Fgf signaling negatively regulates Nodal-dependent endoderm induction in zebrafish

(1)

Fgf signaling negatively regulates Nodal-dependent

endoderm induction in zebrafish

Takamasa Mizoguchi, Toshiaki Izawa, Atsushi Kuroiwa, Yutaka Kikuchi

⁎

Division of Biological Science, Graduate School of Science, Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan Received for publication 7 March 2006; revised 22 August 2006; accepted 25 August 2006

Available online 9 September 2006

Abstract

In zebrafish development, Nodal signaling is critical for the induction of endoderm and mesoderm. Three transcription factors downstream of Nodal, Bonnie and Clyde (Bon), Faust (Fau)/Gata5 and Casanova (Cas), are required for endoderm induction. However, it is not yet fully understood how the Nodal signaling pathway regulates the decision process of endoderm and mesoderm induction. In this study, we focused on Fgf signaling, downstream of Nodal signaling, during endoderm induction. We found that activation of Fgf signaling decreases the number of cas-expressing endodermal cells. Conversely, inhibition of this signaling increases the number of endodermal cells without affecting the expression of Nodal, Nodal antagonists, bon or fau/gata5. Inhibition of Fgf signaling in endoderm mutants suggests that this signaling negatively regulates cas expression by a pathway parallel to Bon and Fau/Gata5 in the molecular cascade leading to endoderm. Furthermore, activation of Fgf signaling can overcome Cas-mediated abrogation of mesodermal gene expression. Altogether, these results suggest that Fgf signaling negatively regulates endoderm induction, possibly through repression of cas expression and down-regulation of Cas function.

Keywords: Zebrafish; Endoderm; Induction; Fgf; bonnie and clyde; faust/gata5; casanova

Introduction

During gastrulation in early vertebrate development, the undifferentiated embryonic cells become committed to the three principal germ layers, ectoderm, mesoderm and endoderm. The endodermal germ layer gives rise to the epithelial lining of the respiratory and gastrointestinal tracts. In zebrafish, fate mapping analyses has shown that the endodermal progenitors arise from the four most marginal blastomeric tiers of the late blastula stage embryo (Warga and Nüsslein-Volhard, 1999). In addition, although at the mid-blastula stage single progenitors give rise to both endodermal and mesodermal cells, by the late-blastula stage these cells have been committed to either an endodermal or mesodermal fate (Kimmel et al., 1990). This indicates that the endoderm and mesoderm are intermingled to some degree along the margin of the blastoderm. In contrast to zebrafish, the vegetal blastomeres of Xenopus embryos contribute specifically to endoderm, and mesoderm is derived from the equatorial

region. Fate mapping of Xenopus embryos has shown that endodermal fate is segregated from both mesodermal and ectodermal fate at a relatively early stage compared to other vertebrates (Fukuda and Kikuchi, 2005).

Nodal, a TGF-β super family member, is a crucial factor for both endodermal and mesodermal induction in Xenopus, mouse and zebrafish (Schier, 2003). In zebrafish, two Nodal-related genes, cyclops (cyc) and squint (sqt), are expressed throughout the marginal domain at both the blastula and early gastrula stages (Erter et al., 1998; Feldman et al., 1998; Rebagliati et al., 1998; Sampath et al., 1998). The EGF-CFC family gene one-eyed-pinhead (oep) encodes a co-receptor of Nodal and is necessary for Nodal signaling (Zhang et al., 1998). All endodermal and most mesodermal lineages are lost in cyc;sqt double mutants and MZoep mutants, which lack both maternal and zygotic oep expression, indicating that Nodal signaling is both essential for endoderm induction and important for mesoderm induction in zebrafish (Feldman et al., 1998; Gritsman et al., 1999). Although the Nodal signaling pathway is activated in marginal blasto-meres, these cells are committed to either endodermal or mesodermal fate at the late blastula stage. The mechanisms

⁎ Corresponding author. Fax: +81 52 789 2995.

E-mail address:[email protected](Y. Kikuchi).

(2)

underlying this segregation of endodermal and mesodermal fates in the marginal blastomeres remain to be elucidated.

Genetic analyses in zebrafish have revealed that endoderm induction is regulated by three genes downstream of Nodal signaling: bonnie and clyde (bon, encoding a member of the Mix family of paired-type homeodomain transcription factors) (Kikuchi et al., 2000), faust (fau, encoding the transcription factor Gata5) (Reiter et al., 1999) and casanova (cas, encoding a Sox-related transcription factor) (Dickmeis et al., 2001; Kikuchi et al., 2001; Sakaguchi et al., 2001). Homozygous embryos containing a bon or fau mutation contain∼10% or ∼60% of the normal number of endodermal cells, respectively, as assessed by early endoderm marker genes, sox17 (encodes a high-mobility-group (HMG) transcription factor) (Alexander and Stainier, 1999), and foxA2 (encodes a winged helix/forkhead transcrip-tion factor, and formerly known as axial) (Strähle et al., 1993; Kikuchi et al., 2000; Reiter et al., 2001). Additionally, another Mix-type transcription factor, Mezzo, has been identified as a partially redundant factor of Bon during endoderm induction (Poulain and Lepage, 2002). In cas mutants, all sox17- and foxA2-expressing endodermal cells, and later all gut-derived organs, are completely lost (Alexander et al., 1999). Moreover, overexpression of cas can transform a mesodermal cell fate into an endodermal fate, indicating that it is sufficient for endoderm induction in zebrafish (Kikuchi et al., 2001). Gene expression and misexpression experiments in both wild-type and mutant embryos have revealed that Bon, Fau/Gata5 and Mezzo function downstream of Nodal signaling and upstream of Cas, which in turn regulates the expression of sox17 and foxA2 (Stainier, 2002; Fukuda and Kikuchi, 2005).

cas is the earliest marker of endoderm cells and cas-expressing endodermal cells are distributed around the marginal domain in a characteristic salt-and-pepper pattern (Alexander and Stainier, 1999; Kikuchi et al., 2001, 2004). This expression pattern suggests that the induction of endoderm and mesoderm is regulated by an inhibition mechanism as well as an activation mechanism and that both mechanisms lead to the segregation of mesendodermal progenitors. We have previously tested whether a lateral inhibition mechanism via Notch signaling plays a role in formation of the salt-and-pepper pattern of cas-, sox17- and foxA2-expressing endodermal cells. We showed that the activation of Notch signaling leads to a reduction in the number of cells expressing cas, sox17 and foxA2, whereas inhibition of this signaling pathway does not increase the number of endodermal cells (Kikuchi et al., 2004). These data suggest that, although Notch signaling has the potential to regulate the endoderm induction, the lateral inhibition of Notch signaling is unlikely to be a critical mechanism for either the specification or segregation of endoderm and mesoderm.

Nodal and its downstream transcription factors, Bon, Fau/ Gata5, Mezzo and Cas, function as positive regulators for endoderm induction (Stainier, 2002; Fukuda and Kikuchi, 2005), whereas negative regulatory mechanisms have not be identified in zebrafish. A recent study using Xenopus embryos has shown that activation and inhibition of Fgf signaling in whole embryos down-regulates and enhances endoderm forma-tion, respectively (Cha et al., 2004). These data, combined with

previous results propose the following model: maternal factors including T-box transcription factor VegT induce TGF-β ligands (Shivdasani, 2002; Stainier, 2002; Fukuda and Kikuchi, 2005) and these ligands subsequently activate Fgf signaling in the equatorial region. Fgf signaling may then function as an inhibitory factor to restrict the expansion of endoderm in the equatorial region (Cha et al., 2004). In zebrafish, both Nodal and Fgf genes are expressed in the marginal domain from which both endodermal and mesodermal cells arise. A recent report has shown that Fgf is required for the induction of oep gene expression, downstream of Nodal signaling, and that a combination of Nodal and Fgf signaling operates to maintain mesodermal cell populations in zebrafish (Mathieu et al., 2004). However, little is known about the function of Fgf signaling in mesendoderm induction and about how Fgf signaling interacts with the molecular cascade leading to the formation of the endoderm in zebrafish embryos.

In this study, we focus on Fgf signaling downstream of Nodal during endoderm induction. We show that the activation of Fgf signaling decreases the number of cas- and foxA2-expressing endodermal cells. Conversely, inhibition of this pathway increases the number of endodermal cells without affecting the expression of Nodal, Nodal antagonists and either bon or fau/ gata5. In addition, inhibition of Fgf signaling can increase the number of cas-expressing endodermal cells in bon and fau mutants, suggesting that it represses the expression of cas by a pathway parallel to Bon and Fau/Gata5 in the molecular cascade leading to endoderm formation. We further found that activation of Fgf signaling in marginal cells suppresses Cas function, thus reducing the expression of mesodermal genes. Our results suggest that Fgf signaling negatively regulates endoderm induction, possibly through repression of cas expression and down-regulation of Cas function in zebrafish embryos.

Materials and methods Zebrafish strains

Adult zebrafish and zebrafish embryos were maintained as described previously by Westerfield (1995). Embryos were incubated in 1/3 Ringer's solution (39 mM NaCl, 0.97 mM KCl, 1. 8 mM CaCl2, 1. 7 mM HEPES, pH 7.2)

at 28.5°C and staged according toKimmel et al. (1995). The following mutant alleles were used; bonm425(Kikuchi et al., 2000) and fautm236a(Chen et al., 1996; Reiter et al., 1999). Genotyping of homozygote mutant embryos was performed as described previously (Kikuchi et al., 2000; Reiter et al., 1999).

Whole-mount in situ hybridization andβ-galactosidase staining Whole-mount in situ hybridization was performed as described previously (Westerfield, 1995) and riboprobes were prepared according to published instructions. β-Galactosidase staining was performed as already described (Sanes et al., 1986) except that Red-Gal® (Research Organics) was used as the substrate. For double staining experiments, whole-mount in situ hybridization was performed following β-galactosidase staining as previously described (Segawa et al., 2001).

mRNA injections

Capped mRNA was synthesized using the SP6 mMESSAGE mMACHINE system (Ambion) from the following previously described templates: fgf8 (Fürthauer et al., 1997), a constitutively active form of Xenopus mek (CA-mek;

(3)

The Ser218 and Ser222 in the Xenopus MEK are mutated to Asp and Glu, respectively) (Fukuda et al., 1997), oep (Zhang et al., 1998), dominant-negative form of Xenopus Fgf receptor (XFD) (Amaya et al., 1991), bon (Kikuchi et al., 2000), fau/gata5 (Reiter, et al., 1999) and cas (Kikuchi et al., 2001). Embryos were injected with the following mRNAs at the 1- to 4-cell stage: fgf8 (10 pg), CA-mek (100 pg), XFD (500 pg), oep (50 pg), bon (100 pg) and fau/gata5 (60 pg). Injection of CA-mek (10 pg) and/or cas (50 pg) mRNAs, with nuclear lacZ (nlacZ) (200 pg), was performed into one blastomere at the 16-cell stage. SU5402 treatments

SU5402 treatments were performed as described previously (Raible and Brand, 2001). Briefly, wild-type (WT) zebrafish embryos were dechorionated in 1/3 Ringer's solution at the one cell stage. These embryos were then treated with 30μM SU5402 (Calbiochem, dissolved in DMSO) or with an equivalent amount of DMSO from the one cell stage until the fixation stage. All DMSO- or SU5402-treated embryos were injected with oep mRNA (50 pg) at the one cell stage.

