In early Ciona embryos, nuclear accumulation of β-catenin is most probably the first step of endodermal cell specification. If β-catenin is mis- and/or overexpressed, presumptive notochord cells and epidermal cells change their fates into endodermal cells, whereas if β-catenin nuclear localization is downregulated by the overexpression of cadherin, the endoderm differentiation is suppressed, accompanied with the differentiation of extra epidermal cells (
Recently convincing evidence has been accumulated to show an involvement of β-catenin in axis determination and embryonic cell specification in a wide range of organisms from cnidarians to vertebrates (Cadigan and Nusse, 1997; Moon and Kimelman, 1998; Sokol, 1999). In Xenopus and zebrafish, β-catenin is concentrated in nuclei of dorsal cells of blastula, and is involved in an establishment of the dorsoventral axis of the embryo (Heasman et al., 1994; Funayama et al., 1995; Guger and Gumbiner, 1995; Kelly et al., 1995; Kelly et al., 2000; Schneider et al., 1996; Larabell et al., 1997; Miller et al., 1999). In sea urchins, the nuclear localization of β-catenin is seen in vegetal blastomeres, which is essential for vegetal fate determination of embryonic cells (Wikramanayake et al., 1998; Logan et al., 1999). β-catenin is also involved in the specification of the endoderm in early C. elegans embryos (Thorpe et al., 1997; Rocheleau et al., 1997). In addition, a recent study demonstrated that β-catenin also works in the axis determination of hydra (Hobmayer et al., 2000). To exert its function, β-catenin enters the nucleus and activates downstream genes together with TCF/LEF1, which is a transcriptional factor with an HMG box (Willert and Nusse, 1998). In vertebrates, several genes have been identified as factors downstream of β-catenin, including siamois (Brannon et al., 1997; Fan and Sokol, 1997), twins (Laurent et al., 1997) and fibronectin (Gradl et al., 1999) in Xenopus embryos, boz/dharma in zebrafish embryos (Fekany et al., 1999), and Nodal-related3 (McKendry et al., 1997) and Brachyury (Yamaguchi et al., 1999; Arnold et al., 2000) in mouse embryos.
The ventral trunk region of an ascidian tailbud embryo is occupied by the endoderm that extends ventrally as the endodermal strand along the notochord of the tail region (Satoh, 1994). This tissue is composed of about 500 cells, and its embryonic lineage is well documented (Conklin, 1905; Nishida, 1987). The endodermal cells are derived from the vegetal A4.1 and B4.1 blastomeres of the 8-cell stage embryo, and A5.1, A5.2, B5.1 and B5.2 of the 16-cell stage (Fig. 1). As early as the 32-cell stage, two pairs of vegetal blastomeres, A6.1 and B6.1, become restricted to generate endoderm only, and at the 64-cell stage, five pairs of vegetal blastomeres (A7.1, A7.2, A7.5, B7.1 and B7.2) become endoderm restricted. These primordial endodermal cells proliferate to form about 500 endodermal cells.
Reflecting such an early fate restriction, presumptive endodermal blastomeres show a high potential for autonomous differentiation when they are isolated from early embryos (Whittaker, 1990; Nishida, 1992). This autonomy is dependent on maternal factors or determinants that are pre-localized in the endoplasm of eggs and early embryos (Nishida, 1993). In a previous study, we have shown that during cleavages of the ascidians Ciona intestinalis and C. savignyi, β-catenin accumulates in the nuclei of vegetal blastomeres by the 32-cell stage, that mis- and/or overexpression of β-catenin induces the development of an endoderm-specific alkaline phosphatase (AP) in presumptive notochord cells and epidermal cells without affecting differentiation of primary lineage muscle cells, and that down regulation of nuclear β-catenin induced by the overexpression of cadherin results in the suppression of endodermal cell differentiation (this suppression was compensated for by the differentiation of extra epidermal cells) (Imai et al., 2000). The accumulation of β-catenin in the nuclei of endoderm progenitor cells is therefore most probably the first step in the process of ascidian embryonic endoderm specification.
To understand the molecular mechanisms of endoderm differentiation in ascidian embryos, it is important to identify the genes that act as direct targets and/or downstream of β-catenin. Because in ascidians β-catenin plays a pivotal role in the endodermal cell specification, it may regulate the expression of genes other than siamois, twins and boz/dharma, these genes being involved in dorsoventral axis determination. In addition, because the ascidian endoderm functions as an inductive source for the differentiation of notochord cells (Nakatani and Nishida, 1994) and mesenchymal cells (Kim et al., 2000), the identification of β-catenin downstream genes will also lead to an elucidation of these induction mechanisms. We took advantage of β-catenin overexpressed embryos and cadherin overexpressed embryos; in the former, β-catenin targets may be upregulated, and in the latter, β-catenin targets may be downregulated. We found that a LIM-homeobox gene Cs-lhx3, an otx homolog Cs-otx and an NK-2 class homeobox gene Cs-ttf1 are downstream genes of β-catenin. Of these genes, an inhibition of possible early embryonic function of Cs-lhx3 resulted in the suppression of endoderm differentiation.
MATERIALS AND METHODS
Ascidian eggs and embryos
Adults of Ciona savignyi were collected near the Otsuchi Marine Research Center, Ocean Research Institute of the University of Tokyo (Iwate, Japan) and maintained under constant light 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 and characterization of candidate cDNA clones
β-catenin overexpressed embryos and cadherin overexpressed embryos were prepared by microinjection of synthetic mRNAs as described previously (Imai et al., 2000). One hundred and twenty former embryos and 129 latter embryos, both at the 110-cell stage, were lysed in 400 μl of GTC solution (4 M guanidinium thiocyanate, 50 mM Tris-HCl, pH 7.5, 10 mM EDTA, 2% sarkosyl, 1% β-mercaptoethanol). Total RNA was isolated from the lysates by the acid guanidinium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi, 1987). cDNA was synthesized from 0.3 μg of total RNA using a SMART PCR cDNA Synthesis kit (Clontech).
The subtraction procedure of Wang and Brown (Wang and Brown, 1991) was adopted with several modifications. Each of the cDNA pools was digested completely with RsaI, ligated with double-stranded linkers prepared from dCTCTTGCTTGAATTAGGACTA and dTAGTCCGAATTCAAGCAAGAGCACA, and electrophoresed through a 1.5% low melting temperature agarose gel to remove the unligated linkers. 0.01 μg of cDNA fragments recovered from the gel were amplified by PCR with dCTCTTGCTTGAATTAGGACTA primer. The optimal number of PCR cycles was determined by electrophoresis of PCR products from every five additional cycles so that the double-stranded cDNA would remain in the exponential phase of amplification. The PCR products were the starting material for subtractive hybridization. The PCR products were divided into the ‘tracer’ fraction and the ‘driver’ fraction. Cs-β-catenin and Ci-cadherin in pBluescript RN3 vector (Lemaire et al., 1995) were amplified by PCR and digested with RsaI. These PCR products were also used as ‘driver’ in order to exclude the specific cDNAs that were derived from injected mRNAs. 20 μg of the ‘driver’ fraction digested with EcoRI and 5 μg of the ‘injected cDNA driver’ were photobiotinylated using a Photoprobe biotin (Vector). This was mixed with 1 μg of the ‘tracer’, denatured and hybridized at 68°C for 20 hours (long hybridization). Then the biotin-containing cDNAs were removed by four cycles of streptavidin (Vector) and phenol-chloroform extraction. The remaining cDNAs were hybridized again with 10 μg of the biotinylated ‘driver’ and 2.5 μg of the biotinylated ‘injected cDNA driver’ for 2 hours (short hybridization). The biotinylated cDNAs were again removed, and the first subtraction cycle was completed. This cycle was repeated three times. The product of the first subtraction cycle was again PCR amplified with the optimal number of PCR cycles determined as previously described and subtracted as the same way as the first cycle. The long hybridization in the second and third cycle used the second and third PCR product, respectively, and the short hybridization in the second and third cycle used the first PCR product. The cDNA fragments amplified by PCR after three subtraction cycles were inserted into pGEM-T vector (Promega).
Isolation of cDNA clones containing the entire coding region and determination of the nucleotide sequences
cDNA clones containing the entire coding region were isolated from a gastrula stage cDNA library by probes derived from the cDNA fragments. Nucleotide sequences were determined for both strands using a Big-Dye Terminator Cycle Sequencing Ready Reaction kit and ABI PRISM 377 DNA sequencer (Perkin Elmer).
Whole-mount in situ hybridization
To determine mRNA distributions in eggs and embryos, RNA probes were prepared using a DIG RNA-labeling kit (Boehringer-Mannheim, Heidelberg, Germany). Whole-mount in situ hybridization was performed using digoxigenin-labeled antisense probes as described previously (Satou and Satoh, 1997). Control embryos that were hybridized with a sense probe did not show signals above background.
In order to examine the endodermal cell differentiation, probes were prepared from cDNA for an endoderm-specific thyroid hormone receptor gene of C. savignyi.
Morpholino oligos and synthetic capped mRNAs
In the present study, we used 24- or 25-mer morpholino oligonucleotides (Gene Tools LLC) for Cs-lhx3, Cs-otx, Cs-ttf1 and lacZ. Nucleotide sequences of the morpholinos are listed in Table 1. Synthetic capped mRNA for Cs-lhx3 was synthesized from Cs-lhx3 cDNA cloned into pBluescript RN3 vector (Lemaire et al., 1995) using a Megascript T3 kit (Ambion, Austin, TX, USA). To obtain a capped mRNA, the concentration of GTP was lowered to 1.5 mM and the cap analog 7mGpppG was added at a final concentration of 6 mM. Capped mRNA for lacZ, delE-Ci-cadherin or delN-β-catenin was synthesized in the same way as Cs-lhx3.
After insemination, fertilized eggs, usually with intact chorion, were microinjected with 15 fmole of morpholinos and/or synthetic capped mRNAs. Each injection contained 30 pl of solution, and microinjection was carried out using a micromanipulator (Narishige Scientific Instruments Laboratory, Tokyo, Japan), as described by Imai et al. (Imai et al., 2000). Injected eggs were reared at about 18°C in Millipore-filtered seawater (MFSW) containing 50 μg/ml streptomycin sulfate.
Histochemical staining for alkaline phosphatase (AP)
Differentiation of endodermal cells was monitored by histochemical detection of AP activity (Whittaker and Meedel, 1989). Embryos were fixed with 5% formaldehyde in seawater for 10 minutes at room temperature. Embryos were washed in AP staining buffer (100 mM NaCl, 50 mM MgCl2, 100 mM Tris-HCl, pH 9.5) three times. For signal detection, specimens were incubated with NBT/BCIP/AP staining buffer (4.5 μl NBT/ml, 3.5 μl BCIP/ml).
Isolation of cDNA clones for β-catenin downstream genes
To identify genes functioning downstream of β-catenin in Ciona embryos, we made a subtractive cDNA library that contains mRNAs, the genes of which are upregulated in β-catenin overexpressed embryos relative to those downregulated in cadherin overexpressed embryos. A winged helix/forkhead gene Cs-HNF3 is a β-catenin downstream gene (Y. S. and N. S., unpublished) and Cs-Epi1 is an epidermis-specific gene of C. savignyi (Chiba et al., 1998). Enrichment of Cs-HNF3 in β-catenin overexpressed library and that of Cs-Epi1 in cadherin overexpressed library were confirmed by Southern blot hybridization (data not shown). We selected candidate cDNAs by the following three criteria. The average length of cDNA fragments obtained by PCR was about 300∼400 bp. First, both the 5′-most and 3′-most ends of cDNAs were completely sequenced, and nucleotide sequence information was used to check the independency of clones and sequence similarity to reported genes. We are particularly interested in genes that encode for transcription factors.
Second, if cDNA clones represent the β-catenin downstream and endoderm-related genes, they may be expressed in the presumptive endodermal blastomeres. The lineage of endodermal cells in early ascidian embryos has been completely documented (Nishida, 1987). The endodermal cells are derived from the vegetal A4.1 and B4.1 blastomeres of the bilaterally symmetrical eight-cell stage embryo. The developmental potential to form endoderm is segregated into A5.1, A5.2, B5.1 and B5.2 cells of the 16-cell stage embryo (Fig. 1A). Because the B5.2 cell contributes only two or so endodermal strand cells, it is not discussed in detail in this report. At the 32-cell stage, A6.1, A6.3, and B6.1 cells are the major endoderm lineages (Fig. 1B), and as early as this stage A6.1 and B6.1 cells become restricted to generate endoderm only. At the 64-cell stage, not only A7.1 and A7.2 cells (A6.1 daughter cells) and B7.1 and B7.2 cells (B6.1 daughter cells) but also the A7.5 cell becomes endoderm-restricted (Fig. 1C). These primordial cells divide five or six times to form the endodermal tissue of about 500 cells (Fig. 1E). Therefore, we examined the embryonic expression of candidate cDNAs by whole-mount in situ hybridization, and selected those expressed in the endoderm lineage.
Third, if cDNA clones represent β-catenin downstream genes, their expression should be upregulated in β-catenin overexpressed embryos and downregulated in Ci-cadherin overexpressed embryos. This was also examined by whole-mount in situ hybridization. If cDNAs satisfy these three criteria, their entire nucleotide sequences were determined. The present screening has yielded cDNAs for twelve different genes including a LIM-homeobox gene (Cs-lhx3), an otx gene (Cs-otx), a ttf1 gene (Cs-ttf1), two forkhead genes, a type II cadherin gene, a protocadherin gene, a netrin gene, a frizzled gene, a lefty/antivin homolog, an EPH gene and a gene encoding a protein with no known sequence similarity. The first three transcription genes were studied in detail.
Characterization of cDNA clones for Cs-lhx3, Cs-otx and Cs-ttf1
A cDNA clone for Cs-lhx3 gene was 1988 bp in length, and the predicted protein consists of 472 amino acids (DDBJ/GenBank/EMBL Accession Number, AB057733). The encoded protein contains two LIM domains and a homeodomain (Fig. 2A), and Cs-LHX3 sequence most resembles that of Halocynthia HrLIM homeoprotein (Wada et al., 1995). The LIM-homeobox genes are categorized, based on the sequence similarities, expression patterns and possible functions, into several groups, including LIN-11, Apterous, Lhx-3, islet, Lhx6/Lhx8 and Lmx groups (Hobert and Westphal, 2000; Fig. 2B). The amino acid sequence of the homeodomain of Cs-LHX3 was compared with other LIM-homeodomains and subjected to molecular phylogenetic analysis using neighbor-joining method. The homeodomains of Cs-OTX and Cs-TTF1 were used as the outgroup. As shown in Fig. 2B, Cs-LHX3 was included within a group of LHX-3 members together with HrLIM. Vertebrate genes of this group are associated with the neuronal cell development (Hobert and Westphal, 2000).
Embryonic expression of Cs-lhx3 became evident as early as the 32-cell stage. As shown in Fig. 1G, the expression begins in A6.1, A6.3, and B6.1 cells of the 32-cell stage embryo. At the 64-cell stage, the expression was inherited to their daughter cells, A7.1, A7.2, A7.5, B7.1 and B7.2 with endodermal fate. In addition, A7.3, A7.7, B7.3 and B7.5 blastomeres of the 64-cell stage embryo expressed Cs-lhx3 (Fig. 1H). This early expression of Cs-lhx3 in endoderm precursor cells (A7.1, A7.2, A7.5, B7.1 and B7.2) persisted until the 110-cell stage (Fig. 1I) and was then downregulated. No detectable in situ hybridization signals were evident in gastrulae (data not shown). Later, neurulae and tailbud embryos expressed Cs-lhx3 again, and one pair of cells in the nerve cord and a subset of brain cells expressed Cs-lhx3. This spatial expression pattern of Cs-lhx3 resembles that of Hrlim (Wada et al., 1995). As shown in Fig. 3A,C, Cs-lhx3 expression was upregulated in the β-catenin overexpressed embryos so that Cs-lhx3 in situ hybridization signals were found in nuclei of almost all blastomeres including animal blastomeres at both the 32-cell (Fig. 3A) and 110-cell stages (Fig. 3C). By contrast, Cs-lhx3 expression was downregulated in the cadherin overexpressed embryos, and no in situ hybridization signals were detectable in these embryos (Fig. 3B,D).
A cDNA clone, which was 2393 bp in length and encoded a 422 amino acid polypeptide (DDBJ/GenBank/EMBL Accession Number, AB057732), predicted a C. savignyi homologue of otx/orthodenticle homeobox genes (Fig. 4A), and therefore this gene was named Cs-otx. The otx gene of ascidians was already identified in H. roretzi (Hroth)(Wada et al., 1996), Herdmania curvata (Hinman and Degnan, 2000) and C. intestinalis (Ci-otx)(Hudson and Lemaire, 2001). The homeodomain amino acid residues were completely conserved between the ascidian Otx proteins, except for one residue between Ciona and Halocynthia (Fig. 4B).
Embryonic expression of Cs-otx also became evident as early as the 32-cell stage. As shown in Fig. 1L, the expression was evident in B6.1, B6.2 and B6.4 blastomeres of the 32-cell stage embryo. At the 64-cell stage, the expression was observed in endoderm precursor cells, A7.1, A7.2, A7.5, B7.1 and B7.2, trunk lateral precursor cells A7.6, presumptive notochord and mesenchyme cells B7.7 and B7.3, and presumptive muscle cells B7.5 (Fig. 1M). This complex spatial expression continued up to the early gastrula stage (Fig. 1N), and then this early embryonic Cs-otx expression in the vegetal blastomeres was downregulated. Cs-otx was also expressed in the neuronal cell lineage (A8.7 and A.8.8) at early gastrula and later stages (Fig. 1O). The embryonic expression pattern of Cs-otx was similar to that of Hroth (Wada et al., 1996) and Ci-otx (Hudson and Lemaire, 2001).
As in the case of Cs-lhx3, Cs-otx expression was upregulated in the β-catenin overexpressed embryos. In situ hybridization signals for Cs-otx expression were found in nuclei of every blastomere, including animal blastomeres of the 110-cell stage embryo (Fig. 3G), although at the 32-cell stage the upregulation of Cs-otx was not so conspicuous (Fig. 3E). By contrast, Cs-otx expression was downregulated in the cadherin overexpressed embryos, and no in situ signals were detectable in these embryos (Fig. 3F,H).
A C. savignyi homolog of mammalian ttf1 (thyroid transcription factor gene 1), which encodes an NK-2 class homeodomain protein, was also identified as a possible β-catenin downstream gene. A cDNA for this gene was 2357 bp in length, and the predicted protein was 473 amino acids long (Fig. 5A) (DDBJ/GenBank/EMBL Accession Number AB057734). The C. intestinalis homologue of ttf1 (Ci-titf1) has been characterized and reported to play a significant role in endodermal cell differentiation (Ristoratore et al., 1999). The homeodomain of Cs-TTF1 showed a significant identity to that of C. intestinalis and of mammals (Fig. 5B).
As in the case of C. intestinalis Ci-titf1, the expression of Cs-ttf1 was first detected in A7.1, A7.2, A7.5, B7.1 and B7.2 blastomeres at the 64-cell stage (Fig. 1R). Cs-ttf1 expression in the endoderm lineage persisted during early gastrulation (Fig. 1S), and was then downregulated. It has been reported that Ci-titf1 is expressed once again in the endostyle primordium during metamorphosis of the larva to the adult (Ogasawara et al., 1999).
As shown in Fig. 3K, Cs-ttf1 expression was upregulated in the β-catenin mRNA-injected 110-cell embryos. However, Cs-ttf1 expression was not detected in β-catenin overexpressed 32-cell embryos (Fig. 3I), suggesting that β-catenin overexpression did not accelerate the timing of initiation of Cs-ttf1 expression. The gene expression was suppressed in the Ci-cadherin mRNA injected embryos at the 110-cell stage (Fig. 3L).
Functional assay of Cs-lhx3, Cs-otx and Cs-ttf1
In Ciona embryos, the nuclear localization of β-catenin in vegetal blastomeres was assessed using a specific antibody that becomes evident by the 32-cell stage, and this localization is most probably the first step in endodermal cell specification (Imai et al., 2000). As shown above, all three transcription factor genes are upregulated in β-catenin overexpressed embryos and downregulated in cadherin overexpressed embryos. In addition, all three genes are expressed in embryonic endodermal cells around the time of their fate restriction, suggesting an involvement of the three genes in endodermal cell differentiation. This issue was investigated by inhibition of possible gene function by morpholino oligonucleotides or by over and/or ectopic expression of the gene by injection of synthetic mRNAs into eggs. Nucleotide sequences of the morpholino oligonucleotides are listed in Table 1, and it has been shown that morpholino oligos specifically suppress certain gene function in Ciona embryos (Satou et al., 2001a).
Fig. 6A-C shows larvae that developed from eggs injected with lacZ morpholino oligo as a control. Although the lacZ morpholino completely inhibits lacZ development in larvae developed from eggs injected with lacZ mRNA (Satou et al., 2001a), this morpholino injection does not affect morphogenesis of Ciona embryos (Fig. 6B,C) and development endoderm-specific alkaline phosphatase (AP) (Fig. 6A).
Fig. 6D-F shows larvae developed from eggs injected with Cs-otx morpholino oligo. Development of AP was evident by histochemical staining (Fig. 6D). This was confirmed in two series of 41 experimental larvae (Table 2). However, Cs-otx-morpholino injected larvae showed a failure of central nervous system formation (Fig. 6E,F). They did not develop the two sensory organs, otolith and ocellus, that are able to be assessed by observing pigment cell differentiation (Fig. 6E,F). The development of the anterior head was also suppressed by Cs-otx morpholino, so that this region looks smooth (Fig. 6E,F). These results suggest that Cs-otx is not directly involved in the larval endodermal cell differentiation but is essential for the neuronal cell differentiation.
Injection of Cs-ttf1 morpholino oligo also did not affect the development of AP activity (Fig. 6G). This was confirmed by two series of 17 manipulated larvae (Table 2). The morphology of resultant larvae looked normal (Fig. 6H,I).
Ristoratore et al. (Ristoratore et al., 1999) showed that injection of Ci-titf1 mRNA resulted in an ectopic expression of AP in C. intestinalis embryos. We confirmed their result: injection of 1.5 pg of Cs-ttf1 mRNA caused an ectopic expression of AP so that many embryonic cells showed AP activity (Fig. 6J). To examine cells with ectopic AP activity, we took advantage of cleavage-arrested embryos. When early ascidian embryos are immersed in seawater containing cytochalasin B, cytokinesis but not nuclear division of blastomeres is blocked. Blastomeres of these cleavage-arrested embryos express differentiation markers depending on their lineage (Whittaker, 1973). Embryos injected with Cs-ttf1 mRNA were arrested at the 110-cell stage, because the developmental fate of almost all of the blastomeres is restricted to one tissue by this stage. AP activity in the cleavage-arrested embryos was examined when untreated control embryos hatched. In uninjected embryos, AP activity was detected in 10 vegetal cells, as expected based on the endoderm lineage (data not shown; cf. Fig. 7E). It became evident that in embryos injected with Cs-ttf1 mRNA, AP activity was evident in almost all of the blastomeres including the primary lineage muscle cells (Fig. 6K,L).
In contrast to Cs-otx and Cs-ttf1, morpholino oligonucleotides against Cs-lhx3 markedly inhibited AP development (Fig. 7A). The Cs-lhx3 morpholino oligo was designed to match 24 nucleotides between −39 and −16 in the 5′ UTR of the mRNA (Table 1) as A of the first ATG was regarded position 0. Fertilized eggs injected with Cs-lhx3 morpholino underwent cleavages that were slightly delayed and failed to undergo normal morphogenesis to form the larvae (Fig. 7A). These experimental larvae showed the expression of an epidermis-specific gene Cs-Epi1 (data not shown). However, as shown in Table 2, among four series of 54 experimental larvae, 32 did not develop AP activity, while 20 showed only reduced activity. Suppression of endodermal cell differentiation by Cs-lhx3 morpholino was examined by another differentiation marker of endodermal cells. We have isolated cDNA clone for a thyroid hormone receptor gene of C. savignyi and have characterized its endoderm-specific expression (Fig. 7B, insert; DDBJ/GenBank/EMBL Accession Number AB057767). Injection of Cs-lhx3 morpholino also inhibited development of this endoderm-specific gene expression (Fig. 7B).
The suppression of endodermal cell differentiation by Cs-lhx3 morpholino was rescued by injection of 9 pg of synthetic mRNA for Cs-lhx3. The synthetic Cs-lhx3 mRNA was designed to be initiated from position −4 and therefore the Cs-lhx3 morpholino does not recognize the synthetic mRNA. As a control, injection of 9 pg of synthetic mRNA for lacZ did not rescue the AP expression (Fig. 7C); AP expression was not detected in 20 out of the 25 experimental embryos injected together with Cs-lhx3 morpholino and lacZ mRNA (Table 3). By contrast, as shown in Fig. 7D and Table 3, more than a half of 23 experimental embryos injected together with Cs-lhx3 morpholino and Cs-lhx3 mRNA rescued the AP activity.
As an ectopic expression experiment, Cs-lhx3 mRNA was injected into fertilized eggs, and this injection promoted ectopic AP development (Fig. 7E,F). When this promotion was examined using cleavage-arrested embryos, AP activity was evident in almost all of the blastomeres, except in the primary lineage muscle cells (Fig. 7F).
As mentioned before, injection of 60 pg of delE-Ci-cadherin into fertilized eggs downregulates the β-catenin nuclear localization, which in turn results in the failure of AP development in resultant larvae (Imai et al., 2000). As shown in Fig. 7G and Table 4, embryos developed from eggs injected with 60 pg of delE-Ci-cadherin mRNA and 9 pg of lacZ mRNA showed a suppression of AP activity. However, when eggs were injected with 60 pg of delE-Ci-cadherin mRNA together with 9 pg of Cs-lhx3 mRNA, resultant embryos recovered the AP activity (Fig. 7H). This was confirmed in all of the 31 experimental embryos (Table 4).
All of these results strongly suggest that the early embryonic expression of Cs-lhx3 is involved in the endodermal cell differentiation in C. savignyi embryos.
Subtraction screening of β-catenin downstream genes in C. savignyi embryos
In the present study, we conducted subtractive hybridization screens of mRNAs between β-catenin overexpressed embryos and cadherin overexpressed embryos to identify potential β-catenin target genes in C. savignyi embryos. In early Ciona embryos, when β-catenin is mis- and/or overexpressed, presumptive notochord cells and epidermal cells change their fates into endodermal cells, whereas when β-catenin nuclear localization is downregulated by the overexpression of cadherin, endodermal cell differentiation is suppressed, which is accompanied by the differentiation of extra-epidermal cells (Imai et al., 2000). The enrichment of an endoderm-related gene Cs-HNF3 (Shimauchi et al., 2001) in β-catenin overexpressed library and that of an epidermis-specific gene Cs-Epi1 in cadherin overexpressed library were confirmed by Southern blot hybridization. In the previous study, we demonstrated that the development of primary muscle cells is not affected by overexpression of β-catenin nor by depletion of nuclear β-catenin, whereas cells with neural fates are transformed into endoderm by overexpression of β-catenin (Imai et al., 2000). Cs-ZICR1 (a ZIC-related gene 1 of C. savignyi) is expressed in both presumptive muscle cells (B6.2 and B6.4) and presumptive nerve cord cells (A6.2 and A6.4) at the 32-cell stage (K. S. Imai, Y. S. and N. S., unpublished). Using this gene probe, we determined whether the transcription activity of genes expressed in the vegetal cells is non-specifically upregulated by β-catenin overexpression and downregulated by depletion of nuclear β-catenin. As expected, the Cs-ZICR1 expression in muscle lineage cells was not altered in experimental embryos, whereas the gene expression in neuronal lineage cells was downregulated by β-catenin injection (K. S. Imai, Y. S. and N. S., unpublished). The present screening thus yielded cDNAs for twelve different genes, including a LIM-homeobox gene (Cs-lhx3), an otx gene (Cs-otx), a ttf1 gene (Cs-ttf1), two forkhead genes, a type II cadherin gene, a protocadherin gene, a netrin gene, a frizzled gene, a lefty/antivin homolog, an EPH gene and a gene encoding a protein with no known sequence similarity; the first three transcription genes were studied in detail.
