In early Ciona savignyi embryos, nuclear localization of β-catenin is the first step of endodermal cell specification, and triggers the activation of various target genes. A cDNA for Cs-FGF4/6/9, a gene activated downstream of β-catenin signaling, was isolated and shown to encode an FGF protein with features of both FGF4/6 and FGF9/20. The early embryonic expression of Cs-FGF4/6/9 was transient and the transcript was seen in endodermal cells at the 16- and 32-cell stages, in notochord and muscle cells at the 64-cell stage, and in nerve cord and muscle cells at the 110-cell stage; the gene was then expressed again in cells of the nervous system after neurulation. When the gene function was suppressed with a specific antisense morpholino oligo, the differentiation of mesenchyme cells was completely blocked, and the fate of presumptive mesenchyme cells appeared to change into that of muscle cells. The inhibition of mesenchyme differentiation was abrogated by coinjection of the morpholino oligo and synthetic Cs-FGF4/6/9 mRNA. Downregulation of β-catenin nuclear localization resulted in the absence of mesenchyme cell differentiation due to failure of the formation of signal-producing endodermal cells. Injection of synthetic Cs-FGF4/6/9 mRNA in β-catenin-downregulated embryos evoked mesenchyme cell differentiation. These results strongly suggest that Cs-FGF4/6/9 produced by endodermal cells acts an inductive signal for the differentiation of mesenchyme cells. On the other hand, the role of Cs-FGF4/6/9 in the induction of notochord cells is partial; the initial process of the induction was inhibited by Cs-FGF4/6/9 morpholino oligo, but notochord-specific genes were expressed later to form a partial notochord.
INTRODUCTION
Fibroblast growth factors (FGFs) comprise a large family of polypeptide growth factors of about 17-34 kDa, and are found in various animal groups (reviewed by Ornitz, 2000; Ornitz and Itoh, 2001). Vertebrates contain 20 or more members of the FGF family, which share 13-71% amino acid identity. During embryonic development, FGFs play diverse roles in regulating cell proliferation, migration and differentiation (reviewed by Vasiliauskas and Stern, 2001). For example, in vertebrate embryos, FGFs play key roles in mesoderm induction; FGFs can induce muscle, mesenchyme, lateral plate mesoderm and blood islands in Xenopus (Slack, 1994; Yamaguchi and Rossant, 1995). Inhibition of Xenopus FGF signaling caused by the truncation of FGF receptors resulted in embryos with reduced mesoderm (Amaya et al., 1991; Yamaguchi et al., 1994; Griffin et al., 1995). In addition, targeting of mouse fgf4 resulted in severe deficiency of the proliferation of the inner cell mass (Feldman et al., 1995).
We are interested in the cellular and molecular mechanisms underlying endomesoderm formation in ascidian embryos (Imai et al., 2000; Satou et al., 2001a) (reviewed by Corbo et al., 2001). As in the case of vertebrate embryos, the endoderm of early ascidian embryos is specified autonomously, and it in turn induces mesenchyme and notochord formation around the 32- to 44-cell stages (Nakatani and Nishida, 1994; Kim et al., 2000). The lineages of the endoderm, mesenchyme and notochord of ascidian embryos are well characterized, and the induction processes are amenable to investigation at the level of single blastomeres. Thus, the ascidian embryo may provide an experimental system for exploring the details of the molecular mechanisms underlying the induction processes (Satoh, 1994; Satoh, 2001).
Nishida and colleagues have shown that bFGF, but not activin A, mimics the induction events at the 32-cell stage in Halocynthia embryos. Immersion of presumptive notochord blastomeres A6.2 and A6.4, dissociated in the early phase of the 32-cell stage in seawater containing human recombinant bFGF resulted in the activation of Brachyury and subsequent differentiation of notochord cells (Nakatani and Nishida, 1994; Nakatani et al., 1996), while the same treatment of presumptive mesenchyme blastomeres B6.2 and B6.4 evoked the differentiation of mesenchyme cells (Kim et al., 2000). In vertebrates, the signal cascade triggered by bFGF binding includes the Ras pathway (Satoh et al., 1992; Pawson, 1995). Nakatani and Nishida (Nakatani and Nishida, 1997) showed that the injection of a dominant-negative form of Ras, which causes the functional inhibition of endogenous Ras, inhibited notochord cell differentiation. In addition, Kim and Nishida (Kim and Nishida, 2001) showed that the 64-cell embryos exhibit a marked increase in extracellular signal-regulated kinase (ERK/MAPK) activity, and that reduction of this ERK activity by treatment with an FGF receptor inhibitor or a MEK (ERK kinase/MAPKK) inhibitor resulted in the failure of notochord and mesenchyme formation.
Although those experiments strongly suggested an involvement of FGF-like signals in the differentiation of notochord and mesenchyme cells in ascidian embryos, the reagents used in those studies were from human, not ascidian, sources. To overcome this problem, in a previous study we isolated cDNA for a gene encoding FGF receptor (HrFGFR) from Halocynthia roretzi embryos and showed that injection of synthetic mRNA for a dominant-negative form of HrFGFR resulted in the suppression of notochord and mesenchyme formation (Shimauchi et al., 2001). However, isolation and characterization of the function of ascidian FGF genes, which are key issues when approaching these problems, remained to be addressed in future studies.