Results

Endodermal cell numbers are increased by the inhibition of Fgf signaling

To investigate the role of Fgf signaling in endoderm induction, we examined both the activation and inhibition of this signaling pathway. We found that the activation of Fgf signaling by overexpression of either fgf8 or CA-mek mRNA decreases the number of both cas- and foxA2-expressing endoderm cells (Figs. 1A–F, S). The percentage reduction of

the average number of cas-expressing endodermal cells in the CA-mek-injected embryos is 60%, whereas the reduction in the average number of foxA2-expressing endoderm cells in the same experiments is only 30% (Fig. 1S). This suggests that the induction of cas is more sensitive than foxA2 to the activating effects of Fgf signaling. To test the effects of the inhibition of Fgf signaling on endoderm formation, we used SU5402, a known pharmacological inhibitor of Fgf receptor 1 ( Moham-madi et al., 1997). A recent report showed that Fgf is required for the induction of oep (Mathieu et al., 2004), encoding an essential co-receptor of Nodal. We therefore injected oep mRNA into SU5402-treated embryos to rescue the Nodal

signaling pathway. Because Oep is known to be a permissive factor for Nodal signaling (Zhang et al., 1998), it was expected that the overexpression of oep mRNA would not alter the number of endodermal cells. Indeed, overexpression of oep mRNA did not change the number of cells expressing cas or foxA2 (Figs. 1G, K).

In contrast, the inhibition of Fgf signaling was found to significantly increase the number of cas- and foxA2-expressing endodermal cells (Figs. 1I, J, M, N, T). We found that inhibition of Fgf signaling by overexpression of a dominant-negative form of Fgf receptor (XFD) also increases the number of endodermal cells, when Oep is supplied (Fig. 1U). High magnification views demonstrate that the number of cas- or foxA2-expressing endodermal cells around the marginal domain is increased in SU5402-treated embryos (Figs. 1J, N). In addition, the cas-expressing endodermal cells were found to be located within 3 cell diameters of the blastoderm margin in DMSO-treated embryos (Fig. 1H), but this was increased to up to 7 cell diameters by SU5402 treatment (Fig. 1J). Conversely, expres-sion of the pan-mesodermal marker, no tail (ntl) ( Schulte-Merker et al., 1994), and the adaxial mesodermal maker, myoD (Weinberg et al., 1996), is down-regulated during late gastrulation in SU5402-treated embryos (Figs. 1P, R).

The expression levels of Nodal, Nodal antagonists, bon and fau/gata5 are unaffected by Fgf signaling

It is possible that an increase in the number of endodermal cells upon the inhibition of Fgf signaling is due to a change in expression of Nodal genes, such as cyc and sqt, or Nodal antagonist genes, lefty1 and lefty2 (Bisgrove et al., 1999; Thisse and Thisse, 1999). Each of these factors can alter the number of endoderm cells when they are overexpressed or inhibited (Agathon et al., 2001; Schier, 2003). To test this possibility, we examined the expression patterns of these genes in the marginal domain of the SU5402 treated or untreated, Oep supplied zebrafish embryo at the 40% epiboly stage. No change in expression was observed for either patterns or the levels of these genes (Figs. 2A–H). We further observed that the expression

Fig. 1. Fgf signaling can regulate the formation of endoderm and mesoderm. (A–F) Zebrafish embryos injected with fgf8 or CA-mek mRNA were examined for cas or foxA2 expression at the 60% epiboly stage (7 h postfertilization [hpf]). Lateral views, dorsal to the right. The number of cas- or foxA2-expressing cells is decreased in embryos in which Fgf signaling is activated. (G–N) Embryos treated with DMSO or SU5402 were examined for cas or foxA2 expression at either the 40% (5 hpf) or 60% epiboly stage (7 hpf), respectively. DMSO, SU5402: oep mRNA was injected into WT embryos at the one cell stage before treatment with DMSO or SU5402. (G–J) Lateral views, anterior to the top. (H and J) Enlarged views of the boxed area in panels G and I, respectively. Arrowheads in panels H and J indicate the number of cas-expressing cells from the margin. SU5402 treatment significantly increases the number of cas-expressing endodermal cells around the marginal domain. (K–N) Lateral views, dorsal to the right. (L and N) Enlarged views of the boxed areas in panels K and M, respectively. The number of foxA2-expressing endodermal cells is also significantly increased by SU5402 treatment. (O–R) Embryos treated with SU5402 were examined for expression of the pan-mesodermal marker ntl and adaxial maker myoD at the 80% epiboly stage (8. 3 hpf). The ntl and myoD expression levels were reduced by SU5402 treatment at the late stages of gastrulation. (S) Effects of the activation of Fgf signaling on the number of cas- or foxA2-expressing endodermal cells. The number of cas- or foxA2-expressing endodermal cells was counted at the 60% epiboly stage on both the left and right lateral sides of the embryos, which we determined to be representative of whole embryos. The embryos analyzed were cas-expressing endodermal cells in WT (n = 37) and injected (n = 45) embryos; foxA2-expressing endodermal cells in WT (n = 52); and CA-mek-injected (n = 38) embryos. Error bars represent the standard error. *P < 0.0005, Student's t test. (T, U) Effects of the inhibition of Fgf signaling by SU5402 treatment (T) or XFD overexpression (U) upon the number of cas- or foxA2-expressing endodermal cells. oep mRNA was injected into WT embryos at the one cell stage before treatment with SU5402 (T) and was co-injected with XFD mRNA at the one cell stage (U). Zebrafish embryos at the 60% epiboly stage (7 hpf) were scored as demonstrating higher cas or foxA2 expression levels if they exhibited a greater than 50% (yellow), 20–50% (red) or 0–20% (blue) increase in endodermal cells. However, because SU5402 treatment or XFD overexpression greatly increases the number of cas- and foxA2-expressing endodermal cells, it is difficult to determine their exact number. The embryos analyzed were (T) cas-expressing endodermal cells in DMSO-treated (n = 37) and SU5402-treated (n = 48) embryos; foxA2-expressing endodermal cells in DMSO-treated (n = 52) and SU5402-treated (n = 51) embryos: (U) cas-foxA2-expressing endodermal cells in WT (n = 28) and XFD-injected (n = 30) embryos; foxA2-expressing endodermal cells in WT (n = 33) and XFD-injected (n = 31) embryos.

(4)

(5)

profiles of these genes were unchanged in CA-mek-injected embryos (data not shown). These data indicate that Fgf signaling does not regulate either Nodal genes or Nodal antagonists, when Oep is supplied.

We additionally found that the expression levels of both bon and fau/gata5, both of which function upstream of cas, are unaffected by either the inhibition or activation of Fgf signaling (Figs. 3A–D, I–L). In contrast, mezzo expression was found to

be absent in SU5402-treated embryos and increased in CA-mek-injected embryos (Figs. 3E–H), suggesting that although Fgf signaling negatively regulates endoderm induction, mezzo expression is positively regulated by this pathway. Recently, we reported that the T-box protein Eomesodermin (Eomes) physically interacts with Bon and Fau/Gata5 to initiate cas transcription (Bjornson et al., 2005). The expression level of eomes was also unchanged following SU5402 treatment and CA-mek injection of late blastula stage embryos (data not shown). These results indicate that Fgf signaling represses

endoderm induction without affecting the expression of Nodal, Nodal antagonists, bon, fau/gata5 and eomes.

Inhibition of Fgf signaling rescues endoderm induction in bon and fau mutants

Our experiments involving the activation or inhibition of Fgf signaling suggest that the negative regulatory pathway mediated by Fgf signaling and the positive regulatory pathway facilitated by Bon and Fau/Gata5 do not interact with each other during the regulation of cas expression. To further address this question, we examined whether the inhibition of Fgf signaling would suppress the endodermal defects in bon and fau mutants. We co-injected oep and XFD mRNAs into either bon or fau hetero-zygote intercrosses and examined cas and foxA2 expression. The number of cas- and foxA2-expressing endodermal cells in both bon and fau mutant embryos was significantly increased by the inhibition of Fgf signaling, compared with that in oep-injected mutant embryos (Fig. 4). However, this inhibition did not restore the number of cas- and foxA2-expressing endodermal cells to a wild-type level in bon mutant embryos (Figs. 4A–C, F–H, K). In contrast, the overexpression of XFD in fau mutants results in an increase in the number of endodermal cells compared with either oep-injected fau mutants or oep-injected WT embryos (Figs. 4A, D, E, F, I, J, L). It thus appears that the different effects of Fgf signaling inhibition between these two mutants reflects the original number of endodermal cells that each contained; bon and fau mutant embryos contain∼10% and ∼60% of the normal number of endodermal cells, respectively (Stainier, 2002).

We next tested whether the overexpression of bon or fau/ gata5 would increase the number of cas-expressing endoder-mal cells in Fgf signaling-inhibited embryos. We found that although cas-expressing endodermal cells are located within 8 cell diameters of the blastoderm margin when Fgf signaling is inhibited (Fig. 5A), co-injection of either bon or fau/gata5 with XFD and oep leads to a further increase in this distance to up to 11 cell diameters (Figs. 5B, C). Moreover, the number of cas-expressing endodermal cells within this 11 cell diameter range was found to be dramatically increased by the over-expression of fau/gata5, XFD and oep (Fig. 5C). Taken together, these results suggest that Fgf signaling represses the expression of cas by a pathway parallel to the Bon and Fau/ Gata5 cascade.

Activation of Fgf signaling restores the expression of mesodermal genes in cas overexpressing cells

cas is an essential gene in endoderm formation and overexpression of cas can transform a mesodermal fate into an endodermal fate (Kikuchi et al., 2001). Consistent with this, embryos injected with both cas and nlacZ in the marginal domain exhibit a reduction in mesodermal marker gene expression (Figs. 6B, E, H, K). It is well known that ntl and spadetail/tbx16 (spt) (Griffin et al., 1998) are pan-mesodermal markers and that paraxial protocadherin (papc) (Yamamoto et al., 1998) and myoD are downstream targets of Ntl and Spt.

Fig. 2. SU5402 treatment does not affect the expression of either Nodal or Nodal antagonists. (A–H) All images are lateral views with anterior to the top. These are magnified views of zebrafish embryos around the marginal region at the 40% epiboly stage (5 hpf). DMSO, SU5402: oep mRNA was injected into WT embryos, and these injected embryos were subsequently treated with either DMSO or SU5402 from the one cell stage to the 40% epiboly stage. (B, D, F, H) The expression of Nodal genes (cyc and sqt) and Nodal antagonists (lefty1 and lefty2) are not affected by SU5402 treatment.

(6)

We tested whether activation of Fgf signaling can rescue the expression of these mesodermal genes in cas misexpressing cells. As shown inFigs. 6C, F, I and L, the expression of all of these mesodermal genes is indeed restored in the marginal domain and adaxial mesoderm following co-injection with cas and CA-mek mRNAs. These data indicate that activation of Fgf signaling can overcome the ability of Cas to abrogate mesodermal gene expression.

Discussion

Fgf signaling represses cas expression by a pathway parallel to Bon and Fau/Gata5

In this study, we show that neither bon nor fau/gata5 expression levels are affected by activation or inhibition of Fgf signaling. Consistent with these results, a previous report has shown that overexpression of either eFgf or XFD does not affect fau/gata5 expression in the marginal domain (Rodaway et al., 1999). In contrast, both bon and fau/gata5 expression are completely absent from the blastomeres at the early gastrulation stage in cyc;sqt double mutants (Alexander and Stainier, 1999; Rodaway et al., 1999). These data indicate that both bon and fau/gata5 expression are regulated by Nodal signaling, but not by the Fgf signaling pathway.

The number of cas-expressing endodermal cells is decreased in both bon and fau mutant embryos and is somewhat increased

around the margin following the overexpression of either bon or fau/gata5 in WT embryos (Kikuchi et al., 2000, 2001; Reiter et al., 2001; Bjornson et al., 2005). This demonstrates that both Bon and Fau/Ggata5 positively regulate cas expression. Our results also clearly indicate that Fgf signaling negatively regulates endoderm induction without affecting the expression of bon or fau/gata5. Moreover, we found that inhibition of Fgf signaling can increase the number of cas-expressing endoder-mal cells in bon and fau mutant embryos and that over-expression of bon or fau/gata5 in Fgf signaling-inhibited embryos further increases the number of cas-expressing endodermal cells. All of these findings indicate that Fgf sig-naling represses cas expression by a pathway that is parallel to the Bon and Fau/Gata5 pathway. Recently, we have shown that Eomes synergistically functions with Bon and Fau/Gata5 to potently induce cas, even in animal pole blastomeres (Bjornson et al., 2005). Eomes physically interacts with both Bon and Fau/ Gata5, and these physical interactions suggest that Eomes promotes endoderm induction in marginal blastomeres by facilitating the assembly of a transcriptional activating complex on the cas promoter (Bjornson et al., 2005). It is therefore possible that Fgf signaling (or downstream effectors of this signaling) inhibits the activation complex containing these transcription factors during cas transcription. Additional analyses of the inhibitory mechanisms mediated by Fgf signaling (or the associated downstream effectors of this signaling) during cas transcription should facilitate further

Fig. 3. Fgf signaling does not affect bon and fau/gata5 expression. (A–L) All images are lateral views with anterior to the top. These are magnified views of zebrafish embryos around the marginal region at the 40% epiboly stage (5 hpf). (A–B, E–F, I–J) DMSO, SU5402: oep mRNAwas injected into WT embryos, and these injected embryos were subsequently treated with either DMSO or SU5402 from the one cell stage to the 40% epiboly stage. Neither the bon nor the fau/gata5 expression levels are affected, whereas mezzo expression is barely detectable, in SU5402-treated embryos. (C–D, G–H, K–L) Overexpression of CA-mek mRNA in WT embryos does not alter the bon or fau/gata5 expression levels, whereas mezzo expression around the marginal region is increased in the CA-mek mRNA-injected embryos.

(7)

investigation and elucidation of the pathways leading to endoderm induction in zebrafish.

Mezzo function in endoderm induction

mezzo encodes a Mix-type transcription factor and is expressed around the marginal domain in the late blastula and gastrula stage embryos. Overexpression of mezzo can ectopically induce cas, sox17 and the pan-mesodermal marker ntl and can partially rescue the bon mutant phenotype (Poulain and Lepage, 2002). In contrast, a previous report has shown that development is not perturbed and that the

expression of sox17 and ntl is normal in mezzo antisense morpholino oligonucleotides-injected embryos (Poulain and Lepage, 2002). In our present study, we found that mezzo expression is absent in Fgf signaling-inhibited embryos, whereas the number of endoderm cells is increased. We also find that the expression of mezzo in the marginal domain is increased in CA-mek-injected embryos, whereas the number of endoderm cells is decreased. These data suggest that although Mezzo has the potential to function during both endoderm and mesoderm formation, its function is not essential for endoderm formation, probably because of a degree of functional redundancy.

Fig. 4. Inhibition of Fgf signaling increases the number of cas- and foxA2-expressing endodermal cells in bon and fau mutants. (A–J) Lateral views of zebrafish embryos with dorsal to the right. bon or fau heterozygote intercrosses were injected with oep and XFD mRNAs at the 1- to 4-cell stage and were examined for cas or foxA2 expression at the 80% epiboly stage (8. 3 hpf). Inhibition of Fgf signaling can increase the number of cas- and foxA2-expressing endodermal cells in both bon and fau mutant embryos. (K) Effects of the inhibition of Fgf signaling on the number of cas- or foxA2-expressing endodermal cells in bon mutant embryos. The number of cas- or foxA2-expressing endodermal cells was counted at the 80% epiboly stage on both the left and right lateral sides of the embryos, which we determined to be representative of whole embryos. The embryos analyzed were cas-expressing endodermal cells in oep-injected WT (n = 21) and oep-injected (n = 15) and oep,XFD-injected (n = 20) bon mutant embryos; foxA2-expressing endodermal cells in oep-injected WT (n = 24) and oep-injected (n = 12) and oep,XFD-injected (n = 23) bon mutant embryos. Error bars represent the standard error. *P < 0.0001, Student's t test. (L) Effects of the inhibition of Fgf signaling upon the number of cas- or foxA2-expressing endodermal cells in fau mutant embryos. Zebrafish embryos at the 80% epiboly stage (8. 3 hpf) were scored as demonstrating higher cas or foxA2 expression levels if they exhibited a greater than 50% (yellow), 20–50% (red) or 0–20% (blue) increase in endodermal cells. However, because oep,XFD mRNAs injection greatly increases the number of cas- and foxA2-expressing endodermal cells, it is difficult to determine their exact number. The embryos analyzed were cas-expressing endodermal cells in oep-injected (n = 14) and oep,XFD-injected (n = 17) fau mutant embryos; foxA2-expressing endodermal cells in oep-injected (n = 12) and oep,XFD-injected (n = 15) fau mutant embryos.

(8)

The expression of cas around the margin

In late blastula zebrafish embryos, the marginal cells in rows 3 to 4 give rise to both endoderm and mesoderm, whereas those between rows 4 and 6 give rise to mesoderm only (Warga and Nüsslein-Volhard, 1999). At the onset of gastrulation, bon transcripts are detectable approximately in marginal cell rows

8 to 10. By contrast, fau/gata5 is expressed in rows 3 to 4 of the marginal cells at the same stage (Reiter et al., 1999; Rodaway et al., 1999), and cas-expressing endodermal cells are seen in only a subset of the fau/gata5-expressing cells (Kikuchi et al., 2001). We have found in our present experiments that Fgf signaling functions to repress cas expression around the margin and that the inhibition of this signaling increases the number of

cas-Fig. 5. Either bon or fau/gata5 overexpression can increase the number of cas-expressing endodermal cells in Fgf signaling-inhibited embryos. (A–C) All images are lateral views with anterior to the top. These are magnified views of zebrafish embryos around the marginal region at the 40% epiboly stage (5 hpf). Arrowheads indicate the number of cas-expressing cells from the margin. (A) Overexpression of XFD and oep in WT embryos increases the number of cas-expressing endodermal cells to up to 8 cell diameters from the margin. (B) The cas-expressing endodermal cells were observed to be increased to 11 cell diameters from the margin by co-injection of XFD and oep with bon into WT embryos. (C) Most of the marginal cells within 11 cell diameters from the margin express cas following co-co-injection of XFD and oep with fau/gata5 into WT embryos.

Fig. 6. The activation of Fgf signaling can restore the expression of mesodermal-specific genes that are down-regulated by cas misexpression. (A–L) Either cas mRNA, or a combination of cas with CA-mek mRNAs, were injected into a marginal blastomere of a 16-cell stage zebrafish embryo. Embryos injected with mRNA (s) were examined for ntl, spt or papc expression at the 50% epiboly stage (5. 3 hpf) and myoD expression at the 80% epiboly stage (8. 3 hpf). Nuclear lacZ (nlacZ) mRNA was used as a lineage tracer. Red staining byβ-galactosidase activity indicates the progeny of injected cells. The progeny of marginal blastomeres gives rise to the marginal clones of the cells. (A–I) All images are of animal pole views. (J–L) Dorsal views with anterior to the top. (A, D, G, J) The nlacZ mRNA injections did not disrupt the ntl, spt, papc or myoD expression profile. (B, E, H, K) The expression of the mesodermal genes ntl, spt, papc or myoD was down-regulated by cas mRNA misexpression (arrows). (C, F, I, L) The down-regulation of the mesodermal genes ntl, spt, papc or myoD by cas mRNA misexpression was restored by co-injection with CA-mek mRNA (arrows).

(9)

expressing endodermal cells to up to 7 or 8 rows of the blastoderm margin. Co-injection of bon or fau/gata5 with XFD and oep leads to a further increase to up to 11 rows, which overlap with the bon expression domain. Moreover, when fau/ gata5, XFD and oep are co-injected into WT embryos, most of the marginal cells within the 11 rows express cas. Our results thus suggest that the marginal cells within these 11 rows have the potential to induce cas and may also have the potential to become endoderm when Fgf signaling is inhibited and Fau/ Gata5 and Oep are supplied. In support of this conclusion, we have shown that Cas can transform a mesodermal fate to an endodermal fate when overexpressed in marginal cells, whereas the fate of ectodermal cells in the animal pole region cannot be changed to an endodermal fate by this overexpression (Kikuchi et al., 2001).

Competition between Cas and Fgf signaling during the induction of endodermal or mesodermal cells

In this study, we have found that the activation of Fgf signaling overcomes the ability of Cas to decrease expression of mesodermal markers. There are two possible explanations for this observation. One is that Fgf signaling directly modifies the Cas protein, which has now been observed to be phosphorylated by MAP kinase (MAPK), leading to a reduction in its transcriptional activity (Poulain et al., 2006). Alternatively, the activation of Fgf signaling may induce downstream targets, which might rescue mesodermal-specific gene expression by inhibition of the transcriptional repression function of Cas. In both of these possible explanations, Fgf signaling negatively regulates Cas function, and Cas and Fgf signaling, both of which are downstream of Nodal, compete with each other to induce either endodermal- or mesodermal-specific gene expression. A model for the specification of endoderm and mesoderm

Based upon our current observations, we propose a model for the mechanisms underlying specification of endoderm and mesoderm in zebrafish embryos (Fig. 7). In this model, Nodal signaling induces fgf via TGF-β receptors and the co-receptor Oep. Fgf signaling is involved in the amplification and propagation of Nodal signaling by the regulation of oep expression (Fig. 7A). Thus, the combination of Nodal and Fgf is necessary for both endoderm and mesoderm induction. Both Fgf and Nodal signaling activate mesodermal genes such as ntl, spt, papc and myoD (Griffin and Kimelman, 2003; Mathieu et al., 2004), whereas Nodal signaling activates bon and fau/gata5, and these two factors and the maternal factor Eomes induce cas expression (Bjornson et al., 2005) (Fig. 7B). Our results suggest that Fgf signaling negatively regulates endoderm induction by inhibiting cas expression and Cas function. In this context, cells that receive much of the Fgf signaling stimuli are fated to become mesoderm, whereas cells in which Fgf signaling is suppressed, and in which Cas is activated, are fated to develop as endoderm. However, in zebrafish embryos, it has been reported by in situ hybridization analyses that endodermal and meso-dermal genes are coexpressed in marginal cells. The expression

of the mesodermal gene ntl overlaps with both bon and fau/ gata5 (Alexander et al., 1999; Rodaway et al., 1999), and double labeling in situ data show that cas is expressed in cells that also express ntl in the marginal domain at the late blastula and early gastrula stage (T. M. and Y. K., unpublished observation). Our current results suggest that the activation levels of Fgf signaling are a key factor during the cell fate decision events between endoderm and mesoderm. Therefore, additional future analyses of the repression mechanisms underlying Fgf signaling at the resolution of a single cell should further our understanding of the molecular events that direct individual marginal blastomeres toward an endodermal fate.

Acknowledgments

We thank Tokuko Niwa, Hiroko Nishii and Mika Yamada for both fish maintenance and technical assistance. We also thank Igor Dawid, Hiroshi Hanafusa, Thierry Lepage, Bernard Thisse, Akihito Yamamoto and Joseph Yost for providing useful DNA templates. We are grateful to Atsuo Kawahara and Hitoshi

Fig. 7. A model for endoderm and mesoderm formation in zebrafish. In late blastula stage zebrafish embryos, there are two steps in endoderm and mesoderm induction. (A) Zygotic oep expression is induced by Fgf signaling. The Nodal-related proteins Cyc and Sqt act through TGF-β type receptors. Oep is maternally supplied and it is an essential co-receptor for Nodal signaling (Gritsman et al., 1999). Nodal signaling induces fgf and this Fgf signaling functions to induce zygotic oep expression (Mathieu et al., 2004). Thus, the combination of Nodal and Fgf signaling is necessary for both endoderm and mesoderm induction. (B) Induction of endoderm and mesoderm by Nodal and Fgf signaling. Nodal signaling activates bon and fau/gata5 expression. Bon, Fau/Gata5 and Eomes cooperatively regulate the expression of cas (Bjornson et al., 2005). *Eomes; eomes is maternally supplied to the zygote. Cas activates sox17 and foxA2 and represses mesodermal genes to determine endodermal fates. On the other hand, both Fgf and Nodal signaling synergistically activate mesodermal genes, such as ntl, spt, papc and myoD. Fgf and MAPK signaling also negatively regulate cas expression and Cas function to determine the mesodermal fate.

(10)

Okamoto for technical advice, Thierry Lepage for sharing his unpublished data and Heather Verkade for careful reading of the manuscript. This work was supported by grants from the Mitsubishi Foundation, Takeda Foundation and Grant-in-Aid for Scientific Research from The Ministry of Education, Science, Sports and Technology (KAKENHI 13138204 and 16027223) and by the Japan Society for the Promotion of Science (KAKENHI 15370094) to Y.K.

References

Agathon, A., Thisse, B., Thisse, C., 2001. Morpholino knock-down of antivin1 and antivin2 upregulates nodal signaling. Genesis 30, 178–182.

Alexander, J., Stainier, D.Y., 1999. A molecular pathway leading to endoderm formation in zebrafish. Curr. Biol. 9, 1147–1157.

Alexander, J., Rothenberg, M., Henry, G.L., Stainier, D.Y., 1999. casanova plays an early and essential role in endoderm formation in zebrafish. Dev. Biol. 215, 343–357.

Amaya, E., Musci, T.J., Kirschner, M.W., 1991. Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66, 257–270.

Bisgrove, B.W., Essner, J.J., Yost, H.J., 1999. Regulation of midline development by antagonism of lefty and nodal signaling. Development 126, 3253–3262.

Bjornson, C.R., Griffin, K.J., Farr III, G.H., Terashima, A., Himeda, C., Kikuchi, Y., Kimelman, D., 2005. Eomesodermin is a localized maternal determinant required for endoderm induction in zebrafish. Dev. Cell 9, 523–533.

Cha, S.W., Hwang, Y.S., Chae, J.P., Lee, S.Y., Lee, H.S., Daar, I., Park, M.J., Kim, J., 2004. Inhibition of FGF signaling causes expansion of the endoderm in Xenopus. Biochem. Biophys. Res. Commun. 315, 100–106. Chen, J.N., Haffter, P., Odenthal, J., Vogelsang, E., Brand, M., van Eeden,

F.J., Furutani-Seiki, M., Granato, M., Hammerschmidt, M., Heisenberg, C.P., et al., 1996. Mutations affecting the cardiovascular system and other internal organs in zebrafish. Development 123, 293–302. Dickmeis, T., Mourrain, P., Saint-Etienne, L., Fischer, N., Aanstad, P., Clark, M.,

Strähle, U., Rosa, F., 2001. A crucial component of the endoderm formation pathway, CASANOVA, is encoded by a novel sox-related gene. Genes Dev. 15, 1487–1492.

Erter, C.E., Solnica-Krezel, L., Wright, C.V., 1998. Zebrafish nodal-related 2 encodes an early mesendodermal inducer signaling from the extraembryonic yolk syncytial layer. Dev. Biol. 204, 361–372.

Feldman, B., Gates, M.A., Egan, E.S., Dougan, S.T., Rennebeck, G., Sirotkin, H.I., Schier, A.F., Talbot, W.S., 1998. Zebrafish organizer development and germ-layer formation require nodal-related signals. Nature 395, 181–185. Fukuda, K., Kikuchi, Y., 2005. Endoderm development in vertebrates: fate

mapping, induction and regional specification. Dev. Growth Differ. 47, 343–355.

Fukuda, M., Gotoh, Y., Nishida, E., 1997. Interaction of MAP kinase with MAP kinase kinase: its possible role in the control of nucleocytoplasmic transport of MAP kinase. EMBO J. 16, 1901–1908.

Fürthauer, M., Thisse, C., Thisse, B., 1997. A role for FGF-8 in the dorsoventral patterning of the zebrafish gastrula. Development 124, 4253–4264. Griffin, K.J., Kimelman, D., 2003. Interplay between FGF, one-eyed pinhead,

and T-box transcription factors during zebrafish posterior development. Dev. Biol. 264, 456–466.

Griffin, K.J., Amacher, S.L., Kimmel, C.B., Kimelman, D., 1998. Molecular identification of spadetail: regulation of zebrafish trunk and tail mesoderm formation by T-box genes. Development 125, 3379–3388.

Gritsman, K., Zhang, J., Cheng, S., Heckscher, E., Talbot, W.S., Schier, A.F., 1999. The EGF-CFC protein one-eyed pinhead is essential for nodal signaling. Cell 97, 121–132.

Kikuchi, Y., Trinh, L.A., Reiter, J.F., Alexander, J., Yelon, D., Stainier, D.Y., 2000. The zebrafish bonnie and clyde gene encodes a Mix family homeodomain

protein that regulates the generation of endodermal precursors. Genes Dev. 14, 1279–1289.

Kikuchi, Y., Agathon, A., Alexander, J., Thisse, C., Waldron, S., Yelon, D., Thisse, B., Stainier, D.Y., 2001. casanova encodes a novel Sox-related protein necessary and sufficient for early endoderm formation in zebrafish. Genes Dev. 15, 1493–1505.