An essential role for Cs-lhx3 in the endoderm differentiation of ascidian embryos
As shown in the present study, Cs-lhx3 is a possible target of nuclear localization of β-catenin in Ciona embryos, because Cs-lhx3 expression was upregulated in the β-catenin overexpressed embryos and downregulated in the cadherin overexpressed embryos (Fig. 3). The first embryonic expression of Cs-lhx3 is seen in A6.1, A6.2 and B6.1 blastomeres of 32-cell stage embryos (Fig. 1G), all three pairs being presumptive endodermal cells, and A6.1 and B6.1 are restricted at this stage to give rise to the endoderm only (Fig. 1B). This early and transient expression of Cs-lhx3 is found mainly in endodermal cells. Cs-lhx3 misexpression by injection of its synthetic mRNA causes development of extra cells with AP activity (Fig. 7F). In addition, its morpholino oligonucleotide blocks development of cells with AP activity (Fig. 7A). The Cs-lhx3 morpholino also completely suppresses expression of an endoderm-specific thyroid hormone receptor gene (Fig. 7B). This suppression was rescued by injection of Cs-lhx3 mRNA (Fig. 7D) but not by injection of lacZ mRNA (Fig. 7C). Furthermore, injection of Cs-lhx3 mRNA rescues development of cells with AP activity in cadherin-overexpressed embryos that would not express AP. These results strongly suggest that a LIM-homeobox gene Cs-lhx3 is one of the β-catenin downstream genes and its early expression in embryonic endodermal cells is responsible for their differentiation.
Future studies therefore should determine whether the early expression of Cs-lhx3 is directly controlled by the complex of β-catenin and Tcf/LEF-1. A preliminary investigation suggests that the 5′ flanking region of the Cs-lhx3 gene contains consensus sequence motifs for Tcf-binding site. One of these motifs is likely to be required for the Cs-lhx3 expression, although other factors are also required for proper Cs-lhx3 expression (K. S. Imai, Y. S. and N. S., unpublished). In addition, future studies should explore the mechanisms of Cs-lhx3 function, whether or not Cs-lhx3 directly regulates the transcription of Cs-AP gene or thyroid hormone receptor gene or other genes required for endodermal cell differentiation. A recent study of gene expression profile of C. intestinalis tailbud embryos has revealed that more than 30 genes are specifically expressed in the endodermal cells (Satou et al., 2001b). Counterparts of these C. intestinalis endoderm-specific genes may be expressed in C. savignyi embryonic endodermal cells. Therefore, it should also be determined how Cs-lhx3 controls the expression of these endoderm-specific genes.
During later embryogenesis, Cs-lhx3 is again expressed in one pair of cells in the nerve cord and a subset of brain cells, as in the case of Hrlim (Wada et al., 1995). Because most of the LIM-homeobox genes are involved in neuronal development and function (reviewed by Hobert and Westphal, 2000), it is conceivable that the ascidian LIM-homeobox genes are also involved in the differentiation of neuronal cells. Injection of Cs-lhx3 morpholino disrupts morphogenesis; therefore, we could not investigate the role of Cs-lhx3 in the differentiation of neuronal cells. This should be examined in future studies by experiments in which this gene is temporally controlled suppressed and/or misexpressed at different times in development. In addition, because the endoderm cells emanate signals required for the specification of notochord cells (Nakatani and Nishida, 1994; Nakatani et al., 1996) and mesenchyme cells (Kim et al., 2000) in Halocynthia embryos, whether Cs-lhx3 regulates genes associated with these cellular interactions is also an intriguing issue.
Roles for Cs-otx and Cs-ttf1 in the endoderm differentiation of ascidian embryos
Wada et al. (Wada et al., 1996) have shown that Hroth of H. roretzi begins to be expressed as early as the 32-cell stage and exhibits a rather complex spatial expression pattern in early embryos, namely, at the 32-cell stage Hroth is expressed in both involuting mesoendoderm and anterior ectoderm during gastrulation, and later expression is restricted to the sensory vesicle and anterior epidermis. C. intestinalis Ci-otx shows a similar but slight different expression pattern (Hudson and Lemaire, 2001). In addition, Wada and Saiga (Wada and Saiga, 1999) have shown that microinjection of Hroth mRNA into fertilized eggs leads to embryos with an expanded trunk and a reduced tail, in which development of muscle and notochord is effected. Hroth overexpression also causes ectopic formation of the anterior neuroectoderm, together with the suppression of epidermis development. They suggested that Hroth plays roles in both specification of mesoendodermal cells and anterior neuroectoderm formation (Wada and Saiga, 1999).
As in the case of Hroth and Ci-otx, Cs-otx is first expressed at the 32-cell stage in vegetal blastomeres of B6.1 (endoderm lineage), B6.2 (mesoderm lineage) and B6.4 (muscle and mesenchyme lineages), and at the 64-cell stage A-line endodermal cells begin to express Cs-otx. However, Cs-otx morpholino did not block AP expression in endodermal cells, and differentiation of mesenchyme, muscle and notochord cells appeared normal (Fig. 6D). Instead, Cs-otx morpholino oligo severely inhibited the formation of the central nervous system in that the two sensory pigment cells did not develop and the anterior head structure was affected. These results suggest that the primary function of otx during ascidian embryogenesis is associated with the patterning of the CNS, as in Drosophila and vertebrate embryos (Finkelstein and Boncinelli, 1994; Matsuo et al., 1995; Simone, 1998).
