β-catenin signaling marks the prospective site of primitive streak formation in the mouse embryo

Authors

  • Othman A. Mohamed,

    1. Department of Biology, McGill University Health Center, Royal Victoria Hospital, Montreal, QC, Canada
    Search for more papers by this author
  • Hugh J. Clarke,

    1. Department of Biology, McGill University Health Center, Royal Victoria Hospital, Montreal, QC, Canada
    2. Department of Obstetrics and Gynecology, Division of Experimental Medicine, McGill University Health Center, Royal Victoria Hospital, Montreal, QC, Canada
    Search for more papers by this author
  • Daniel Dufort

    Corresponding author
    1. Department of Obstetrics and Gynecology, Division of Experimental Medicine, McGill University Health Center, Royal Victoria Hospital, Montreal, QC, Canada
    • Royal Victoria Hospital, Department of Obstetrics and Gynecology, 687 Pine Avenue West, Room F3-24, Montreal, QC, Canada H3A 1A1
    Search for more papers by this author

Abstract

β-Catenin signaling has been shown to be involved in triggering axis formation in several organisms, including Xenopus and zebrafish. Genetic analysis has demonstrated that the Wnt/β-catenin signaling pathway is also involved in axis formation in the mouse, since a targeted deletion of β-catenin results in embryos that have a block in anterior–posterior axis formation, fail to initiate gastrulation, and do not form mesoderm. However, because β-catenin is ubiquitously expressed, the precise time and cell types in which this signaling pathway is active during early embryonic development remain unknown. Thus, to better understand the role of the Wnt/β-catenin signaling pathway in axis formation and mesoderm specification, we have examined both the distribution and signaling activity of β-catenin during early embryonic development in the mouse. We show that the N-terminally nonphosphorylated form of β-catenin as well as β-catenin signaling is first detectable in the extraembryonic visceral endoderm in day 5.5 embryos. Before the initiation of gastrulation at day 6.0, β-catenin signaling is asymmetrically distributed within the epiblast and is localized to a small group of cells adjacent to the embryonic–extraembryonic junction. At day 6.5 and onward, β-catenin signaling was detected in the primitive streak and mature node. Thus, β-catenin signaling precedes primitive streak formation and is present in epiblast cells that will go on to form the primitive streak. These results support a critical role for the Wnt/β-catenin pathway in specifying cells to form the primitive streak and node in the mammalian embryo as well as identify a novel domain of Wnt/β-catenin signaling activity during early embryogenesis. Developmental Dynamics 231:416–424, 2004. © 2004 Wiley-Liss, Inc.

INTRODUCTION

β-Catenin is a multifunctional protein that plays at least two roles in cells. One is in cell adhesion, where its role is mediated through interactions with cadherins (McCrea and Gumbiner, 1991; McCrea et al., 1991; Gumbiner and McCrea, 1993). The second is as a downstream effector of the canonical Wnt signaling pathway (reviewed in Cadigan and Nusse, 1997). In the absence of Wnt signaling, β-catenin associates with a large cytoplasmic protein complex containing glycogen synthase kinase 3β (GSK3β), the tumor suppressor adenomatous polyposis coli (APC), and axin (Behrens et al., 1998; Kishida et al., 1998). In this complex, GSK3β phosphorylates sites within the N-terminal domain of β-catenin and subsequently targets it for degradation by the ubiquitin-dependent proteasome pathway (Hart et al., 1999; Kitagawa et al., 1999; Latres, 1999). Binding of Wnt ligands to their Frizzled receptors and the low-density lipoprotein receptor-related protein (LRP) coreceptors, LRP-5 and LRP-6, results in the activation of Dishevelled protein, which inhibits GSK-3β activity, thus leading to the stabilization and accumulation of the hypophosphorylated form of β-catenin (Yost et al., 1996; Willert and Nusse, 1998; Pinson et al., 2000). β-Catenin then translocates to the nucleus, where it interacts with members of the TCF/Lef family of transcription factors and activates downstream target genes (reviewed in Barker and Clevers, 2000; Sharpe et al., 2001). In the absence of β-catenin, TCF/Lef proteins bind to their target DNA and act as transcriptional repressors (Brannon et al., 1997; Riese et al., 1997; Merrill et al., 2003) by interacting with Groucho-related corepressors (Roose et al., 1998; Brantjes et al., 2001, 2002).

In both Xenopus and zebrafish, β-catenin signaling plays an important role in specifying the embryonic axis (reviewed in Moon and Kimelman, 1998; Sokol, 1999; Schier, 2001). In Xenopus, the dorsal–ventral axis is specified upon fertilization, when sperm entry initiates a cortical rotation relative to the cytoplasm of the fertilized egg leading to the stabilization and nuclear accumulation of β-catenin in the prospective dorsal side of the embryo. Upon embryonic genome activation, β-catenin interacts with TCF/Lef proteins to activate the expression of genes, such as Siamois and Twin, that participate in the formation of the Organizer, which is essential for axis specification (Lemaire et al., 1995; Carnac et al., 1996; Brannon et al., 1997). Overexpression of several components of this pathway leads to axis duplication (reviewed in Moon and Kimelman, 1998). Similarly, in zebrafish, nuclear β-catenin accumulates in nuclei on the dorsal side of the blastulae (Schneider et al., 1996), and overexpression of β-catenin in early zebrafish embryos induces the formation of a complete secondary axis (Kelly et al., 1995).

In the mouse, the first morphological sign of axis formation occurs at day 6.5 of embryogenesis with the formation of the primitive streak. The region where the primitive streak forms becomes the posterior end of the embryo (reviewed in Beddington and Robertson, 1999). Several lines of evidence demonstrate that, as in Xenopus and zebrafish, the Wnt/β-catenin signaling pathway plays an essential role in axis specification in the mouse. Most dramatically, a targeted deletion of β-catenin produces embryos that have a block in anterior–posterior axis formation at embryonic day (E) 6.0, fail to initiate gastrulation and do not form mesoderm (Haegel et al., 1995; Huelsken et al., 2000). Moreover, overexpression of Wnt8c (Popperl et al., 1997) or, conversely, deletion of genes encoding proteins involved in the degradation of cytoplasmic β-catenin, including Axin, APC, and the LIM domain-binding protein 1, all produce embryos with at least partially duplicated axes (Zeng et al., 1997; Ishikawa et al., 2003; Merrill et al., 2003; Mukhopadhyay et al., 2003). β-Catenin is required at the time of axis formation for correct positioning of the anterior visceral endoderm. Although some anterior visceral endoderm genes are expressed in embryos lacking β-catenin, the expressing cells remain at the distal tip of the embryo and fail to migrate to the future anterior region of the embryo (Huelsken et al., 2000).

Although these results clearly indicate an obligatory role for Wnt/β-catenin signaling during formation of the primitive streak and gastrulation, the precise time and cell types in which the signaling pathway is active remain unknown. Furthermore, some of the phenotypic effects observed in β-catenin mutant embryos may reflect its role in cell adhesion. Thus, to better understand the role of Wnt/β-catenin signaling in axis formation and mesoderm specification, it is essential to identify when and where this pathway is active during these processes. Although several transgenic lines that can report Wnt/β-catenin signaling have been generated (DasGupta and Fuchs, 1999; Maretto et al., 2003), the domains of active Wnt/β-catenin signaling during the pre- and early postimplantation period have not been reported. We have examined both the distribution and signaling activity of β-catenin during early embryonic development. As recent evidence indicates that N-terminally nonphosphorylated β-catenin interacts with TCF/Lef (van Noort et al., 2002), we used immunohistochemistry to examine the distribution of this nonphosphorylated isoform of β-catenin. We have also used a transgenic reporter approach to identify where β-catenin cooperates with TCF/Lef to activate gene expression during early embryonic development.

RESULTS

Nonphosphorylated β-Catenin Accumulates in Extraembryonic Visceral Endoderm and in Epiblast Cells Located Where the Primitive Streak Will Form

To investigate in which cell types of the early mouse embryo β-catenin was N-terminally nonphosphorylated and, therefore, could be translocated to the nucleus to regulate gene expression, we stained embryos at different stages of development using two different antibodies. The first is an antibody that recognizes β-catenin independently of its phosphorylation state, which we refer to as total β-catenin. The second is an antibody that recognizes only the nonphosphorylated S37 and T41 isoform of the protein, which we refer to as nonphosphorylated (van Noort et al., 2002). This nonphosphorylated isoform of β-catenin has been shown to accumulate in response to Wnt signaling (van Noort et al., 2002). Preimplantation embryos from the two-cell to blastocyst stage stained for total β-catenin revealed strong immunoreactivity localized at the cell surfaces, but no nuclear staining was detected, consistent with previous reports (Fig. 1a–e; Pauken and Capco, 1999; Rogers and Varmuza, 2000).Furthermore, nonphosphorylated β-catenin was not detectable in these embryos (Fig. 1f–j). These results suggest that the Wnt/β-catenin signaling pathway is not active in embryos up to the blastocyst stage.

