T-brain homologue (HpTb) is involved in the archenteron induction signals of micromere descendant cells in the sea urchin embryo

In the sea urchin, the fourth cleavage produces a 16-cell stage embryo with eight mesomeres in the animal hemisphere and four macromeres and four micromeres in the vegetal hemisphere. Micromeres are specified autonomously,as isolated micromeres give rise to skeletogenic cells in vitro(Okazaki, 1975), and no other fate is ever observed for micromeres when transplanted to ectopic location in the embryo. In addition, micromeres play an important role in axis formation,as shown by deletion of micromeres or their transplantation of micromeres into the animal pole region (Hörstadius,1973; Ransick and Davidson,1993). Removal of micromeres during the period from fourth and fifth cleavage also impairs expression of an endoderm specific gene(Ransick and Davidson, 1995). Furthermore, signal(s) emanating from micromere descendants at late blastula stages are important for gastrulation itself(Minokawa and Amemiya, 1999;Ishizuka et al., 2001).

Recently, progress has been made in identifying molecular mechanisms that underlie the specification and subsequent differentiation of the micromere-primary mesenchyme cell (PMC) lineage (reviewed byDavidson et al., 1998;Angerer and Angerer, 2000). First, genes encoding proteins involved in the formation of spicule have been identified, including SM50 (Benson et al., 1987) and SM30(George et al., 1991). Thecis-regulatory systems controlling the expression of SM50(Makabe et al., 1995) andSM30 (Akasaka et al.,1994; Frudakis and Wilt,1995; Yamasu and Wilt,1999) were analysed in detail. One of the transcription factors responsible for SM50 and SM30 expression is Ets;HpEts induces the expression of HpSM50 and loss ofHpEts function results in the failure of spicule formation(Kurokawa et al., 1999). Second, nuclear localization of β-catenin is essential for the autonomous specification of micromere (Wikramanayake et al., 1998; Logan et al.,1999; Emily-Fenouil et al.,1998). Third, Delta, which is expressed by micromere descendants,plays an essential role in the Notch-dependent specification of SMCs(Sweet et al., 1999;McClay et al., 2000; Sherwood et al., 2001; Sweet et al.,2002).

It has been demonstrated that transcription factors containing a T-domain,the DNA-binding domain homologous to the mouse brachyury (orT) gene product, play important roles in various aspects of animal development (reviewed by Herrmann and Kispert, 1994; Smith,1997; Papaioannou and Silver,1998). T-domains fall into a number of subfamilies, such asbrachyury, Tbx and T-brain. T-brain-1, which is expressed in the cerebral cortex (leading to the name T-brain) was first isolated from mouse (Bulfone et al., 1995). A related T-box gene, referred to as Eomesodermin, isolated fromXenopus laevis is first expressed in the mesoderm and then expressed in the most anterior part of the brain at the tadpole stage(Ryan et al., 1996;Ryan et al., 1998). Recently,invertebrate homologues of T-Brain-1 have been isolated from a hemichordate acorn worm (Tagawa et al.,2000), a starfish (Shoguchi et al., 2000) and a sea cucumber(Maruyama, 2000).

We report the isolation and characterization of sea urchin homologue ofT-brain-1, referred to as HpTb. We suggest thatHpTb is involved in the production of signals from micromere progeny responsible for gastrulation. We also propose that HpTb is involved in the cascade responsible for the production of signals required for the spicule formation, and for signals from the vegetal hemisphere required for the differentiation of oral-aboral ectoderm.

Gametes of the sea urchin (H. pulcherrimus) were obtained by coelomic injection of 0.55 M KCl, and fertilized eggs were cultured in the artificial sea water Jamarin U (Jamarin Lab) at 16°C.

