In Xenopus, Wnt signals and their transcriptional effectorβ-catenin are required for the development of dorsal axial structures. In zebrafish, previous loss-of-function studies have not identified an essential role for β-catenin in dorsal axis formation, but the maternal-effect mutation ichabod disrupts β-catenin accumulation in dorsal nuclei and leads to a reduction of dorsoanterior derivatives. We have identified and characterized a second zebrafish β-catenin gene,β-catenin-2, located on a different linkage group from the previously studied β-catenin-1, but situated close to the ichabod mutation on LG19. Although the ichabod mutation does not functionally alter the β-catenin-2 reading frame, the level of maternal β-catenin-2, but not β-catenin-1,transcript is substantially lower in ichabod, compared with wild-type, embryos. Reduction of β-catenin-2 function in wild-type embryos by injection of morpholino antisense oligonucleotides (MOs)specific for this gene (MO2) results in the same ventralized phenotypes as seen in ichabod embryos, and administration of MO2 to ichabod embryos increases the extent of ventralization. MOs directed against β-catenin-1 (MO1), by contrast, had no ventralizing effect on wild-type embryos. β-catenin-2 is thus specifically required for organizer formation and this function is apparently required maternally, because the ichabod mutation causes a reduction in maternal transcription of the gene and a reduced level of β-catenin-2 protein in the early embryo. A redundant role of β-catenins in suppressing formation of neurectoderm is revealed when both β-catenin genes are inhibited. Using a combination of MO1 and MO2 in wild-type embryos,or by injecting solely MO1 in ichabod embryos, we obtain expression of a wide spectrum of neural markers in apparently appropriate anteroposterior pattern. We propose that the early, dorsal-promoting function ofβ-catenin-2 is essential to counteract a later, dorsal- and neurectoderm-repressing function that is shared by both β-catenin genes.

Wnt signaling pathways are involved in both formation of the major dorsal signaling centers in vertebrate embryos (i.e. Nieuwkoop center and Spemann organizer and cognate structures) (Harland and Gerhart, 1997; Sokol,1999; Tao et al.,2005; Schier and Talbot,2005) and also, somewhat later in development, in the posteriorizing and ventralizing signals that derive from more lateral and ventral embryonic regions (Erter et al.,2001; Lekven et al.,2001). Indeed, a GFP reporter assay for β-catenin-responsive gene activation in the zebrafish reveals Wnt signaling activity in a localized, presumably dorsal, domain at dome stage, and in both dorsal and ventrolateral discrete domains at shield stage(Dorsky et al., 2002). These two phases of Wnt signaling have been considered (e.g. Kim et al., 2002; Momoi et al., 2003) as components of sequential activator and transformer signals that fit Nieuwkoop's two signal model of neural induction and patterning(Nieuwkoop et al., 1952; Nieuwkoop and Nitgevecht,1954).

Wnt signaling mediated by the transcriptional effector β-catenin plays an essential role in axis formation and neural induction. A functionalβ-catenin gene is essential for formation of the Nieuwkoop center and Spemann organizer in amphibians (Heasman et al., 1994; Wylie et al.,1996; Heasman, 2000) and for gastrulation and normal axis formation in the mouse (Haegel et al., 1995; Huelsken et al., 2000). Accumulation of nuclear β-catenin is first observed on the dorsal side of pregastrula Xenopus (Schneider et al., 1996; Larabell et al.,1997; Rowning et al.,1997) and zebrafish (Schneider et al., 1996; Kelly et al.,2000; Dougan et al.,2003) embryos, consistent with the key role of β-catenin in initiating the activation of dorsalizing genes.

In zebrafish embryos bred from females homozygous for ichabod, a spontaneous maternal effect mutation resulting in severe ventralization,β-catenin fails to localize to dorsal yolk syncytial layer (YSL) and blastomere nuclei (Kelly et al.,2000). RNA injection experiments showed that Wnt pathway components upstream of β-catenin failed to rescue these embryos, but RNAs for β-catenin and for the downstream nodal-related factor Squint(Znr2/Ndr1) and homeodomain factor Bozozok (Dharma/Nieuwkoid) can rescue the embryos to wild-type phenotype (Kelly et al., 2000). These results indicated that failure to regulateβ-catenin properly in the zebrafish embryo results in loss of dorsal axial structures.

The Wnt pathway also promotes posterior and ventral fates and inhibits formation of anterior neural tissue. wnt8, expressed in the ventrolateral germ ring (Kelly et al.,1995b), is required for formation of ventrolateral and posterior mesoderm and spinal cord and posterior brain(Lekven et al., 2001; Erter et al., 2001; Momoi et al., 2003; Ramel and Lekven, 2004). Similarly, Xenopus wnt8 is expressed in the ventrolateral marginal zone (and in a wider vegetal region)(Christian et al., 1991; Smith and Harland, 1991; Christian and Moon, 1993) and may have a similar posteriorizing role(Christian and Moon, 1993; Hoppler et al., 1996; Fredieu et al., 1997). Ventralizing and posteriorizing Wnt signals also appear to be dependent on β-catenin (Dorsky et al.,2002; Weidinger et al.,2005) and thus, may be impaired in embryos with β-catenin deficiencies.

We report here that the zebrafish genome contains a second β-catenin gene, β-catenin-2 (ctnnb2 - Zebrafish Information Network). Considering that both the formation of the organizer and the somewhat later posteriorizing and ventralizing effects of Wnt signaling are both mediated by β-catenin, it was of interest to explore the degree of redundancy of function of the two β-catenins in this organism. Moreover,we report that β-catenin-2 maps close to the ichabodmutation and that maternal expression of β-catenin-2 is impaired in ichabod mutant embryos, and we provide evidence thatβ-catenin-2 is the sole β-catenin gene required for formation of the dorsal organizer in the zebrafish. When onlyβ-catenin-1 (ctnnb1 - Zebrafish Information Network)function was inhibited, we observed no early patterning defects, indicating that this factor was not solely required for Wnt signaling activities that control the extent and nature of posterior and ventral tissue formation. Inhibiting the expression of both β-catenin genes, however, resulted in an unexpected phenotype in which neurectoderm of both anterior and posterior identity develop with at least some appropriate anteroposterior patterning. Thus, the two β-catenin genes normally act with functional redundancy to restrict formation of neurectoderm. In the absence of bothβ-catenin-1 and β-catenin-2 function, the embryo fails to form a recognizable organizer, but as a consequence of the loss of repressive effects on neurectoderm formation, the embryo is still capable of expressing both posterior and anterior neural makers.

Fish husbandry

Embryos of the brass strain of zebrafish (originally obtained from EkkWill Waterlife Resources, Gibbonston, FL), selected as `wild-type' embryos because their delayed onset of pigmentation is advantageous for hybridization,were raised under standard conditions at 28.5°C(Westerfield, 1993). ichabod p1 embryos were obtained by breeding homozygous ichabod females with brass males. Only ichabodfemales that reproducibly gave severely ventralized phenotypes were used in this work. Throughout this report, we will use the expression`ichabod embryos' to mean those embryos derived from homozygous ichabod mothers.

DNA constructs and phylogenetic analysis

Sequence searches of the zebrafish EST database identified ESTs that defined fragments of a second β-catenin gene in zebrafish, which we termβ-catenin-2. Sequences from these ESTs were used to identify a PCR-based polymorphism (primers: 5′-CCTACCTGGATTCAGGGATTC-3′ and 5′-ATGAGCAGAGTCGAACTGGGT-3′) for genetic mapping and to obtain a full-length cDNA clone by screening an embryonic cDNA library. The sequence of the β-catenin-2 cDNA has been deposited in GenBank (Accession Number AF329680). β-catenin-2 was mapped to the telomeric region of LG 19 by scoring the polymorphism in the Heat Shock mapping panel(Woods et al., 2000; Woods et al., 2005).

The following clones were used to prepare antisense probes for hybridization and/or sense RNAs for embryo injection:β-catenin-1 (Kelly et al.,1995a), β-catenin-2, krox20(Oxtoby and Jowett, 1993), emx1 (Morita et al.,1995), hoxb6b (previously named hoxa7)(Prince et al., 1998), islet-1 (Inoue et al.,1994), myoD (Weinberg et al., 1996), bozozok/nieuwkoid/dharma(Koos and Ho, 1998), squint/znr2/ndr1 (Erter et al.,1998; Rebagliati et al.,1998), chordin(Miller-Bertoglio et al.,1997) and goosecoid(Stachel et al., 1993).β-catenin-2* RNA contained nucleotides -100 to +15 ofβ-catenin-1 (Accession Number NM_131059) substituted for -167 to+3 of β-catenin-2 (Accession Number AF329680) andβ-catenin-1* RNA contained nucleotides -120 to +3 ofβ-catenin-2 substituted for -195 to +3 ofβ-catenin-1, where +1 is the A of the initiating ATG codon. The coding sequence of β-catenin-1* was identical toβ-catenin-1, but the coding sequence ofβ-catenin-2* contained an additional four amino acids after the initiating methionine.

For phylogenetic analysis, translated cDNA sequences were aligned using MUSCLE (Edgar, 2004) and based on this alignment, a maximum parsimony tree of nucleic acid coding sequence was performed using the protpars program from the PHYLIP package(Felsenstein, 1989). For this analysis, 1000 bootstrap replicates were performed. Accession Numbers for theβ-catenin genes of rat, human, chicken, Pelodiscus sinensis (a turtle), Xenopus laevis, goldfish and Ciona intestinalisare, respectively, NM_053357, NM_001904, NM_205081, AB124575, BC082826,AY336093 and AB031543. Sequences of the two β-catenin genes of Tetraodon and Takifugu were obtained from the data bases of the Tetraodon(http://www.genoscope.cns.fr/externe/English/Projets/Projet_C/C.html)and Takifugu(http://fugu.hgmp.mrc.ac.uk/)genome projects.

RNA probe synthesis, hybridization, mRNA synthesis and injection

Antisense RNA probes were synthesized and hybridization procedures were performed as previously described (Kelly et al., 2000) except that BM purple alkaline phosphatase substrate(Roche) was used as chromogen. Capped mRNAs were synthesized using the appropriate mMessage mMachine Kit (Ambion), following the manufacturer's protocol. RNAs were stored in distilled sterile H2O at -80°C,and injection solutions were prepared by adjusting the RNA concentration to twice that desired and then adding an equal volume of Dulbecco's modified phosphate-buffered saline containing 0.5% Phenol Red (Sigma). Approximately 1 nl of RNA solution was injected through the intact chorion into the yolk at the base of the blastomeres of one- to four-cell stage embryos, using an agar mold to hold the embryos.

Morpholino antisense oligonucleotide injections

The following custom-designed morpholino antisense oligonucleotides were obtained from Gene Tools (Philomath, OR): β-catenin-1 MO1,CTGGGTAGCCATGATTTTCTCACAG; β-catenin-1 MO1mis,CTcGGaAGCCATcATTTTCaCAgAG; β-catenin-1 MO1-2,CTGTGTCAAAAGCTGTATATTCCTG; β-catenin-1 MO1-3,CAGCACGTAAAACGCGAATAATGGC; β-catenin-2 MO2,CCTTTAGCCTGAGCGACTTCCAAAC; β-catenin-2 MO2mis,CgTTTaGCCTcAGCGACTTgCAtAC; and β-catenin-2 MO2-2,GTGTCTTGTAAGCAGTGAATCCACC. The morpholino oligonucleotides were diluted and injected as described above for the RNAs.

RT-PCR

Total RNA was isolated from 20 whole embryos using TRIzol reagent(Invitrogen). Oligo-dT-primed cDNA was synthesized and RT-PCR reactions were performed using SuperScript First-Strand Synthesis System for RT-PCR(Invitrogen) from 14 μg of total RNA, following the manufacturer's protocol. Separate reactions were set up with primer pairs derived fromβ-catenin-1, β-catenin-2 and ef1αusing 25 cycles at an annealing temperature of 55°C (in the semi-quantitative range). The following primers were used for the amplifications: β-catenin-1,5′-CGCACACATTCACTCTCAGC-3′ and 5′-TGGGTAGCCATGATTTTCTCA-3′; β-catenin-2,5′-ACGCTCAGGATCTGATGGAC-3′ and 5′-AGGCACTTTCTGAACCTCCA-3′; ef1α,5′-ACCGGCCATCTGATCTACAA-3′ and 5′-CAATGGTGATACCACGCTCA-3′.

Identification of a second zebrafish β-cateningene

Sequence searches identified zebrafish ESTs (e.g. AI657717) that represented fragments of a second β-catenin gene in the zebrafish genome,which we term β-catenin-2. A full-length cDNA clone (GenBank Accession Number AF329680) was isolated by screening a zebrafish embryo cDNA library with probes derived from these EST sequences. The β-catenin-2 open reading frame (ORF) is predicted to encode a protein of 778 amino acids. The nucleotide sequence of the protein coding region is 82% identical to the previously characterized zebrafish β-catenin gene(Kelly et al., 1995a),hereafter named β-catenin-1. No significant homology was found in the 5′ UTR and 3′ UTR when comparing theβ-catenin-1 and β-catenin-2 cDNAs. Theβ-catenin 1 and 2 proteins are 92.9% identical and 96.4% similar. Most of the protein is highly conserved, with 96.1% identity from amino acids 1-686. Only the C-terminal 92 amino acids (#687-778), comprising part of the C-terminal activation domain, showed marked divergence, and even here the sequence identity was 70.2% (Fig. 1A). Although these proteins are highly conserved, zebrafishβ-catenin-2 is more diverged from zebrafish β-catenin-1 than the latter is from any of the tetrapod β-catenins. Maximum parsimony phylogenetic analysis of nucleotide sequences shows that zebrafishβ-catenin-1 is part of a highly conserved branch that includes human,mouse, Xenopus, goldfish and pufferfish β-catenins, with zebrafish β-catenin-2 and a second β-catenin gene (not clearly orthologous) of pufferfish (both Fugu and Tetraodon)branching off somewhat earlier (Fig. 1B). All of these vertebrate β-catenins are more related to one another than to ascidian, D. melanogaster or C. elegansβ-catenins, or to vertebrate plakoglobins (data not shown).

We mapped β-catenin-2 near the telomere of LG19 (this work)(Woods et al., 2005), whereasβ-catenin-1 had already been mapped to LG16(Postlethwait et al., 1998). Although the phenotype of the ichabod mutation was consistent with a loss of function of a β-catenin gene, we had ruled out the involvement ofβ-catenin-1, the only known zebrafish β-catenin gene at the time of the initial characterization of this mutant, because the mutation mapped to LG19 (Kelly et al.,2000). The finding of a second β-catenin gene on the same region as ichabod prompted us to re-examine whether the ichabod phenotype was a consequence of a loss of function ofβ-catenin-2.

The ichabod mutation maps near β-catenin-2 but does not functionally change its ORF

Map crosses were set up between ichabod homozygous females and brass males and the resulting fertilized eggs were injected withβ-catenin-2 RNA, which allowed rescue of these embryos to fertile adult fish. These heterozygotes were mated with each other and the F2 generation raised to adulthood. DNA was prepared from the adult female ichabod homozygotes (identified from the ichabod phenotype of their F3 offspring embryos), and tested for linkage of simple sequence repeat markers by standard techniques(Talbot and Schier, 1999). As shown in Fig. 2A, ichabod is located near the telomere of LG19, 1.1 cM distal to a CA repeat marker (CA85.6-3, a marker polymorphic for the map cross, found from genomic sequence to be closely linked to the non-polymorphic z26695 marker),and close to the position of β-catenin-2, which had been mapped to the same region in a separate map cross(Woods et al., 2000; Woods et al., 2005). We found that six markers distal to CA85.6-3, extending to a genetic distance of 10 cM from this marker, failed to show any recombination with the ichabodlocus. One of these markers was a CA repeat marker (CA91.6-1) located within the 3′ UTR of β-catenin-2. Because these results suggested a suppression of recombination in the telomeric region of ichabodLG19, we retested recombination frequencies of a number of markers from this region in an independent map cross (Fig. 2B). These markers were found to recombine with each other, as expected from their previously reported map positions(Shimoda et al., 1999)(http://zebrafish.mgh.harvard.edu/zebrafish/index.htm). Thus, it appears that the ichabod chromosome contains a region of suppressed recombination, perhaps owing to an inversion or other chromosomal rearrangement, in proximity to both the ichabod locus and theβ-catenin-2 gene. It is interesting in this regard that the cycb16 mutation, a deletion of the cyclops region near the telomere of LG12, also suppresses recombination on that chromosome(Talbot et al., 1998).

Sequence comparison of RT-PCR β-catenin-2 products from wild-type and ichabod homozygous fish revealed one amino acid change in the conceptually translated sequence. This change, a valine to methionine at position 450, however, does not diminish the ability of injectedβ-catenin-2 RNA to rescue the ventral phenotypes of ichabod embryos, as efficient rescue was obtained in two separate experiments.

Maternal expression of β-catenin-2, but not β-catenin-1, is decreased in ichabod embryos

The mapping data, and the possibility that the ichabod mutation involves a chromosomal rearrangement near β-catenin-2, prompted us to test whether ichabod embryos had decreased levels ofβ-catenin-2 transcript. We examined this by RT-PCR and by hybridization. Semi-quantitative RT-PCR (Fig. 3A) revealed no differences in β-catenin-1 transcript levels between wild-type and ichabod embryos at all stages examined, from the one-cell stage to 24 hpf. By contrast, β-catenin-2 transcript levels were markedly reduced in ichabod embryos from the one-cell stage to 50% epiboly. At bud stage, the level of β-catenin-2 RNA was only slightly lower in mutant embryos and by 24 hpf, the levels of this transcript were about the same in the two types of embryos. These results indicate that ichabod embryos exhibit a specific deficit in maternal (but not zygotic) expression of β-catenin-2.

In situ hybridization also revealed that β-catenin-2 mRNA is reduced in ichabod mutant embryos at early stages(Fig. 3B-M). In wild-type embryos, β-catenin-1 and β-catenin-2 transcripts were observed at the one-cell stage (Fig. 3B,H)and ubiquitously at sphere (Fig. 3C,I) and 90% epiboly stages(Fig. 3D,J). At the 90% epiboly stage, there was somewhat more pronounced expression of both genes in the dorsal midline. In ichabod embryos, there was little difference in expression of β-catenin-1 at the two early stages(Fig. 3E,F), and at 90%epiboly, expression in the animal half of the embryo was reduced and expression in the vegetal half of the embryo was more intense(Fig. 3G). The results were very different with β-catenin-2 expression at the one-cell and sphere stage embryos (Fig. 3K,L), which revealed a very low expression of this gene in ichabod embryos. By 90% epiboly(Fig. 3M), someβ-catenin-2 transcript can be detected, mainly in the vegetal half of the embryo. These findings support the idea that ichabodembryos are defective in maternal expression of β-catenin-2.Results from microarray hybridization experiments also were consistent with these findings. Using a Sigma-Compugen oligonucleotide microarray and RNA targets prepared from wild-type or ichabod embryos at 30% epiboly, we found that β-catenin-2 showed 3.7-fold downregulation in the mutant embryos (W. Wang, S.M. and E.S.W. unpublished). As will be described later, indirect assays using western immunoblotting show that pre-gastrula ichabod embryos are deficient in β-catenin-2, but notβ-catenin-1, protein. Thus, four lines of evidence all reveal a deficit in maternal expression of β-catenin-2 in ichabodembryos, but indicate no downregulation of zygotic expression of this gene or of maternal or zygotic expression of β-catenin-1 in mutant embryos.

β-Catenin-2, but not β-catenin-1, is required for organizer formation

To determine if β-catenin-2 is essential for development of dorsal axial structures, we injected morpholino antisense oligonucleotides(MOs) into one- to two-cell stage wild-type embryos. Injection of an MO directed against the translation initiation region ofβ-catenin-2 (MO2) phenocopied ichabod(Fig. 4D-H), whereas an MO directed against the translation initiation region ofβ-catenin-1 (MO1) had no ventralizing effect(Fig. 4A-C). After injection,embryos were allowed to develop to 24 hpf and then scored for degree of ventralization using a slight modification of previously defined criteria(Kelly et al., 2000). C1-C4 embryos all lack notochord and have increasing amounts of neural tube. C1 embryos (e.g. Fig. 4F) lack an embryonic axis, have a protruding tail-like structure, often with multiple tail fins, and a protruding mass of necrotic cells on the opposite end of the embryo. C1a embryos have spinal cord tissue, often reduced in length and girth, but do not form more anterior neural structures. C2 embryos have spinal cord, hindbrain and midbrain, but do not have forebrain or eyes (e.g. Fig. 4E). C3 embryos have defects in forebrain and/or eye development but have an intact neural tube posterior to the forebrain. C4 embryos have a normal appearing neural tube but lack notochord. C5 ichabod embryos, which are very rare, appear to be normal at 24 hpf, but do not survive. ichabod embryos are usually C1,C1a and C2, but as the expressivity of the mutation is somewhat variable, some crosses also yield substantial numbers of C3 embryos and, more rarely, C4 and C5 embryos.

Wild-type embryos injected with MO2 show all of the ventralized phenotypes found in progeny of female ichabod homozygotes(Fig. 4D-F,H). A second MO,MO2-2, designed against a non-overlapping region of β-catenin-2,gave similar distributions of ventralized embryos, when injected into wild-type embryos, whereas a mismatched MO, MO-2mis, with five base pair differences, had no effect (Fig. 4G). Injection of MO2 into ichabod embryos resulted in a shift of distribution of phenotypes to the more severely ventralized classes(Fig. 4I), suggesting that someβ-catenin-2 function remains in ichabod embryos and that this function is further reduced by the MO. In contrast to the results obtained with β-catenin-2 MOs, no ventralizing effect was seen in embryos injected with MOs designed against β-catenin-1 (e.g. Fig. 4A-C), even at high concentrations that caused bent tails and head necrosis(Fig. 4C). These phenotypes caused by high concentrations of MO1 were not rescued by injection ofβ-catenin-1* mRNA (in which the 5′-UTR ofβ-catenin-1 RNA was altered so that it was no longer complementary to the sequence of MO1, see Materials and methods), indicating that they were probably not due to specific interactions of the MO withβ-catenin-1 mRNA (data not shown). Injection into wild-type embryos of one or another of two additional non-overlapping MOs (MO1-2, MO1-3), designed against the 5′ UTR of β-catenin-1, also failed to yield ventralized phenotypes.

To test whether the MO effects observed at 24 hpf were a consequence of the disruption of early signaling centers, we examined expression of the early dorsal markers, bozozok (boz) and squint(sqt), in wild-type embryos injected with specific and control MOs(Fig. 5). Both of these markers are absent from dorsal blastomeres of severely ventralized batches of ichabod embryos at sphere stage(Kelly et al., 2000). Injection of MO2, but not MO1, is effective in eliminating bozexpression from wild-type sphere stage embryos(Fig. 5B,C). MO2 also decreases sqt expression in these embryos, although some transcript is still produced, but MO1 has no effect on expression of this gene(Fig. 5F,G). Thus, reducing expression of maternal β-catenin-2, but notβ-catenin-1, in wild-type embryos eliminates or reduces the expression of very early markers of the dorsal organizer. Although injection of MO2 ventralizes wild-type embryos, it does not result in the majority of embryos showing the most severe phenotypes(Fig. 4H), and thus, the failure to obtain complete inhibition of expression of early dorsal markers(e.g. sqt) in MO2-injected embryos is not unexpected. boz,however, is extremely sensitive to levels of β-catenin-2 but is completely insensitive to inhibition of β-catenin-1 expression. These results strongly support a specific requirement for β-catenin-2 in the early dorsal signaling center.

In addition to obtaining similar phenotypes with multiple MOs and the use of mismatched MO controls, we performed a number of additional tests to show that the MOs specifically act on their target transcripts. MO1 and MO2 were each shown to specifically inhibit the translation of their cognate lacZ fusion targets when the RNA targets and MO were injected together into wild-type embryos (data not shown). More importantly, we were able to rescue the ventralizing effects of MO2 on wild-type embryos(Fig. 4H). For this experiment,the 5′-UTR of β-catenin-2 RNA was altered so that it was no longer complementary to the sequence of MO2 (see Materials and methods). Injection of this RNA (β-cat-2* RNA) into ichabod embryos was effective in rescuing these embryos (data not shown). When this RNA was co-injected with MO2 into wild-type embryos, we observed effective rescue of the ventralized phenotype obtained with MO2 alone in the very same batch of embryos (Fig. 4H). Injection of the β-cat-2* RNA,alone, did not alter the wild-type phenotype. Additional controls, to show that the MO1 and MO2 that were used in these experiments were effective in decreasing the level of expressed β-catenin proteins, will be described in the next section along with results indicating that there is a deficiency in the amount of β-catenin-2 protein in pre-gastrula ichabodembryos.

Expression of the two β-catenin proteins in wild-type and ichabod embryos

The two β-catenin zebrafish proteins are of identical size and are very highly conserved in sequence. Although it would be extremely useful to obtain antibodies that could distinguish the two β-catenins, we have thus far been unsuccessful in raising antibodies against the specific C-terminal peptides. However, we were able to employ an indirect immunological approach to identify the contribution of each of the β-catenins to total cellularβ-catenin protein in wild-type and ichabod embryos. To achieve this, we used a pan-β-catenin antibody on western blots to compare levels of β-catenin in embryos injected with specific MOs against eachβ-catenin transcript or with control mismatched MOs (see Fig. S1 in the supplementary material). By comparing the effects of the MO treatment, it is possible to determine the relative amounts of β-catenin derived from the two types of β-catenin RNA. In wild-type and ichabod embryos at 100% epiboly, MO2 or MO1 treatment each decreases the amount of β-catenin and a combination of MO1 and MO2 reduces the level of β-catenin even further. We conclude that at this developmental stage when most of theβ-catenin transcript is presumably zygotically derived, the twoβ-catenins are expressed in both types of embryos. At 30% epiboly, the MO treatments gave different results for wild-type and ichabod embryos. In wild-type embryos, the results of MO treatment at 30% and 100% epiboly are essentially the same. Treatment of ichabod embryos at 30% epiboly,however, shows that MO2 has no effect on β-catenin protein levels,whereas MO1 greatly reduced the protein and, importantly, no further reduction is observed when the two MOs are injected together. We conclude that ichabod embryos produce little or no β-catenin-2 protein from maternal transcripts, although maternal expression of β-catenin-1 is normal. These results reinforce the conclusion that the ichabodmutation causes a specific downregulation of maternal expression ofβ-catenin-2.

The two β-catenins function redundantly to repress neurectoderm

To test if β-catenin-1 and β-catenin-2 might have overlapping functions in early embryonic patterning, we analyzed the phenotype of embryos in which expression of both β-catenin genes is inhibited. We carried out these studies in two ways: first, by injecting MO1 into ichabod embryos; and second, by injecting MO1 and MO2 into wild-type embryos (Fig. 6). Both types of embryos exhibited a new phenotype, which we termed `ciuffo' (a tuft or quiff of hair, in Italian) because of the appearance of a protrusion of the embryo away from the yolk (Fig. 6B,D). Examination of expression of a series of mesodermal and neurectodermal markers revealed that instead of a ventralized phenotype that might have been expected if both β-catenins functioned redundantly to promote dorsalization, the `ciuffo' embryos were dorsalized and expressed all neurectodermal markers that were tested. To determine whether this dorsalizing and neuralizing effect of injection of MO1 into ichabod embryos was due to the specific inhibition of β-catenin-1 transcript, we performed an RNA rescue experiment (Fig. 6E-H). Typical severe ichabod embryos(Fig. 7E) were mostly converted into `ciuffo' embryos by injection of MO1(Fig. 6F), but reverted back to C1, C1a and C2 ventralized embryos when MO1 and β-catenin-1*RNA were co-injected (Fig. 6H). When just the RNA was injected, full and partial rescue of the ichabod phenotype was obtained(Fig. 6G), a result expected from our finding that injection of all β-catenin RNAs tested, including Xenopus β-catenin RNA (Kelly et al., 2000), can induce organizer and rescue ichabod.The main conclusion of this experiment, however, is that the `ciuffo'phenotype can be rescued, indicating that the phenotype is not a result of a non-specific effect of MO1 on a non-β-catenin transcript.

We tested for expression of emx1(Fig. 6L-N), krox20(Fig. 7I-K) and hoxb6b(Fig. 6O-Q) to determine if`ciuffo' embryos were able to express a wide range of anterior and posterior neurectodermal markers. Whereas emx1 and krox20 were not expressed in the severe ichabod embryos used for these experiments(Fig. 6J,M), they are expressed at localized regions of the protrusion in `ciuffo' embryos(Fig. 6K,N), which developed as a result of injection of MO1 into ichabod embryos of the same breeding. Most ichabod embryos also fail to express hoxb6b(Fig. 6P), although expression is observed by hybridization in a small number of 24 hpf embryos (not illustrated) and confirmed by RT-PCR (data not shown). This low level of expression is consistent with the finding that ichabod mutants do express earlier posterior neurectodermal markers(Kudoh et al., 2004). hoxb6b is expressed in all `ciuffo' embryos (e.g. Fig. 6Q) and RT-PCR indicates a much greater amount of this transcript is present than in ichabodembryos. Thus, even for the most posterior marker tested, the absence of expression of the two β-catenins resulted in an increase in transcript levels. Interestingly, the more posterior the marker, the further its expression site was from the yolk. This pattern, as well as the double ring of krox20 expression (found in some, but not all embryos stained for this marker), indicated that the `ciuffo' protrusion had some degree of appropriate anteroposterior patterning, although the morphology of the embryo was abnormal. We also tested for expression of the neuronal markers, islet1 (Fig. 6R-T), tlxA (Andermann and Weinberg,2001) (data not shown) and HuC (data not shown). None of these genes was expressed in severely mutant ichabod embryos (e.g. Fig. 6S), but each was expressed in the `ciuffo' embryos obtained after MO1 injection. Thus, the neurectoderm that is formed by reducing expression of the two β-catenins does generate neurons, although they are not organized in a recognizably patterned way. We also assayed the expression of a number of mesodermal markers. Notably, myoD, which is expressed in concentric rings at the posterior end of ichabod embryos(Fig. 6V), is also expressed in the `ciuffo' protrusion (Fig. 6Z). ntl, although expressed during gastrula stages in ichabod embryos (Kelly et al.,2000) and `ciuffo' embryos (data not shown) as a pan-mesodermal marker, is not expressed in `ciuffo' gastrulae in any notochord-like structure or at post-gastrula stages. We also found that `ciuffo' embryos fail to express sonic hedgehog, another notochord marker. Thus, loss of function of β-catenin-1 in embryos already deficient inβ-catenin-2 leads to a restoration of neurectoderm, but not notochord. It is noteworthy that none of these effects observed in `ciuffo'embryos occurs merely by inhibiting expression of β-catenin-1 in wild-type embryos, indicating a genetic redundancy of the two β-catenins in repression of neurectoderm.

β-Catenin-1 and β-catenin-2 function redundantly to repress expression of chordin and goosecoid

To better understand how loss of function of both β-catenin genes can give rise to a wide range of anterior and posterior neural markers, we examined the expression in `ciuffo' embryos of genes normally expressed in dorsal regions at gastrula and pregastrula stages(Fig. 7). At 30% epiboly, the prospective organizer expresses high levels of boz, sqt, goosecoid(gsc) and chordin (chd)(Fig. 5A,D; Fig. 7A,D,G,M), while ichabod embryos do not express boz, gsc and chd,and express sqt circumferentially around the germ ring without any sign of a dorsal expression domain (Fig. 7B,E,H,N) (Kelly et al.,2000). In `ciuffo' embryos produced by MO1 treatment of ichabod embryos, there is no indication of dorsal expression of any of these markers at 30% epiboly, with results identical to ichabodembryos for boz, sqt and chd, and a faint radial symmetric expression of gsc around the germ ring(Fig. 7C,F,I,O). At 50%epiboly, this germ ring expression of gsc becomes more intense and a similar expression of chd is now observed(Fig. 7J,R). By this stage, for all three types of embryos, boz expression disappeared and sqt expression was greatly decreased (data not shown). We also tested for expression of fgf3 and fgf8, and their patterns of expression in the three types of embryos were identical to sqt at these stages. We conclude that a defined organizer does not form in `ciuffo'embryos, as there is no expression of boz and no dorsal expression of sqt, chd, fgf3 and fgf8. Although radial germ ring expression was observed for sqt, fgf3 and fgf8 in `ciuffo'embryos, the pattern was unchanged qualitatively and quantitatively from that of ichabod embryos. The late onset of gsc and chdexpression around the germ ring, however, indicates that β-catenin-1 andβ-catenin-2 normally function redundantly to repress expression of these genes in this domain. None of the genes tested was ectopically expressed when wild-type embryos were treated with MO1 (data not shown). Thus, the ventrolateral ectopic expression of a subset of genes normally expressed only dorsally is observed only in the absence of the two β-catenins. Interestingly, ectopic chd expression has also been observed in embryos expressing a dominant-negative form of Tcf3, which would be expected to inhibit the function of both β-catenins(Pelegri and Maischein, 1998). The ectopic expression of chd in the germ ring ventrolateral domain would be expected to result in inhibition of BMP signaling, thus explaining the precocious expression of neural markers in `ciuffo' embryos.

ichabod embryos lack maternal expression ofβ-catenin-2

ichabod, a spontaneous maternal effect mutation affecting organizer formation was found to disrupt Wnt pathway signaling and lead to a failure to localize β-catenin in dorsal YSL and blastomere nuclei(Kelly et al., 2000). Our more recent discovery of a second β-catenin gene, β-catenin-2,led us to consider that the phenotypes of ichabod mutant embryos were a result of failure of proper regulation of this gene. ichabod was found to map close to the position of β-catenin-2 in the telomeric region of LG19. Very suggestive of a chromosome rearrangement in this region of the ichabod chromosome was the failure to observe any recombination between markers extending from LG19 map positions 91.5 to 95.4 cM. Still ongoing fluorescent hybridization studies with interphase nuclei clearly show a different order of markers in this region of the ichabod chromosome compared with wild-type chromosomes (S.M., M. Nimmakayalu, B. Emanuel and E.S.W., unpublished). RT-PCR and whole mount hybridization experiments presented above show that ichabod mutant embryos are defective in maternal, but not zygotic, expression ofβ-catenin-2. Treatment of ichabod embryos with MO2 results in a shift of phenotypes to more severely ventralized classes(Fig. 4I), suggesting that ichabod is a hypomorphic β-catenin-2 mutant, which is shifted towards a null phenotype by further depletion of the protein product. Moreover, evidence from western blots of MO2- and MO1-treated embryos also indicated that β-catenin-2 (but not β-catenin-1) protein levels are low in pre-gastrula ichabod embryos. As MO β-catenin-2loss-of-function phenotypes are clearly the same as the ventralized ichabod phenotypes, we suggest that the ichabod mutation is a regulatory mutation affecting the maternal expression ofβ-catenin-2.

Maternal expression of β-catenin-2 is required for organizer formation

The results presented here clearly show that both zebrafish β-catenin genes are expressed maternally and zygotically in wild-type embryos. Although mRNA from both genes is expressed ubiquitously, we have uncovered a specific role of β-catenin-2 in organizer formation and early dorsal signaling. Wild-type embryos treated with MO2, or embryos bred from homozygous ichabod mothers both show a decrease in β-catenin-2transcripts and protein (Fig. 3, Fig. 6), fail to express markers of the organizer and anterior neurectoderm(Fig. 5, Fig. 7J,M), and exhibit similar ventralized phenotypes (Fig. 6). MOs directed against β-catenin-1 have no ventralizing effects on wild-type embryos and do not enhance the ventralization of ichabod embryos.

The specific requirement for β-catenin-2 for organizer formation was unexpected. Both β-catenin transcripts are expressed ubiquitously at stages prior to and during organizer formation. Moreover, our experiments revealed that both proteins were expressed from maternal transcripts in wild-type embryos. However, as we have not yet had success in raising antibodies that can specifically recognize each of the two proteins, we were not able to determine whether both proteins are present in the dorsal region of the embryo where β-catenin acts to establish DV asymmetry. At least three formal possibilities exist to explain the requirement forβ-catenin-2. First, the level of β-catenin-1 protein in dorsal territories may not be sufficient to compensate functionally for the loss ofβ-catenin-2. Second, the protein itself may be under post-translational control resulting in rapid inactivation of β-catenin-1 on the dorsal side of the embryo. Third, the β-catenin 1 and 2 proteins might have different activities (e.g. different transactivation domains or different requirements for co-factors), such that endogenous levels ofβ-catenin-1 are not able to activate early dorsal gene expression even though it can do so when overexpressed. A model in which β-catenin-2 promotes organizer formation by sequestering components that otherwise would facilitate degradation of β-catenin-1 is unlikely because MO1 depletion of β-catenin-1 in wild-type embryos does not ventralize and, in ichabod embryos, fails to increase the severity of ventralization. Most of the differences between the two β-catenins are found in the C-terminal 92 amino acids, corresponding to the major region of a transactivation domain known to interact with CBP and p300(Hecht et al., 2000; Takemaru and Moon, 2000). This region also can interact with the armadillo repeat region of the protein(Cox et al., 1999; Piedra et al., 2001) and has been implicated in regulating selective binding to cadherin or Tcf proteins(Gottardi and Gumbiner, 2004). However, to explain the lack of sufficiency of β-catenin-1 for organizer formation, such differences between the two β-catenins would also have to explain why both β-catenins appear to be absent in dorsal nuclei of ichabod embryos (Kelly et al.,2000).

We performed two types of experiments to judge if there were indeed functional differences between the two β-catenins in the context of organizer formation. First, we determined if expression of eitherβ-catenin from injected RNA was more efficient in rescuing ichabod embryos. The results showed that the two β-catenin RNAs were equally efficient in rescue, with a precipitous decline in the ability to rescue when the concentration of injected RNA for each of the twoβ-catenins was reduced from 10 ng/μl to 5 ng/μl. We also carried out an experiment in which tagged forms of each β-catenin were expressed from RNAs co-injected into ichabod embryos, and the subcellular and embryonic localization of the two proteins were then determined at the 500-cell stage using antibodies against the tagged epitopes(GFP-β-catenin-1 and myc-β-catenin-2) (data not shown). Our initial results showed that β-catenin-2 had a nuclear localization in a small group of marginal cells on one side of the embryo. However, taggedβ-catenin-1 protein was also found in the very same nuclei, offering no support for a difference in subcellular distribution between the twoβ-catenins as the basis for the requirement for β-catenin-2 for organizer formation. As each of the β-catenins is capable of localizing in blastoderm marginal nuclei, it is not surprising that rescue of ichabod embryos is achieved by injection of either β-catenin,but we still do not know the basis of the specific requirement for endogenousβ-catenin-2 for organizer formation.

Zebrafish β-catenins redundantly inhibit neurectoderm

In contrast to formation of the dorsal axial structures, loss of function of one or the other of the two zebrafish β-catenins does not appear to affect the role of the Wnt pathway in posteriorization and ventralization of neurectoderm and mesoderm (Lekven et al.,2001; Erter et al.,2001; Momoi et al.,2003; Ramel and Lekven,2004). We found that reduction of β-catenin-2expression alone results in ventralization, not dorsalization, and reduction of β-catenin-1 expression alone has no effect on dorsoventral or anteroposterior patterning. However, when expression of both β-catenins is inhibited, we found a robust formation of neurectodermal tissue in comparison with wild-type embryos treated with MO2 or with ichabodembryos, which only occasionally express posterior neural markers in their tail-like appendage. Although the neurectoderm formed when bothβ-catenins are inhibited is not organized into a normal neural tube, a substantial amount of tissue expressing both anterior and posterior neurectodermal markers is incorporated into a large protrusion extending out from the yolk (e.g. Fig. 7K,N,Q). The neurectoderm of these `ciuffo' embryos is able to form neurons (Fig. 7T) and appears to express neural markers in correct anteroposterior order. In the absence of organizer and β-catenin-2 expression,β-catenin-1 expression alone is able to repress the expression of all neural markers tested. When β-catenin-2 expression is unperturbed, inhibition of β-catenin-1 does not lead to an expansion of neural tissue or patterning defects. Thus, only when expression of both β-catenin genes is inhibited do we find a de-repression of neural markers.

It is very likely that this activation of neural marker expression is due to an impairment of Wnt pathway signaling normally involved in posteriorization and ventralization. A posteriorizing activity of zebrafish ventrolateral germring tissue has been demonstrated by transplantation studies(Woo and Fraser, 1997), and wnt8, which is expressed in this region of the embryo(Kelly et al., 1995b), is required for formation of ventrolateral and posterior mesoderm and spinal cord and posterior brain (Lekven et al.,2001; Erter et al.,2001; Momoi et al.,2003; Ramel and Lekven,2004). The posteriorizing effect of zebrafish Wnt signaling, at least in part, functions by negating the repression of posterior neural factors by Tcf3a (Kim et al.,2000; Dorsky et al.,2003). Other important targets of ventroposteriorizing Wnt signals include the vox/vent/ved, related homeodomain transcriptional repressors that restrict dorsal gene expression to the proper territories, and cdx4, which regulates the action of posterior Hox genes(Kawahara et al., 2000a; Kawahara et al., 2000b; Imai et al., 2001; Ramel and Lekven, 2004; Shimizu et al., 2005).

In zebrafish embryos lacking expression of both β-catenins, we hypothesize that the dorsolateral Wnt8 signals that normally posteriorize the neurectoderm are not transduced. Indeed, we find that ichabod embryos treated with MOs targeted against both translated forms of wnt8(Lekven et al., 2001) develop the `ciuffo' phenotype and, moreover, such embryos, as well as wild-type embryos injected with both MOs, express large amounts of chordintranscript around the germ ring (M.V., S.M. and E.S.W., unpublished)(Ramel and Lekven, 2004). Our finding that MO1 treatment of ichabod embryos also results in germ ring chordin expression (Fig. 7R) suggests that chordin repression by Wnt8 and vox/vent/ved is mediated by both β-catenin-1 and -2 (M.V., S.M. and E.S.W., unpublished). A true dorsal signaling center is not established in these embryos as boz fails to be expressed, and otherwise dorsal markers are expressed later than they normally would be, consistent with a response to the later Wnt8 signaling. Although some ichabod embryos express hoxb6b in their tail extension [an observation consistent with expression of earlier posterior neural markers in ichabodembryos (Kudoh et al., 2004)],the amount of transcript of this factor is also markedly increased in MO1-treated ichabod embryos. Thus, not only do the twoβ-catenins act to posteriorize and ventralize the embryo, but they also act to repress formation of neurectoderm. In a normal embryo, such repression may act to restrict the amount of organizer-induced neurectoderm formation.

Neurectoderm patterning in the absence of an organizer

The studies presented above add to an array of evidence indicating that AP patterning of the neurectoderm can occur in embryos lacking organizer tissue. Ectoderm of embryos lacking expression of both β-catenins forms neurectoderm that expresses posterior, hindbrain and forebrain markers in apparently correct AP order. Wnt/β-catenin signaling is thus not required in the early embryo for either neural induction or for at least some degree of proper AP patterning. A similar conclusion has recently been reached by simultaneous elimination of the Spemann organizer and BMP signaling in Xenopus (Reversade et al.,2005). In this case, BMP signaling was eliminated by MO treatment against three different BMP ligands. In the zebrafish, as shown above, the depletion of the two β-catenins results in ectopic expression of chordin, which would in effect reduce or eliminate BMP signaling in the absence of organizer. A significant difference between the two organisms is that elimination of the single β-catenin in Xenopus does not result in chordin expression(Reversade et al., 2005), and totally eliminates neural induction(Heasman et al., 1994; Heasman et al., 2000), whereas in zebrafish both β-catenins appear to be required to prevent chordin expression in ventrolateral regions of the germ ring. In both organisms, however, it is now clear that extensive neurectoderm can be formed with a degree of proper AP pattern in the absence of β-catenin and organizer. Further study of this condition may provide insight into factors that could underlie the anteroposterior pattern of the neurectoderm in the absence of both Wnt/β-catenin and BMP signaling.

Supplementary material

We thank Joshua Bradner, Jiangyan He, Richard Gill, Jr and Carol Hu for fish care and general laboratory assistance. We gratefully acknowledge Lucía Peixoto for help with the phylogenetic analysis. This work was supported by NIH grants HD39272 (to E.S.W.) and RR12349 (to W.S.T.).

Andermann, P. and Weinberg, E. S. (
2001
). Expression of zTlxA, a Hox-11-like gene, in early differentiating neurons and cranial sensory ganglia of the zebrafish embryo.
Dev. Dyn.
222
,
595
-610.
Christian, J. L. and Moon, R. T. (
1993
). Interactions between Xwnt-8 and Spemann organizer signaling pathways generate dorsoventral pattern in the embryonic mesoderm of Xenopus.
Genes Dev.
7
,
13
-28.
Christian, J. L., McMahon, J. A., McMahon, A. P. and Moon, R. T. (
1991
). Xwnt-8, a Xenopus Wnt-1/int-1-related gene responsive to mesoderm-inducing growth factors,may play a role in ventral mesodermal patterning during embryogenesis.
Development
111
,
1045
-1055.
Cox, R. T., Pai, L.-M., Kirkpatrick, C., Stein, J. and Peifer,M. (
1999
). Roles of the C-terminus of armadillo in wingless signaling in Drosophila.
Genetics
153
,
319
-332.
Dorsky, R. I., Sheldahl, L. C. and Moon, R. T.(
2002
). A transgenic Lef1/β-catenin-dependent reporter is expressed in spatially restricted domains throughout zebrafish development.
Dev. Biol.
241
,
229
-237.
Dorsky, R. I., Itoh, M., Moon, R. T. and Chitnis, A.(
2003
). Two tcf3 genes cooperate to pattern the zebrafish brain.
Development
130
,
1937
-1947.
Dougan, S. T., Warga, R. M., Kane, D. A., Schier, A. F. and Talbot, W. S. (
2003
). The role of the zebrafish nodal-related genes squint and cyclops in patterning of mesendoderm.
Development
130
,
1837
-1851.
Edgar, R. C. (
2004
). MUSCLE: a multiple sequence alignment method with reduced time and space complexity.
BMC Bioinformatics
5
,
1
-19.
Erter, C. E., Solnica-Krezel, L. and Wright, C. V. E.(
1998
). The zebrafish nodal-related 2 encodes and early mesendodermal inducer signaling from the extraembryonic yolk syncytial layer.
Dev. Biol.
204
,
361
-372.
Erter, C. E., Wilm, T. P., Basler, N., Wright, C. V. E. and Solnica-Krezel, L. (
2001
). Wnt8 is required in lateral mesendodermal precursors for neural posteriorization in vivo.
Development
128
,
3571
-3583.
Felsenstein, J. (
1989
). PHYLIP-Phylogeny Inference Package (Version 3.2).
Cladistics
5
,
164
-166.
Fridieu, J. R., Cui, Y., Maier, D., Danilchik, M. V. and Christian, J. L. (
1997
). Xwnt-8 and lithium can act upon either dorsal mesodermal or neuroectodermal cells to cause a loss of forebrain in Xenopus embryos.
Dev. Biol.
186
,
100
-114.
Gottardi, C. J. and Gumbiner, B. M. (
2004
). Distinct molecular forms of β-catenin are targeted to adhesive or transcriptional complexes.
J. Cell Biol.
167
,
339
-349.
Haelgel, H., Larue, L., Ohsugi, M., Federov, L., Herrenknecht,K. and Kemler, R. (
1995
). Lack of beta-catenin affects mouse development at gastrulation.
Development
121
,
3529
-3537.
Harland, R. and Gerhart, J. (
1997
). Formation and function of Spemann's organizer.
Annu. Rev. Cell Dev. Biol.
13
,
611
-667.
Heasman, J., Crawford, A., Goldstone, K., Garner-Harnick, P.,Gumbiner, B., McCrea, P., Kintner, C., Noro, C. Y. and Wylie, C.(
1994
). Overexpression of cadherins and underexpression of bet-catenin inhibit dorsal mesoderm induction in early Xenopusembryos.
Cell
79
,
791
-803.
Heasman, J., Kofron, M. and Wylie, C. (
2000
).β-catenin signaling activity dissected in the early Xenopusembryo: a novel antisense approach.
Dev. Biol.
222
,
124
-134.
Hecht, A., