Role of crescent in convergent extension movements by modulating Wnt signaling in early Xenopus embryogenesis

Mechanisms of Development 122 (2005) 1322–1339 www.elsevier.com/locate/modo

Role of crescent in convergent extension movements by modulating Wnt signaling in early Xenopus embryogenesis Mikihito Shibata, Mari Itoh, Hiroki Hikasa1, Sumiko Taira, Masanori Taira* Department of Biological Sciences, Graduate School of Science, University of Tokyo; Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan Received 9 November 2004; received in revised form 26 April 2005; accepted 9 June 2005 Available online 19 August 2005

Abstract The Xenopus gene crescent encodes a member of the secreted Frizzled-related protein (sFRP) family and is expressed in the head organizer region. However, the target and function of Crescent in early development are not well understood. Here, we describe a role of Crescent in the regulation of convergent extension movements (CEMs) during gastrulation and neurulation. We show that overexpression of Crescent in whole embryos or animal caps inhibits CEMs without affecting tissue specification. Consistent with this, Crescent efficiently forms complexes with Xwnt11 and Xwnt5a, in contrast to another sFRP, Frzb1. As expected, the inhibitory effect of Crescent or Xwnt11 on CEMs is cancelled when both proteins are coexpressed in the neuroectoderm. Interestingly, when coexpressed in the dorsal mesoderm, the activity of Xwnt11 is rather enhanced by Crescent. Supporting this finding, the inhibition of CEMs by Crescent in mesodermalized but not neuralized animal caps is reversed by the dominant-negative form of Cdc42, a putative mediator of Wnt/Ca2C pathway. Antisense morpholino oligos for Crescent impair neural plate closure and elicit microcephalic embryos with a shortened trunk without affecting early tissue specification. These data suggest a potential role for Crescent in head formation by regulating a non-canonical Wnt pathway positively in the adjacent posterior mesoderm and negatively in the overlying anterior neuroectoderm. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Xenopus laevis; Head organizer; Convergent extension movements; Crescent; Frzb1; Wnt11; Wnt5a; Wnt8; Cdc42; Activin; BF2/FoxD1

1. Introduction In early vertebrate embryogenesis, the Spemann organizer in amphibians or the equivalent tissue in other organisms plays a central role in establishing the embryonic body axes (Harland and Gerhart, 1997; Smith and Schoenwolf, 1998). In this process, anteroposterior patterning and morphogenetic movements of the neural and mesodermal tissues occur cooperatively. It has been proposed that the morphogenetic movements during gastrulation in Xenopus embryos are driven by a dramatic coordinated cell rearrangement, involving involution and convergent extension movements * Corresponding author. Tel./fax: C81 3 5841 4434. E-mail address: [email protected] (M. Taira). 1 Present address: Department of Microbiology and Molecular Genitics, Harvard Medical School, and Molecular Medicine Unit, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA

0925-4773/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2005.06.002

(CEMs). During involution, dorsal marginal zone (DMZ) cells start to intercalate, resulting in mediolateral narrowing (convergence) and subsequently in anterior–posterior lengthening (extension) (Keller et al., 1992). The DMZ includes a subpopulation of cells in the Spemann organizer that correspond to the posterior dorsal mesoderm, functionally called the trunk organizer. In contrast, cells in the deep layer of the Spemann organizer, including the anterior endomesoderm, which is thought to correspond to the head organizer, show negligible CEMs (Keller et al., 1989). Therefore, in terms of CEMs, the organizer region can also be divided into two regions: the chordal region, which actively undergoes CEM, and the anterior endomesoderm region, which does not. However, the mechanisms by which CEMs are excluded from the anterior region to allow proper head formation are poorly understood. Various factors involved in CEMs have been isolated, and their functions have been extensively studied. In Xenopus and zebrafish, Wnt5a, Wnt4, and Wnt11 affect

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

the morphogenetic movements of the ectodermal and mesodermal tissues at gastrula and neurula stages, and inhibit the elongation of animal caps treated with a mesodermalizing factor, activin, in the Xenopus system (Christian and Moon, 1993; Du et al., 1995; Heisenberg et al., 2000; Ungar et al., 1995). It has been shown that Wnt5a stimulates the Wnt/Ca2C pathway in which Ca2C release and protein kinase Ca (PKCa) activation are involved (Sheldahl et al., 1999; Slusarski et al., 1997). The analysis of zebrafish mutant silberblick/wnt11 has shown that Wnt11 is required for CEMs (Heisenberg et al., 2000). Whereas the canonical Wnt signaling is transduced through b-catenin, the regulation of CEMs by Wnt11 takes place through a non-canonical Wnt signaling pathway similar to that involved in the planar cell polarity (PCP) signaling pathway in Drosophila (Heisenberg et al., 2000). The components of non-canonical Wnt signaling include Frizzled 7 (Xfz7) (Djiane et al., 2000; Winklbauer et al., 2001), the Rho family GTPase, Cdc42 (Djiane et al., 2000), Rho, and Rac (Habas et al., 2003; Tahinci and Symes, 2003), Dishevelled (Dsh) (Sokol, 1996; Wallingford et al., 2000), and so on, most of which have been suggested to mediate the regulation of CEMs in Xenopus embryos. Dsh is a multifunctional protein involved not only in the canonical Wnt signaling pathway that regulates cell fates but also in the non-canonical Wnt signaling pathways. In addition, we have shown that a receptor tyrosine kinase, Xror2, which is expressed in the organizer and also in the posterior regions of both the axial mesoderm and neuroectoderm, forms complexes with several Wnts, and affects CEMs through Cdc42, implying that Xror2 is a newly identified modulator of Wnt signaling (He, 2004; Hikasa et al., 2002; Oishi et al., 2003). It has been noted that dominant negative Xwnt11 inhibits CEMs and blastopore closure, as does wild-type Xwnt11 (Heisenberg et al., 2000). Similarly, overexpression of wild-type or dominant negative Dsh leads to a similar phenotype with a shortened body axis (Wallingford et al., 2000), suggesting that a certain level of PCP signaling must be maintained for proper CEMs in the posterior region. In the previous study, we identified the Xenopus orthologue of chick crescent, a member of the secreted Frizzled-related protein (sFRP) gene family, together with another two laboratories who also independently identified the Xenopus crescent gene (Bradley et al., 2000; Pera and De Robertis, 2000; Pfeffer et al., 1997; Shibata et al., 2000, 2001) (also designated frzb2). The Xenopus gene crescent is first expressed in the deep layer of the organizer in the early stages of gastrulation, and later, expressed in the anterior endomesoderm until the late neurula stages. Crescent is a secreted protein with a Frizzled-like cysteine-rich domain (CRD), and is expected to bind to some Wnts to inhibit their activity, as has been shown for other sFRP family members. Although Frzb1, FrzA, and sFRP1 have been shown to inhibit the canonical Wnt signaling by binding to Wnt1, Wnt2, Wnt8, and Drosophila Wingless (Dennis et al., 1999; Finch et al., 1997; Leyns et al., 1997; Lin et al., 1997;

1323

Xu et al., 1998), sFRP family members that inhibit noncanonical Wnt signaling have not been described yet. Crescent has an activity distinct from that of Frzb1, because overexpression of Crescent in the animal pole region or the DMZ of embryos leads to reduced head structures (Bradley et al., 2000; Pera and De Robertis, 2000). It was also reported that crescent is involved in cardiogenesis in Xenopus (Pandur et al., 2002; Schneider and Mercola, 2001). However, the biochemical features of Crescent and its role in early embryogenesis have not been elucidated. In this paper, we show that Xenopus Crescent forms complexes efficiently with several Wnts, including Xwnt11 and Xwnt5a as well as Xwnt8, and is required for the convergent extension, neural tube closure, and head development. We also show that, interestingly, Crescent does not always inhibit Wnt signaling, but modulates the activity of Xwnt11 in different ways in the mesoderm and neuroectoderm.

2. Results 2.1. Crescent affects gastrulation movements but not the expression of axial mesoderm or neural markers in the whole embryo Overexpression of Crescent in the animal pole region or the DMZ of embryos leads to severe dorsal flexure phenotypes at tailbud stages (data not shown), implying that these phenotypes may be due to the inhibition of proper morphogenetic movements during gastrulation and neurulation. To assess this possibility, we analysed how cells expressing Crescent are distributed. A mixture of crescent and nb-gal mRNAs was coinjected on the left side of the dorsal equatorial region at the four-cell stage. As shown in Fig. 1A, nb-gal-positive globin-expressing control cells were restricted to the region along the midline of the trunk, indicating that normal CEMs had occurred (Fig. 1Aa). In contrast, Crescent-expressing cells remained stacked anterior to the blastopore (Fig. 1Ab). Moreover, blastopore closure was delayed, and neural tube closure was significantly impaired in those embryos, whereas Frzb1expressing cells were distributed almost normally along the dorsal midline (Fig. 1Ac). The tissue specification of dorsal structures in Crescentoverexpressing embryos was analyzed by whole-mount in situ hybridization for a notochord marker, XPA26 (Hikasa and Taira, 2001), a somite marker, XmyoDa (Hopwood et al., 1989), and a neural marker, sox2 (Mizuseki et al., 1998), at late neurula stages (stages 17–18) (Fig. 1B). To compare the activity of Crescent with that of Frzb1, we used the FLAG-tagged constructs, Crescent-FLAG (Cres-F) and Frzb1-FLAG (Frzb1-F), to assess that those proteins are expressed at equivalent levels. Using western blotting with anti-FLAG antibody, equivalent amounts of Cres-F and Frzb1-F proteins were detected in embryos injected with