Recently, convincing evidence has been accumulated for the pivotal role of β-catenin in axis determination and embryonic cell specification in a wide range of organisms from cnidarians to vertebrates (reviewed by Cadigan and Nusse, 1997; Moon and Kimelman, 1998; Sokol, 1999). In early Ciona savignyi embryos, β-catenin accumulates in the nuclei of endoderm precursor cells by the 32-cell stage, and this nuclear accumulation of β-catenin is the first step of endodermal cell specification (Imai et al., 2000). If β-catenin is mis- and/or overexpressed, the fate of presumptive notochord cells and epidermal cells changes so that they become endodermal cells, while if β-catenin nuclear localization is downregulated by the overexpression of cadherin, which binds to cytoplasmic β-catenin, endodermal cell differentiation is suppressed and the differentiation of extra epidermal cells occurs (Imai et al., 2000). To identify β-catenin target genes and to determine their function in endomesoderm differentiation of Ciona embryos, we previously conducted subtractive hybridization screening of mRNAs between β-catenin-overexpressing embryos and cadherin-overexpressing embryos (Satou et al., 2001a). We found that one of the β-catenin target genes encodes an FGF protein. The characterization of this gene reported here strongly suggests that Ciona FGF4/6/9 is expressed transiently in endodermal cells of early embryos and functions as an inductive signal for the differentiation of mesenchyme cells.
MATERIALS AND METHODS
Ascidian embryos
Adults of Ciona savignyi were collected near the Otsuchi Marine Research Center, Ocean Research Institute of the University of Tokyo, Iwate, Japan, and the Maizuru Fisheries Research Station of Kyoto University, Maizuru, Japan. Adults were maintained under constant illumination to induce oocyte maturation. Eggs and sperm were obtained surgically from the gonoduct. After insemination, embryos were reared at about 18°C in Millipore-filtered seawater (MFSW) containing 50 μg/ml streptomycin sulfate.
Isolation of cDNA clones for Ciona eFGF gene
cDNA clones for Cs-FGF4/6/9 (one of the potent β-catenin target genes) were isolated as described in detail elsewhere (Satou et al., 2001a). Briefly, β-catenin-overexpressing embryos and cadherin-overexpressing embryos were prepared by microinjection of synthetic mRNAs. Total RNAs were isolated from 120 of the former type of embryos and 129 of the latter type, both at the 110-cell stage, and cDNAs were synthesized from 0.3 μg of total RNAs using a SMART PCR cDNA Synthesis kit (Clontech). The subtraction procedure of Wang and Brown (Wang and Brown, 1991) was adopted, with several modifications (Satou et al., 2001a). The cDNA fragments amplified by PCR after three subtraction cycles were inserted into pGEM-T vector (Promega). Both the 5′-most and 3′-most ends of the cDNA fragments were completely sequenced, and nucleotide sequence information was used to check the independency of clones and sequence similarity to reported genes. BLASTX search of the 5′ sequences suggested that several clones encoded amino acid sequences which resembled those of FGF family members. cDNA clones containing complete coding regions were isolated from a gastrula cDNA library using probes derived from the cDNA fragments. Nucleotide sequences were determined for both strands using a Big-Dye Terminator Cycle Sequencing Ready Reaction kit and an ABI PRISM 377 DNA sequencer (Perkin Elmer).
Whole-mount in situ hybridization and histochemical staining for alkaline phosphatase (AP)
In situ hybridization of whole-mount specimens was carried out to determine the spatial pattern of expression of Cs-FGF4/6/9 according to the method described by Satou and Satoh (Satou and Satoh, 1997). An antisense RNA probe corresponding to the entire coding region of Cs-FGF4/6/9 cDNA was made and used at a concentration of 0.2 μg/μl in hybridization buffer. Control embryos that were hybridized with a sense probe did not show signals above background. We have already isolated cDNA clones for FGF8/18 and FGF11/12/13 from Ciona savignyi (K. S. Imai and Y. Satou, unpublished data), and confirmed that the probe for Cs-FGF4/6/9 does not crossreact to transcripts of the other FGF genes.
To examine the function of Cs-FGF4/6/9, the differentiation of epidermal cells in experimental embryos was monitored with probes for an epidermis-specific gene, Cs-Epi1 (Chiba et al., 1998), of muscle cells with probes for muscle actin gene Cs-MA1 (Chiba et al., 1998), of mesenchyme cells with probes for a mesenchyme-specific gene, Cs-mech1 (DDBJ/GenBank/EMBL accession number, AB073374; K. S. I., N. S. and Y. S., unpublished results), and of neuronal cells with probes for nervous system-specific genes, Cs-Otx (Satou et al., 2001a) and Cs-ETR (DDBJ/GenBank/EMBL accession number, AB073375; K. S. I., N. S. and Y. S., unpublished results). Notochord differentiation was assessed with probes for the C. savignyi Brachury gene (Cs-Bra) (Imai et al., 2000) and notochord-specific fibronectin-like gene Cs-fibrn (DDBJ/GenBank/EMBL accession number, AB073373; K. S. I., N. S. and Y. S., unpublished results).
The differentiation of endodermal cells was monitored by histochemical detection of AP activity (Whittaker and Meedel, 1989).
Morpholino oligos and synthetic capped mRNAs
In the present study, we used 25-mer morpholino oligo (Gene Tools, LLC) for Cs-FGF4/6/9. The nucleotide sequence of the Cs-FGF4/6/9 morpholino oligo contained the first methionine codon. For rescue experiments, synthetic capped mRNA for Cs-FGF4/6/9 was synthesized from Cs-FGF4/6/9 cDNA cloned into pBluescript RN3 vector (Lemaire et al., 1995) using a Megascript T3 kit (Ambion). To obtain capped mRNA, the concentration of GTP was lowered to 1.5 mM and the cap analog 7mGpppG was added at 6 mM. The synthetic Cs-FGF4/6/9 mRNA was designed to lack the morpholino oligo sequence, and therefore the Cs-FGF4/6/9 morpholino oligo does not recognize the synthetic mRNA. In the present study, we also used a morpholino oligo against β-catenin (Satou et al., 2001b), one against Cs-FoxD (S. I., N. S. and Y. S., unpublished results) and one against lacZ as control (Satou et al., 2001b).