Nevertheless, expression of Cs-otx in mesoendoderm cells of early Ciona embryos suggests its role in the differentiation of the mesoendoderm cells. This possibility is also suggested in sea urchin embryos (Angerer and Angerer, 2000). In sea urchins, the nuclear localization of β-catenin is seen in vegetal blastomeres and is essential for vegetal fate determination of embryonic cells (Wikramanayake et al., 1998; Logan et al., 1999). This nuclear localization of β-catenin, in complex with Tcf (Vonica et al., 2000), may activate several transcription factor genes, including Otx (SpOtx) (Li et al., 1999). Because the Engrailed-SpOtx fusion protein reduced the expression of endoderm- and aboral ectoderm-specific genes, the SpOtx gene is likely to be involved in the endoderm differentiation (Li et al., 1999). The sea urchin Otx gene also controls the aboral-specific gene Ars (Kiyama et al., 2000). In sea urchin embryos, a lim1-related homeobox gene, Hplim1 is expressed transiently in the vegetal plate cells of hatching blastulae (Kawasaki et al., 1999). Overexpression experiments suggest that Hplim1 is involved in endoderm development. Therefore, whether Hplim1 is a downstream gene of β-catenin remains an intriguing issue.
Furthermore, an ascidian otx gene is highly likely to play an essential role in the pharynx formation after metamorphosis (Hinman and Degnan, 2000). Ascidian tadpole larvae are nonfeeding larvae and the gut does not fully differentiate until after metamorphosis of the larvae. One of the primary tissues that the endoderm differentiates into after metamorphosis is the pharynx. Hinman and Degnan (Hinman and Degnan, 2000) have shown that otx (Hec-Otx) of another ascidian Herdmania curvata is expressed in the pharynx. In addition, treatment of juveniles with retinoic acid suppresses expression of otx after metamorphosis, which results in development of adult ascidians lacking the pharynx (Hinman and Degnan, 1998; Hinman and Degnan, 2000). These results strongly suggest that the ascidian otx gene functions during very early embryogenesis in endomesoderm specification, then in neuroectoderm differentiation during larval formation, and later in the pharynx formation of juveniles after metamorphosis.
Ci-titf1 of C. intestinalis (Ristoratore et al., 1999; Ogasawara et al., 1999), Cs-ttf1 of C. savignyi (the present study) and Hrtitf1 of H. roretzi (Ogasawara et al., 1999) are ascidian members of the NK-2 gene family. These genes are expressed in the endodermal cells of the 64-cell to early gastrula stages, and later in the endostyle of young adults (Ristoratore et al., 1999; Ogasawara et al., 1999). Ristoratore et al. (Ristoratore et al., 1999) have shown that misexpression of these genes causes the development of extra cells with AP activity. As shown in the present study, the expansion of AP activity in extra cells by Cs-ttf1 is as strong as that of β-catenin or Cs-lhx3, suggesting a role of the ascidian NK-2 family gene in the embryonic endoderm differentiation, as suggested by Ristoratore et al. (Ristoratore et al., 1999). However, the expression of this gene is initiated at the 64-cell stage, and its morpholino oligonucleotide does not block the appearance of AP activity in the endodermal cells (Fig. 6G). These results suggest that the ascidian ttf1 gene is not involved in the initial phase of endodermal cell specification and/or differentiation but in the process of the maintenance of differentiated state. Because ascidian ttf1 seems to upregulate the AP gene activity, it is interesting to ask whether the Cs-TTF1 binds to cis-elements of endoderm-specific genes of Ciona embryos.
β-Catenin target downstream genes
In Xenopus and zebrafish, a key step in the establishment of the dorsoventral axis is nuclear localization of β-catenin, which is supplied by maternal transcript (Heasman et al., 1994; Funayama et al., 1995; Guger and Gumbiner, 1995; Kelly et al., 1995; Schneider et al., 1996; Larabell et al., 1997). Dorsally localized β-catenin forms a complex with Tcf-like transcription factor, and this complex is thought to trigger the zygotic activation by transcription factor genes: siamois (Brannon et al., 1997; Fan and Sokol, 1997) and twins (Laurent et al., 1997) in Xenopus embryos, and boz/dharma in zebrafish embryos (Fekany et al., 1999). These transcription factor genes are likely to act upstream of transforming growth factor β, which plays a pivotal role in the development of both axial mesoderm and anteroposterior neural patterning (Cadigan and Nusse, 1997; Moon and Kimelman, 1998; Sokol, 1999). Therefore, it is expected that the nuclear localization of β-catenin activates several other transcription factor genes that in turn may regulate other genes required for the differentiation of specific cell types in sea urchin and ascidian embryos or for cell-cell signaling cascades involved in axis formation in frog and fish embryos.
In conclusion, the present attempt to identify β-catenin target downstream genes in C. savignyi embryos demonstrates that a LIM-homeobox gene Cs-lhx3 is probably a direct target of β-catenin and that the expression of this gene is required for endodermal cell differentiation.
We thank 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 Grant-in-Aids from the Ministry of Education, Science, Sports and Culture, Japan to Y. S., K. S. I. and N. S.