1324

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 1. Crescent inhibits CEMs without affecting mesodermal or neural specification in whole embryos. (A) Lineage tracer analysis of CEMs after overexpression of Crescent or Frzb1. (B) Analysis with FLAG-tagged proteins, Crescent-FLAG (Cres-F) and Frzb1-FLAG (Frzb1-F). (a) Comparison of expression levels of Cres-F and Frzb1-F in injected embryos by Western blot analysis. Homogenates from embryos injected with cres-F or frzb1-F mRNA (100 pg/embryo) were electrophoresed (0.5 embryo equivalent per lane), FLAG-tagged proteins were detected with an anti-FLAG antibody. a and d nonspecific bands; b, Cres-F band overlapping a non-specific band; g, Frzb1-F band. (a 0 ) Intensities of bands were quantified with an image analyzer. The net volume of Cres-F band (b–a) is equivalent to that of Frzb1-F band (g). Vertical axis, band intensity in arbitrary units. (b–d) FLAG-tagged proteins, Cres-F and Frzb1-F show activities similar to the wild type (see A). (e–m) Embryos at neurula stages 16–17 were stained for b-gal (red), and subjected to whole-mount in situ hybridization for XmyoDa (e–g), XPA26 (h–j), or sox2 (k–m). Dorsal view; anterior is upwards. (C) and (D) Expression of early marker genes in injected embryos. Four-cell-stage embryos were injected with mRNA, as indicated, with nb-gal mRNA on the left (C) or both sides (D) of the dorsal equatorial region. Embryos at the gastrula (stage 10.5 or 11) or mid neurula stage (stage 14), as indicated, were stained for b-gal and subjected to whole-mount in situ hybridization for goosecoid (gsc) and Xotx2 (C), or for Xbra (D). An arrowhead indicates the spot-like expression of Xbra. (E) Crescent inhibits axis-inducing activity of Xwnt8. Four-cell-stage embryos were injected into the ventral equatorial region with mRNA as indicated. Embryos at the tailbud stage (stage 40) are shown. Numbers in each panel indicate the frequency of the represented phenotype. Amounts of injected mRNAs (pg/embryo): b-globin, 100 (A,Ba,D,E) or 25 (Be,h,k,C); crescent, 100; frzb1, 100; cres-F, 100 (Bb), 25 (Bf,i,l,C); frzb1-F, 100 (Bc), 25 (Bg,j,m); Xwnt8, 2; dkk1, 50.

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

the same amounts of mRNAs (Fig. 1Ba,a 0 ). Embryos injected with these mRNAs showed phenotypes quite similar to those injected with mRNAs for the wild-type proteins at the same concentrations (Fig. 1Bb–d). Although injection of cres-F mRNA at higher doses (more than 50 pg per blastomere) reduced the expression of XmyoDa in whole embryos (not shown), injection at the lower dose of 25 pg per blastomere led to negligible reduction in XmyoDa expression in the area in which nb-gal and Cres-F are expressed (Fig. 1Bf), but severe dorsal flexure still occurred (not shown). In contrast, injection of frzb1-F mRNA did not seem to affect the expression level or pattern of XmyoDa at a dose of 25 pg per blastomere (Fig. 1Bg), nor did it affect the CEMs of injected cells. In Cres-F-expressing embryos, the notochord (marked by XPA26 expression) formed but did not fully elongate anteriorly (Fig. 1Bi). Most of the notochord remained just anterior to, or extended ventrolaterally along, the blastopore. Neural specification, visualized by sox2 expression, also appeared to occur but the neural plate was not correctly shaped and did not elongate (Fig. 1Bl). In Frzb1-F-expressing embryos, the sox2 expression domain converged medially as normal embryos, although the level of sox2 expression was slightly reduced (Fig. 1Bm). These results suggest that Crescent but not Frzb1 regulates CEMs during gastrulation and neurulation without affecting dorsal tissue specification or differentiation. We also analyzed early marker genes. During the gastrula to mid neurula stages, the expression levels of the head organizer marker goosecoid (Cho et al., 1991) and the head organizer and anterior neural marker Xotx2 (Pannese et al., 1995), or the early posterior mesodermal marker Xbra (Smith et al., 1991), were not dramatically affected in the nb-gal-positive and Crescent-overexpressing region (Fig. 1C,D). This data suggests again that the dorsal flexure caused by Crescent is not due to the impairment of the mesoderm formation and patterning (Fig. 1C,D). In some embryos overexpressing Crescent at a higher dose (100 pg/ embryo), spotted expression of Xbra was observed (Fig. 1D, arrowhead; eight of 25 embryos), probably owing to either ectopic activation of Xbra or disorganized cell migration. In contrast, overexpression of Frzb1 (100 pg/embryo) greatly reduced the expression of Xbra (Fig. 1D), probably interfering with the canonical Wnt ligand(s) necessary for Xbra expression (Vonica and Gumbiner, 2002). These data imply that Crescent but not Frzb1 affects a noncanonical Wnt signaling regulating the CEMs. However, as shown in Fig. 1E, Crescent could also inhibit the canonical Wnt signaling, because it inhibited Xwnt8-induced secondaryaxis formation, similar to that with dkk1 mRNA, though embryos rescued by Crescent showed dorsally kinked phenotypes as observed above, in contrast to normallooking embryos rescued by Dkk1. These data will be in good agreement with coimmunoprecipitation data with Crescent (or Frzb1) and Wnts as shown below.

1325

2.2. Crescent inhibits CEMs in animal cap assays We further examined the inhibitory activity of Crescent on CEMs using animal cap assays, as a simpler system than that using whole embryos. As shown in Fig. 2A, Crescent effectively blocked the animal cap elongation induced by activin treatment, whereas Frzb1 did not. Mesoderm specification was monitored with XPA26. XPA26 was induced in the activin-treated animal caps as well as in the activin-treated and Crescent-overexpressing animal caps (Fig. 2B). Xwnt11 also inhibited CEMs as reported (Djiane et al., 2000), but did not abolish XPA26 expression. We next examined whether Crescent affects CEMs in neural tissue using XBF2/FoxD1-overexpressing animal caps. XBF2 upregulates several neural markers and induces morphological elongation without mesoderm induction when overexpressed in animal caps, which mimics the CEMs in the posterior neural plate (Mariani and Harland, 1998; Wallingford and Harland, 2001). Crescent, as well as Xwnt11, strongly inhibits the elongation of neuralized animal caps without inhibiting neuralization, monitored as the expression of nrp1 (Fig. 2C). To examine whether Crescent can diffuse across cells and affect cell movements, we used an animal cap conjugation assay as outlined in Fig. 2D. The elongation of XBF2-overexpressing animal caps was effectively blocked by conjugation with Crescent-expressing caps, but not by conjugation with Xwnt11-expressing caps (Fig. 2D). These data suggest that Crescent can diffuse from the Crescent-expressing cap to the adjacent BF2-expressing cap to inhibit its CEMs, but Xwnt11 cannot. 2.3. CRD is necessary for crescent activity Because the CRDs of sFRPs are thought to be responsible for their inhibition of target Wnts through direct physical interactions (Finch et al., 1997; Hsieh et al., 1999; Rattner et al., 1997), we investigated whether Crescent activity and the affinity of Crescent for Wnts in the embryo depend on the CRD. For this purpose, we constructed Cres-FZD1-F, a FLAG-tagged Crescent mutant that lacks 20 amino acids, including the second cysteine in the CRD, based on the study of Hsieh et al. (1999) and our data for Xror2 (Hikasa et al., 2002). Overexpression of Cres-FZD1-F failed to elicit embryos with dorsal flexure (Fig. 3A), and failed to block the elongation of activintreated animal caps (Fig. 3B), suggesting that Crescent action on CEMs depends on the integrity of the CRD. 2.4. Physical interactions between crescent and the Wnts that transduce non-canonical Wnt signaling To identify the binding targets of Crescent, Myc-tagged Wnts that transduce canonical Wnt signaling (Xwnt8 and Xwnt3a) or non-canonical Wnt signaling (Xwnt11, Xwnt5a, and mouse Wnt4) were coimmunoprecipitated with Cres-F,

1326

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 2. Crescent affects the elongation of animal caps induced by activin treatment or XBF2 overexpression without inhibiting mesodermal or neural specification. (A) Crescent but not Frzb1 inhibits animal cap elongation. (B) Neither Crescent nor Xwnt11 abolish notochord specification as assayed by expression of the notochord marker XPA26. Two-cell-stage embryos were injected with mRNA, as indicated above each panel, in the animal pole region of both blastomeres. Animal caps were dissected and treated with or without activin as indicated. (C) Crescent and Xwnt11 inhibit the animal cap elongation elicited by XBF2, but do not interfere with expression of the pan-neural marker nrp1. Two-cell-stage embryos were injected with mRNA, as indicated, together with or without XBF2 mRNA in the animal pole region of both blastomeres. XPA26 and nrp1 mRNA expression was visualised by whole-mount in situ hybridization. (D) Crescent but not Xwnt11 acts over a long distance. Two-cell-stage embryos were injected with mRNA for animal cap conjugation assays. Animal caps from embryos injected with globin, crescent, or Xwnt11 mRNA were conjugated with those from embryos injected with mRNAs for XBF2 plus nb-gal, and the conjugates were cultured until sibling stage 19. Amounts of injected mRNAs (pg/embryo): b-globin, 100 (A), 1000 (B) and (D), 500 (C); crescent, 50 (A) and (B), 100 (C), 20 (D); frzb1, 100; Xwnt11, 1000 (B) and (D), 500 (C); XBF2, 200.