After insemination, fertilized eggs were dechorionated and microinjected with 15 pmole of morpholino oligos and/or synthetic capped mRNAs. Each injection contained 30 pl of solution, and microinjection was carried out using a micromanipulator (Narishige Sci. Instru. Lab., Tokyo, Japan), as described (Imai et al., 2000). Injected eggs were reared at about 18°C in MFSW containing 50 μg/ml streptomycin sulfate. Cleavage of some embryos was arrested at the 110-cell stage with cytochalasin B and the embryos were then further cultured for about 12 hours, when control embryos reached the early tailbud stage.
RESULTS
Isolation and characterization of β-catenin-downstream FGF family gene
Subtractive hybridization screening of mRNAs from β-catenin -overexpressing embryos versus cadherin-overexpressing embryos in C. savignyi yielded several potential β-catenin-downstream genes (Satou et al., 2001a). One of them encoded a polypeptide with amino acid sequence similarity to FGF, and we have fully characterized this clone. The cDNA clone consisted of 1,559 nucleotides that encoded a predicted polypeptide of 301 amino acids (DDBJ/GenBank/EMBL database accession number, AB073372). The predicted amino acid sequence most closely resembled the sequences of vertebrate FGF4, FGF6 and FGF9, and therefore the gene was tentatively named Cs-FGF4/6/9.
Fig. 1 shows an alignment of the amino acid sequences of Cs-FGF4/6/9, human FGF4 (Delli Bovi et al., 1987), human FGF6 (Marics et al., 1989), human FGF9 (Miyamoto et al., 1993) and human FGF20 (Kirikoshi et al., 2000). Although Cs-FGF4/6/9 contained three insertions (at amino acid positions 52-62, 76-118 and 141-156) in its amino acid sequence, Cs-FGF4/6/9 showed 42-50% amino acid identity to human FGF4, FGF6, FGF9 and FGF20. The identities between Cs-FGF4/6/9 and FGF4, FGF6, FGF9 and FGF20 calculated within the conserved region were 50%, 50%, 42% and 46%, respectively. This indicates that Cs-FGF4/6/9 is a member of the FGF family.
Interestingly, it was noticed that Cs-FGF4/6/9 has features characteristic of both FGF4/6 and FGF9/20. FGF4 and FGF6 contain a signal peptide sequence in the N terminus, which is absent in FGF9 and FGF20 (Fig. 1). On the other hand, FGF9 and FGF20 have an additional 17 residues at the C terminus, which are not found in FGF4 and FGF6. Cs-FGF4/6/9 contains the signal peptide sequence in the N terminus and shares 15 residues with FGF9 and FGF20 at the C terminus (Fig. 1). It is thus likely that an ancestral gene such as Cs-FGF4/6/9 evolved into FGF4/6 and FGF9/20 in a lineage leading to vertebrates.
Early embryonic expression of Cs-FGF4/6/9
Maternal transcription of Cs-FGF4/6/9 was not detected (Fig. 2A,B). Zygotic expression of Cs-FGF4/6/9 began as early as the 16-cell stage, when its in situ hybridization signal was evident in the A5.1, A5.2 and B5.1 pairs of cells in the vegetal hemisphere (Fig. 2C). These three pairs of cells belong to the endodermal lineage. During the next cleavage to form the 32-cell stage embryo, the A-line expression was inherited by daughter-cell pairs A6.1, A6.2, A6.3 and A6.4 (Fig. 2D), while the B-line expression was inherited by only one pair of daughter cells, the B6.1 pair (Fig. 2D). A6.1 and B6.1 are destined to give rise to endoderm only, while A6.3 has the potential to form endoderm and trunk lateral cells. A6.2 and A6.4 are cells of the notochord/nerve cord lineage (A6.4 also forms two muscle cells). At the 64-cell stage, however, an in situ signal was seen in the A7.4, A7.8 and B7.4 pairs (Fig. 2E). A7.4 develops into the nerve cord and the visceral ganglion, and A7.8 to the nerve cord and muscle. B7.4 is a primordial muscle cell. At the 110-cell stage, a signal was evident in A-line nerve cord cells and primordial muscle cells, B8.7 and B8.8 (Fig. 2F). This early embryonic expression of Cs-FGF4/6/9 was transient, and became downregulated by the gastrula stage (Fig. 2G).
The early expression of Cs-FGF4/6/9 is therefore found in multiple types of embryonic cells; the expression appears to expand spatially from the center to the marginal region of the vegetal hemisphere as development proceeds (Fig. 2) (cf. Fujiwara et al., 2002). The expression first occurred in the endodermal cells at the 16- and 32-cell stages (Fig. 2C,D), and this expression became downregulated by the 64-cell stage (Fig. 2E). At the 64-cell stage, one pair of primordial muscle cells (B7.4) expressed this gene (Fig. 2E). This expression in the muscle cells was seen by the 110-cell stage (Fig. 3F) but was downregulated thereafter (Fig. 2G). The gene expression in A-line cells was inherited by cells of the nerve cord lineage (Fig. 2D-F). All of this early embryonic expression became downregulated by gastrulation (Fig. 2G).
After neurulation, an in situ signal appeared again in a pair of two or three dorsal cells at the central region of the embryo (Fig. 2H). This signal was inherited by their progeny cells in the tailbud embryos (Fig. 2I). Judging from their position, these Cs-FGF4/6/9-positive cells may be cells of the visceral ganglion. During further development of the tailbud embryo, weak signals also appeared in the anterior region of the embryo, in the posterior endoderm and in the caudal tip region (Fig. 2I).
Embryonic expression of Cs-FGF4/6/9 is controlled by β-catenin nuclear localization
The expression of Cs-FGF4/6/9 was ectopically induced by overexpression of β-catenin; in situ hybridization signals were detected in every blastomere of the 32-cell stage embryos developed from eggs injected with in vitro synthesized β-catenin mRNA (Fig. 3A). In contrast, the expression of Cs-FGF4/6/9 was downregulated by overexpression of cadherin; no in situ hybridization signal was detected in the 32-cell stage embryos developed from eggs injected with in vitro-synthesized cadherin mRNA (Fig. 3B). This indicates that the early and transient expression of Cs-FGF4/6/9 in the vegetal blastomeres is controlled either directly or indirectly by β-catenin nuclear localization.
Possible roles of Cs-FGF4/6/9 in differentiation of Ciona embryonic cells
The transient expression of Cs-FGF4/6/9 in endoderm precursor cells suggests its role in the induction of embryonic cell differentiation. We examined this issue by suppressing Cs-FGF4/6/9 gene function with a specific antisense morpholino oligo. Recently we found that morpholino oligos inhibit Ciona gene function without any detectable toxicity to the embryo and that their functional specificity can be ascertained by rescue experiments with in vitro synthesized mRNAs that lack the target sequences of the morpholino oligos (Satou et al., 2001a,b). As a control, we injected morpholino against lacZ, which showed no effects on C. savignyi embryogenesis (data not shown).
Epidermis and endoderm
Fertilized eggs injected with Cs-FGF4/6/9 morpholino oligo cleaved normally, and gastrulation took place as in normal embryos, but formation of tailbud embryos was usually affected (Fig. 4). The resultant tailbud embryos, however, showed expression of an epidermis-specific gene, Cs-Epi1 (Fig. 4A′; Table 1) and an endodermal marker, alkaline phosphatase (Cs-AP) (Fig. 4B′; Table 1). Therefore, suppression of Cs-FGF4/6/9 function did not affect the differentiation of epidermis or endoderm, two types of ascidian embryonic cells that differentiate autonomously without cell-cell interaction with other cell types.
Muscle
The suppression of Cs-FGF4/6/9 function appeared not to affect the differentiation of muscle cells, as far as could be seen by examination of the global expression of a muscle-specific actin gene, Cs-MA1, in the manipulated tailbud embryos (Fig. 4C′; Table 1). However, when the effect was examined in cleavage-arrested 110-cell stage embryos (Fig. 4D), we noticed that the actin gene expression was evident in B-line muscle cells (Fig. 4D′) while A-line muscle cells frequently failed to express Cs-MA1 (Fig. 4D′).
Notochord cells
The effects of the suppression of Cs-FGF4/6/9 function on the differentiation of notochord cells were examined with probes for notochord-specific genes, Cs-Bra (Fig. 4E,F) and Cs-fibrn (Fig. 4G,H). Injection of Cs-FGF4/6/9 morpholino oligo into fertilized eggs resulted in the development of 110-cell stage embryos or early gastrulae that failed to express Cs-Bra (the insert of Fig. 4E′; Table 1). When the left-side blastomere of the 2-cell stage embryos was injected with morpholino oligo, the resultant gastrulae showed Cs-Bra expression in the right-side notochord cells (Fig. 4E′, white arrow) but not in the left-side notochord cells (Fig. 4E′, black arrow).
However, when Cs-FGF4/6/9-morpholino-injected embryos were allowed to develop further, the resultant tailbud embryos eventually showed Cs-Bra expression (Fig. 4F′; Table 1). This was confirmed in 23 of the 27 experimental embryos (Table 1). This delayed expression of Cs-Bra resulted in the retardation of notochord formation, and only a partial notochord was formed (Fig. 4F′). This suggests that Cs-FGF4/6/9 is involved in the initial step of notochord induction, but other factors can eventually induce formation of the notochord.
As shown in Fig. 4G′, suppression of Cs-FGF4/6/9 function with the specific morpholino oligo affected the expression of Cs-fibrn. As in the case of Cs-Bra, the inhibition of Cs-fibrn expression was partial, and all of the experimental embryos showed expression of the gene in a smaller number of notochord cells (Fig. 4G′; Table 1) as compared with normal control embryos (Fig. 4G). When Cs-FGF4/6/9 morpholino-injected embryos were arrested at the 110-cell stage, the number of blastomeres expressing Cs-fibrn was reduced (compare Fig. 4H′ with 4H).
Although Cs-FGF4/6/9 is expressed in endodermal cells as well as presumptive notochord cells at the 32- and 64-cell stages, effects of Cs-FGF4/6/9 morpholino oligo on notochord differentiation were not absolute. We therefore examined the effects of overexpression of Cs-FGF4/6/9 by microinjection of in vitro-synthesized Cs-FGF4/6/9 mRNA into fertilized eggs. As shown in Fig. 5A′, the overexpression of Cs-FGF4/6/9 did not promote the differentiation of extra notochord cells; instead, the number of blastomeres expressing Cs-fibrn was reduced in manipulated embryos (compare Fig. 5A′ with 5A). Thus, Cs-FGF4/6/9 is likely to be involved in the initial process of notochord induction, but the nature of the function of Cs-FGF4/6/9 in the notochord formation could not be fully determined in the present study.
Mesenchyme cells
On the other hand, examination of the differentiation of mesenchyme cells in Cs-FGF4/6/9 morpholino-injected embryos with probe for a mesenchyme-specific gene, Cs-mech1 (Fig. 4I), showed that the expression of Cs-mech1 was completely inhibited (Fig. 4I′). No Cs-FGF4/6/9 morpholino-injected tailbud embryos expressed this marker gene (Table 1). This inhibition of expression of a mesenchyme marker gene was due to suppression of Cs-FGF4/6/9 function, because the inhibition was abrogated by injection of in vitro-synthesized Cs-FGF4/6/9 mRNA, which lacked the sequence targeted by the morpholino oligo (Fig. 4I′′). Cs-mech1 gene expression was seen in the region corresponding to mesenchyme in morphologically disordered experimental embryos (Fig. 4I′′). These results strongly suggest that Cs-FGF4/6/9 plays an essential role as an inductive signal for mesenchyme cell differentiation.