The amino acid sequences of T domain of the T gene products are highly conserved among mouse (Herrmann et al., 1990), Xenopus(Smith et al., 1991),zebrafish (Schulte-Merker et al.,1992), ascidians (Yasuo and Satoh, 1994) and sea urchins(Harada et al., 1995). The sense-strand oligonucleotide that corresponds to the amino acid sequence YIHPDSP and the antisense oligonucleotide that corresponds to the amino acid sequence NPFAKG(A)L(F) were synthesized using an automated DNA synthesizer(Applied Biosystems). Using these oligonucleotides as primers, we amplified target fragments from an H. pulcherrimus gastrula cDNA library by means of PCR. Probing with candidate cDNA fragments random-labelled with[³²P]-dCTP (Amersham), we screened the library at high stringency(hybridization, 6×SSPE, 0.1% SDS, 1×Denhardt's solution, 50%formamide at 42°C; washing, 2×SSC, 0.1% SDS at 65°C). The isolated clones were subcloned into pBluescriptII SK(+) (Stratagene). The clones were sequenced by dideoxy chain termination(Sanger et al., 1992). In order to obtain a cDNA clone that contains entire open reading frame, we re-screened H. pulcherrimus hatched blastula cDNA library with an RNA probe synthesized from the obtained cDNA. The RNA probe was labelled with digoxigenin (DIG)-11-UTP (Roche) using T3 Megascript kit (Ambion) as described in the instruction manual. An antibody against digoxigenin that had been conjugated to alkaline phosphatase was used to probe the membrane (Roche). The chemiluminescent signal produced by enzymatic dephosphorylation of CSPD(TROPIX) by alkaline phosphatase was detected by X-ray film.

The RNA was extracted from H. pulcherrimus embryos at various developmental stages as described by Chomczynski and Sacchi(Chomczynski and Sacchi, 1987). The total RNA (2 μg) was electrophoresed on each lane of a denaturing formaldehyde-1% agarose gel, transferred to a Nytran membrane (Schleicher and Schuell), and hybridized to the antisense RNA of HpTb labelled with DIG-11-UTP. The signal was detected as described above.

Whole-mount in situ hybridization was performed as described previously(Kurokawa et al., 1999). DIG-labelled antisense RNA probe was prepared with an Ambion's Megascript kit using DIG-11-UTP. Riboprobes were hydrolyzed with alkali to sizes of about 150-400 nucleotides, as described by Cox et al.(Cox et al., 1984).

To generate DNA templates for in vitro RNA synthesis, the plasmids that contain cDNA for HpTb, truncated HpEts-FLAG(Kurokawa et al., 1999) and the intracellular domain of sea urchin LvG-cadherin(Logan et al., 1999) were linearized with restriction enzymes. 5′ capped mRNA was synthesized by using the T7 Megascript kit (Ambion) and Cap Analog[m7G(5′)ppp(5′)G; Ambion] as described in the instruction manual. Microinjection of sea urchin eggs was done as described by Gan et al.(Gan et al., 1990). Morpholino oligonucleotides complementary to sequence containing the translation start site of HpTb AAATTCTTCTCCCATCATGTCTCCT and the control lacZmorpholino were obtained from Gene Tools (Corvallis). Oligonucleotides were dissolved in 40% glycerol at a concentration of 5 pg/pl(3.5×10⁸ molecules/pl). Two picolitres of the solution was injected into each fertilized egg.

The cDNA fragment coding for the N-terminal region of HpEts protein,corresponding to codons 450 bp to 881 bp, and the N-terminal region of HpTb protein, corresponding to codons 291 bp to 707 bp and 807 bp to 1373 bp, were fused downstream to the malE gene in the pMAL-cRI vector (New England Biolabs), which encodes the E. coli maltose-binding protein (MBP). The fusion proteins, HpEts-MBP and HpTb-MBP, were produced in E. coli, affinity purified using an amylose resin, and used for immunization of rabbits to generate anti-HpEts and anti-HpTb, respectively.

Antibodies were purified using affinity column containing specific antigens(Harlow and Lane, 1988). Embryos were fixed and stained with affinity-purified anti-HpEts polyclonal sera or affinity-purified anti-HpTb polyclonal sera as described by Logan et al. (Logan et al., 1999). These primary antibodies were detected with Oregon green-conjugated goat anti-rabbit secondary antibodies (Molecular Probes). Embryos were blocked in TBS-T (5 mM Tris-HCl, pH 7.5, 70 mM NaCl, 1.3 mM KCl, 0.5% Tween20) containing 40 mg/ml goat serum. For the staining with the Hpoe antibody(Yoshikawa, 1997), embryos were fixed as described by Coffman and McClay(Coffman and McClay, 1990) and the primary antibodies were detected with Texas Red-conjugated secondary antibodies (Molecular Probes).

