β-Catenin is known to play important roles in the establishment of embryonic axes in various animals. In vertebrates, β-catenin functions in the establishment of the axis that is perpendicular to the animal-vegetal axis. Xenopus and zebrafish embryos show accumulation of β-catenin in the nucleus of cells on the organizer side of the blastula (Schneider et al.,1996; Bellipanni et al.,2006). As a consequence, the organizer is defined on this side and the axis is established (Larabell et al.,1997; Kelly et al.,2000). Sea urchin embryos form a gradient of nuclear β-catenin concentration with the highest density at the vegetal pole, and this nuclear β-catenin is crucial for specification of vegetal cell fates (Logan et al.,1999). In cnidarian, a more primitive organism, it has been shown that the asymmetric localization of nuclear β-catenin within the embryo is essential for establishing the oral–aboral axis and defining the site of gastrulation (Wikramanayake et al.,2003). Therefore, accumulation of nuclear β-catenin at the gastrulation site (vegetal region in sea urchin and oral region in cnidarian) appears to be a remarkably ancestral phenomenon in the animal kingdom.
β-Catenin was originally identified as a factor that interacts intercellularly with a cell adhesion molecule, cadherin (Ozawa et al.,1989; McCrea et al.,1991; Aberle et al.,1996), but was later found to play an important role in the canonical Wnt signaling pathway. Activity of β-catenin is regulated by a complex mechanism that includes phosphorylation (Amit et al.,2002; Liu et al.,2002; Yanagawa et al.,2002). Phosphorylated β-catenin is polyubiquitinated and degraded by proteasomes. Wnt is a secreted ligand and a trigger of the Wnt signaling pathway. Wnt transfers its signal intercellularly through its receptor, Frizzled. The signal activates Dishevelled (Dsh), which results in inhibition of phosphorylation of β-catenin by the Axin/GSK3β complex through direct interaction between Dsh and Axin. Unphosphorylated and stabilized β-catenin moves into the nucleus where it binds to TCF/LEF transcription factor to trigger the transcription of its target genes. This signaling pathway is conserved in a wide range of animals (Logan and Nusse,2004).
In ascidians, endoderm originates from the anterior-vegetal (A4.1) blastomere pair and the posterior-vegetal (B4.1) pair of the 8-cell-stage embryo, from three pairs of blastomere (A6.1, A6.3, and B6.1) at the 32-cell stage, and from five pairs (A7.1, A7.2, A7.5, B7.1, and B7.2) at the 64-cell stage. Every descendant cell of these five pairs of 64-cell-stage embryo contributes only to endoderm. In the ascidians Ciona intestinalis and C. savignyi, β-catenin is required for formation of mesoderm and endodermal tissues that are derived from the vegetal hemisphere. Ci-β-catenin protein is distributed in the cytoplasm of all blastomeres at the 16-cell stage, but at the 32-cell stage, it begins to accumulate in the nucleus only in the vegetal hemisphere. At the 110-cell stage, the accumulation is most evident in endoderm cells and decreases in other vegetal cells (Imai et al.,2000). When cadherin mRNA is overexpressed in the embryo to sequester Ci-β-catenin in the cytoplasm, thus preventing β-catenin from moving into the nucleus, the embryo fails to form endoderm and mesoderm tissues, suggesting that β-catenin is a regulatory factor for mesendodermal specification (Imai et al.,2000). Thus, β-catenin is used to specify vegetal cell fates in ascidians in the same manner as in echinoderms, but not in vertebrates. In Ciona, Cs-lhx3 and Cs-ttf1 (Ci-titf1 in C. intestinalis), downstream target genes of β-catenin, have been identified as zygotic genes that encode key transcription factors for endoderm specification (Ristoratore et al.,1999; Satou et al.,2001).
The cue that prompts β-catenin to move into the nucleus of specific cells in embryos has also been studied extensively. In sea urchins, maternal Dsh protein localized to the vegetal cortex is involved in the stabilization of β-catenin, although overexpression of Dsh in the whole embryo cannot lead to accumulation of β-catenin in the nuclei of the animal hemisphere (Weitzel et al.,2004). In Xenopus, two controversial models have been proposed. GSK3β binding protein (GBP) and Dsh translocate from the vegetal to the prospective organizer side through involvement of kinesin protein, a motor protein, borne on microtubules, which form at the vegetal hemisphere after fertilization (Miller et al.,1999; Weaver et al.,2003). In the organizer side of the embryo, GBP and Dsh inhibit the activities of GSK3β and Axin, respectively. As a result, β-catenin is stabilized and transported into the nuclei there (Weaver and Kimelman,2004). Recently, another model has been proposed. The translated product of maternal Wnt11 mRNA located at the vegetal pole is secreted extracellularly and acts as a signal molecule in establishing the organizer side (Tao et al.,2005). In these two models, the cues for nuclear localization of β-catenin in Xenopus early embryos are entirely different, and remain to be elucidated.
In the ascidian, Halocynthia roretzi, some components of the Wnt pathway have been isolated by previous studies. Two Wnt molecules have been identified. One is Wnt5α, whose mRNA is maternally localized to the posterior-vegetal cortex of fertilized eggs after the cytoplasmic and cortical movement, the so-called ooplasmic segregation (Sasakura et al.,1998; Sasakura and Makabe,2001). The other is Wnt5β, which is expressed zygotically in muscle precursor blastomeres (Miya and Nishida,2002). Dsh mRNA of H. roretzi is present maternally and distributed uniformly (Iida and Nishida, unpublished). However, the cue that prompts β-catenin to move into the nucleus in the vegetal hemisphere has not been elucidated. Previous egg cytoplasm transfer experiments have suggested that endodermal determinants are localized in the egg (Nishida,1993; Yamada and Nishida,1997). It is likely that these determinants are the cue for accumulation of β-catenin in the nuclei. These determinants appear to be distributed in the form of a gradient, with the highest activity at the vegetal pole in the unfertilized egg, and then concentrated at the vegetal pole during the first phase of cytoplasmic and cortical reorganization after fertilization. Endoderm formation is inhibited by ultraviolet irradiation from the vegetal pole at this stage (Marikawa and Satoh,1995). The distribution of the determinants extends to the entire vegetal hemisphere during the second phase of reorganization. Deletion of the ooplasm at various stages of the first cell cycle indicates that the determinants for gastrulation move in a similar manner (Nishida,1996).
The mechanism for nuclear localization of β-catenin has been studied extensively in sea urchins and Xenopus. Ascidians are nonvertebrate chordate and positioned between them in phylogeny. Therefore, it is interesting to study these mechanisms in ascidian embryos to gain insights into the evolution of the regulatory system for β-catenin nuclear accumulation. In the ascidian, Ciona, it has been reported that β-catenin is crucial for the specification of mesendoderm. However, the cue for nuclear accumulation of β-catenin is still unknown. In the present study, we first confirmed that Hr-β-catenin is required for the formation of various tissues in the vegetal hemisphere including endoderm. We then investigated the cue for the nuclear localization of β-catenin and the mechanism underlying the expression of target genes downstream from β-catenin using a specific antibody against Hr-β-catenin and various probes. Considering the results of previous studies with sea urchins and Xenopus, we investigated the roles of Wnt and Dsh in nuclear accumulation of Hr-β-catenin and expression of its target genes. We found that accumulation of nuclear Hr-β-catenin after the 64-cell stage requires zygotic Hr-Wnt5α and that expression of some β-catenin target genes is modulated by Hr-Wnt5α and Hr-Dsh.
Hr-β-Catenin Is Required for Formation of Mesendoderm in Halocynthia roretzi
We first confirmed whether β-catenin is involved in the formation of mesendoderm in H. roretzi, as reported in C. savignyi and C. intestinalis (Imai et al.,2000). First, we searched the H. roretzi expressed sequence tag (EST) database MAGEST (Kawashima et al.,2000,2002; Makabe et al.,2001) for a β-catenin homologue, and found an EST clone, 003B14. This clone is designated as Hr-β-catenin, consisting of a single and full open reading flame (ORF) encoding 786 amino acids, including characteristic armadillo repeats and putative sites of phosphorylation by CK1 and GSK3β, which is involved in its degradation by the ubiquitin–proteasome system (Kikuchi et al.,2006; Fig. 1A). Homology searches for the full ORF using NCBI BLAST revealed that the Hr-β-catenin protein had the highest homology (78%) with that of C. intestinalis and the second highest (76%) with that of C. savignyi. To identify the spatial and temporal pattern of expression of the Hr-β-catenin gene, we carried out in situ hybridization with embryos from the egg to the tail bud stage. In situ signals were observed almost uniformly in cytoplasm from the fertilized egg to the 32-cell stage (Supplementary Figure S1A–G, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). Relatively higher signals were detected in mesenchyme precursors at the 64-cell and 110-cell stages (Supplementary Figure S1E,F). In the tail bud, mesenchyme, trunk lateral cells and nerve cord showed evident signals (Supplementary Fig. 1G). The sense probe did not give any significant staining (Supplementary Figure S1A′–G′).
To determine the function of Hr-β-catenin, we injected fertilized eggs with antisense morpholino oligonucleotide (MO) against Hr-β-catenin to inhibit its translation. The injected embryos showed defects in gastrulation and developed a permanent blastula-like morphology lacking a tail and morphologically evident notochord (Fig. 1B,C). We next investigated whether the endoderm was formed in MO-injected embryos by detection of alkaline phosphatase (AP) activity, an endoderm marker (Whittaker,1977). In the injected embryos, AP activity was not detected (Fig. 1J,K). In contrast, the entire surface of the embryos showed expression of a specific epidermis marker, Epi-2 (Fig. 1D,E). These embryos also lacked notochord and mesenchyme, as revealed using specific antibodies, Not-1 and Mch-3 (Fig. 1F–I). This phenotype is almost identical to that of larvae in which the vegetal pole cytoplasm has been removed from eggs just after the first phase of ooplasmic segregation (Nishida,1996). These results are consistent with the loss of function of Cs-β-catenin in Ciona. Thus, Hr-β-catenin also plays conserved roles in mesendoderm formation in H. roretzi.
Nuclear Localization of Hr-β-catenin Protein During the Cleavage Stage
A certain amount of maternal protein was already present in eggs, as detected by Western blotting (Supplementary Figure S2A). The amount of the protein increased gradually and then decreased during the cleavage stages. The specificity of the antibody was confirmed by lack of nuclear staining in MO-injected embryos (compare Supplementary Figure S2B–C′ with Fig. 2). Probably because of the maternal protein already present before MO injection, staining of the cell boundary persisted in the MO-injected embryos. Western blotting (Supplementary Figure S2D) showed that the intensities of the bands were significantly reduced in MO-injected embryos and enhanced in embryos that had been injected with Hr-β-catenin mRNA.
We observed localization of Hr-β-catenin protein during the cleavage stage in detail by immunostaining of embryos from the 8-cell to 110-cell stages using this antibody. In every embryo, Hr-β-catenin protein was evident at the cell–cell boundary (Fig. 2), consistent with its conserved role in cell adhesion by binding to the intracellular domain of cadherin and, thus, providing additional evidence of the specificity of the antibody. In 8-cell and 16-cell-stage embryos, Hr-β-catenin protein was faintly detected in a subcellular structure known as the centrosome-attracting body (CAB; Fig. 2A–E,A′–E′, arrows; Hibino et al.,1998), but not in the nucleus. In 32-cell-stage embryos, Hr-β-catenin protein first appeared in all the nuclei of vegetal blastomeres (Fig. 2F–H,F′–H′,O). In 64-cell-stage embryos, the nuclear accumulation of Hr-β-catenin protein in the vegetal blastomeres was restricted to A7.1, A7.2, A7.5, and B7.1, which give rise only to endoderm (Fig. 2I–L,I′–L′,O). In 110-cell-stage embryos, Hr-β-catenin protein was no longer evident in the nuclei of vegetal blastomeres (Fig. 2M–N,M′–N′). These results indicate that Hr-β-catenin protein accumulates transiently in the nucleus of only vegetal blastomeres at the 32-cell and 64-cell stages. The nuclear accumulation of Hr-β-catenin protein only in the vegetal hemisphere supports the role of Hr-β-catenin in the specification of mesendoderm. In Ciona, nuclear β-catenin first appears at the 32-cell stage, but persists until at least the 110-cell stage (Imai et al.,2000). This difference is probably due to differences in species or the specificity of the antibody because an antibody against sea urchin β-catenin was used in Ciona experiment.
Hr-Wnt5α Is Required for Maintaining Nuclear Localization of Hr-β-catenin at the 64-Cell Stage and Expression of Hr-lim in the Anterior Vegetal Blastomeres
In the ascidian H. roretzi, there is a maternally expressed Wnt, Hr-Wnt5α (Sasakura et al.,1998; Sasakura and Makabe,2001). Therefore, we investigated whether Hr-Wnt5α involves nuclear accumulation of Hr-β-catenin protein.
First, we injected fertilized eggs with antisense MO against Hr-Wnt5α, and observed the accumulation of Hr-β-catenin protein by immunostaining using the specific antibody at the 32- and 64-cell stages. Injection of the MO had no effect on Hr-β-catenin protein at the cell–cell boundary in the 32- and 64-cell embryos. Hr-β-catenin protein appeared in nuclei of the vegetal blastomeres in most MO-injected 32-cell embryo in 95% of cases (n = 19, Fig. 3A–B′). In contrast, the 64-cell embryos injected with the MO did not show nuclear accumulation of Hr-β-catenin protein in 92% of cases (n = 12, Fig. 3C–E′). These results suggest that Hr-Wnt5α regulates nuclear accumulation of Hr-β-catenin protein in the A7.1, A7.2, A7.5, and B7.1 blastomeres at the 64-cell stage, but not at the 32-cell stage.
Next, we investigated the expression of β-catenin downstream target genes by in situ hybridization. We examined the expression of FGF9/16/20, FoxA, FoxDa, and lim, which are known as β-catenin target genes in Ciona and Halocynthia (Satou et al.,2001; Imai et al.,2002a,b; Imai,2003). The embryos injected with control MO showed no difference in the expression of any of the genes examined (Fig. 4A–E). Hr-FGF9/16/20, Hr-FoxA, and Hr-FoxDa were expressed normally in embryos injected with Hr-Wnt5α MO at the 32-cell stages (Fig. 4F–H). On the other hand, expression of Hr-lim was eliminated in the anterior-vegetal blastomeres A6.1 and A6.3 at the 32-cell stage (40%; n = 15) and A7.2, A7.3, and A7.7 at the 64-cell stage (100%; n = 9). Expression in the other blastomeres was reduced to some extent (Fig. 4I,J). These results suggest that Hr-Wnt5α is essential for the expression of Hr-lim in the anterior-vegetal blastomeres.