Figure 1.

Distribution of total and nonphosphorylated β-catenin in preimplantation stage embryos. a–j: Immunohistochemistry was performed using either an antibody that detects all forms of β-catenin (a–e) or an antibody that recognizes only the nonphosphorylated form of β-catenin (f–j). Immunohistochemistry was performed on two-cell stage (a,f), four-cell stage (b,g), eight-cell stage (c,h), morula stage (d,i), and blastocyst stage (e,j).

The primitive streak first develops at the future posterior region of the embryo, near the junction of the embryonic and extraembryonic ectoderm and is morphologically detectable beginning at day 6.5 (Beddington and Robertson, 1999; Lu et al., 2001). Accordingly, we analyzed the distribution of β-catenin in early postimplantation stage embryos, ranging in age from 5.5 to 6.75 days. At each stage examined, total β-catenin was detected throughout the embryos and showed no evidence of asymmetrical distribution (Fig. 2a,c). In contrast to the apparently uniform distribution of total β-catenin, the nonphosphorylated form of the protein showed a tissue-restricted and dynamic pattern of expression during this period of embryogenesis. In 5.5-day embryos, nonphosphorylated β-catenin was readily detected in the extraembryonic visceral endoderm. However, no immunostaining could be detected in other tissues, notably including those of the embryo proper (Fig. 2b). At day 6.0, although staining remained largely restricted to the extraembryonic visceral endoderm, a small group of immunoreactive cells could be identified that projected down toward the embryo proper (Fig. 2d, red arrow). Moreover, by day 6.75, this group of cells had expanded both laterally and along the proximal–distal axis (Fig. 2e, red arrow).

Figure 2.

Distribution of total and nonphosphorylated β-catenin in postimplantation stage embryos. a,c: Immunohistochemistry using an antibody that detects all forms of β-catenin was performed on day 5.5 (a) and day 6.0 (c) embryos. b,d,e: Immunohistochemistry using an antibody that specifically detects only nonphosphorylated β-catenin was performed on day 5.5 embryos (b), day 6.0 embryos (d), and day 6.75 embryos (e). f,g: Confocal microscopy of day 6.25 embryos showing the surface of an embryo (f) or a longitudinal section through a different embryo (g). β-Catenin is shown in green, 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) -stained nuclei are shown in red. Red arrows indicate the presence of nonphosphorylated β-catenin in epiblast cells. eve, extraembryonic visceral endoderm; ee, extraembryonic ectoderm; epc, ectoplacental cone.

To examine this transition in more detail, 6.25-day embryos were stained for nonphosphorylated β-catenin and examined using confocal microscopy. As illustrated in Figure 2f,g, nonphosphorylated β-catenin could be detected in extraembryonic visceral endoderm cells. In addition, no immunoreactivity could be detected in the extraembryonic ectoderm or in the majority of the embryonic ectoderm cells. However, a small group of embryonic ectodermal cells adjacent to the embryonic–extraembryonic junction were clearly stained (Fig. 2g, longest arrow). Furthermore, accumulation of nonphosphorylated β-catenin was also detected in the visceral endoderm overlying this region of the epiblast. Although nonphosphorylated β-catenin was detected in the epiblast starting at E6.25, staining in the visceral endoderm could be detected slightly earlier, at E5.75–E6.0 (data not shown). These results indicate that, several hours before the formation of a morphologically identifiable primitive streak, nonphosphorylated β-catenin accumulates in a subset of cells within the region where this structure will form.

β-Catenin Responsive Reporter Gene Is First Expressed in Extraembryonic Visceral Endoderm Cells and Then in Cells Located Where the Primitive Streak Will Form

β-Catenin activates transcription of target genes through association with members of the TCF/Lef protein family, which bind to DNA at a well-characterized consensus sequence (Korinek et al., 1997; van Beest et al., 2000). The presence of β-catenin converts the complex, which in its absence represses transcription, to a transcriptional activator. To map sites in the embryo of transcription regulated by β-catenin, we generated lines of transgenic mice that carried a lacZ reporter gene linked to six copies of the TCF/Lef binding site together with a minimal promoter element (Rossant et al., 1991). Six independent TCF/Lef-lacZ transgenic lines that were tested for expression all gave similar expression patterns, as described below.