Embryos were dissolved in sample buffer [final concentration: 290 mM Tris-HCl (pH 6.8), 8.3% SDS, 30% glycerol, 0.01% Bromphenol Blue, 4%2-mercaptoethanol], and boiled for 5 minutes. Proteins were analysed on 8%acrylamide gels by SDS-PAGE and transblotted on to a PVDF membrane (Immobilon Transfer Membranes; Millipore). The membrane was reacted with affinity-purified polyclonal anti-HpTb antibodies, followed by horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (1:1000000; KPL),followed by detection with Super Signal West Dura Extended Duration Substrate(PIERCE) as an enzymatic substrate. The chemiluminescent signal was detected by X-ray film.

Total RNA was isolated from 50 control and 50 mRNA-injected embryos or 50 embryos derived from animal cap mesomeres using ISOGEN (Wako). The extracted RNAs were used to synthesize cDNA using RNA PCR kit (AMV) (Takara). An aliquot of the RT reaction was then used for PCR containing 0.2 μM concentrations of appropriate primers. All comparisons were performed in the linear range of amplification. The products were resolved on 2% agarose gels and then transferred to a Nytran membrane (Schleicher and Schuell). To visualize the PCR products, hybridization with appropriate DIG-labelled RNA probes was followed by commercial Fab fragments of antibody to DIG conjugated to alkaline phosphatase (Roche). The chemiluminescent signal produced by enzymatic dephosphorylation of CSPD by alkaline phosphatase was detected by X-ray film.

Micromere and animal cap isolation, and cell transplantations were performed by hand using a glass needles as described by Kurokawa et al.(Kurokawa et al., 1999).

Using oligonucleotide primers corresponding to a conserved sequence in theT-box gene family, we amplified target fragments from H. pulcherrimus gastrula cDNA by the PCR reaction. Sequencing the amplified 300 bp-long fragments revealed that the sea urchin contains a T-box gene different from HpTa, which we previously reported as abrachyury homologue (Harada et al., 1995). Thus, we designated this second T-box gene asHpTb. Screening hatched blastula cDNA library (H. pulcherrimus) with DIG-labelled RNA probe yielded a clone that consisted of 4975 nucleotides. As shown in Fig. 1, the cloned fragment contains a single open reading frame of 2817 nucleotides that encodes a polypeptide of 939 amino acids and a calculated molecular mass of 105 kDa.

We performed a molecular phylogenetic analysis using the 136 confidently aligned sites of the T-domain amino acid residues. The resultant phylogenetic tree indicates that HpTb is a T-box gene belongs to the subfamily of T-brain (Fig. 2A).

Fig. 2B shows a comparison of the amino acid sequences of the T domain of HpTb with proteins encoded by the human hu-Tbr-1 (Bulfone et al., 1995), mouse m-T-brain-1(Bulfone et al., 1995),zebrafish zf-tbr 1 (Yonei-Tamura et al., 1999), X-Eomesodermin(Ryan et al., 1996) and starfish Ap-Tbr (Shoguchi et al.,2000). Although the overall degree of amino acid identity is not very high, in the T domain shown in Fig. 2 the extent of amino acid identity was 61% (sea urchin/mouse),60% (sea urchin/frog), 60% (sea urchin/zebrafish) and 72% (sea urchin/starfish). The relatively high degree of identity in the T domain between the sea urchin protein and the mouse T-brain-1, X-Eomesodermin,ZF-tbr 1 and Ap-Tbr proteins demonstrates that this cDNA clone corresponds to a sea urchin homologue of the chordate T-braingene.

Northern blotting analysis revealed that the HpTb transcripts are transiently present during embryogenesis of H. pulcherrimus. The probe produced from the entire cDNA for HpTb hybridized to a 6 kb RNA. Although the length of the cloned HpTb cDNA isolated from gastrula is ∼5 kb, 6 kb RNA seems to be actual size mRNA of HpTb. The distinct hybridization signal for HpTb was first detected in blastulae and the level of the signal was almost constant until gastrula stage. Thereafter, the signal diminished rapidly(Fig. 3A).

The Northern blotting also detected a very weak band of about 4.5 kb in the egg and cleavage stages (Fig. 3A). We isolated the cDNA clones from cleavage stage embryos using the entire HpTb cDNA as a probe. Sequencing of these clones revealed that the mRNAs of the cleavage stage embryos are shorter, being truncated at the C-terminal region (Fig. 1).

We studied which territories or cell types express HpTb by means of in situ hybridization of whole-mount specimens. At the hatched blastula stage, the distinct signal of HpTb was detected in the presumptive PMCs (Fig. 3B), which form a ring around vegetal pole (Fig. 3C). At the mesenchyme blastula stage, these HpTbpositive cells migrate into the blastocoel to give rise to PMCs(Fig. 3D). The primary skeletogenic mesenchyme cells are derived from (large) micromeres. No hybridization signals were detected in embryonic cells other than PMCs. After the gastrula stage, the HpTb whole-mount hybridization signal is no longer detectable (data not shown).