In the embryos injected with Hr-Wnt5α MO, the anterior expression of Hr-lim was inhibited in 40% of cases at the 32-cell stage, whereas nuclear localization of Hr-β-catenin was normal at this stage. There are two possible explanations for this. One is that the quantity of nuclear Hr-β-catenin might have been reduced to some extent and did not reach the threshold for active transcription of Hr-lim in the 32-cell embryos injected with Hr-Wnt5α MO, although this slight decrease of nuclear Hr-β-catenin could not be detected by nonquantitative immunostaining. The other is that the expression of Hr-lim is regulated by a mechanism independent of Hr-β-catenin. To address these possibilities, we attempted to rescue Hr-lim expression by overexpression of Hr-β-catenin by injecting a stable form of Hr-β-catenin mRNA (Fig. 1A). After co-injection of the stable form of Hr-β-catenin mRNA and MO against Hr-Wnt5α, expression of Hr-lim was monitored. The signals in the anterior-vegetal blastomeres were rescued at the 32-cell (86%; n = 8) and 64-cell stages (91%; n = 11; Fig. 4K,L). This finding suggests that Hr-Wnt5α promotes the transcription of Hr-lim mRNA through the canonical Wnt/β-catenin pathway, thus supporting the former possibility. If this is the case, then Hr-Wnt5α appears to be involved in the accumulation of Hr-β-catenin not only at the 64-cell stage but also at the 32-cell stage to some extent. In contrast, embryos injected with Hr-Wnt5α MO showed normal expression of Hr-FGF9/16/20, Hr-FoxA, and Hr-FoxDa at the 32-cell stage (Fig. 4F–H). One possible explanation for this is that the thresholds for promoting the transcription of these genes are lower than that of Hr-lim, although other possibilities cannot be ruled out.
Zygotic Hr-Wnt5α Regulates Nuclear Accumulation of Hr-β-catenin Protein and the Expression of Hr-lim
Maternal Hr-Wnt5α mRNA is localized to the posterior-vegetal region of fertilized eggs and partitioned into posterior-most vegetal (B-line) blastomeres during cleavage stages. In addition, it has been reported that the zygotic expression is first detectable in some vegetal blastomeres at the early 64-cell stage (Sasakura et al.,1998). We re-examined the zygotic expression of Hr-Wnt5α mRNA by in situ hybridization, and found that it was detectable in the vegetal blastomeres, A6.1 and B6.1, at the 32-cell stage (data not shown). Because maternal Hr-Wnt5α RNA is inherited only by the B4.1 blastomere, to test whether maternal or zygotic expression of Hr-Wnt5α is involved in the nuclear accumulation of Hr-β-catenin protein and expression of Hr-lim, we injected each of the A4.1 or B4.1 blastomeres with Hr-Wnt5α MO at the eight-cell stage and detected the localization of Hr-β-catenin protein and Hr-lim expression.
When both the left and right blastomeres of the A4.1 pair were injected, nuclear accumulation of Hr-β-catenin was lost in the entire embryo at the 64-cell stage in 80% of cases (n = 10; Fig. 5B,B′). Nuclear accumulation of Hr-β-catenin was also lost in the B7.1 blastomere in these embryos. The expression of Hr-lim was eliminated in the anterior-vegetal blastomeres A7.1, A7.2, A7.3, A7.5, and A7.7 at the 64-cell stage in 73% of cases (n = 11; Fig. 5F). This result was somewhat different from the previous result for unknown reason in which the expression was retained in A7.1 and A7.5 in embryos resulting from injection of MO into fertilized eggs (Fig. 4J). In the embryos injected with control MO, Hr-lim expression was quite normal (Fig. 5E). Interestingly, embryos in which the B4.1 pair had been injected showed normal nuclear accumulation of Hr-β-catenin protein in 83% of cases (n = 6), and normal Hr-lim expression in 91% (n = 11; Fig. 5C,D,H). As maternal Hr-Wnt5α mRNA is partitioned into B-line cells, these results suggest that zygotic Hr-Wnt5α expression in the A-line blastomeres plays a crucial role as a signaling source that triggers activation of the Wnt/β-catenin pathway for nuclear accumulation of Hr-β-catenin protein and expression of Hr-lim.
Localization of Hr-Dsh Protein During Cleavage Stages
Dsh is a component of the Wnt/β-catenin pathway for stabilization of β-catenin protein, resulting in β-catenin nuclear accumulation. It has been reported that in sea urchin embryos, stabilization of β-catenin protein requires localization of Dsh to the vegetal region (Weitzel et al.,2004). To reveal the localization of Hr-Dsh protein, we searched the H. roretzi EST database (MAGEST) for a Dsh homologue. We found an EST clone, 176E02, which is predicted to include a full-length ORF and then determined its entire nucleotide sequence (GenBank accession no. AB290435). The expression of Hr-Dsh was maternal and ubiquitous during cleavage stages (data not shown). We generated a polyclonal antibody against Hr-Dsh. A certain amount of maternal protein is already present in eggs, and this was detected by Western blotting (Supplementary Figure S3A). The amount of the protein quickly decreased during the very early stages and then increased again from the four-cell stage. The specificity of the antibody was confirmed by Western blotting with MO-injected and mRNA-injected embryos (Supplementary Figure S3A), and immunostaining of MO-injected embryos (Supplementary Figure S3B–D′).