We also examined the effects of overexpression of Cs-FGF4/6/9. As shown in Fig. 5B′, injection of synthetic Cs-FGF4/6/9 mRNA induced the expression of Cs-mech1 in a few extra cells. In line with the rather wide distribution of cells showing embryonic expression of Cs-FGF4/6/9, the overexpression of this gene appeared not to cause conspicuous inductive effects on embryonic cells.
Nervous system
Whether Cs-FGF4/6/9 functions in the development of the nervous system was examined with two specific probes, Cs-Otx (Fig. 4J) and Cs-ETR (Fig. 4K). Both genes were expressed in Cs-FGF4/6/9-morpholino-injected embryos (Fig. 4J′,K′; Table 1). This suggests that Cs-FGF4/6/9 does not play a critical role in the differentiation of the nervous system. However, it remains to be elucidated whether Cs-FGF4/6/9 plays a role in the patterning of the nervous system, because the expression domains of the two marker genes were perturbed in Cs-FGF4/6/9-morpholino-injected tailbud embryos (Fig. 4J′,K′).
Possible cascades of Cs-FGF4/6/9 expression and function
As mentioned above, Cs-FGF4/6/9 is essential for the differentiation of mesenchyme cells. The gene is expressed in the endoderm, notochord and nerve cord in the anterior vegetal hemisphere, and in the endoderm and muscle in the posterior vegetal hemisphere; that is, the gene is not expressed in mesenchyme cells that would receive an inductive signal. To gain insight into the functional cascade of Cs-FGF4/6/9, we therefore carried out further experiments.
First, we addressed the question of whether presumptive mesenchyme cells change their fate to that of some other cell-types when the induction of their differentiation is inhibited by Cs-FGF4/6/9 morpholino oligo. The answer to this question was obtained by examining a muscle marker gene, Cs-MA1, in Cs-FGF4/6/9-morpholino-injected, cleavage-arrested embryos. As shown in Fig. 4D′, the number of B-line Cs-MA1-positive cells increased compared with the number in the control (Fig. 4D). These extra cells with Cs-MA1 expression were situated inside the B-line muscle cells, indicating that they were mesenchyme cells B8.5 and B7.7 (Fig. 4D′). This result strongly suggests that when the inductive signal from the endoderm is blocked, the presumptive mesenchyme cells change their fate into that of muscle cells, as suggested previously by Kim et al. (Kim et al., 2000).
Second, we examined the relationship between the nuclear accumulation of β-catenin and the role of Cs-FGF4/6/9. Nuclear accumulation of β-catenin is essential for endodermal cell specification in Ciona embryos (Imai et al., 2000), and therefore when β-catenin nuclear localization was downregulated by the overexpression of cadherin, the endoderm failed to differentiate, which in turn resulted in the failure of expression of Cs-FGF4/6/9 (Fig. 3B). Therefore, it would be expected that the differentiation of mesenchyme cells would not occur in β-catenin-depleted embryos, and indeed Fig. 5D shows a lack of expression of Cs-mech1 in β-catenin-morpholino-injected tailbud embryos. Injection of synthetic Cs-FGF4/6/9 mRNA in β-catenin-morpholino-injected eggs rescued the expression of Cs-mech1 (Fig. 5D′); this was confirmed in all 17 experimental embryos. Taken together, these data suggest that Cs-FGF4/6/9 is downstream of β-catenin signaling and that expression of Cs-FGF4/6/9 is necessary and sufficient for mesenchyme induction in Ciona savignyi embryos.
We also examined the relationship between β-catenin, Cs-FGF4/6/9 and notochord differentiation. Injection of morpholino oligo against β-catenin into fertilized eggs resulted in the failure of endodermal cell differentiation (Satou et al., 2001b), which in turn caused no notochord differentiation in all of the 15 experimental embryos (Fig. 5C). This inhibition of notochord differentiation was not rescued by injection of in vitro-synthesized Cs-FGF4/6/9 mRNA (Fig. 5C′); this was confirmed in all 14 experimental embryos.
Finally, we examined the relationship between Cs-FGF4/6/9 and Cs-FoxD. Cs-FoxD is also one of the β-catenin target genes, and the gene is expressed transiently in endodermal cells at the 16-cell and 32-cell stages (Imai et al., unpublished observations). The suppression of its function does not block the differentiation of endodermal cells themselves, but completely blocks notochord differentiation (Imai et al., unpublished observations). Cs-FoxD has no role in mesenchyme induction. We therefore examined the functional relationship between Cs-FoxD and Cs-FGF4/6/9. As shown in Fig. 5E, injection of Cs-FGF4/6/9 morpholino oligo into fertilized eggs resulted in the development of embryos with Cs-FoxD expression. On the other hand, injection of Cs-FoxD morpholino oligo did not affect Cs-FGF4/6/9 expression (Fig. 5F). Therefore, Cs-FGF4/6/9 and Cs-FoxD function independently of each other.
DISCUSSION
FGF genes in Ciona
In the present study, we characterized a gene for an FGF family member, Cs-FGF4/6/9, which is most similar to FGF4, FGF6, FGF9 and FGF20, and also contains features characteristic of both FGF4/6 and FGF9/20 of vertebrates (Fig. 1). This suggests that vertebrate FGF4, FGF6, FGF9 and FGF20 evolved from a common ancestral gene resembling urochordate FGF4/6/9/20 by gene duplication, and first diverged into FGF4/6 and FGF9/20 and then into FGF4, FGF6, FGF9 and FGF20. Therefore, identification and characterization of ascidian FGF genes may yield insights relevant to the expression and function of vertebrate members of this family.