Western blot analysis, using affinity-purified anti-HpTb antibodies,revealed that HpTb protein is detected as a single band with an estimated molecular weight of 105 kDa. As the shorter, processed HpTb-mRNAs detected in the egg and cleavage stage embryos lack part of the coding region, we would expect the molecular mass of the proteins translated from cleavage stage HpTb-mRNAs to be smaller than the HpTb expressed after the blastula stage. However, no such smaller band was detected, suggesting that the processed HpTb-mRNA is not translated. The full-length HpTb protein is present in the egg and cleavage stage, decreases in abundance before hatching and then increases at the hatched blastula stage. Thereafter the level of protein remains almost constant until the pluteus stage(Fig. 4A). Taken together with the results of northern blots (Fig. 3A), the results indicate that the HpTb protein is accumulated maternally, destroyed before hatching, and then produced zygotically after the blastula stage.

Immunostaining of the embryos with anti-HpTb antibodies revealed that HpTb is present in the cytoplasm, but is absent from nuclei of all blastomeres in the cleavage stage. This suggests that the maternally stored HpTb does not function as a transcription factor (Fig. 4B). After the hatching blastula stage, the HpTb disappears from blastomeres except for PMCs, and HpTb is accumulated in the nuclei of(presumptive) PMCs (Fig. 4C).

In order to gain the insight into the role of HpTb during development, we designed experiments to perturb the embryo by inhibiting the translation of HpTb by injecting fertilized eggs with HpTb morpholino antisense oligonucleotides. When the control lacZ morpholino(7×10⁸ molecules/egg) or low amounts of HpTbmorpholino (1×10⁸ molecules/egg) were injected, most of the injected embryos developed almost normally to the pluteus stage(Fig. 5A,B). The embryos injected with 7×10⁸ molecules of HpTb morpholino seemed to be morphologically normal until hatched blastula stage (data not shown). When control embryos, which were injected with lacZmorpholino, had reached the early gastrula stage(Fig. 5C), embryos injected with HpTb morpholino showed suppressed and delayed gastrulation;however, the ingression of PMCs into blastocoel occurred normally in such embryos (Fig. 5D). When the control embryos reached the prism stage(Fig. 5E), embryos injected with HpTb morpholino showed retarded gastrulation. Furthermore, the differentiation of oral-aboral ectoderm seemed to be repressed, and formation of spicules was also suppressed (Fig. 5F). As judged by cell morphology and their localization in the animal hemisphere, secondary mesenchyme cells (SMCs) were formed in 48 hour prism embryos injected with HpTb morpholino. At 72 hours after fertilization, when the control embryos had reached the pluteus stage, only a limited archenteron and a reduced number of pigment cells (approx. one-third compared with the control embryos) were observed in the embryo injected withHpTb-morpholino (Fig. 5H). Most of the embryos injected with 7×10⁸molecules (10 pg) of HpTb morpholino (n=256/267) showed this phenotype. We confirmed that the HpTb morpholino antisense oligonucleotides inhibited the translation of HpTb by immunostaining with anti-HpTb antibodies (Fig. 5I,J). These results suggest that the expression of HpTbin (developing) PMCs is required for the gastrulation, spicule formation and the normal development of the oral-aboral axis in sea urchin development. The formation of an archenteron was rescued in more than half of embryos injected with HpTb morpholino by co-injection of HpTb mRNA(n=72/128; Fig. 5K). However the inhibition of skeletogenesis was barely rescued by co-injection ofHpTb mRNA. We cannot explain the reason for the inefficient rescue of skeletogenesis at this point.

We performed quantitative RT-PCR to determine the expression level of various cell type-specific genes in the embryos injected with HpTbmorpholino. The RNA was isolated from embryos at 28 hours after fertilization,when the control embryos had reached late gastrula stage. As a control, the level of ubiquitin mRNA, which is almost spatially ubiquitous(Nemer et al., 1991), was unaffected by injection of the HpTb morpholino.