and, for comparison, with Frzb1-F and the FLAG-tagged ectodomain of Frizzled 7 (Exfz7-F) using extracts from injected whole embryos. Of the Wnts examined, Xwnt8 was preferentially coimmunoprecipitated with Frzb1 (Fig. 4A), consistent with a previous report (Lin et al., 1997), whereas Xwnt11 was coimmunoprecipitated with Frzb1 only weakly, and coimmunoprecipitated Xwnt5a was hardly detectable (Fig. 4A). Xwnt11 was also coimmunoprecipitated with Exfz7-F, as reported previously (Djiane et al., 2000) (Fig. 4A). In contrast, all the Wnt proteins examined were coimmunoprecipitated with Crescent, and the complex formed between Xwnt11 and Crescent was particularly

evident (Fig. 4A,B). Unexpectedly, Myc-tagged Xwnt11 protein was coimmunoprecipitated with Cres-FZD1-F more efficiently than with Cres-F protein (Fig. 4C), suggesting that the intact CRD of Crescent is not required for it to form complexes with Wnt but is required for its activity, as reported previously for Frzb1 and Wnt1 (Finch et al., 1997). To compare the Wnt-binding affinities of Crescent and Frzb1 more precisely, we injected several doses (250, 500, 1000, or 1500 pg/embryo) of cres-F or frzb1-F mRNA together with a fixed dose (500 pg/embryo) of myc-tagged Xwnt11, Xwnt5a, or Xwnt8 mRNA. The amounts of coimmunoprecipitated Wnts (Fig. 4D, upper panels) were

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

1327

Fig. 3. Crescent exerts its effect on CEMs via the CRD. (A) Cres-FZD1-F does not elicit embryos with dorsal flexure. mRNA was injected, as indicated, into the dorsal region at the four-cell stage. Injected embryos were observed at the tailbud stage (stage 42). (B) Cres-FZD1-F does not inhibit the elongation of animal caps induced by activin treatment. Amounts of injected mRNAs (pg/embryo): b-globin, 50 (A) and (B), cres-F, 50 (A) and (B); cres-FZD1-F, 50 (A) and (B).

then plotted against the amounts of immunoprecipitated Cres-F or Frzb1-F (bottom panels) to produce line graphs (bottom) that show clearly that Crescent has stronger affinity in complexes with Xwnt11 and Xwnt5a than does Frzb1, and has an affinity for Xwnt8 almost equal to that of Frzb1. These data suggest that Crescent blocks the CEMs (Figs. 1, 2) by affecting non-canonical Wnt signaling mediated by Xwnt11 and Xwnt5a and also inhibits Xwnt8-induced secondary-axis formation (Fig. 1E) by interacting with Xwnt8, and that Frzb1 reduces Xbra and perhaps sox2 expression (Fig. 1) by inhibiting canonical Wnt signaling mediated by Xwnt8 or its equivalent. 2.5. Comparison of the expression patterns of crescent with those of Xwnt11 and Xwnt5a As described above, Xwnt11 and Xwnt5a are primary candidates for the binding targets of Crescent. To assess the timing of Crescent in regulating the activities of Wnts during gastrulation, we compared the spatial and temporal expression patterns of crescent with those of Xwnt11 and Xwnt5a using double whole-mount in situ hybridization. As shown in Fig. 5A, the expression domains of crescent and Xwnt11 are situated close together at the early gastrula stage. A detailed comparison was made with a pair of hemisectioned early-gastrula embryos; one hemisection was probed with crescent and the other with Xwnt11. As reported previously (Heisenberg et al., 2000; Ku and Melton, 1993), Xwnt11 is expressed mainly in the surface epithelial cells of the dorsal marginal zone (Fig. 5B, right), whereas crescent is expressed in the deep layer of the endomesoderm through the surface layer of the dorsal blastopore lip (Fig. 5B, left). Therefore, the expression domains of crescent and Xwnt11 do not overlap but are adjacent at the early gastrula stages (stages 10.25 and 10.5).

As gastrulation proceeds, the expression domains of crescent and Xwnt11 progressively separate into the anterior endomesoderm and the surface layer of the circumblastoporal tissues, respectively, from the mid gastrula stage (stage 11; data not shown) to neurula stages (Fig. 5C,D). These data imply that Crescent can effectively interact with Xwnt11 in the early to mid gastrula stages. The expression patterns of Xwnt5a during the lategastrula to early-neurula stages have not been reported. The expression of Xwnt5a at the early gastrula stage was only faintly and broadly detected in ectodermal tissues (data not shown). A comparison with a pair of hemisectioned earlyto-mid-neurula-stage embryos showed that like Xwnt11, Xwnt5a expression is restricted to circumblastoporal tissues far from the crescent expression domain (Fig. 5E,F). This expression pattern coincides well with that of zebrafish pipetail/Wnt5 (Kilian et al., 2003). These data indicate that Xwnt11 is a good candidate for a Crescent target during the gastrula stages. 2.6. Crescent antagonizes the action of Xwnt11 on CEMs in the neuroectoderm, but enhances it in mesodermal tissue The physical interactions and expression patterns of Crescent and Xwnt11 demonstrated above prompted us to investigate the functional interactions between these two proteins. For this purpose, we injected crescent or Xwnt11 mRNA alone or together to express the respective proteins in the neuroectoderm or axial mesodermal region. When crescent mRNA or Xwnt11 mRNA (at a range of 50–100 and 50–500 pg per embryo, respectively) alone was injected into the dorsal animal pole region that is destined to become the neuroectoderm, CEMs were similarly impaired, resulting in embryos with dorsal flexure (Fig. 6A,B; not shown). Such phenotypes were partially abolished by the coinjection

1328

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 4. Physical interactions between Crescent and Wnts as assayed by coimmunoprecipitation. Experimental procedures are described in Materials and Methods. (A) and (B) Crescent interacts with Wnts that transduce non-canonical Wnt signaling (Xwnt11, Xwnt5a, and mouse Wnt4) as well as those that transduce canonical signaling (Xwnt8 and Xwnt3a). Xwnt11-Myc, Xwnt5a-Myc, or Xwnt8-Myc (A) or Xwnt8-Myc, Xwnt3a-Myc, or mouse Wnt4-Myc (B) was coimmunoprecipitated with either Cres-F (A) and (B) or Frzb1-F (A). As a positive control, Xwnt11-Myc was coimmunoprecipitated with the FLAGtagged ectodomain of Xfz7 (Exfz7-F). (C) Cres-FZD1-F forms a complex with Xwnt11-Myc. (D) Comparison of complex formation between Crescent and Frzb1 for Xwnt11, Xwnt5a, and Xwnt8. Cres-F shows a higher activity to form a complex with Xwnt11-Myc and Xwnt5a-Myc than does Frzb1-F, whereas Cres-F and Frzb1-F show similar activities for Xwnt8-Myc. Numbers in the panels indicate the amounts of cres-F or frzb1-F mRNA (ng/embryo) injected. Note that frzb1-F mRNA was injected at higher doses than cres-F mRNA to immunoprecipitate equivalent amounts of proteins. The horizontal axes indicate the amounts of Cres-F or Frzb1-F protein immunoprecipitated with anti-FLAG antibody, in arbitrary units. The vertical axes indicate the amounts of Myc-tagged Wnt proteins coimmunoprecipitated with either Cres-F (red line) or Frzb1-F (blue line), in arbitrary units. The amount of Wnt mRNAs injected was fixed at 500 pg/embryo in all experiments.

of crescent and Xwnt11 mRNAs (Fig. 6A,B; not shown), suggesting that Crescent can antagonize Xwnt11 signaling by physical interaction with it, and that the inhibition of CEMs by Crescent is due to the depletion of Xwnt11 or its equivalent(s) present in the neuroectoderm. We next performed similar experiments except that mRNA was injected into the dorsal equatorial region to express Crescent or Xwnt11 proteins in the dorsal mesoderm. Unexpectedly, there was no sigh of antagonism between Crescent and Xwnt11 (at a range of 50–100 and

20–100 pg per embryo, respectively), but rather synergistic or additive effects on CEMs by the coexpression of both mRNAs. Therefore, to observe this more clearly, we used the lowest doses of Xwnt11 and crescent mRNA (50 and 20 pg per embryo, respectively), at which a single injection of Xwnt11 or crescent mRNA elicited embryos with moderate dorsal flexure. Under this condition, we found that coinjection of crescent and Xwnt11 mRNAs elicited embryos with more severe dorsal flexure than the injection of either transcript alone (Fig. 6C,D).

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

1329

Fig. 5. Comparison of the expression domain of crescent with those of Xwnt11 and Xwnt5a at early gastrula to mid neurula stages. (A) Dorsal view of early gastrula (animal pole is upwards). Expression domains of crescent (cyan) and Xwnt11 (purple) are situated close to each other at this stage. (B) An earlygastrula-stage embryo was bisected sagittally into halves, which were then hybridized for crescent (left panel) or Xwnt11 (right panel; flipped horizontally for comparison). The expression domains of crescent and Xwnt11 do not overlap but are adjacent. Dotted lines indicate the dorsal border of crescent expression. (C,D) Hemisections of early-neurula- (C) or mid-neurula-stage (D) embryos. The expression domains of crescent (cyan) and Xwnt11 (purple) progressively separate into the anterior endomesoderm and the surface layer of the circumblastoporal tissue, respectively. Anterior is to the left. (E) and (F) Hemisections of late gastrula (E) or mid neurula (F). The expression domains of crescent (cyan) and Xwnt5a (purple, indicated by arrows) progressively separate into the anterior endomesoderm and the surface layer of the circumblastoporal tissue, respectively. Anterior is to the left. Arrowheads, blastopores. Photographs of the hemisections stained for Xwnt5a were flipped horizontally for comparison.

These opposite effects of Crescent on CEMs in the neuroectoderm and dorsal mesoderm were further examined using animal caps either overexpressing XBF2 or treated with activin. In XBF2-expressing caps, injection of either crescent or Xwnt11 mRNA (20 or 500 pg per embryo, respectively) alone resulted in the inhibition of elongation, whereas coinjection of both mRNAs restored the animal cap elongation (Fig. 7A,B), supporting the observation that Crescent and Xwnt11 are competitive when expressed in the neuroectodermal region of the whole embryo (Fig. 6A,B). We next tested the interactions between Crescent and Xwnt11 using activin-treated animal caps. Similar to dorsal equatorial injections, we detected no sign of elongation rescue by coexpression of Crescent and Xwnt11 after injection of the several different sets of mRNA doses and ratios at a range of 5–500 and 100– 1000 pg per embryo, respectively (not shown). As shown in Fig. 7C,D, whereas crescent or Xwnt11 mRNA injected at the lowest doses (5 and 100 pg per embryo, respectively) caused partial inhibition of cap elongation, coinjection of both mRNAs at these doses inhibited animal cap elongation more effectively than injection of a single transcript (Fig. 7C,D). Supporting this, coexpression of Dsh-DN, a mediator of non-canonical Wnt signal pathway (Heisenberg et al., 2000), with Crescent failed to restore the elongation of activin-treated animal caps (not shown). Thus, the effects of Crescent and Xwnt11 on the elongation of mesodermalized animal caps are not competitive, but rather enhancing, supporting the effect observed when they were coexpressed in the dorsal mesoderm (Fig. 6C,D).