So far, we have isolated cDNA clones for FGF4/6/9, FGF8/18 and FGF11/12/13 from Ciona savignyi, and FGF4/6/9, FGF3/7/10 and FGF11/12/13 from C. intestinalis (K. S. Imai and Y. Satou, unpublished data). This suggests that urochordates have at least four members of the FGF family: FGF4/6/9, FGF3/7/10, FGF8/18 and FGF11/12/13. cDNAs for Cs-FGF4/6/9 (present study), Cs-FGF8/18, Ci-FGF4/6/9, Ci-FGF3/7/10 and Ci-FGF11/12/13 have been fully characterized (K. S. Imai and Y. Satou, unpublished data). The present study showed that Cs-FGF4/6/9 is expressed in various vegetal blastomeres of early embryos and cells of the nervous system of the tailbud embryos. Preliminary observations showed that, among the other three types of Ciona FGF genes, Cs-FGF8/18 is expressed specifically in trunk lateral cells at the 64-cell to 110-cell stages. The expression and function of all Ciona FGF genes are now under investigation.
Function of Cs-FGF4/6/9
The present study demonstrated that Cs-FGF4/6/9 functions in the induction of mesenchyme formation, and is likely to be involved in the differentiation of notochord cells.
Mesenchyme differentiation
The present study provided evidence that FGF4/6/9 plays a pivotal role in the inductive cell-cell communication between signal-emitting endodermal cells and signal-receiving mesenchyme cells in ascidian embryos. FGF4 and FGF6 subfamily members are sometimes called embryonic FGF (eFGF) because their genes are expressed during early embryonic stages, and play important roles in mesoderm formation in vertebrate embryos (Slack, 1994; Feldman et al., 1995). We have isolated Cs-FGF4/6/9 as a putative downstream gene of β-catenin. As shown in Fig. 3, increased accumulation of nuclear β-catenin in embryonic cells resulted in upregulation of Cs-FGF4/6/9 expression while a decreased level of nuclear β-catenin resulted in downregulation of Cs-FGF4/6/9 expression. Therefore, we should determine in future studies whether nuclear localization of β-catenin directly activates the expression of Cs-FGF4/6/9.
In early Ciona embryos, Cs-FGF4/6/9 was shown to be expressed in endodermal cells and muscle cells which surround mesenchyme cells, while mesenchyme cells themselves were shown not to express the gene. When the gene function was suppressed by morpholino oligo, a marker gene, Cs-mech1, for mesenchyme differentiation was not expressed (Fig. 4I′), and this inhibition was abrogated by coinjection of in vitro-synthesized Cs-FGF4/6/9 mRNA (Fig. 4I′′). Therefore, it is highly likely that Cs-FGF4/6/9 produced by endodermal cells around the 32-cell stage acts as an inducing signal to promote the neighboring cells to differentiate into mesenchyme. Muscle cells, B7.4 at the 64-cell stage and B7.7 and B7.8 at the 110-cell stage, also express Cs-FGF4/6/9, and these cells are situated beside mesenchyme cells. Therefore, it is possible that these muscle cells are also involved in the induction event, presumably not in the initial step, but rather in a later accelerating or maintenance step for mesenchyme differentiation.
In a previous study investigating the role of FGF receptor in the induction events in Halocynthia embryos (Shimauchi et al., 2001), we noticed that the inhibitory effect of a dominant-negative form of HrFGFR on notochord differentiation was not always absolute, but rather notochord-specific gene expression was seen after a delay in experimental embryos. In contrast, the inhibitory effect of the dominant-negative form of HrFGFR on mesenchyme differentiation was more nearly absolute. Together with the present results, it is highly likely that the primary function of embryonic FGF signals in the ascidian embryo is to induce mesenchyme but not the notochord.
It appeared that the presumptive mesenchyme cells change their developmental fate into that of muscle cells when they fail to receive the FGF signal from the endoderm. Kim et al. (Kim et al., 2000) showed that human recombinant bFGF mimics the induction event for mesenchyme cells in Halocynthia embryos. Furthermore, they showed that the differentiation pathway responsible for the development of mesenchyme or notochord cells depends on whether the precursor cells contain the posterior-vegetal cytoplasm (PVC) of the eggs. PVC is present in mesenchyme precursor cells, whereas notochord precursor cells do not contain PVC. In addition, those authors also showed that if isolated presumptive mesenchyme cells are not immersed in seawater that contains bFGF, they eventually differentiate into muscle cells. Therefore, the present results confirm their experimental findings, although the details of the molecular mechanisms remain to be elucidated.
Notochord differentiation
Expression of Cs-FGF4/6/9 was found in presumptive notochord cells A5.1 and A5.2 at the 16-cell stage and A6.2 and A6.4 at the 32-cell stage (Fig. 2), although the gene expression was downregulated in the primordial notochord cells, A7.3 and A7.7 at the 64-cell stage (Fig. 2). Because the inductive signal from the endoderm for notochord differentiation occurs around the 32-cell stage, and is emitted from A6.1 and A6.3 to A6.2 and A6.4 (Nakatani and Nishida, 1994; Nakatani et al., 1996), the expression pattern of Cs-FGF4/6/9 indicates that the gene is expressed in both signal-emitting endodermal cells and signal-receiving notochord/nerve cord cells. Injection of synthetic Cs-FGF4/6/9 mRNA into fertilized eggs resulted in suppression of notochord-specific gene expression (Fig. 5C′). In addition, suppression of the function of Cs-FGF4/6/9 with specific morpholino oligo resulted in the inhibition of the initial expression of Cs-Bra; however, Cs-Bra was expressed later in embryogenesis (Fig. 4E,F). Similarly, suppression of Cs-FGF4/6/9 function resulted in partial inhibition of notochord-specific Cs-fibrn gene expression (Fig. 4G). Together with the results of the studies of the dominant-negative form of HrFGFR on notochord formation (Shimauchi et al., 2001), the present results indicate that it is likely that FGF signals are involved in the initial step of notochord induction in ascidian embryos, but other factors may overcome the effect of the lack of FGF signals, eventually resulting in formation of the notochord. The BMP4 signaling pathway may be one of such factors (Darras and Nishida, 2001).