We previously showed that HpEts, an ets-related transcription factor, is expressed exclusively in micromere descendants after blastula stage, and that HpEts is involved in the differentiation of PMCs(Kurokawa et al., 1999). Recently, Sweet et al. (Sweet et al.,2002) reported that a sea urchin Delta homolog(LvDelta) is also expressed in micromere descendants at blastula stage, and that LvDelta is responsible for the SMC-inducing activity of micromere descendants. In order to determine the functional relationship ofHpTb to HpEts and HpDelta, we examined the expression of HpEts and HpDelta in embryos injected withHpTb morpholino. As shown in Fig. 6, HpEts and HpDelta were unaffected in the injected embryos. This was supported by the observation that PMCs, which ingressed into the blastocoel of embryos injected with HpTbmorpholino, were immunologically positive using the anti-HpEts antibodies(Fig. 7B). Furthermore, some pigment cells, which are derived from SMCs, formed in morpholino injected embryos. In addition, the expression of PMC-specific HpSM50 was not affected by the injection with HpTb morpholino(Fig. 6). These results suggest that HpTb is not involved in the specification of PMCs. By contrast,another PMC-specific gene HpSM30(Kitajima et al., 1996) was suppressed in the embryos injected with HpTb morpholino(Fig. 6). It has been reported that expression of SM30 requires signal(s) from non-PMCs(Urry et al., 2000). It is possible that HpTb is indirectly involved in the production of signal(s)responsible for the expression of HpSM30.

Expression of the aboral ectoderm-specific gene HpArs(Akasaka et al., 1990) was also suppressed in the embryos injected with HpTb morpholino, whereas Hpoe, which is an oral ectoderm-specific epitope(Yoshikawa, 1997), was expressed over almost all the surface of epithelial cells of the injected embryos (Fig. 7D). Wikramanayake et al. have shown that the activation of aboral ectoderm-specific Spec1 in L. pictus requires signals from vegetal hemisphere (Wikramanayake et al.,1995). Recent studies also suggest that vegetal signals are involved in the establishment of oral-aboral axis(Wikramanayake and Klein,1997; Li et al.,1999; Angerer and Angerer,2000). In order to confirm that the activation of aboral ectoderm-specific Ars also requires interaction with vegetal blastomeres in H. pulcherrimus embryos, we performed quantitative RT-PCR to determine the expression level of Ars in the embryos derived from animal cap mesomeres. The RNA was isolated from control embryos and from embryos derived from animal cap mesomeres at 28 hours after fertilization when the control embryo had reached the late gastrula stage. The expression level of the Ars was significantly lower in the embryo derived solely from animal cap mesomeres, suggesting that signals from vegetal hemisphere are required for the activation of Ars(Fig. 8). These results raise a possibility that HpTb is involved in the production of vegetal signal(s)involved in aboral ectoderm differentiation.

The levels of HpTa (Harada et al., 1995) and HpEndo16(Akasaka et al., 1997), both of which are expressed in the vegetal plate at blastula stage, were not affected by the injection of HpTb morpholino at the developmental stage we examined, suggesting that HpTb is not involved in the initial specification of endoderm. It is important to note that the level of expression ofHpTb was enhanced by the repression of translation of HpTb(Fig. 6).

Micromere-progeny induction signals have been shown to play an important role in the specification of SMCs (McClay et al., 2000) and initiation of gastrulation(Minokawa and Amemiya, 1999;Ishizuka et al., 2001). As repression of HpTb in (presumptive) PMCs leads a significant delay of gastrulation, we predicted that HpTb might be involved in regulating(presumptive) PMC signaling. To test this hypothesis, we prevented the translation of HpTb in micromere descendant cells by injecting HpTbmorpholino antisense oligonucleotides. We combined animal cap mesomeres from a normal embryo with a micromere quartet isolated from a 16-cell stage embryo that had developed from a zygote injected with 7×10⁸molecules of HpTb morpholino (Fig. 9A,B). In order to follow HpTb deficient micromeres, donor embryos were labelled by co-injecting the morpholino with 10 pg of rhodamine-dextran. In all experiments, over 100 injected embryos were also cultured in parallel for 48 hours to confirm that the archenteron formation was suppressed in the donor embryos.

