCdx-B was obtained by screening a λExlox stage 11 chick embryo cardiac enriched library by using the homeodomain sequence for cCdx-B as a probe (Wei et al., 1996). The 1.4 kb clone contains the full length coding region for cCdx-B. PCR and library screening conditions are as described in (Searcy and Yutzey, 1998). The 1.4 kb cCdx-B clone was verified by sequencing and was submitted to GenBank with Accession number AF353624.

Embryo Cultures

Fertilized White Leghorn chicken eggs (Spafas Inc., Roanoke, IL) were incubated at 38°C under high humidity. The embryos were staged as described by Hamburger and Hamilton (1951). Before adenovirus infection, anterior lateral mesoendoderm was isolated from stage 5 chicken embryos with pulled glass capillary tubes (Yutzey et al., 1995). The explanted mesoendoderm was manually dissociated, placed in 8-well Permanox chamber slides coated with 0.01% collagen and cultured in high glucose DMEM supplemented with 10% fetal bovine serum, 1% L-glutamine, penicillin (100 U/mL), and streptomycin (100 μg/mL) (Gibco-BRL). Explants were incubated at 37°C with high humidity and 5% CO2. After the initial 3 hr, the explants were cultured in infection media consisting of DMEM supplemented with 2% fetal bovine serum, 1% L-glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL) and containing 1.2 × 109 PFU/mL of AdCdx-B, Adβgal, or no adenovirus. Anterior mesoendoderm derived from eight embryos was treated for each experimental group. After 24 hr, an equal amount of DMEM supplemented with 10% fetal bovine serum, 1% L-glutamine, penicillin (100 U/ml), and streptomycin (100 μg/ml) was added to the infection media. Forty-eight hours after explantation, RNA was isolated for RT-PCR analysis (Searcy and Yutzey, 1998).

In Situ Hybridization

Embryos were removed from the eggs at stage 4–10 on paper rings, rinsed in sterile PBS, dissected away from extraembryonic membranes, and fixed in 4% paraformaldehyde in PBS at 4°C overnight. Fixed embryos were dehydrated through a graded methanol series in PTW (PBS, 0.01% Tween-20) and stored in 100% methanol at −20°C. Digoxigenin UTP-labelled antisense RNA probes were generated for cCdx-B, cWnt-8c, and Hoxa-7. The pGEM-7Zf plasmid containing 2.6-kb cCdx-B clone was linearized with Xba1, and the antisense probe was synthesized by using Sp6 polymerase. cWnt-8c was amplified from stage 7–8 posterior chicken RT by using the 5′ primer- 5′ GCT CCGCTG GTG TTA TG 3′ and the 3′ primer- 5′ CCT CCT CTC CTC CAC TTT 3′ based on the reported sequence (Hume and Dodd, 1993). cWnt-8c probe was synthesized with Sp6 polymerase from a SphI linearized pGEM-T plasmid (Promega) containing cWnt-8c cDNA (278–919 bp). Hoxa-7 was amplified from stage 7–8 posterior chicken RT by using a degenerate PCR screening strategy. The 5′ degenerate primer was designed to conserved 5′ coding sequences for mouse Hoxa-7 and Hoxb-7 (Parikh et al., 1995). The 3′ primer was designed to the mouse Hoxb-7 homeodomain sequence. An approximately 550-bp sequence was amplified and inserted into pBluescript T-vector (Marchuk et al., 1991). After sequencing and GenBank analysis, the clone was identified as chicken Hoxa-7. Hoxa-7 antisense probe was synthesized with T7 by using a BamHI linearized pBluescript plasmid containing Hoxa-7 cDNA sequence. In situ hybridizations were performed essentially as described by Wilkinson (1993) with modifications (Ehrman and Yutzey, 1999).

Recombinant Adenovirus Construction and Infection

The 1.4-kb clone containing the full-length coding region for cCdx-B was inserted into the HindIII/ EcoRI site of the pACCMVpLpA plasmid (Gomez-Foix et al., 1992). The preparation of recombinant adenovirus was performed as previously described (Gomez-Foix et al., 1992). The supernatant from the initial adenoviral production was amplified into a stock and titered by using a plaque forming assay (Sallenave et al., 1998). Adβgal was kindly provided by Dr. Mark Sussman. Northern analysis was performed to confirm the size of the transcript generated from AdCdx-B. 10T1/2 cells were infected with the supernatant from the HEK 293 cells for 24 hr. 10T1/2 cells were maintained in high glucose DMEM supplemented with 10% fetal bovine serum, 1% L-glutamine, penicillin (100 U/mL), and streptomycin (100 μg/mL). After infection, total RNA was isolated by using guanidinium thiocyanate/phenol/chloroform extraction and isopropanol precipitation (Chomczynski and Sacchi, 1987). Northern analysis was performed as previously described (Yutzey et al., 1994).