Kikuchi, Y., Verkade, H., Reiter, J.F., Kim, C.H., Chitnis, A.B., Kuroiwa, A., Stainier, D.Y., 2004. Notch signaling can regulate endoderm formation in zebrafish. Dev. Dyn. 229, 756–762.

Kimmel, C.B., Warga, R.M., Schilling, T.F., 1990. Origin and organization of the zebrafish fate map. Development 108, 581–594.

Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B., Schilling, T.F., 1995. Stages of embryonic development of the zebrafish. Dev. Dyn. 203, 253–310. Mathieu, J., Griffin, K., Herbomel, P., Dickmeis, T., Strähle, U., Kimelman, D., Rosa, F.M., Peyriéras, N., 2004. Nodal and Fgf pathways interact through a positive regulatory loop and synergize to maintain mesodermal cell populations. Development 131, 629–641.

Mohammadi, M., McMahon, G., Sun, L., Tang, C., Hirth, P., Yeh, B.K., Hubbard, S.R., Schlessinger, J., 1997. Structures of the tyrosine kinase domain of fibroblast growth factor receptor in complex with inhibitors. Science 276, 955–960.

Poulain, M., Lepage, T., 2002. Mezzo, a paired-like homeobox protein is an immediate target of Nodal signalling and regulates endoderm specification in zebrafish. Development 129, 4901–4914.

Poulain, M., Fürthauer, M., Thisse, B., Thisse, C., Lepage, T., 2006. Zebrafish endoderm formation is regulated by combinatorial Nodal, FGF and BMP signalling. Development 133, 2189–2200.

Raible, F., Brand, M., 2001. Tight transcriptional control of the ETS domain factors Erm and Pea3 by Fgf signaling during early zebrafish development. Mech. Dev. 107, 105–117.

Rebagliati, M.R., Toyama, R., Haffter, P., Dawid, I.B., 1998. cyclops encodes a nodal-related factor involved in midline signaling. Proc. Natl. Acad. Sci. U.S.A. 95, 9932–9937.

Reiter, J.F., Alexander, J., Rodaway, A., Yelon, D., Patient, R., Holder, N., Stainier, D.Y., 1999. Gata5 is required for the development of the heart and endoderm in zebrafish. Genes Dev. 13, 2983–2995.

Reiter, J.F., Kikuchi, Y., Stainier, D.Y., 2001. Multiple roles for Gata5 in zebrafish endoderm formation. Development 128, 125–135.

Rodaway, A., Takeda, H., Koshida, S., Broadbent, J., Price, B., Smith, J.C., Patient, R., Holder, N., 1999. Induction of the mesendoderm in the zebrafish germ ring by yolk cell-derived TGF-α family signals and discrimination of mesoderm and endoderm by FGF. Development 126, 3067–3078. Sakaguchi, T., Kuroiwa, A., Takeda, H., 2001. A novel sox gene, 226D7, acts

downstream of Nodal signaling to specify endoderm precursors in zebrafish. Mech. Dev. 107, 25–38.

Sampath, K., Rubinstein, A.L., Cheng, A.M., Liang, J.O., Fekany, K., Solnica-Krezel, L., Korzh, V., Halpern, M.E., Wright, C.V., 1998. Induction of the zebrafish ventral brain and floorplate requires cyclops/nodal signalling. Nature 395, 185–189.

Sanes, J.R., Rubenstein, J.L., Nicolas, J.F., 1986. Use of a recombinant retrovirus to study post-implantation cell lineage in mouse embryos. EMBO J. 5, 3133–3142.

Schier, A.F., 2003. Nodal signaling in vertebrate development. Annu. Rev. Cell Dev. Biol. 19, 589–621.

Schulte-Merker, S., van Eeden, F.J., Halpern, M.E., Kimmel, C.B., Nüsslein-Volhard, C., 1994. no tail (ntl) is the zebrafish homologue of the mouse T (Brachyury) gene. Development 120, 1009–1015. Segawa, H., Miyashita, T., Hirate, Y., Higashijima, S., Chino, N., Uyemura, K.,

Kikuchi, Y., Okamoto, H., 2001. Functional repression of Islet-2 by disruption of complex with Ldb impairs peripheral axonal outgrowth in embryonic zebrafish. Neuron 30, 423–436.

Shivdasani, R.A., 2002. Molecular regulation of vertebrate early endoderm development. Dev. Biol. 249, 191–203.

Stainier, D.Y., 2002. A glimpse into the molecular entrails of endoderm formation. Genes Dev. 16, 893–907.

Strähle, U., Blader, P., Henrique, D., Ingham, P.W., 1993. Axial, a zebrafish gene expressed along the developing body axis, shows altered expression in cyclops mutant embryos. Genes Dev. 7, 1436–1446.

(11)

Thisse, C., Thisse, B., 1999. Antivin, a novel and divergent member of the TGFα superfamily, negatively regulates mesoderm induction. Development 126, 229–240.

Warga, R.M., Nüsslein-Volhard, C., 1999. Origin and development of the zebrafish endoderm. Development 126, 827–838.

Weinberg, E.S., Allende, M.L., Kelly, C.S., Abdelhamid, A., Murakami, T., Andermann, P., Doerre, O.G., Grunwald, D.J., Riggleman, B., 1996. Developmental regulation of zebrafish MyoD in wild-type, no tail and spadetail embryos. Development 122, 271–280.

Westerfield, M., 1995. The Zebrafish Book. University of Oregon Press, Eugene, OR.

Yamamoto, A., Amacher, S.L., Kim, S.H., Geissert, D., Kimmel, C.B., De Robertis, E.M., 1998. Zebrafish paraxial protocadherin is a downstream target of spa-detail involved in morphogenesis of gastrula mesoderm. Development 125, 3389–3397.

Zhang, J., Talbot, W.S., Schier, A.F., 1998. Positional cloning identifies zebrafish one-eyed pinhead as a permissive EGF-related ligand required during gastrulation. Cell 92, 241–251.