To verify that the lacZ transgene could be activated by β-catenin, we isolated embryonic fibroblast cells from hemizygous embryos derived from one transgenic line. β-Galactosidase activity was measured after transfection with different constructs and normalized against the activity measured in cells transfected with a CMV-lacZ under the control of the strongly active cytomegalovirus promoter (Fig. 3, column 1). Very low activity was detected in mock-transfected transgenic fibroblasts (Fig. 3, column 2) or in those transfected with a construct encoding Lef-1 (Fig. 3, column 3). In contrast, high β-galactosidase activity was measured after transfection of a construct encoding a mutant form of β-catenin in which serine-37 has been replaced by alanine (S37A; Zorn et al., 1999) rendering the protein nonphosphorylatable at this site and, therefore, constitutively active (Fig. 3, column 4). Cotransfection of Lef-1 and S37A did not further increase β-galactosidase activity (Fig. 3, column 5). These results confirmed that in these transgenic lines the nonphosphorylatable form of β-catenin induced endogenous lacZ expression, strongly suggesting that the activity of the TCF/Lef-lacZ transgene accurately reflects endogenous activity of nuclear β-catenin.

Figure 3.

β-Catenin–dependent activation of the endogenous TCF/Lef-LacZ transgene in transgenic mouse embryonic fibroblasts. Mouse embryonic fibroblast cells were isolated from hemizygous day 12.5 transgenic embryos and transfected using the indicated constructs. Transfections were performed in triplicate, and the average value is indicated. The lacZ gene driven by the CMV promoter was used as a positive control, whereas the empty pCDNA vector (mock transfection) was used as a negative control. S37A, constitutively active form of β-catenin cloned in pCDNA; Lef-1, Lef-1 cDNA cloned in pCDNA.

Next, hemizygous transgenic embryos were stained for β-galactosidase activity at different stages of development. First, we examined whether β-galactosidase activity could be detected in tissues where Wnt expression, which lies upstream of β-catenin activity, has been reported previously. Activity was detected in the midbrain–hindbrain region of day 8.25 embryos (Fig. 4a), and at day 8.75 and 9.5, high levels of activity were detected in the dorsal region of the neural tube and somites (Fig. 4b–d). This activity is consistent with expression of Wnt1 and Wnt3a in the roof plate and their essential role in neuronal specification of the dorsal neural tube (Muroyama et al., 2002). β-Galactosidase activity was also present in the apical ectodermal ridge of developing limb buds (Fig. 4e,f). This finding is consistent with the recent demonstration that ectodermal Wnt3/β-catenin signaling is required for both the establishment and maintenance of this structure (Barrow et al., 2003).

Figure 4.

Expression of the endogenous TCF/Lef-LacZ transgene in day 8.25 to day 10.5 embryos. Transgenic embryos were isolated at the indicated stages and stained for β-galactosidase activity. a–d: Day 8.25 (a), day 8.75 (b), day 9.5 (c), section through a day 9.5 embryo as indicated in c (d). e,f: Day 10.5 (f) section through the limb bud (lb) of a day 10.5 day embryo as indicated in e, showing staining in the apical ectodermal ridge (aer). g: Day 9.0 heterozygous (+/vt) and homozygous(vt/vt) vestigial tail, a Wnt3a hypomorphic allele stained for lacZ activity. mh, midbrain–hindbrain region; s, somite; dnt, dorsal neural tube; c, caudal region.

To further test whether the TCF/Lef-lacZ transgenic line reflects endogenous Wnt signaling, we placed the transgene onto a homozygous vestigial tail (vt) mutant background, which is an hypomorphic allele of Wnt3a (Greco et al., 1996). Wnt3a is normally expressed in the dorsal region of the neural tube of day 9.0 embryos (Takada et al., 1994). Thus, if the transgene responds to Wnt signaling, reduced β-galactosidase activity on the vt background would be expected in the dorsal region of the neural tube but not in more anterior regions of the embryo where other Wnt family members other than Wnt3a are expressed. As shown in Figure 4g, significantly less β-galactosidase activity was detectable in the dorsal neural tube of day 9.0 vt/vt embryos as compared with +/vt. In more anterior regions, where Wnt3a is not expressed, β-galactosidase activity was equal between embryos of the two genotypes. Taken together, these results demonstrate that the TCF/Lef lacZ transgene is expressed when the Wnt/β-catenin signaling pathway is activated.