Involvement of FGF signals in the notochord induction has been suggested in Halocynthia embryos (Nakatani et al., 1996; Nakatani and Nishida, 1997; Kim and Nishida, 2001). Therefore, it is possible that FGF signals play an essential role in the entire process of notochord induction of Halocynthia embryos, but in Ciona embryos the signals play in the initial step of induction. In addition, Corbo et al. (Corbo et al., 1998) suggested an involvement of Notch-Su(H) pathway in the notochord induction in Ciona embryos. We recently found that nuclear localization of β-catenin directly triggers the activity of a gene (Cs-FoxD) for forkhead transcription factor (S. I., N. S. and Y. S., unpublished observations). The gene is expressed transiently in endoderm cells A5.1 and A5.2 at the 16-cell stage and A6.1 and A6.3 at the 32-cell stage, and is downregulated by the 64-cell stage. Suppression of the function of this gene with specific morpholino oligos resulted in complete loss of notochord formation, although endoderm and mesenchyme differentiated normally (S. I., N. S. and Y. S., unpublished observations). As shown in the present study, Cs-FoxD and Cs-FGF4/6/9 function independently (Fig. 5E,F). Therefore, elucidation of signaling events downstream of Cs-FoxD may be a key to identifying the molecules essential for notochord induction in Ciona embryos.
Differentiation of nervous system
Possible roles of FGF signals in the differentiation of the nervous system in ascidian larvae have been studied extensively (e.g. Ohtsuka et al., 2001; Darras and Nishida, 2001; Kim and Nishida, 2001; Hudson and Lemaire, 2001); these studies suggest an involvement of FGF signals in this differentiation process. As far as the differentiation of the nervous system is assessed by the expression of specific genes, Cs-Otx and Cs-ETR, Cs-FGF4/6/9 does not play a critical role in the differentiation of the nervous system, because the suppression of its function with morpholino did not diminish expression of the nervous system-specific genes. However, it remains to be determined whether FGF signals other than FGF4/6/9 play an essential role in the differentiation of the nervous system in ascidian larvae. As mentioned before, we have already characterized cDNA clones for Cs-FGF8/18, Ci-FGF3/7/10 and Ci-FGF11/12/13, and we are now examining the role of these FGF genes in the nervous system formation.
In conclusion, in Ciona savignyi, embryonic FGF or Cs-FGF4/6/9 is activated as a consequence of the nuclear localization of β-catenin, and the activity of this gene is necessary and sufficient for mesenchyme formation in ascidian embryos, and essential for the first step of notochord induction.
Alignment of amino acid sequences of Cs-FGF4/6/9, human FGF4, human FGF6, human FGF9 and human FGF20. Identical residues are enclosed by black boxes, and similar residues by grey boxes. Identities between Cs-FGF4/6/9 and FGF4, FGF6, FGF9 and FGF20 are shown at the ends of the respective sequences. The putative signal peptide sequence of Cs-FGF4/6/9 is underlined.
Alignment of amino acid sequences of Cs-FGF4/6/9, human FGF4, human FGF6, human FGF9 and human FGF20. Identical residues are enclosed by black boxes, and similar residues by grey boxes. Identities between Cs-FGF4/6/9 and FGF4, FGF6, FGF9 and FGF20 are shown at the ends of the respective sequences. The putative signal peptide sequence of Cs-FGF4/6/9 is underlined.
Expression of Cs-FGF4/6/9 in early Ciona savignyi embryos, revealed by whole-mount in situ hybridization. (A) A fertilized egg. (B) An 8-cell stage embryo, lateral view. (C) A 16-cell stage embryo, (D) 32-cell stage embryo, (E) 64-cell stage embryo, (F) 110-cell stage embryo and (G) late gastrula-stage embryo, vegetal pole view; anterior is up and posterior is down. Blastomeres containing Cs-FGF4/6/9 transcripts are named in C-F according to Conklin (Conklin, 1905). In ascidians, in situ signals for zygotic gene expression are first detected in the nuclei of embryonic cells. (H) A neurula, dorsal view; (I) mid tailbud embryo, lateral view. Arrows indicate Cs-FGF4/6/9 expression in cells of the nervous system. Scale bar, 50 μm.
Expression of Cs-FGF4/6/9 in early Ciona savignyi embryos, revealed by whole-mount in situ hybridization. (A) A fertilized egg. (B) An 8-cell stage embryo, lateral view. (C) A 16-cell stage embryo, (D) 32-cell stage embryo, (E) 64-cell stage embryo, (F) 110-cell stage embryo and (G) late gastrula-stage embryo, vegetal pole view; anterior is up and posterior is down. Blastomeres containing Cs-FGF4/6/9 transcripts are named in C-F according to Conklin (Conklin, 1905). In ascidians, in situ signals for zygotic gene expression are first detected in the nuclei of embryonic cells. (H) A neurula, dorsal view; (I) mid tailbud embryo, lateral view. Arrows indicate Cs-FGF4/6/9 expression in cells of the nervous system. Scale bar, 50 μm.