Localization of Hr-Dsh protein during cleavage stages was examined by immunostaining. No localization to a specific region was recognized in eggs. At the eight-cell stage, Hr-Dsh protein was present in the CAB (Fig. 6A–B′, arrows). From the 16-cell to 64-cell stages, Hr-Dsh was detected in the nuclei of all animal blastomeres and almost all vegetal blastomeres, with the exception of B7.3 and B7.4 at the 64-cell stage and their progenitors (Fig. 6C–L′). At the 110-cell stage, it was present in the nucleus of all animal blastomeres and in the A8.7, A8.8, A8.15, A8.16, and A7.6 blastomeres of the vegetal hemisphere (Fig. 6M–N′). Recently, it has been shown that Dsh functions in the nucleus to stabilize β-catenin in Xenopus (Itoh et al.,2005). Therefore, nuclear localization of Hr-Dsh is consistent with this report. In addition, a weak signal was observed at the cell–cell boundary from the 16-cell to the 110-cell stages. This may be consistent with the observation that Dsh is localized to the cell membrane in response to Frizzled activation in noncanonical PCP pathway (Axelrod et al.,1998). However in general, Dsh is known to function within the cytoplasm to prevent the phosphorylation of β-catenin that leads to its degradation.
Hr-Dsh Is Required for Expression of Hr-lim in the Posterior-Vegetal Blastomeres of the 64-Cell Embryo
To determine the role of Hr-Dsh in early embryogenesis of H. roretzi, we injected fertilized eggs with antisense MO against Hr-Dsh, and then examined the localization of Hr-β-catenin protein and the expression of β-catenin target genes. In the embryos injected with the MO, Hr-β-catenin protein showed normal nuclear accumulation at the 32- and 64-cell stages in 100% (n = 12) and 89% (n = 9) of cases, respectively (Fig. 7A–D). The expression of Hr-FoxDa was decreased in 67% (n = 12) of cases, whereas that of Hr-FGF9/16/20, and Hr-FoxA was normal at the 32-cell stage in 100% (n = 13 and n = 14, respectively; Fig. 7E–G). Expression of Hr-lim was detected at the 32-cell stage in only 46% (n = 13) of cases. However, at the 64-cell stage, the anterior-vegetal blastomeres, A7.1, A7.2, A7.3, A7.5, and A7.7, regained Hr-lim expression, whereas the posterior-vegetal blastomeres, B7.1, B7.2, and B7.3, did not (100%, n = 7, Fig. 7H,I). These results suggest that Hr-Dsh is not involved in the accumulation of Hr-β-catenin protein in nuclei. Therefore, the expression of both Hr-FoxDa and Hr-lim especially in the posterior blastomeres is regulated by Hr-Dsh in parallel with the β-catenin pathway. It is also suggested that the mechanism by which Hr-lim expression is initiated differs between the anterior and posterior regions.
β-Catenin plays pivotal roles in mesendoderm specification in the vegetal hemisphere of early embryos of the ascidians, Ciona and Halocynthia. Nuclear localization of β-catenin was observed in the entire vegetal hemisphere at the 32-cell stage, and then became restricted to endoderm precursor blastomeres at the 64-cell stage. We analyzed the roles of Hr-Wnt5α and Hr-Dsh in the Wnt/β-catenin pathway and in the expression of genes downstream from Hr-β-catenin during early embryogenesis in Halocynthia.
Zygotic Hr-Wnt5α Is Required for Maintaining Nuclear Localization of Hr-β-catenin at the 64-cell Stage
The present results obtained by Hr-Wnt5α knockdown show that Hr-Wnt5α is essential for the accumulation or maintenance of Hr-β-catenin in the nuclei of vegetal cells at the 64-cell stage. However, nuclear localization of Hr-β-catenin is initiated independently of Hr-Wnt5α at the 32-cell stage. The mechanisms responsible for initiation of nuclear accumulation of Hr-β-catenin at the 32-cell stage are still unknown. One possibility is that maternally supplied Hr-Wnt5α protein or the protein translated before the MO injection is sufficient to regulate Hr-β-catenin nuclear localization at the 32-cell stage. We also cannot rule out the possibility that Hr-Wnt5α MO may not sufficiently suppress Hr-Wnt5α translation to exert its effect at the 32-cell stage. In other possibility, it is likely that unknown factors localized in the egg cytoplasm would be involved in this process (Nishida,1993; Yamada and Nishida,1997). Wnt5 is known to function in the noncanonical pathway (Tada et al.,2002; Zhu et al.,2006). However, recent studies have found that Wnt5a promotes nuclear localization of β-catenin through the canonical pathway in cultured cells (Mikels and Nusse,2006). In addition, Xenopus Wnt11, which has been shown to be a noncanonical pathway ligand, plays a role in the canonical pathway to modulate the accumulation of β-catenin in the nucleus during early embryogenesis (Tao et al.,2005). Therefore, it is possible that ascidian Hr-Wnt5α may also function in the canonical pathway.