2.7. Functional interactions of Crescent with Dsh and Cdc42 The different effects of Crescent on Xwnt11 in the neuroectoderm and mesoderm described above could be accounted for by a difference in the affinity between Crescent and Xwnt11 in these tissues. We tested this possibility with coimmunoprecipitation assays. As shown in Fig. 8A, however, almost the same amounts of Xwnt11-Myc were coimmunoprecipitated with Cres-F in both activintreated and XBF2-overexpressing animal caps, suggesting that Crescent shows equal affinity for Xwnt11 in both tissues. To clarify whether or not Crescent actually activates intracellular Wnt signal transduction together with Xwnt11 in the mesoderm, we next analysed the level of phosphorylation of Dsh as an index of intracellular signaling in the PCP pathway. For this purpose, Myc-tagged Dsh was expressed in animal caps, and then a phosphorylated form of Dsh was detected by immunoprecipitation followed by western blotting, using the method of Tada et al. (Heisenberg et al., 2000). Hyperphosphorylated Dsh was detected when Xwnt11 was overexpressed in animal caps treated with activin or overexpressing XBF2 (Fig. 8B; lanes 3 and 6). This Xwnt11-induced hyperphosphorylated Dsh was diminished by the coexpression of Crescent in both activin-treated and XBF2-expressing animal caps (Fig. 8B; lanes 4 and 7). Although the relationship between the hyperphosphorylation of Dsh induced by Xwnt11 and the regulation of CEMs is still unclear, these results suggest that Crescent inhibits the signal transduction pathway between Xwnt11 and Dsh in both the neuroectoderm and mesoderm.

1330

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 6. Crescent antagonizes the action of Xwnt11 on CEMs in the neural plate, but enhances it in the dorsal mesoderm. (A) and (B) Coexpression of Crescent and Xwnt11 cancels out their action on CEMs in the neural plate. (A) Morphological appearance. Four-cell-stage embryos were injected in the dorsal animal pole region with mRNAs, as indicated, together with nb-gal mRNA. Embryos at the tailbud stage (stage 28) are shown. (B) Summary of the phenotypes shown in (A). (C) and (D) Dorsal flexure elicited by Crescent or Xwnt11 is enhanced by coexpression of these proteins in the dorsal mesoderm. (C) Morphological appearance. Four-cell-stage embryos were injected with mRNA, as indicated, in the dorsal equatorial region. Embryos at the tailbud stages (stages 33–34) are shown. (D) Summary of the phenotypes shown in (C). The extent of dorsal flexure elicited by dorsal animal pole or equatorial injection was classified by blind scoring as follows: K, normal; C, weak; CC moderate; CCC, severe; s.a, short axis without dorsal flexure. Amounts of injected mRNAs (pg/embryo): bglobin, 100 (A), 600 (C); crescent, 100 (A), 50 (C); Xwnt11, 500 (A), 20 (C).

The enhancement of Xwnt11 activity and inhibition of the hyperphosphorylation of Dsh by Crescent in the mesodermal tissue observed above raises the possibility that Crescent positively regulates a non-canonical Wnt pathway in a Dsh-independent manner. Such non-canonical Wnt signaling pathway is the Wnt11/Fz7-dependent G-protein signaling pathway that is mediated by Ca2C, PKCa, and the small GTPase Cdc42 (Choi and Han, 2002; Kuhl et al., 2000; Penzo-Mendez et al., 2003; Sheldahl et al., 1999; Slusarski et al., 1997). In addition, a dominantnegative Cdc42 mutant (Cdc42T17N) rescues the inhibitory effects of Xwnt11, Xfz7, or Xror2 on the activin-induced elongation of animal caps (Djiane et al., 2000; Hikasa et al., 2002), but either dominant-negative Rho or Rac does not (Djiane et al., 2000). Therefore, we next examined the involvement of Cdc42 in Crescent-mediated regulation of cell movements. As shown in Fig. 8C,D, inhibition of animal cap elongation by Crescent (bars 4, 6, 11 and 13 in

Fig. 8D) was rescued by coexpression of Cdc42T17N (bars 5 and 7) in activin-treated caps (bars 2–9), but not in XBF2expressing caps (bars 10–14), whereas inhibition by Xwnt11 was rescued in both mesodermalized (Hikasa et al., 2002) and neuralized animal caps (Fig. 8D; bars 15 and 16). Overexpression of Cdc42T17N alone did not affect the elongation of animal cap treated with activin (Fig. 8D; bars 2 and 3) as reported previously (Djiane et al., 2000). These data support the possibility that Crescent positively regulates the non-canonical Wnt pathways via Cdc42 in the mesoderm. 2.8. Crescent is required for the convergent extension, neural tube closure and head development To analyze, the role of endogenous Crescent, we performed loss-of-function analysis of Crescent by injecting antisense morpholino oligos (MO) targeting the exon–intron

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

1331

Fig. 7. Effects of Crescent on the activity of Xwnt11 in neuralized and mesodermalized animal caps. (A) and (B) Crescent and Xwnt11 antagonize each other in the inhibition of XBF2-induced elongation of animal caps. (A) Morphological appearance of animal caps. (B) Summary of the animal cap elongation assays shown in (A). (C) and (D) The inhibitory action on activin-induced elongation of animal caps by Crescent and Xwnt11 is enhanced when the proteins are coexpressed. (C) Morphological appearance of animal caps. (D) Summary of animal cap elongation assays shown in (C). See Fig. 2 for the experimental procedures. The extent of animal cap elongation induced by activin treatment or overexpression of XBF2 was classified by blind scoring as follows: K, no elongation; C, weak elongation; CC moderate elongation; CCC, strong elongation. Amounts of injected mRNAs (pg/embryo): b-globin, 600 (A), 100 (C); crescent, 20 (A), 5 (C); Xwnt11, 500 (A), 100 (C); XBF2, 250.

boundary of crescent pre-mRNA (cresMO). We first examined whether cresMO inhibits splicing of the crescent mRNA precursor as assayed by RT-PCR to amplify DNA fragments spanning the targeted exon 1–intron 1 splice junction. In cresMO-injected embryos, wild-type 410-bp products (Fig. 9Aa, band a) decreased dose-dependently and sequence-specifically as compared to that in the control embryos, whereas approximately 2-kb products supposedly containing intron 1 increased. In addition, two extra bands were detected in cresMO-injected embryos (Fig. 9Aa, bands b and g). Isolation and sequencing of these bands revealed that they were 362-bp (band b) and 232-bp (band g) products derived from aberrantly spliced RNA as shown in Fig. 9Ab. The sequences around these possible cryptic splice donor sites were ATT/Ile TGC/Cys tttgat and GAC/ Asp CAC/His G gcatgt (lower cases indicate spliced sequences), which correspond to the nucleotide numbers 332–337 and 462–467, respectively (GenBank accession

number AF260729). Curiously, these cryptic splice donor sites do not have the invariant gt splice sequence downstream of the exon, although the latter sequence gcatgt has some similarity to the consensus sequence gt(a/g)agt. These aberrant splicing events caused 16 amino acid-deletion in the region C-terminal to the Frizzled-like domain and a frameshift mutation in the Frizzled-like domain, respectively. Thus, cresMO is very likely to impair the function of Crescent by inhibiting proper splicing of the crescent mRNA precursor. We then examined the effects of cresMO on embryonic development at the tailbud stage as well as the specificity of cresMO by rescue experiments with coinjected crescent mRNA. As shown in Fig. 9Ba,b, cresMO-injected embryos showed the reduction of head structures and a shortened body axis, compared to the control embryos. These phenotypes were almost rescued by coinjection of crescent mRNA (Fig. 9Bc). Consistent with head reduction in tailbud

1332

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 8. (A) Physical interactions of Cres-F and Xwnt11-Myc are similar in activin-treated and XBF2-expressing animal caps. Coimmunoprecipitation assays with Xwnt11-Myc and Cres-F were performed using activin-treated and XBF2-expressing animal caps. (B) Hyperphosphorylation of Dsh by Xwnt11 is downregulated by Crescent in both activin-treated and XBF2-expressing animal caps (upper band). (C) and (D) Inhibition of animal cap elongation by Crescent was rescued by coexpression of Cdc42T17N, a dominant-negative construct of Cdc42, in the activin-treated, but not in XBF2-expressing caps. (D) Summary of animal cap elongation assays. The extent of animal cap elongation induced by activin treatment or XBF2 overexpression was classified by blind scoring as follows: K, no elongation; C, weak elongation; CC moderate elongation; CCC, strong elongation. Amounts of injected mRNAs (pg/embryo): b-globin, 250 (A), 2,000 (B); Xwnt11-Myc, 500; Xwnt11, 1000 (B), 500 (D); cres-F, 250; crescent, 500 (B), 10 (C), 50 or 10 (D); Myc-Dsh, 250; XBF2, 250; Xror2, 200; Cdc42T17N, 600.