(A) Upregulation of Cs-FGF4/6/9 expression in β-catenin-overexpressing embryos and (B) downregulation of Cs-FGF4/6/9 expression in cadherin-overexpressing embryos at the 32-cell stage. Vegetal pole view.
Effects of suppression of Cs-FGF4/6/9 function with a specific antisense morpholino oligo on the expression of (A,A′) epidermis-specific gene Cs-Epi1, (B,B′) endoderm-specific histochemical activity of alkaline phosphatase, (C,C′,D,D′) muscle-specific actin gene Cs-MA1, (E,E′,F,F′) notochord-specific Cs-Bra, (G,G′,H,H′) notochord-specific gene Cs-fibrn, (I,I′,I′′) mesenchyme-specific gene Cs-mech1, and nervous system-specific genes (J,J′) Cs-otx and (K,K′) Cs-ETR. Arrows in D′ indicate extra cells with Cs-MA1 expression. The black arrow in E′ indicates no Cs-Bra expression in the left-side notochord cells while the white arrow points to Cs-Bra expression in the right-side notochord cells. This embryo developed from the 2-cell stage in which the left-side blastomere was injected with morpholino oligo. Two embryos in the insert were developed from eggs injected with morpholino oligo, and they do not show the expression of Cs-Bra. (I′′) Recovery of expression of Cs-fibrn in tailbud embryos developed from eggs injected with Cs-FGF4/6/9 morpholino oligo together with in-vitro-synthesized Cs-FGF4/6/9 mRNA.
Effects of suppression of Cs-FGF4/6/9 function with a specific antisense morpholino oligo on the expression of (A,A′) epidermis-specific gene Cs-Epi1, (B,B′) endoderm-specific histochemical activity of alkaline phosphatase, (C,C′,D,D′) muscle-specific actin gene Cs-MA1, (E,E′,F,F′) notochord-specific Cs-Bra, (G,G′,H,H′) notochord-specific gene Cs-fibrn, (I,I′,I′′) mesenchyme-specific gene Cs-mech1, and nervous system-specific genes (J,J′) Cs-otx and (K,K′) Cs-ETR. Arrows in D′ indicate extra cells with Cs-MA1 expression. The black arrow in E′ indicates no Cs-Bra expression in the left-side notochord cells while the white arrow points to Cs-Bra expression in the right-side notochord cells. This embryo developed from the 2-cell stage in which the left-side blastomere was injected with morpholino oligo. Two embryos in the insert were developed from eggs injected with morpholino oligo, and they do not show the expression of Cs-Bra. (I′′) Recovery of expression of Cs-fibrn in tailbud embryos developed from eggs injected with Cs-FGF4/6/9 morpholino oligo together with in-vitro-synthesized Cs-FGF4/6/9 mRNA.
Functional cascade of Cs-FGF4/6/9. (A,A′) Expression of Cs-fibrn in (A) control uninjected tailbud-embryos and (A′) experimental tailbud embryos developed from eggs injected with in vitro-synthesized Cs-FGF4/6/9 mRNA. (B,B′) Expression of Cs-mech1 in (B) control embryos and (B′) experimental embryos developed from eggs injected with in vitro-synthesized Cs-FGF4/6/9 mRNA. (C,C′) Expression of Cs-fibrn in experimental tailbud embryos developed from eggs injected with (C) β-catenin morpholino oligo or (C′) β-catenin morpholino oligo together with in vitro-synthesized Cs-FGF4/6/9 mRNA. (D,D′) Expression of Cs-mech1 in experimental tailbud embryos developed from eggs injected with (D) β-catenin morpholino oligo or (D′) β-catenin morpholino oligo together with in vitro-synthesized Cs-FGF4/6/9 mRNA. (E) Cs-FoxD expression at the 32-cell stage in embryos developed from eggs injected with Cs-FGF4/6/9 morpholino oligo. (F) Cs-FGF4/6/9 expression at the 32-cell stage in embryos developed from eggs injected with Cs-FoxD morpholino oligo.
Functional cascade of Cs-FGF4/6/9. (A,A′) Expression of Cs-fibrn in (A) control uninjected tailbud-embryos and (A′) experimental tailbud embryos developed from eggs injected with in vitro-synthesized Cs-FGF4/6/9 mRNA. (B,B′) Expression of Cs-mech1 in (B) control embryos and (B′) experimental embryos developed from eggs injected with in vitro-synthesized Cs-FGF4/6/9 mRNA. (C,C′) Expression of Cs-fibrn in experimental tailbud embryos developed from eggs injected with (C) β-catenin morpholino oligo or (C′) β-catenin morpholino oligo together with in vitro-synthesized Cs-FGF4/6/9 mRNA. (D,D′) Expression of Cs-mech1 in experimental tailbud embryos developed from eggs injected with (D) β-catenin morpholino oligo or (D′) β-catenin morpholino oligo together with in vitro-synthesized Cs-FGF4/6/9 mRNA. (E) Cs-FoxD expression at the 32-cell stage in embryos developed from eggs injected with Cs-FGF4/6/9 morpholino oligo. (F) Cs-FGF4/6/9 expression at the 32-cell stage in embryos developed from eggs injected with Cs-FoxD morpholino oligo.
Acknowledgements
We are thankful to Dr Yasuaki Takagi, Mr Kouichi Morita and all of the members in the Otsuchi Marine Research Center, Ocean Research Institute of the University of Tokyo, Iwate, Japan for collecting Ciona savignyi. This research was supported by Grants-in-Aid from the MEXT, Japan to Y. S., K. S. I. and N. S., and by a grant from the Human Frontier Science Program to N. S. (RG0290/2000-MR).