When HpTb morpholino-injected donor micromeres were combined with the animal cap mesomeres of uninjected embryo, almost all the transplanted micromere descendant cells ingressed into the blastocoel, but the injected donor micromeres only induced a very limited archenteron, and the micromeres did not form spicules (n=5/5; Fig. 9B). SMCs were either not formed at all, or very small number of SMCs were formed, in the chimaeric embryos. By contrast, when uninjected micromeres were combined with animal cap mesomeres of uninjected embryo, the micromeres ingressed and formed spicules (n=15/15). In addition, the control donor micromeres were able to induce an archenteron, oral-aboral ectoderm and SMCs (n=15/15; Fig. 9C). These data suggest that one of the functions of HpTb is to provide the micromere descendant cells with the ability to produce a signal that induces neighbouring cells to develop archenteron and SMCs. When the animal cap mesomeres derived from zygotes injected with the morpholino antisense oligonucleotides were combined with micromere quartet derived from normal embryos, the mesomeres developed archenteron, SMCs and both oral and aboral ectoderms (n=14/14; 9D,E). The micromere progeny developed spicules in the chimaeric embryos. These data support the hypothesis thatHpTb morpholino antisense oligonucleotides do not affect the responsiveness of mesomeres to the signals emanating from micromere progeny.

Recent studies provide convincing evidence that β-catenin, a molecule of Wnt signaling cascade, plays an essential role in specification of micromere-derived PMCs (Wikramanayake et al., 1998; Logan et al.,1999). Because HpTb is expressed exclusively in the(presumptive) PMCs, it is probable that the HpTb expression is regulated by nuclear localization of β-catenin. In order to examine this issue, we performed overexpression of intracellular domains of cadherin(ΔLvG-cadherin) to deplete β-catenin from the nuclei of blastomeres of early embryos. We injected 2 pg of ΔLvG-cadherin mRNA; this has been shown to abolish vegetal development of Lytechinus embryo(Logan et al., 1999). As shown in Fig. 10, this injection ofΔ LvG-cadherin mRNA resulted in the suppression ofHpTb; the HpTb band is barely detectable in experimental embryos. This is consistent with the idea that HpTb is in a downstream component of a Wnt signalling cascade that functions in the micromere primary mesenchyme lineage of the sea urchin embryo. We examined the expression of HpEts, which is also expressed exclusively in the(presumptive) PMCs after blastula stage, in the embryo injected withΔLvG-cadherin mRNA. The expression of HpEts was also repressed in these injected embryos (Fig. 10). Thus, we might assume that HpTb and HpEtsare in the same cascade of gene regulation. We examined the functional relationship of HpTb to HpEts. PMC-specific expression ofHpEts was not affected by the injection with HpTb morpholino(Fig. 6,Fig. 7B). Conversely, the overexpression of the dominant negative, ΔHpEts, suppressed the expression of HpTb (Fig. 11A). These results suggest that both HpEts andHpTb are downstream components of the Wnt signalling cascade, and that HpTb is regulated by HpEts. The expression of dominant negative ΔHpEts also resulted in a significant delay and repression of gastrulation, as well as the suppression of differentiation of PMCs(Fig. 11C).

As shown in the present study, the T-brain gene is conserved in the sea urchin embryo and its expression is restricted to PMCs. The mouse T-Brain-1 gene (Tbr1 — Mouse Genome Informatics) was initially isolated from a subtracted cDNA library enriched for genes transcribed at higher levels in the stage E14.5 mouse telencephalon than in the adult telencephalon (Bulfone et al.,1995). Examination by in situ hybridization of mouse T-Brain-1 gene expression demonstrated that the transcript is first detected around E10.5 in the mantle zone of the telencephalon, and defines molecularly distinct domains within the cerebral cortex(Bulfone et al., 1995). Interestingly, molecular phylogenetic analyses based on comparison of the T-domain amino acid sequences suggest that Eomesodermin should be included in the T-Brain subfamily (Fig. 2) (Papaioannou and Silver,1998). Eomesodermin was originally isolated fromXenopus laevis embryos as a novel T-box gene(Ryan et al., 1996). The gene is expressed in mesodermal cells in a ventral-to-dorsal gradient of increasing concentration during gastrulation, and it is critical for mesoderm differentiation (Ryan et al.,1996), and later it was revealed to be expressed in the forebrain of tadpoles (Ryan et al.,1998). Recently, a mouse homologue (Eomes; also known asTbr2) of eomesodermin was shown to be expressed in the early primitive streak, nascent mesoderm and the anterior visceral endoderm. This early expression disappears at later stages, and a second domain ofEomes expression is observed in the telencephalon around E10.5(Ciruna and Rossant, 1999).