When MO against Hr-Wnt5α was injected only into A-line cells, accumulation of Hr-β-catenin in the nucleus was inhibited in the entire vegetal hemisphere, including the B7.1 blastomere, at the 64-cell stage (Fig. 5B). As maternal Hr-Wnt5α mRNA is preferentially partitioned into B-line cells, Hr-Wnt5α that is translated from zygotically expressed Hr-Wnt5α mRNA and secreted from A-line cells is of primary importance, probably transferring the signal also to the B7.1 blastomere that is in contact with the A-line. Hr-Wnt5α translated from maternal and zygotic mRNA in the B-line is dispensable, as injection of Hr-Wnt5α MO into B-line cells resulted in embryos showing normal nuclear localization of Hr-β-catenin. It is still unknown why nuclear localization of Hr-β-catenin is not observed in the animal hemisphere where zygotic Hr-Wnt5α is also expressed, although its expression in the animal hemisphere starts slightly later than that in the vegetal hemisphere.
Functions of Hr-β-catenin in Gene Expression
Accumulation of Hr-β-catenin in the nucleus was observed in all the vegetal blastomeres at the 32-cell stage, but not at earlier stages. Concomitantly, transcription of Hr-β-catenin target genes became evident at the 32-cell stage. However, detailed studies have shown that Hr-FGF9/16/20 first becomes detectable at the 8-cell stage and Hr-FoxA and Hr-FoxDa at the 16-cell stage (Shimauchi et al.,1997). This situation is the same in Ciona. Expression of these genes at the 16-cell stage was actually suppressed in embryos injected with MO against Hr-β-catenin (data not shown), showing that this early expression also depends on Hr-β-catenin. Therefore, these early transcription events might be initiated by a low level of nuclear Hr-β-catenin that is undetectable by immunostaining, and the expression of these genes may be accelerated by a higher level of nuclear Hr-β-catenin at the 32-cell stage.
Nuclear Hr-β-catenin was detected only in the presumptive endoderm blastomeres at the 64-cell stage, whereas expression of the target genes persisted in various vegetal blastomeres, including nonendoderm blastomeres. This result suggests that a high level of nuclear Hr-β-catenin is no longer required for maintaining the expression of Hr-β-catenin target genes, except for Hr-lim, from the 64-cell stage onward. Therefore, nuclear accumulation of Hr-β-catenin mainly functions to promote the expression of target genes and determine cell fates up to the 32-cell stage. These events are consistent with our observation that nuclear localization of Hr-β-catenin is eliminated at the 64-cell stage in Hr-Wnt5α MO-injected embryos, whereas expression of Hr-FGF9/16/20, Hr-FoxA, and Hr-FoxDa is initiated and maintained normally.
Regulation of Hr-lim Expression
Hr-Wnt5α and Hr-Dsh knockdown experiments revealed that regulation of Hr-lim expression was unexpectedly complex and dependent on positioning within the embryo. Hr-lim expression in cells of the anterior region (A-line, especially the A7.2, A7.3, and A7.7 blastomeres) was regulated by nuclear Hr-β-catenin activated by the Hr-Wnt5α signal. Knockdown of Hr-Wnt5α resulted in the absence of Hr-lim expression in the anterior region at the 32-cell stage, although nuclear Hr-β-catenin appeared to be unaffected at that stage. One possible explanation is that the amount of nuclear Hr-β-catenin was slightly reduced to below the threshold necessary for supporting Hr-lim expression, although this was undetectable by immunostaining, and the thresholds for expression of the other genes downstream from Hr-β-catenin might be lower than that of Hr-lim. In contrast to the anterior region, Hr-lim expression in the posterior region (B-line) required Hr-Dsh even if nuclear accumulation of Hr-β-catenin was normal in Hr-Dsh MO-injected embryos. Therefore, Hr-lim expression in the posterior region depends on nuclear Hr-β-catenin as well as a yet unknown mechanism that involves Hr-Dsh and is independent of the Hr-β-catenin pathway. It is also likely that Hr-FoxDa expression is regulated in a similar way.
Knockdown of Hr-β-catenin by MO injection led to complete loss of endoderm, indicating that the function of Hr-β-catenin in endoderm formation is conserved between echinoderms and ascidians. However, our previous study showed that embryos injected with Hr-Wnt5α MO formed endoderm normally (Nakamura et al.,2006). In these embryos, nuclear Hr-β-catenin was observed at the 32-cell stage but not at the 64-cell stage. Therefore, nuclear Hr-β-catenin before the 32-cell stage would be sufficient for endoderm formation. However, there is an inconsistency between Halocynthia and Ciona, in which lim is necessary and sufficient for endoderm formation (Satou et al.,2001). Our results in Halocynthia indicated that embryos injected with Hr-Wnt5α MO lost their expression of Hr-lim in the anterior region, but the anterior blastomeres were nevertheless able to develop into endoderm. Therefore, it is necessary to examine directly whether or not Hr-lim is dispensable for endoderm formation in Halocynthia.