A transcriptional response assay was performed to confirm that functional cCdx-B protein was generated from AdCdx-B. A luciferase reporter plasmid containing 5′ sequences from human guanylyl cyclase C (-257 GCC/luc) (Mann et al., 1996), is transactivated by mCdx-2 in HepG2 liver cells (Park et al., 2000). The -257 GCC/luc reporter plasmid was kindly provided by Dr. Elizabeth Mann at the University of Cincinnati. HepG2 cells were cultured in MEM (Gibco-BRL) supplemented with 10% fetal bovine serum, penicillin (100 U/ml), and streptomycin (100 μg/ml). Cells were plated in 60-mm dishes and transiently transfected 24 hr later with either the reporter plasmid, -257 GCC/luc, or the empty backbone plasmid pGL3 (Promega) by using Superfect transfection reagent (Qiagen). Cells were infected 24 hr after transfection with 2.3 × 108 PFU/mL of AdCdx-B or Adβgal for 90 min and incubated overnight. Luciferase activity was measured by using the luciferase assay system (Tropix) in a Monolight 2010 luminometer. Protein concentration of each sample was determined by a protein assay (Bio-Rad) and used to normalize luciferase activity.

RT-PCR

RNA was isolated from infected anterior explants as previously described (Searcy and Yutzey, 1998). First-strand cDNA was synthesized from parallel infected and control cultures with oligo(dT) priming and Superscript reverse transcriptase (Gibco-BRL). Reverse transcription reactions included RNA derived from infected or control explants from eight embryo equivalents and each experiment was repeated four times with similar results. Previously published primer sets were used for the amplification of gapd (Gannon and Bader, 1995) and cNkx-2.5 (Schultheiss et al., 1995). Additional primers and amplified sequences were as follows: cCdx-B 5′ CAC TGC AAC AGA TAC ATT AC 3′ and 5′ CAC GAA CTC CCT GAG ATT TC 3′ (473 bp); cWnt-8c, 5′ GCT CCG CTG GTG TTA TG 3′ and 5′ CCT CCT CTC CTC CAC TTT 3′ (640 bp); Hoxa-7, 5′ TGC GGA GCC TAC TTC TT 3′ and 5′ GGT TTG GTT AGG TCA TTG 3′ (224 bp). The number of cycles for each primer set was chosen to be well within the linear range as determined by PhosphorImager analysis of PCR products from sequential cycles. Polymerase chain reactions (PCR) contained PCR buffer with 15 mM MgCl2 (Roche Molecular Biochemical), 75 μM each dNTP, 1 μCi [32P]α-dCTP (NEN), 0.5 pM each primer, 1.25 Unit Taq polymerase in a 50-μl reaction volume. For PCR amplification of gapd, cCdx-B, cWnt-8c, and cNkx-2.5, 1 μl from a 20 μl RT reaction was used as template. Five microliters from a 20-μl RT reaction was used as template for PCR amplification of HoxA7. For gapd amplification, reactions were subjected to 18 cycles of 94°C for 30 sec, 50°C for 30 sec, 72°C for 90 sec. For Hoxa-7 and cWnt-8c amplification, reactions were subjected to 27 cycles of 94°C for 30 sec, 58°C for 30 sec, 72°C for 90 sec. For cNkx-2.5 and cCdx-B amplification, reactions were subjected to 22 cycles of 94°C for 30 sec, 50°C for 30 sec, 72°C for 90 sec. A total of 2 μl of the gapd PCR reaction and 5 μl of the PCR reaction for cCdx-B, cWnt-8c, and cNkx-2.5 was loaded on a 5% acrylamide gel. The dried gel was imaged on a Molecular Dynamics PhosphorImager and quantitated by using ImageQuant software.

Additional homeodomain sequences were identified by a degenerate PCR screen by using primers specific to the conserved first and third helix of the homeobox DNA binding domain (Searcy and Yutzey, 1998). Homeodomain sequences (180bp) were amplified from cDNA derived from AdCdx-B infected anterior explants (Searcy and Yutzey, 1998). After ligation of the PCR product into a pBluescript T-vector, 20 homeodomain clones were sequenced and classified through a blast search of the GenBank database.

Acknowledgements

We thank Mark Sussman and Elizabeth Mann for reagents and Bruce Trapnell for guidance in adenovirus preparation. Technical support was provided by Robin Searcy-Schrick. K.E.Y. was supported by the Ohio Valley Affiliate of the American Heart Association and the NIH. L.A.E. was supported by the NIH and the Albert J. Ryan Foundation.

Ancillary