stage embryos, the expression domain of an anterior neural marker, Rx2A (forebrain/eye anlagen) (Mathers et al., 1997), was reduced by cresMO at the mid neurula stage (Fig. 9Ca,b,g), and this reduction was rescued by coinjection with crescent mRNA (Fig. 9Cc,g). At the gastrula stage, cresMO-injected embryos showed the retardation of gastrulation (not shown). To examine CEMs and neurulation in the morphants, the expression of Dlx3 and Xhairy2a was examined. The neural plate was

negatively stained with Dlx3, which is expressed in the epithelial layer of neural fold and epidermis (Feledy et al., 1999). As shown in Fig. 9Cd,e,h, neural tube closure was impaired in the morphants, resulting in a shorter and laterally wider neural plate. Cross sections revealed that, in cresMO-injected embryos, nb-gal-positive cells expand laterally, and the notochord is slightly enlarged and roundshaped, which differs from the small oval-shaped notochord in the control embryos (Fig. 9Cd 0 ,e 0 ). In some of the

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

morphants, the archenteron is shrunk as shown in Fig. 9Ce 0 , probably due to defects in CEMs as observed previously (Hikasa et al., 2002). Xhariy2a expression marks the neural plate border, cranial neural crest, thin intermediate region, and floor plate (Mizuseki et al., 1998), as shown in Fig. 9Ci, and these expression patterns basically remained in the morphants (Fig. 9Cj). However, the neural plate did not extend longitudinally, and the floor plate expression of Xhairy2a did not fully extend anteriorly in the morphants (Fig. 9Cj), whereas this phenotype could be rescued by coinjection with crescent mRNA (Fig. 9Ck). This data also suggest that reduction of Crescent function by cresMO impairs the convergent extension of neural tissue. Expression patterns of Xhairy2a in the morphant imply that tissue differentiation of the neural plate is not grossly affected by cresMO. To assess this possibility, we examined the expression of the mesodermal markers Xbra and XmyoDa, the organizer markers chordin and goosecoid, and the neural marker sox2 at the early gastrula and early neurula stages. As shown in Fig. 9D, expression levels and patterns of these markers were not largely affected in cresMO-injected embryos compared to those of the control. Taken together, these data suggest that Crescent plays a role in the regulation of morphogenetic movements for head formation.

3. Discussion In this study, we have shown that (1) overexpression of Crescent inhibits mesodermal and neural CEMs without affecting tissue specification in both whole-embryo and animal cap assay systems (Figs. 1,2); (2) Crescent forms a complex with Xwnt11 and Xwnt5a, which are ligands for non-canonical Wnt pathways and expressed in the region near the Crescent expression domain (Figs. 4,5); (3) the second cysteine-containing region of the CRD is required for Crescent activities but not for complex formation with Wnts (Figs. 3,4); (4) Crescent exerts both antagonistic and enhancing effects on the action of Xwnt11 in CEMs in the neuroectoderm and mesoderm, respectively (Figs. 6,7), (5) Effects of Crescent on CEMs is rescued by Cdc42T17N in the mesoderm but not in the neuroectoderm (Fig. 8); and (6) knockdown of the endogenous Crescent by morpholino oligos impairs head development and dorsal morphogenesis including neural tube closure and anterior extension of the floor plate without affecting neural and mesodermal specification (Fig. 9). Taken together, these data suggest that Crescent regulates cell movements during gastrulation and neurulation by modulating non-canonical signaling by Wnt11 in a context-dependent manner possibly through the Wnt/Cdc42 pathway. However, the possibility still remains that Crescent also acts as an inhibitor for canonical Wnt ligands as discussed below.

1333

3.1. Crescent is a context-dependent modulator of mesodermal and neural CEMs Similar to Crescent, several modulators of Wnt signaling pathways, including sFRP1, Xenopus Cyr61 and Wise (Itasaki et al., 2003; Latinkic et al., 2003; Uren et al., 2000), have been suggested as playing roles in both the activation and inhibition of the canonical Wnt pathway in a dose- or context-dependent manner. However, the molecular basis of those biphasic or dual activities has not been clarified yet. In the case of Crescent, its dual role is due neither to a change in complex forming ability for Xwnt11 (Fig. 8A) nor to enhancement of the phosphorylation of Dsh in the mesoderm (Fig. 8B). Instead, the phosphorylation of Dsh by Xwnt11 is reduced by Crescent, probably resulting from the binding of Crescent to Xwnt11. Because the effects of Crescent and Xwnt11 on CEMs in mesodermal tissue are additive or even synergistic (Figs. 6,7) and the inhibition of activin-treated animal cap elongation by Crescent was rescued by coexpression of dominant negative Cdc42 (Fig. 8C,D; bars 4–7), we think that Crescent may inhibit some pathways of Xwnt11 that include Dsh, but may enhance other pathway(s) downstream from or parallel with Xwnt11 signaling that mediate small GTPase(s) including Cdc42. Such non-canonical Wnt signaling pathways are the Wnt/Ca2C pathway and the Wnt11/Fz7-dependent G-protein signaling pathway, both of which activate PKCa and Cdc42 (Choi and Han, 2002; Penzo-Mendez et al., 2003; Sheldahl et al., 1999; Slusarski et al., 1997) independent of Dsh. Similarly, it has been shown that both Xwnt11 and Crescent can promote cardiogenesis in the mesoderm by stimulating the c-Jun N-terminal kinase (JNK) pathway (Pandur et al., 2002; Schneider and Mercola, 2001). Habas et al. have shown that dominant negative Cdc42 crossinhibits Rac, a downstream component of Dsh in the PCP pathway, in dorsal marginal explants, probably by titrating Rac-GEF (Habas et al., 2003). Therefore, it is possible that Crescent activates Rac in the mesoderm. However, this seems less likely because Crescent inhibits signal transduction between Xwnt11 and Dsh, as assayed by the phosphorylation of Dsh (Fig. 8B). In contrast, Xwnt11 but not Crescent stimulates this pathway in the neuroectoderm, because dominant negative Cdc42 can rescue the inhibition of neuralized animal cap elongation by Xwnt11 but not that by Crescent (Fig. 8C,D; bars 11–16). The data presented above raise the question of how the activity of Crescent is modulated in the mesoderm and neuroectoderm. One explanation is the existence of mesoderm- or neuroectoderm-specific cofactor(s) that modulates Crescent activity according to the tissue context. However, strong candidates for such cofactors have not yet emerged. 3.2. Interactions between Crescent and Wnts Comparison of the spatiotemporal expression patterns of crescent and Xwnt11 suggests that Crescent probably

1334

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

Fig. 9. Morpholino-mediated knockdown of Crescent. Four-cell-stage embryos were injected into the dorsal equatorial region with MOs as indicated on the left side of panels. (A) Efficacy of cresMO to block the splicing of crescent transcript. (a) RT-PCR showing aberrant splicing in MO-injected embryos. RTK, RT minus. (b) Normal and cryptic splice donor sites in exon 1 (indicated by arrowheads). (B) Morphological appearances of injected embryos at the tailbud stage (stage 33/34). Compared to normal embryos (a), the morphants showed short body axis and small head with small eyes (b), whereas these phenotypes were weakened by co-injection with crescent mRNA (c). Anterior is to left. (C) Anterior development and neurulation are affected at the neurula stage. The expression of Rx2A (a–c,g) at stage 17, or that of Dlx3 (d–f,d 0 –f 0 ,h) and Xhairy2a (i–k) at stage 15 was examined. Anterior view; dorsal is upwards in (a–c). Dorsal view, anterior is upwards in (d–f,i–k). Cross sections at the position indicated by dashed lines in (d–f) are shown in (d 0 –f 0 ), respectively. DAPI staining revealed that wider neural plate and notochord. Notochord was indicated by white dots. (g,h) Width of Rx2A domain (indicated in (a–c) and the width per length

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

interacts with Xwnt11 effectively at least until the mid gastrula stage (Fig. 5A–D). Different from Xwnt11, Xwnt5a is first detected faintly at the mid gastrula stage (not shown), mainly in the ectodermal region, and specifically in the circumblastoporal tissues from the late gastrula and early neurula stages onwards (Fig. 5E,F). The expression of Xwnt4 reportedly begins in the lateral ridge of neural plate and diencephalon/mesencephalon region, and Xwnt3a in the anterior end of neural fold from the mid neurula stage (McGrew et al., 1992; Wolda et al., 1993), when the expression of crescent starts to wane. Thus, Xwnt11 is more likely to be a target of Crescent, than the other Wnts. However, it may be still possible that diffused Crescent interacts with those Wnts. In the case of the pronephros, crescent and Xwnt4 are coexpressed during the tailbud stages (McGrew et al., 1992; Shibata et al., 2000), implying that physical and functional interactions between them would be expected in that organ. Interestingly, whereas a small deletion in the CRD abolishes the activity of Crescent (Fig. 3), the mutant protein can still form a complex with Xwnt11 with somewhat higher affinity than can the wild-type protein (Fig. 4C). This data indicates that the complex formation with Crescent is not enough to inhibit Xwnt11, and that the N-terminal region of the CRD is critical for Crescent functions. Similar results were reported by Lin et al., (Finch et al., 1997), who examined the binding affinity of several CRD-deletion constructs of Frzb1 for Wnt1. They showed that while some CRD-deletion constructs form complexes with Wnt1, they lack the inhibitory activity for Wnt1. Uren et al. also reported that the CRD of sFRP1 is not necessary for Wnt binding, while the C-terminal region of sFRP1 interacts with Wingless (Uren et al., 2000). These data suggest a possibility that sFRP protein can be functionally subdivided into the domains for the interaction with Wnts and that for the inhibition or activation of Wnt activities by binding other proteins. Those putative interacting proteins might be Wnt receptors or Crescent itself, because sFRP family members have been shown to form complexes with Frizzled receptors (Bafico et al., 1999) and CRD domains have been suggested to form hetero- or homodimers (Dann et al., 2001). 3.3. Roles of crescent in the head organizer region The biological activity of Crescent agrees well with biochemical data indicating that Crescent forms complexes