Nuclear Accumulation of Hr-Dsh Protein
In sea urchin embryos, it is proposed that maternal Dsh protein localized to the vegetal cortex of the egg is responsible for nuclear accumulation of β-catenin in the vegetal hemisphere (Weitzel et al.,2004). However in the ascidian, such vegetal localization in eggs and embryos was not observed, although some maternal Dsh protein was detected by Western blotting. Hr-Dsh is localized to the nucleus of almost every blastomere after the 16-cell stage. It has generally been considered that Dsh is localized in the cytoplasm and interacts with Axin complex to inhibit its function (Wallingford and Habas,2005). A recent study has demonstrated that Xenopus Dsh protein has a nuclear localization signal and moves into the nucleus, where it stabilizes β-catenin (Itoh et al.,2005). Hr-Dsh also possesses a sequence similar to the nuclear localization signal (304-IKLT-307). However, ascidian embryos injected with Hr-Dsh MO showed normal nuclear localization of Hr-β-catenin. One possibility is that maternal Hr-Dsh protein that is already present in unfertilized eggs or that is translated before the MO injection might be enough to support canonical Wnt signaling pathway. We cannot also rule out a possibility that the effect of Hr-Dsh MO may not be sufficient to completely inhibit Hr-Dsh translation. Another possibility is that there might be other Hr-Dsh genes in the Halocynthia genome than we analyzed in the present study. In Hr-Dsh MO-injected embryos, expression of Hr-lim in the posterior region and overall Hr-FoxDa expression was down-regulated. Therefore, Hr-Dsh is likely to promote transcription of these genes through an unknown mechanism that is probably independent of β-catenin nuclear localization because nuclear localization of β-catenin did not seem to be affected by Hr-Dsh MO injection.
The present study has revealed the presence of complex mechanisms governing the nuclear localization of β-catenin and regulation of genes downstream from β-catenin in ascidian early embryos. Such complexity would be necessary for promoting the expression of many target genes that have similar but distinct temporal and spatial expression patterns. In future studies, it will be of primary importance to search for and identify factors that initiate the nuclear accumulation of Hr-β-catenin at the 32-cell stage. Transcriptional activation by β-catenin is known to be regulated by various cofactors, such as kaiso, p120-catenin, canopy-1, and TCF (Liu et al.,2005; Park et al.,2005; Hirate and Okamoto,2006). Clarification of transcriptional regulation by these cofactors, including Hr-Dsh in H. roretzi, will help to yield a complete understanding of animal–vegetal axis establishment in ascidian embryos.
Animals and Embryos
Adults of Halocynthia roretzi were purchased from fishermen near the Otsuchi International Coastal Research Center, Ocean Research Institute, University of Tokyo, Iwate, Japan, and near the Asamushi Research Center for Marine Biology, Tohoku University, Aomori, Japan. Naturally spawned eggs were fertilized with a suspension of non-self sperm. Embryos were cultured in Millipore-filtered seawater containing 50 μg/ml streptomycin sulfate and 50 μg/ml kanamycin sulfate at 9–13°C.
Preparation of Antibody Against Hr-β-catenin and Hr-Dsh
A 614-basepair (bp) fragment, consisting of 1 to 204 amino acid residues of Hr-β-catenin and a 393-bp fragment, consisting of 1 to 131 amino acid residues of Hr-Dsh, were subcloned in-frame into the pGEX-4T-3 expression vector (Amersham) with the correct orientation. Escherichia coli DH5α expressing glutathione S-transferase (GST)-Hr-β-catenin or -Hr-Dsh fusion protein after IPTG induction was solubilized in phosphate-buffered saline (PBS), and the extract was applied to glutathione-immobilized agarose beads to trap the fusion proteins. To release the Hr-β-catenin, the GST-Hr-β-catenin was digested with thrombin protease (Amersham). The Hr-β-catenin fragment was further purified by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis and dialyzed against PBS. GST-Hr-Dsh was eluted with elution buffer (50 mM Tris/HCl, pH 8.8, and 20 mM glutathione reduced form) without thrombin digestion. The isolated Hr-β-catenin or GST-Hr-Dsh was emulsified in TiterMax Gold (CytRx Corporation, Los Angeles, CA) and injected subcutaneously into rabbits followed by six successive booster injections. The antisera were purified by affinity chromatography using each of the proteins as an antigen, and the immunoreactivity was verified by Western blotting of the fusion proteins.