Homologs of the Tbr1 gene have been identified from hemichordates(Tagawa et al., 2000),starfish (Shoguchi et al.,2000) and sea cucumbers(Maruyama, 2000). In these deuterostome embryos, T-brain is first expressed in the region which eventually forms endoderm and mesoderm. The hemichordate T-brain is expressed later in the apical organ or light sensory organ(Tagawa et al., 2000). Therefore, it is likely that the T-brain gene has two distinct expression domains: in the mesoendoderm of early embryos and in the nervous system of later embryos. The sea urchin HpTb expression in the micromere may correspond to the first expression domain of this family, and analogous to Eomesodermin, HpTb is likely to be critical for endoderm and mesoderm differentiation.

Inhibition of translation of specific gene products with morpholine-substituted antisense oligonucleotides is a useful way to gain insight into the function of the gene(Howard et al., 2001).HpTb is expressed specifically in (presumptive) PMCs during blastula and mesenchyme blastula stage. The expression pattern of HpTb, and suppression of archenteron formation caused by inhibition of the HpTb translation in (presumptive) PMCs are both consistent with what is known of the importance of presumptive PMCs in development(Minokawa and Amemiya, 1999;Ishizuka et al., 2001). Not only do micromeres provide cells for the skeleton, but it is known micromere descendants are an important source of developmental signals that affect other tissues. Removal of micromeres during the period from forth and fifth cleavage impairs expression of the endoderm specific Endo16 and results in significant delay of archenteron formation(Ransick and Davidson, 1995). Recently, it has been shown that signal(s) emanating from micromere-descendants at late blastula stages are important for gastrulation itself (Minokawa and Amemiya,1999; Ishizuka et al.,2001). Interference with (presumptive) PMC function by inhibiting translation of HpTb with HpTb morpholino oligonucleotides led to significant delay of gastrulation, as we report here. This is consistent with the known important role of micromere-descendants in these processes.

The repression of HpTb did not cause the inhibition ofHpEndo16 expression. This is also consistent with the previous report that the expression of Endo16 is induced by micromere-descendant cells during forth to sixth cleavage stages(Ransick and Davidson, 1995). It seems likely that at least three distinct signals are provided by micromere descendant cells. The first signal(s) is produced during forth to sixth cleavage stages, as Ransick and Davidson reported(Ransick and Davidson, 1995);the second signal is Delta which is responsible for the SMC specification(McClay et al., 2000;Sweet et al., 2002) and the third signal(s) which is produced at blastula stage when the HpTb is expressed (Minokawa and Amemiya,1999; Ishizuka et al.,2001). The present data favour the idea that HpTb is involved in gastrulation itself, but not in the initial specification of the vegetal plate.

The development of embryos from aggregates between mesomeres and micromeres(Hörstadius, 1973;Amemiya, 1996) demonstrates the organizing activity of micromeres. Normally, the mesomeres isolated from the 16-cell stage embryo form thin an epithelial ball(Hörstadius, 1973;Henry et al., 1989). Chimeras composed of animal cap mesomeres and micromere quartet from normal embryo developed almost normal embryos containing an archenteron, PMCs and SMCs. Conversely, HpTb morpholino-injected donor micromere quartet combined with the animal cap mesomeres of uninjected embryo induced only partially invaginated archenteron, and no or few SMCs (although the donor micromere descendant cells ingressed into blastocoel)(Fig. 9A,B). These results strongly support the idea that HpTb is involved in the regulation of signal(s) from (presumptive) PMCs. Although the chimeras composed of animal cap mesomeres and micromere quartet derived from zygotes injected withHpTb morpholino formed no or few SMCs, the embryos injected withHpTb morpholino formed SMCs (Fig. 5F,H) and expressed HpDelta(Fig. 6), suggesting thatHpTb is not substantially involved in the specification of SMCs in normal development. Injection of HpTb morpholino resulted in the decreased number (approx. one third) of pigment cells. This raises the possibility that HpTb is involved in the differentiation of pigment cells indirectly.

Croce et al. have reported that the T-brain homolog referred to asske-T is expressed in P. lividus embryos(Croce et al., 2001). They showed that transcripts hybridized with probes that detect ske-Texist ubiquitously in egg and the early cleavage stage embryos. They also showed that the transcripts appear after blastula stage, as we have shown in the present study. As Croce et al. pointed out, the early transcripts are smaller than the ske-T cDNA fragment they isolated(Croce et al., 2001). In our present work, we barely detected the small transcript in the eggs and cleavage stage embryos of H. pulcherrimus. We have also shown that the early transcripts are not translated (Fig. 4A) and that the maternally stored HpTb protein is not present in the nucleus, suggesting that the HpTb does not function as a transcription factor during clevage stage (Fig. 4B). Furthermore, injection of HpTb morpholino antisense oligo into embryos of H. pulcherrimus did not affect early development.

When the animal cap mesomeres derived from zygotes injected with the morpholino antisense oligo were combined with a micromere quartet isolated from normal embryos, the mesomeres developed an archenteron, oral and aboral ectoderm and SMCs. Hence, even if low level of processed T-braintranscripts exist in the embryonic cells other than progeny of micromeres,they are not involved in the archenteron inducing activity, they are probably not responsible for the differentiation of oral-aboral ectoderm in H. pulcherrimus. It is also possible, of course, that differences exist between H. pulcherrimus and P. lividus.

Wikramanayake et al. (Wikramanayake et al., 1995) have shown that the aboral ectoderm specific genes were not expressed in animal hemisphere explants from L. pictus,suggesting that the ectoderm differentiation in L. pictus embryos requires interaction with vegetal blastomeres. Using H. pulcherrimusembryos, we also demonstrated that the embryos derived from the animal cap mesomeres did not express aboral ectoderm specific Ars(Fig. 8). The repression of translation of HpTb in (presumptive) PMCs resulted in the disturbances of patterning of oral and aboral ectoderm (overexpression of Hpoe epitope,reduction of Ars expression). It is possible that HpTb is involved in the production of the vegetal signals responsible (at least in part) for patterning oral and aboral ectoderm, although we do not know if theHpTb functions in this process directly or not.

HpTb morpholino oligonucleotides did not suppress the expression of HpSM50, HpDelta and HpEts, suggesting that HpTbis not involved in the specification of PMCs. In the previous paper, we have demonstrated that ectopic expression of HpEts altered the fate of many cells, transforming them into migrating PMCs. However, the transformed PMCs did not form spicules without addition of serum(Kurokawa et al., 1999). The expression of HpEts is necessary for the spicule formation, but it is not sufficient for the spicule formation of PMCs. It has been reported that another PMC-specific protein, SM30, requires signals produced by non-PMC cells (Urry et al.,2000). The repression of HpSM30 after injection withHpTb morpholino (Fig. 6) supports the idea that the differentiation of vegetal regions,which are induced by a signal(s) emanated from (presumptive) PMCs, is required for the production of signal(s) responsible for the spicule formation.

We have shown that both HpEts and HpTb are downstream components of the Wnt signalling cascade. The injection of HpTbmorpholino antisense oligonucleotides did not affect the expression ofHpEts. Conversely, the overexpression of dominant negativeΔHpEts repressed the expression of HpTb, the gastrulation and the differentiation of oral-aboral ectoderm, suggesting that HpEtsregulates HpTb. The injection of the HpTb morpholino also did not affect the expression of HpDelta, suggesting thatHpDelta is not a downstream component of HpTb.

The repression of translation of HpTb caused enhancement of HpTbexpression. There is experimental evidence that midcleavage stage blastomeres have an extensive capacity to change their states of specification in response to cell interactions (reviewed byDavidson, 1989). It has been thought that micromere descendant cells repress the capacity of neighbouring macromeres to change their cell fate into micromere progeny. It is possible that the expression of HpTb in (presumptive) PMCs downregulates the expression of HpTb in neighbouring cells.

The authors express their thanks to Dr Fred Wilt for his advice in preparation and critical reading of the manuscript, to Dr D. McClay (Duke University) for the kind gift of cDNA for ΔLvG-cadherin, to Dr C. Ettensohn (Carnegie Mellon University) for the kind gift of cDNA for LvDelta,and to Dr S. Yoshikawa for the kind gift of monoclonal antibody to Hpoe. The authors also thank Dr H. Katow (Asamushi Marine Biological Station) for supplying live sea urchins. Part of this work was carried out in the Center for Gene Research of Hiroshima University. We thank the Cryogenic Center,Hiroshima University for supplying liquid nitrogen. This work was supported in part by Grants-in-Aid for Scientific Research (C2) (number 11680724) and for Scientific Research on Priority Areas (number 11152227) to K. A. and (B1)(number 12145101) to Y. T.; and by (A1) (number 13044001) to S. N. from the Ministry of Education, Science, Sports and Culture, Japan, and the Program for Promotion of Basic Research Activities for Innovative Biosciences.