1335

with Xwnt11 and Xwnt5a more efficiently than does Frzb1, whereas Crescent and Frzb1 show similar complex forming ability for Xwnt8 (Fig. 4). Although Crescent does form a complex with Xwnt8 and inhibits the secondary axis inducing activity of Xwnt8 (Fig. 1E), it does not downregulate Xbra expression at least at relatively low doses, whereas Frzb1 does (Fig. 1D), as do other artificial inhibitors of canonical Wnt signaling, such as dominant negative Xwnt8 or dominant negative Tcf3 (Vonica and Gumbiner, 2002). Moreover, whereas Crescent blocks activin-induced animal cap elongation, Frzb1 fails to do so (Fig. 2). On the other hand, knockdown of Crescent by MO leads to the reduction of anterior marker gene expression (Fig. 9Cb). These data suggest three possibilities that: (1) inhibition of CEMs leads to reduction of head structure possibly due to incomplete separation of the head organizer from the trunk organizer; (2) while Crescent strongly interacts with non-canonical Wnt signaling, it inhibits the activity of canonical Wnt ligands other than Xwnt8; and (3) Crescent acts as a canonical Wnt antagonist by interacting with Frzb-1 in the head organizer, because we observed that Crescent and Frzb-1 synergistically anteriorize the embryos when co-expressed (M.S. and M.T., unpublished data). As for possibility (2), it was recently reported that maternal Xwnt11 is also an activator for canonical Wnt signaling required for dorsal axis formation upstream of b-catenin, suggesting that Xwnt11 changes its activity in a developmental context dependent fashion probably through some cofactors such as FRL1 (Tao et al., 2005). Therefore, it is possible that zygotic Wnt11 can activate the canonical Wnt pathway to some extent, which can be modulated by Crescent. However, because crescent expression is not detected before midblastula stage by Northern blot analysis (Shibata et al., 2001), it is unlikely that maternal Wnt11 is an inhibitory target of Crescent. As for possibility (3), it is interesting to speculate that Crescent and Frzb1 form a complex, as discussed above (Bafico et al., 1999; Dann et al., 2001), in the head region to inhibit canonical Wnt signaling, whereas Crescent also acts as a non-canonical Wnt modulator across the cell in the posterior region as suggested by the fact that Crescent, unlike typical sFRPs including Frzb1 (Jones and Jomary, 2002), lacks a Netrin-like domain which is thought to bind heparan–sulfate proteoglycans, components of the extracellular matrix, and by a previous study showing that Crescent can be secreted effectively into cell culture medium (Pera and De Robertis, 2000).

3 ratio of the Dlx3-negative region (neural plate) were measured. The vertical axes in (g) and (h) indicate arbitrary units and the width per length ratio, respectively. The width and length of the Dlx3-negative region were defined as indicated in the bottom panel (h). Because similar results were obtained from at least two independent experiments, a representative experiment is shown. The number of samples (no) was indicated below the bar graph. Arrows indicate floor plate expression (i–k). (D) cresMO does not affect on the expression of the mesoderm and neural markers. The expression of Xbra (a,b), XmyoDa (c,d), and chordin at stage 11 (e,f), goosecoid at stage 10.5 (g,h), and sox2 at stage 14 (i,j) were examined. nb-gal mRNA was coinjected as a lineage tracer (C)–(E), but distributions of nb-Gal activity does not necessarily correspond to those of MOs because MOs could be more diffusible than mRNAs. crescent mRNA was coinjected for rescue experiments (B) and (C). Amounts of MOs (ng/embryo), or mRNAs (pg/embryo): standard MO (stdMO), 200 (Ca,d,i; D) or 400 (A); cresMO, 200 (A; B; Cb,c,e,f,j,k; E) or 400 (A); 5 mm MO, 400 (A): crescent mRNA, 5–10; nb-gal mRNA, 30. Numbers in each panel indicate the frequency of the represented phenotype from 1 (D), 2 (B, Ck), or 3 (Ci,j) experiments.

1336

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

The anterior neuroectoderm displays negligible CEMs relative to the posterior neural plate (Keller et al., 1989; Winklbauer and Keller, 1996). Because Crescent is expressed in the anterior endomesoderm underlying the anterior neuroectoderm, and can form a complex with Xwnt11 and negatively regulate its activity in neural tissue as described above, it is likely to play a role in the inhibition of CEMs in this region. A similar scenario to explain how CEMs are excluded from the anterior region was proposed previously by Morgan et al. (1999), in which XclpH3, the Xenopus orthologue of mammalian calponin, is expressed in the anterior neural plate and inhibits CEMs, probably by preventing the sliding of actin filaments over a myosin substrate (Shirinsky et al., 1992). Expression of XclpH3 is directly upregulated by Xotx2 and its expression domain overlaps that of Xotx2 in the anterior region (Morgan et al., 1999). Crescent is not activated by Xotx2 alone nor by a combination of Xotx2 and Xlim1/Ldb1 (M.S., T. Mochizuki, and M.T., unpublished data), whereas it is synergistically upregulated by a combination of Xlim1/Ldb1 and Siamois (Shibata et al., 2000). These data imply that the mechanisms that inhibit CEMs in the anterior neuroectodermal region are generated by several distinct sets of transcription factors. Another possible function of Crescent in the axial mesoderm is to convey Xwnt11 or Xwnt5a anteriorly from the most posterior region (see Fig. 5), and to enhance their activity in the entire posterior region. Enhancement of Xwnt11 activity by Crescent is seemingly contradictory to the morphological observation that the anterior endomesoderm also displays negligible CEMs (Keller et al., 1989; Winklbauer and Keller, 1996). Recently, it was reported that interactions between anterior and posterior mesodermal tissues per se initiate CEMs (Ninomiya et al., 2004), suggesting that the anterior–posterior polarity in addition to the Wnt PCP pathway regulates CEMs in the mesoderm. However, the molecular basis of the anterior–posterior polarity for CEMs is not known yet, and this model does not explain why CEMs is excluded from the anterior mesoderm. Because crescent is specifically expressed in the head organizer region (Shibata et al., 2000), and MO-mediated knockdown of Crescent function leads to shortened body axis with a reduced head (Fig. 9), it is still possible to speculate that Crescent has a role in the segregation of the head organizer from truck organizer. If so, it is interesting to examine the mechanisms in which Crescent inhibits CEMs in the anterior region and enhances them in the posterior region probably by regulating Dsh-independent and Cdc42midiating pathways of non-canonical Wnt signaling. We have shown here for the first time the evidence for functional and physical interactions between the sFRP family members and those Wnts that transduce noncanonical signaling, and for context-dependent modulation of non-canonical signaling and cell movements by Crescent. Further comprehensive analysis of Crescent and Frzb1 should provide clues to better our understanding of the

molecular mechanisms by which the anterior territory is established, and of how CEMs in the mesoderm and neuroectoderm are regulated along the anterior–posterior axis. The role of Crescent in the formation of the heart and pronephros also remains to be clarified.

4. Experimental procedures 4.1. Manipulation of Xenopus embryos Artificial fertilization, culture of embryos, and animal cap assays were performed as described previously (Hikasa et al., 2002; Shibata et al., 2001). Embryos were staged according to the criteria of Nieuwkoop and Faber (Nieuwkoop and Faber, 1967). Animal caps were treated with human activin A (190 pM; Genzyme Techne) until sibling stage 19. 4.2. Plasmid constructs for mRNA injection experiments To generate pCS2C constructs for making mRNA, coding sequences were PCR-amplified and cloned into the pCS2C or pCS2CMT vector. The names of constructs, and the sequences around cloning sites are as follows (cloning sites are underlined, capital letters indicate the first/last coding sequences): pCS2CXcrescent, 5 0 -gaattccacc ATG/ TAG tctaga-3 0 ; pCS2CXwnt3a-Myc, 5 0 -ggatccacc ATG/AAG ggatcc-Myc tags-3 0 ; pCS2CmWnt4-Myc, 5 0 -ggatccacc ATG/CGG aatcgat-Myc tags-3 0 . C-terminally FLAG-tagged constructs, pCS2CXcres-FLAG and pCS2C frzb1-FLAG, were generated by introducing the FLAGcoding sequence into pCS2CXcrescent and pCS2Cfrzb1 constructs, respectively, using an in vitro site-directed mutagenesis system (GeneEditor, Promega). A deletion (nucleotide nos 146–205/amino acid nos 40–59 in the sequence of accession number AF260729) was introduced by the two-round PCR method into the CRD in pCS2C Xcres-FLAG to generate pCS2CXcres-FZD1-FLAG. The absence of PCR-generated mismatch sequences was confirmed by sequencing. We also used pCS2CXwnt5aMyc, pCS2CXwnt8-Myc, and pCS2CXwnt11-Myc, pCS2CExfz7-FLAG (a C-terminally FLAG-tagged Xfz7 ectodomain construct) (Hikasa et al., 2002), pCS2CMTDsh (an N-terminally Myc-tagged Dsh construct) (Sokol, 1996), and pCS105-XBF2, which was isolated as a clone with neuralizing activity in animal cap assays during screening by Osada et al. (2003). The constructs in the pCS2C and pCS2CMT vectors were linearized with NotI, pCS105-XBF2 was linearized with AscI, and mRNA was synthesized as described (Hikasa et al., 2002). mRNA encoding nuclear b-galactosidase (nb-gal) was coinjected at 30 pg per embryo as a lineage tracer, and the enzyme activity of nb-gal was visualized using Red-Gal (Research Organics) as substrate. Some embryos stained with the b-gal

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

reaction were subjected to whole-mount in situ hybridization. 4.3. Whole-mount in situ hybridization Antisense digoxigenin (DIG)- or fluorescein isothiocyanate (FITC)-labelled RNA probes were synthesized using DIG RNA Labeling Mix (Roche) or Fluorescein RNA Labeling Mix (Roche), respectively, with T7 or T3 RNA polymerase. An XmyoDa cDNA clone (1.2-kb insert), which was isolated in our previous study (Shibata et al., 2001), was used as a template for the probe. Whole-mount in situ hybridization was performed manually or with the automated in situ hybridization systems, AIH101 and AIH201 (Aloka), according to the method of Harland (Harland, 1991). Hemisections of embryos were produced as previously described (Shibata et al., 2001). For doublestaining analysis, mixed DIG- and FITC-labelled RNA probes were used. The first staining was performed with anti-FITC antibody conjugated with alkaline phosphatase (AP) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP; Roche), followed by methanol treatment to inactivate AP. The second staining was performed with anti-DIG antibody conjugated with AP and BM Purple (Roche). To photograph specimens, dehydrated embryos were placed in small holes on a 1% agarose plate that had been dehydrated by soaking several times in methanol. 4.4. Sectioning Stained embryos were embedded in paraffin and sectioned at 15 mm. Nuclei were stained with 400 ng/ml DAPI (Sigma) in 2X SSC.

1337

4.6. Design of antisense morpholino oligo The antisense morpholino oligo used to knockdown Crescent expression, cresMO (5 0 -GAAGAACTACTTACTTCTACGGGTA-3 0 ), was purchased from GeneTools LLC. It spans the exon 1–intron 1 boundary of X. laevis crescent genome to block pre-mRNA splicing (Draper et al., 2001). Determination of sequence encompassing exon 1–intron 1 boundary is as follows: after deducing the position of exon 1–exon 2 junction in the X. laevis crescent cDNA by the comparison of Xenopus/ Silurana tropicalis crescent genome (TKS377742.x1 and TKS311489), X. laevis crescent genome DNA fragment covering intron 1 was PCR-amplified with primers specific to exon 1 and exon 2 (forward, 5 0 -TGGCTCTACCCGTAGAA-3 0 ; reverse, 5 0 -AGTCTCTACAGCTGGCT-3 0 ) and sequenced. We used standard oligo (stdMO: CCTCTTACC TCAGTTACAATTTATA) and five mismatched oligo (5 mm: GAAcAAaTACTTAgTTCTAaGGcTA) of cresMO as control.

4.7. RNA isolation and RT-PCR Total RNA was extracted at stage 10 or 11 from MOinjected embryos by proteinase K-phenol extraction as described (Itoh and Sokol, 1997). cDNA was made from DNase-treated RNA using a Superscript first strand synthesis system (Invitrogen). RT-PCR was carried out as previously described (Itoh and Sokol, 1997) to amplify the region spanning exon1–intron1 boundary of crescent cDNA with the following primers: forward, 5 0 -AGCTCCTCAAGATGCATGC; and reverse, 5 0 -AGTCTCTACAGCTG GCT-3 0 .

4.5. Immunoprecipitation and western blotting Acknowledgements Coimmunoprecipitation and western blotting of Wnt proteins were performed according to the method of Djiane et al. (2000) with some modifications (Hikasa et al., 2002). Embryos were coinjected with 500 pg mRNA encoding Myc-tagged Wnts together with mRNA encoding CrescentFLAG, Cres-FZD1-F, Frzb1-FLAG, or Exfz7-FLAG, in the animal pole region at the two-cell stage. Detection of phosphorylated Dsh by immunoprecipitation and western blotting was carried out basically according to Tada et al. (Heisenberg et al., 2000), except that anti-mouse Ig monoclonal antibody conjugated with Alexa Fluor 680 (Molecular Probes) was used as secondary antibody, and blotted membranes were visualized with the infrared imaging system, Odyssey (Aloka). The intensities of FLAG-tagged Crescent and Frzb1, and Myc-tagged Wnt protein bands visualized by chemiluminescence were estimated using a film scanner (Epson) and Basic Quantifier software (Genomic Solutions).

We thank C. Niehrs for providing us with pRN3Xwnt11; M. Moos for pCS2Cfrzb1; D. Turner, R. Rupp, and J. Lee for pCS2C, pCS2CMT, and pCS2Cnb-gal; R. Moon for PE7-Xwnt5a, pGEM2-Xwnt3a, and pCS2C Xwnt8; X. He for pCS2CmWnt4; E. De Robertis for SK(-) gsc; M. Pannese for pGEM3-Xotx2; P. Good for nrp1; Y. Sasai for SK(K)chordin, pCS2Csox2, and Xhairy2a; S. Sokol for pCS2CMT-Xdsh; D.-L. Shi for pSP64TCdc42T17N; J. Smith for pSP73-Xbra; M. Jamrich for pBS-Rx2A; F. Watt for the MZ15 antibody; T. Sargent for SK(K)Dlx3, and T. Nohno and Y. Sasai for double in situ hybridization protocols. This work was supported in part by a Grant-in-Aid for Science Research from the Ministry of Education, Science, Sports, and Culture of Japan; and by the Toray Science Foundation, Japan. M.S. and H.H are Research Fellows of the Japan Society for the Promotion of Science.

1338

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339

References Bafico, A., Gazit, A., Pramila, T., Finch, P.W., Yaniv, A., Aaronson, S.A., 1999. Interaction of frizzled related protein (FRP) with Wnt ligands and the frizzled receptor suggests alternative mechanisms for FRP inhibition of Wnt signaling. J. Biol. Chem. 274, 16180–16187. Bradley, L., Sun, B., Collins-Racie, L., LaVallie, E., McCoy, J., Sive, H., 2000. Different activities of the frizzled-related proteins frzb2 and sizzled2 during Xenopus anteroposterior patterning. Dev. Biol. 227, 118–132. Cho, K.W., Blumberg, B., Steinbeisser, H., De Robertis, E.M., 1991. Molecular nature of Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid. Cell 67, 1111–1120. Choi, S.C., Han, J.K., 2002. Xenopus Cdc42 regulates convergent extension movements during gastrulation through Wnt/Ca2C signaling pathway. Dev. Biol. 244, 342–357. Christian, J.L., Moon, R.T., 1993. Interactions between Xwnt-8 and Spemann organizer signaling pathways generate dorsoventral pattern in the embryonic mesoderm of Xenopus. Genes Dev. 7, 13–28. Dann, C.E., Hsieh, J.C., Rattner, A., Sharma, D., Nathans, J., Leahy, D.J., 2001. Insights into Wnt binding and signalling from the structures of two Frizzled cysteine-rich domains. Nature 412, 86–90. Dennis, S., Aikawa, M., Szeto, W., d’Amore, P.A., Papkoff, J., 1999. A secreted frizzled related protein, FrzA, selectively associates with Wnt1 protein and regulates wnt-1 signaling. J Cell Sci. 21 (Pt 21), 3815– 3820. Djiane, A., Riou, J., Umbhauer, M., Boucaut, J., Shi, D., 2000. Role of frizzled 7 in the regulation of convergent extension movements during gastrulation in Xenopus laevis. Development 127, 3091–3100. Draper, B.W., Morcos, P.A., Kimmel, C.B., 2001. Inhibition of zebrafish fgf8 pre-mRNA splicing with morpholino oligos: a quantifiable method for gene knockdown. Genesis 30, 154–156. Du, S.J., Purcell, S.M., Christian, J.L., McGrew, L.L., Moon, R.T., 1995. Identification of distinct classes and functional domains of Wnts through expression of wild-type and chimeric proteins in Xenopus embryos. Mol. Cell. Biol. 15, 2625–2634. Feledy, J.A., Beanan, M.J., Sandoval, J.J., Goodrich, J.S., Lim, J.H., Matsuo-Takasaki, M., et al., 1999. Inhibitory patterning of the anterior neural plate in Xenopus by homeodomain factors Dlx3 and Msx1. Dev. Biol. 212, 455–464. Finch, P.W., He, X., Kelley, M.J., Uren, A., Schaudies, R.P., Popescu, N.C., et al., 1997. Purification and molecular cloning of a secreted, Frizzledrelated antagonist of Wnt action. Proc. Natl Acad. Sci. USA 94, 6770– 6775. Habas, R., Dawid, I.B., He, X., 2003. Coactivation of Rac and Rho by Wnt/Frizzled signaling is required for vertebrate gastrulation. Genes Dev. 17, 295–309. Harland, R.M., 1991. In situ hybridization: an improved whole-mount method for Xenopus embryos. In: Kay, B.K., Peng, H.B. (Eds.), Methods in Cell Biology, vol. 36. Academic Press, San Diego, CA, pp. 685–695. Harland, R., Gerhart, J., 1997. Formation and function of Spemann’s organizer. Annu. Rev. Cell Dev. Biol. 13, 611–667. He, X., 2004. Wnt signaling went derailed again: a new track via the LIN18 receptor? Cell 118, 668–670. Heisenberg, C.P., Tada, M., Rauch, G.J., Saude, L., Concha, M.L., Geisler, R., et al., 2000. Silberblick/Wnt11 mediates convergent extension movements during zebrafish gastrulation. Nature 405, 76–81. Hikasa, H., Taira, M., 2001. A Xenopus homolog of a human p53-activated gene, PA26, is specifically expressed in the notochord. Mech. Dev. 100, 309–312. Hikasa, H., Shibata, M., Hiratani, I., Taira, M., 2002. The Xenopus receptor tyrosine kinase Xror2 modulates morphogenetic movements of the axial mesoderm and neuroectoderm via Wnt signalling. Development 129, 5227–5239.

Hopwood, N.D., Pluck, A., Gurdon, J.B., 1989. MyoD expression in the forming somites is an early response to mesoderm induction in Xenopus embryos. Eur Mol. Biol. Org. J. 8, 3409–3417. Hsieh, J.C., Rattner, A., Smallwood, P.M., Nathans, J., 1999. Biochemical characterization of Wnt-frizzled interactions using a soluble, biologically active vertebrate Wnt protein. Proc. Natl Acad. Sci. USA 96, 3546–3551. Itasaki, N., Jones, C.M., Mercurio, S., Rowe, A., Domingos, P.M., Smith, J.C., Krumlauf, R., 2003. Wise, a context-dependent activator and inhibitor of Wnt signalling. Development 130, 4295–4305. Itoh, K., Sokol, S.Y., 1997. Graded amounts of Xenopus dishevelled specify discrete anteroposterior cell fates in prospective ectoderm. Mech. Dev. 61, 113–125. Jones, S.E., Jomary, C., 2002. Secreted Frizzled-related proteins: searching for relationships and patterns. Bioessays 24, 811–820. Keller, R., Cooper, M.S., Danilchik, M., Tibbetts, P., Wilson, P.A., 1989. Cell intercalation during notochord development in Xenopus laevis. J. Exp. Zoolog. 251, 134–154. Keller, R., Shih, J., Domingo, C., 1992. The patterning and functioning of protrusive activity during convergence and extension of the Xenopus organiser. Dev. Suppl., 81–91. Kilian, B., Mansukoski, H., Barbosa, F.C., Ulrich, F., Tada, M., Heisenberg, C.P., 2003. The role of Ppt/Wnt5 in regulating cell shape and movement during zebrafish gastrulation. Mech. Dev. 120, 467–476. Ku, M., Melton, D.A., 1993. Xwnt-11: a maternally expressed Xenopus wnt gene. Development 119, 1161–1173. Kuhl, M., Sheldahl, L.C., Malbon, C.C., Moon, R.T., 2000. Ca(2C) /calmodulin-dependent protein kinase II is stimulated by Wnt and Frizzled homologs and promotes ventral cell fates in Xenopus. J. Biol. Chem. 275, 12701–12711. Latinkic, B.V., Mercurio, S., Bennett, B., Hirst, E.M., Xu, Q., Lau, L.F., et al., 2003. Xenopus Cyr61 regulates gastrulation movements and modulates Wnt signalling. Development 130, 2429–2441. Leyns, L., Bouwmeester, T., Kim, S.H., Piccolo, S., De Robertis, E.M., 1997. Frzb-1 is a secreted antagonist of Wnt signaling expressed in the Spemann organizer. Cell 88, 747–756. Lin, K., Wang, S., Julius, M.A., Kitajewski, J., Moos Jr.., M., Luyten, F.P., 1997. The cysteine-rich frizzled domain of Frzb-1 is required and sufficient for modulation of Wnt signaling. Proc. Natl Acad. Sci. USA 94, 11196–11200. Mariani, F.V., Harland, R.M., 1998. XBF-2 is a transcriptional repressor that converts ectoderm into neural tissue. Development 125, 5019– 5031. Mathers, P.H., Grinberg, A., Mahon, K.A., Jamrich, M., 1997. The Rx homeobox gene is essential for vertebrate eye development. Nature 387, 603–607. McGrew, L.L., Otte, A.P., Moon, R.T., 1992. Analysis of Xwnt-4 in embryos of Xenopus laevis: a Wnt family member expressed in the brain and floor plate. Development 115, 463–473. Mizuseki, K., Kishi, M., Matsui, M., Nakanishi, S., Sasai, Y., 1998. Xenopus Zic-related-1 and Sox-2, two factors induced by chordin, have distinct activities in the initiation of neural induction. Development 125, 579–587. Morgan, R., Hooiveld, M.H., Pannese, M., Dati, G., Broders, F., Delarue, M., et al., 1999. Calponin modulates the exclusion of Otx-expressing cells from convergence extension movements. Nat. Cell Biol. 1, 404– 408. Nieuwkoop, P.D., Faber, J., 1967. Normal Table of Xenopus laevis (Daudin). North Holland, Amsterdam. Ninomiya, H., Elinson, R.P., Winklbauer, R., 2004. Antero-posterior tissue polarity links mesoderm convergent extension to axial patterning. Nature 430, 364–367. Oishi, I., Suzuki, H., Onishi, N., Takada, R., Kani, S., Ohkawara, B., et al., 2003. The receptor tyrosine kinase Ror2 is involved in non-canonical Wnt5a/JNK signalling pathway. Genes Cells 8, 645–654.

M. Shibata et al. / Mechanisms of Development 122 (2005) 1322–1339 Osada, S., Ohmori, S.Y., Taira, M., 2003. XMAN1, an inner nuclear membrane protein, antagonizes BMP signaling by interacting with Smad1 in Xenopus embryos. Development 130, 1783–1794. Pandur, P., Lasche, M., Eisenberg, L.M., Kuhl, M., 2002. Wnt-11 activation of a non-canonical Wnt signalling pathway is required for cardiogenesis. Nature 418, 636–641. Pannese, M., Polo, C., Andreazzoli, M., Vignali, R., Kablar, B., Barsacchi, G., Boncinelli, E., 1995. The Xenopus homologue of Otx2 is a maternal homeobox gene that demarcates and specifies anterior body regions. Development 121, 707–720. Penzo-Mendez, A., Umbhauer, M., Djiane, A., Boucaut, J.C., Riou, J.F., 2003. Activation of Gbetagamma signaling downstream of Wnt11/Xfz7 regulates Cdc42 activity during Xenopus gastrulation. Dev. Biol. 257, 302–314. Pera, E.M., De Robertis, E.M., 2000. A direct screen for secreted proteins in Xenopus embryos identifies distinct activities for the Wnt antagonists Crescent and Frzb-1. Mech. Dev. 96, 183–195. Pfeffer, P.L., De Robertis, E.M., Izpisua-Belmonte, J.C., 1997. Crescent, a novel chick gene encoding a Frizzled-like cysteine-rich domain, is expressed in anterior regions during early embryogenesis. Int. J. Dev. Biol. 41, 449–458. Rattner, A., Hsieh, J.C., Smallwood, P.M., Gilbert, D.J., Copeland, N.G., Jenkins, N.A., Nathans, J., 1997. A family of secreted proteins contains homology to the cysteine-rich ligand-binding domain of frizzled receptors. Proc. Natl Acad. Sci. USA 94, 2859–2863. Schneider, V.A., Mercola, M., 2001. Wnt antagonism initiates cardiogenesis in Xenopus laevis. Genes Dev. 15, 304–315. Sheldahl, L.C., Park, M., Malbon, C.C., Moon, R.T., 1999. Protein kinase C is differentially stimulated by Wnt and Frizzled homologs in a G-protein-dependent manner. Curr. Biol. 9, 695–698. Shibata, M., Ono, H., Hikasa, H., Shinga, J., Taira, M., 2000. Xenopus crescent encoding a Frizzled-like domain is expressed in the Spemann organizer and pronephros. Mech. Dev. 96, 243–246. Shibata, M., Itoh, M., Ohmori, S.Y., Shinga, J., Taira, M., 2001. Systematic screening and expression analysis of the head organizer genes in Xenopus embryos. Dev. Biol. 239, 241–256. Shirinsky, V.P., Biryukov, K.G., Hettasch, J.M., Sellers, J.R., 1992. Inhibition of the relative movement of actin and myosin by caldesmon and calponin. J. Biol. Chem. 267, 15886–15892. Slusarski, D.C., Corces, V.G., Moon, R.T., 1997. Interaction of Wnt and a Frizzled homologue triggers G-protein-linked phosphatidylinositol signalling. Nature 390, 410–413.

1339

Smith, J.L., Schoenwolf, G.C., 1998. Getting organized: new insights into the organizer of higher vertebrates. Curr. Top. Dev. Biol. 40, 79–110. Smith, J.C., Price, B.M., Green, J.B., Weigel, D., Herrmann, B.G., 1991. Expression of a Xenopus homolog of Brachyury (T) is an immediateearly response to mesoderm induction. Cell 67, 79–87. Sokol, S.Y., 1996. Analysis of Dishevelled signalling pathways during Xenopus development. Curr. Biol. 6, 1456–1467. Tahinci, E., Symes, K., 2003. Distinct functions of Rho and Rac are required for convergent extension during Xenopus gastrulation. Dev. Biol. 259, 318–335. Tao, Q., Yokota, C., Puck, H., Kofron, M., Birsoy, B., Yan, D., et al., 2005. Maternal wnt11 activates the canonical wnt signaling pathway required for axis formation in Xenopus embryos. Cell 120, 857–871. Ungar, A.R., Kelly, G.M., Moon, R.T., 1995. Wnt4 affects morphogenesis when misexpressed in the zebrafish embryo. Mech. Dev. 52, 153–164. Uren, A., Reichsman, F., Anest, V., Taylor, W.G., Muraiso, K., Bottaro, D.P., et al., 2000. Secreted frizzled-related protein-1 binds directly to Wingless and is a biphasic modulator of Wnt signaling. J. Biol. Chem. 275, 4374–4382. Vonica, A., Gumbiner, B.M., 2002. Zygotic Wnt activity is required for Brachyury expression in the early Xenopus laevis embryo. Dev. Biol. 250, 112–127. Wallingford, J.B., Harland, R.M., 2001. Xenopus Dishevelled signaling regulates both neural and mesodermal convergent extension: parallel forces elongating the body axis. Development 128, 2581–2592. Wallingford, J.B., Rowning, B.A., Vogeli, K.M., Rothbacher, U., Fraser, S.E., Harland, R.M., 2000. Dishevelled controls cell polarity during Xenopus gastrulation. Nature 405, 81–85. Winklbauer, R., Keller, R.E., 1996. Fibronectin, mesoderm migration, and gastrulation in Xenopus. Dev. Biol. 177, 413–426. Winklbauer, R., Medina, A., Swain, R.K., Steinbeisser, H., 2001. Frizzled-7 signalling controls tissue separation during Xenopus gastrulation. Nature 413, 856–860. Wolda, S.L., Moody, C.J., Moon, R.T., 1993. Overlapping expression of Xwnt-3A and Xwnt-1 in neural tissue of Xenopus laevis embryos. Dev. Biol. 155, 46–57. Xu, Q., D’Amore, P.A., Sokol, S.Y., 1998. Functional and biochemical interactions of Wnts with FrzA, a secreted Wnt antagonist. Development 125, 4767–4776.