In Situ Hybridization and Immunostaining
Whole-mount in situ hybridization was performed according to the standard protocol (Miya et al.,1997). Specimens were hybridized with digoxigenin-labeled Hr-β-catenin, Hr-FGF9/16/20 (Kumano et al.,2006), Hr-FoxA (Shimauchi et al.,1997), Hr-FoxDa (Kumano et al.,2006), and Hr-lim (Wada et al.,1995) probes. Whole-mount immunostaining was carried out as described previously (Miyawaki et al.,2003), with slight modifications. Embryos were fixed by gentle shaking in 2% paraformaldehyde in seawater for 10 min at room temperature. Fixed specimens were permeabilized with 100% methyl alcohol and hydrated with PBST (PBS containing 0.01% Triton-X). The fixed specimens were first immersed in blocking solution (PBST containing 2% goat serum) to block nonspecific signals. The specimens were then sequentially incubated with the anti–Hr-β-catenin antibody or anti–Hr-Dsh antibody solution as the primary antibody solution. Primary antibodies were diluted in blocking solution at 7.5 μg/ml. After incubation with each antibody solution (1 hr at room temperature), the specimens were thoroughly washed with PBST. They were then incubated with horseradish peroxidase (HRP) -conjugated anti-rabbit IgG secondary antibody (Simple Stain MAX PO; Nichirei, Tokyo, Japan) diluted twofold with PBST. Immunofluorescence detection was performed according to the standard protocol using a TSA-Plus Cyanin 3 system (Perkin-Elmer Life Science). The nuclei of the specimens were counterstained with 1 mM SYTOX Green (Molecular Probes) dissolved in Tris buffer (50 mM Tris/HCl, pH 7.4, 150 mM NaCl) for 1 h at room temperature. Then, the specimens were mounted in 80% glycerol. Fluorescent images were obtained with a confocal scanning microscope (CSU10, Yokogawa, Musashino, Tokyo).
Microinjection of MOs and Synthetic mRNA
To suppress the functions of Hr-β-catenin, Hr-Wnt5α, and Hr-Dsh proteins, we used antisense morpholino oligonucleotides (MOs; Gene Tools). The nucleotide sequence of Hr-β-catenin MO was 5′-GGCTCATTAACATCTCGGCCATGAT-3′, that of Hr-Wnt5α MO was 5′-ATTGTATTCTTGTCATTCCGACCAT -3′, and that of Hr-Dsh MO was 5′-TATGGCGGAAGAAACCAAGATAGTT-3′. As a control MO, we used standard control oligo (5′-CCTCTTACCTCAGTTACAATTTATA-3′; Gene Tools). The stable form of Hr-β-catenin mRNA, which is similar to that used in Xenopus (Yost et al.,1996; Fig. 1A), was transcribed from pBluescriptHTB (Akanuma et al.,2002) containing its open reading frame lacking the N-terminal (1–38 amino acid reduces) that includes putative phosphorylation sites, using a mMessage mMachine kit (Ambion) and a Poly(A) Tailing kit (Ambion). MO and synthetic mRNA were dissolved in sterile water and injected into fertilized eggs or eight-cell stage embryos as described by Miya et al. (1997). Solution containing 1–1.5 μg/μl Hr-β-catenin MO, 5 μg/μl Hr-Wnt5α Hr-Dsh, and control MOs was injected into a fertilized egg or blastomeres. Synthetic mRNA was injected at 1–2 μg/μl. The injected amount was one fourth to one fifth of the diameter of an egg or a blastomere (approximately one hundredth of the volume).
Monitoring of Tissue Formation and Inhibition of Cell Division
Differentiation of endoderm was monitored by histochemical staining for AP. The reaction deposits a purple product, as described previously (Whittaker and Meedel,1989). Formation of notochord was monitored by staining with the Not-1 monoclonal antibody. It recognizes a notochord-specific antigen at the middle tail bud stage (Nishikata and Satoh,1990; Nakatani and Nishida,1994). The monoclonal antibody Mch-3 was used for monitoring mesenchyme formation. It specifically recognizes small particles in mesenchyme cells of Halocynthia larvae (Kim and Nishida,1998). The differentiation of epidermis was monitored by staining with the Epi-2 monoclonal antibody. It recognizes an antigen that is specific for differentiated epidermal cells and the larval tunic, which is secreted by epidermal cells (Nishikata et al.,1987). Samples were fixed in cold methanol for 10 min at adequate developmental stages. Indirect immunofluorescence detection was done according to standard methods using an HRP-conjugated anti-mouse IgG secondary antibody (Simple Stain MAX PO) and a TSA fluorescein system (Perkin-Elmer Life Science). To permanently arrest cleavage, the 110-cell stage embryos were transferred to seawater containing 2.5 μg/ml cytochalasin B (Sigma).
We thank the members of the Asamushi Research Center for Marine Biology and the Otsuchi International Coastal Research Center for help in collecting live ascidian adults, and the members of the Seto Marine Biological Laboratory for help in maintaining them. We also thank Dr. Hidetoshi Saiga for providing Hr-β-catenin MAGEST plasmids.