Based on these results, we then stained preimplantation and postimplantation transgenic embryos for β-galactosidase activity. No activity was detected at any preimplantation stage (data not shown), consistent with the absence of detectable nonphosphorylated β-catenin in these embryos. β-Galactosidase activity was first detected at day 5.5, where it was restricted to the extraembryonic visceral endoderm cells (Fig. 5a, and data not shown). This activity was transient and at day 6.0 had been lost from these cells (Fig. 5b). At this stage of embryogenesis, however, β-galactosidase activity was detected in a subset of cells located within the embryo proper near the embryonic–extraembryonic junction. Moreover, the intensity of staining became considerably stronger as development progressed to day 6.25, suggesting that more cells were expressing the transgene (Fig. 5c).

Figure 5.

Expression of the endogenous TCF/Lef-LacZ transgene in gastrulation stage embryos. Transgenic embryos were isolated at the indicated stages and stained for β-galactosidase activity. a–d: Day 5.5 (a), day 6.0 (b), day 6.25 (c), section through a day 6.25 embryo as illustrated in c (d), showing β-galactosidase activity in the epiblast. e–h: Day 6.5 (e), day 6.75 (f), day 7.5 (g), day 7.75 (h). i: Section through a day 7.75 embryo as illustrated in h. eve, extraembryonic visceral endoderm; ve, visceral endoderm; ps, primitive streak; e, ectoderm; mw, mesodermal wing.

To identify the cells expressing LacZ, 6.25-day embryos were stained for β-galactosidase activity then paraffin-embedded and sectioned. As shown in Figure 5d, β-galactosidase activity was present in a portion of the epiblast. Importantly, the morphology of these cells indicates that they were ectodermal (epithelial) rather than mesenchymal in nature, and the absence of mesenchymal cells confirms that gastrulation had not yet begun. Unexpectedly, no β-galactosidase activity was detected in the overlying visceral endoderm, although nonphosphorylated β-catenin was present in these cells (Fig. 2f,g).

β-Galactosidase activity remained present in the posterior epiblast at day 6.5 (Fig. 5e), and no activity was detected in the posterior visceral endoderm at this stage. By day 6.75, the domain of β-galactosidase in the epiblast began to expand toward the distal tip of the embryo (Fig. 5f). Between days 7.5 and 7.75, which correspond to the headfold stage, β-galactosidase activity was restricted to the region of the primitive streak and node (Fig. 5g,h). Sectioning of headfold stage embryos demonstrated staining in the posterior embryonic ectoderm, the nascent mesoderm and mesodermal wings, as well as the posterior visceral endoderm (Fig. 5i).

Thus, before morphological manifestation of the primitive streak, activity of a β-catenin–responsive transgene becomes detectable first in extraembryonic visceral endoderm and subsequently in a subset of epiblast cells near the border with the extraembryonic tissue. When the primitive streak is present, β-catenin/TCF transcriptional activity is found in cells of this structure as well as the overlying visceral endoderm.

DISCUSSION

Previous work using targeted gene deletion established an essential role for β-catenin in the formation of the primitive streak and embryonic mesoderm as well as specification of the anterior visceral endoderm during early postimplantation embryogenesis in the mouse. However, this approach could not identify the specific cell types in which β-catenin activity was required. Moreover, as β-catenin has an extracellular function in mediating cell adhesion in addition to its nuclear function in regulating transcription from TCF/Lef-binding sites, the nature of the β-catenin requirement was not established. To identify when and where β-catenin exerts its nuclear function during this period of embryogenesis, we mapped the distribution of the nonphosphorylated isoform and the activity of a β-catenin–regulated transgene in pre- and early postimplantation embryos. We found that neither nonphosphorylated β-catenin nor β-catenin signaling through TCF/Lef were detectable in preimplantation stage embryos. Near the time of primitive streak formation, we identified two regions of β-catenin activity: the extraembryonic visceral endoderm and the future posterior portion of the embryonic ectoderm.

Nonphosphorylated β-catenin was first detected in the extraembryonic visceral endoderm in day 5.5 embryos. Moreover, β-catenin transcriptional activity was also detected in this tissue at day 5.5. This finding provides the first evidence that β-catenin signaling is active in the extraembryonic visceral endoderm of early postimplantation embryos. Studies using chimeric embryos have suggested that β-catenin may not be required in the extraembryonic tissues during early postimplantation development (Huelsken et al., 2000). However, a detailed marker analysis of these chimeric embryos has not been done to determine whether β-catenin signaling in the extraembryonic tissues is required for early embryonic development. It is also possible that genes activated through β-catenin at day 5.5 are required in later development or that other gene products can compensate for the absence of β-catenin. Our results also indicate that β-catenin signaling is only transiently activated in the extraembryonic visceral endoderm. By day 6.0, although nonphosphorylated β-catenin was still present, β-galactosidase activity was no longer detectable. Transient expression may be due to the loss of other factors required for β-catenin–regulated gene expression. Indeed many studies have demonstrated that the Wnt signaling pathway is subject to numerous levels of regulatory control, both at the extracellular and intracellular levels (Yamaguchi, 2001; Brantjes et al., 2002).

Nonphosphorylated β-catenin became asymmetrically distributed within the epiblast at day 6.25, before the formation of the primitive streak. At this stage, nonphosphorylated β-catenin was detected in a small group of epiblast cells located adjacent to the embryonic–extraembryonic junction. As development proceeded, the presence of nonphosphorylated β-catenin in the epiblast expanded both laterally and along the proximal–distal axis in a manner similar to primitive streak elongation. Consistent with the presence of nonphosphorylated β-catenin in the epiblast, β-catenin transcriptional activity was detected in a small group of epiblast cells before the formation of the primitive streak. At day 6.5 and onward, β-catenin transcriptional activity was detected in the primitive streak and mature node. These results demonstrate that β-catenin signaling precedes primitive streak formation and is present in epiblast cells that will go on to form the primitive streak.

The factor most likely responsible for activating β-catenin signaling in the epiblast is Wnt3. Wnt3 is expressed before primitive streak formation in the proximal epiblast adjacent to the embryonic–extraembryonic junction and at the time of gastrulation in the posterior epiblast and associated visceral endoderm (Liu et al., 1999). Thus, in the epiblast there is a strong temporal and spatial correlation between Wnt3 expression and β-catenin signaling activity. Moreover, Wnt3 is known to signal through β-catenin (Shimizu et al., 1997) and targeted deletion of Wnt3 results in embryos that fail to form a primitive streak, similar to β-catenin mutant embryos (Liu et al., 1999). However, despite the expression of Wnt3 in the posterior visceral endoderm, nuclear β-catenin signaling is not detected in these cells at this time. Although unphosphorylated β-catenin accumulates in the posterior visceral endoderm, it is possible that signaling inhibitors functioning downstream of β-catenin are present in this region or that β-catenin functions through a pathway other than by interaction with TCF/Lef. Although it is possible that the transgene is unable to report Wnt/β-catenin signaling in the visceral endoderm, this possibility seems unlikely in view of the finding that activity of the transgene is detected in the extraembryonic visceral endoderm cells at earlier stage as well as in the posterior visceral endoderm at the late primitive streak stage. Although wnt3 is initially expressed in the proximal epiblast, we have neither detected accumulation of nonphosphorylated β-catenin or β-catenin signaling in this region of the embryo. These results suggest that, although Wnt3 is initially expressed in this region, it does not activate the Wnt/β-catenin signaling pathway. Because we have detected both the accumulation of nonphosphorylated β-catenin and β-catenin signaling only in posterior epiblast cells before the initiation of gastrulation, our results suggest that Wnt3 signaling is only active in this region of the embryo. Recently, a similar TCF/Lef-LacZ reporter line has been reported (Maretto et al., 2003). Although a detailed analysis of the expression of the transgene in prestreak embryos has not been done, whole-mount in situ hybridization using LacZ as a probe support our observation that activation of the Wnt/β-catenin pathway is restricted to the posterior region of the embryo and is absent from epiblast cells located at the embryonic–extraembryonic junction. At the early primitive streak stage, we have shown that β-catenin signaling is restricted to the primitive streak, whereas Maretto et al. have demonstrated by whole-mount staining that activity, aside from being present in the primitive streak, is also found in the proximal epiblast region. Upon overstaining of our embryos, we have occasionally observed similar staining in the proximal epiblast region. However, sectioning of these embryos demonstrated that staining was present in the nascent mesoderm and not in the epiblast. Thus, it is possible that the proximal epiblast staining reported by Maretto et al. (2003) may also be staining of the nascent mesoderm. Sectioning of these embryos will need to be performed to determine in which cell types β-catenin signaling is active.

β-catenin has been shown to be required at the time of axis formation for correct positioning of the anterior visceral endoderm. Although some anterior visceral endoderm genes are expressed in embryos lacking β-catenin, the expressing cells remain at the distal tip of the embryo and fail to migrate to the future anterior region of the embryo (Huelsken et al., 2000). However, in Wnt3 mutant embryos, the anterior visceral endoderm is correctly positioned suggesting that β-catenin signaling is either required in a domain other than the domain of Wnt3 expression or that this process does not require β-catenin signaling but rather the cell adhesion function of β-catenin. We have identified two domains where nonphosphorylated β-catenin accumulates at the time of axis specification. The first is in the extraembryonic visceral endoderm and the second in the posterior visceral endoderm. Whether β-catenin is required in these domains for correct positioning of the anterior visceral endoderm still needs to be determined.

We found that β-catenin transcriptional activity was maintained throughout the primitive streak at later stages of embryonic development. Several Wnt family members such as Wnt3a, Wnt2b, and Wnt5a are expressed in the primitive streak (Gavin et al., 1990; Takada et al., 1994; Zakin et al., 1998) and may contribute to β-catenin signaling in the primitive streak. Consistent with the β-catenin/TCF signaling activity we observed in the primitive streak, TCF/Lef binding sites have been identified in the promoter region of at least two genes that are expressed in the primitive streak, Brachyury and Cdx1 (Arnold et al., 2000; Prinos et al., 2001; Lickert and Kemler, 2002). In addition, both of these genes appear to be downstream target of Wnt3a (Yamaguchi et al., 1999; Prinos et al., 2001). We detected no signaling through TCF/Lef in anterior structures until the late headfold stage. Inhibition of Wnt signaling in the future anterior side of the embryo is thought to be important for proper axis formation (Beddington and Robertson, 1999). Expression of Wnt inhibitors such as Dickkopf-1 (Dkk1; Glinka et al., 1998; Mukhopadhyay et al., 2001) and secreted Frizzled-related protein 5 (Sfrp5; Finley et al., 2003) in the anterior visceral endoderm cells are thought to protect adjacent ectoderm from caudalizing or mesoderm-inducing influences (Beddington and Robertson, 1999).

In this study, we focused our analysis on the domains of β-catenin/Tcf transcriptional activity during the pre- and early postimplantation stages of mouse embryos. Although we have not performed a complete analysis of this pathway at later stages of development, we have found that this pathway is active in multiple sites within the embryo that correlate with domains where Wnt signaling is known to be required (O.A. Mohamed et al., unpublished observations). At later stages of embryonic development, our transgenic line reports activity of this pathway in identical domains as previously reported by Maretto et al. This TCF/Lef transgenic line as well as others (Maretto et al., 2003) will be useful in identifying cells where the Wnt/β-catenin signaling pathway is active during later embryonic development and organogenesis as well as in adult mice.

EXPERIMENTAL PROCEDURES

Immunofluorescence

Embryos were obtained from CD-1 female mice (Charles River, Canada). The day on which the plug was observed was designated E0.5. Preimplantation embryos were collected by flushing the oviducts of pregnant females. Postimplantation embryos were dissected from the uteri of pregnant females. Embryos were fixed in freshly prepared 4% paraformaldehyde in phosphate-buffered saline (PBS) either for 15 min at room temperature for preimplantation embryos or overnight at 4°C for postimplantation embryos. They were permeabilized in blocking solution consisting of PBS, 3% bovine serum albumin, 0.1% Triton X-100, for at least 30 min at room temperature and either stored at 4°C in blocking solution or processed immediately for immunofluorescent staining. Immunostaining was carried out in a blocking solution containing either a monoclonal mouse antibody against total β-catenin (Transduction Laboratory) diluted 1:500 in blocking solution or a monoclonal mouse antibody against nonphosphorylated β-catenin (Upstate Biotechnology) diluted 1:500. After overnight incubation at 4°C with agitation, the embryos were washed three times for 15 min each in blocking solution, then incubated for 1 hr at room temperature with agitation in the presence of a fluorescein isothiocyanate–conjugated anti-mouse antibody (Jackson Immunoresearch) diluted 1:100 in blocking buffer. The embryos were then washed and mounted on a microscope slide in the presence of 1 μg/ml of 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) to stain the DNA. Specimens were examined by using either a Leitz Laborlux S microscope or a Zeiss confocal laser scanning microscope.

Generation of Reporter Construct and Transgenic Animals

To generate the β-catenin–responsive reporter construct, six copies of the Tcf/Lef response element (CCTTTGATC; Korinek et al., 1997) were cloned in the XhoI site of plasmid 1-11 that contains the hsp68 minimum promoter driving a LacZ reporter gene (Rossant et al., 1991).

Transgenic mice were produced by using standard protocols. Approximately 10 pl of 10 ng/μl linear DNA was injected into the male pronucleus of one-cell stage embryos. After microinjection, embryos were incubated overnight and the following day, those that had cleaved to the two-cell stage were transferred into the oviducts of pseudopregnant females that were plugged on the day of transfer. Between 20 and 30 embryos were transferred per mouse. Transgenic animals were genotyped by using the polymerase chain reaction (PCR) technique. Briefly, 1 cm of tail was cut and digested in 300 μl of lysis buffer (100 mM Tris · HCl (pH 8.0), 10 mM ethylenediaminetetraacetic acid, 0.5% Tween 20, 0.5% NP-40, 0.4 mg/ml proteinase K) overnight at 55°C. The following day, DNA was phenol/chloroform extracted and 1 μl of DNA was used for each PCR reaction. Primers used for the PCR reaction were LacZ-F1 5′CAGTGGCGTCTGGCGGAAAACCTC 3′ and LacZ-B1 5′AAACAGGCGGCAGTAAGGCGGTCGG 3′. The PCR conditions were 94°C for 1 min, 62°C for 1 min, 72°C for 1 min for 28 cycles. Homozygous Vestigial tail mice were generously provided by Dr. David Lohnes (University of Montreal).

Detection of lacZ Activity

Embryos were dissected in PBS (pH 7.3), rinsed in 100 mM sodium phosphate pH 7.3, and then fixed in 0.2% glutaraldehyde, 2 mM MgCl2, 5 mM ethyleneglycoltetraacetic acid, 100 mM sodium phosphate pH 7.3 for 5 min at room temperature. Embryos were then washed three times in wash buffer (0.02% NP-40, 0.01% deoxycholate, 2 mM MgCl2, 100 mM sodium phosphate pH 7.3) for 15 min each at room temperature. To reveal lacZ activity, they were incubated in 1 mg/ml X-gal, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 0.02% NP-40, 0.01% deoxycholate, 2 mM MgCl2, 100 mM sodium phosphate pH 7.3 overnight at 37°C. They were then rinsed with wash buffer and PBS and post-fixed overnight in 4% paraformaldehyde at 4°C.

Transfection of Embryonic Fibroblasts

Embryonic fibroblasts were isolated from 12.5-day embryos following standard protocols. Transfections were carried out by using Lipofectamine (Invitrogen) according to the manufacturer's recommendations. Plasmids used for this study were S37A (coding for a mutated form of β-catenin in which serine-37 has been changed to alanine, a gift from Dr. S. W. Byers, Georgetown University School of Medicine, Washington, D.C.), pCMV5B-Lef1-HA (coding for mouse lef-1, a gift from Dr. D. Lohnes, University of Montreal), and pCMVβ (coding for β-galactosidase, Clontech, Inc.).

Acknowledgements

We thank Cassandre Labelle-Dumais for critical reading of the manuscript and assistance in the maintenance of mouse lines, and Drs. Byers and Lohnes for providing plasmids and mice as well as members of the Clarke and Dufort labs for helpful discussions. O.A.M. was supported by a scholarship from the Government of Libya. D.D. was funded by the Canadian Institute for Health Research. D.D. is a Chercheur Boursier du Fonds de la Recherche en Sante du Quebec (FRSQ).

Ancillary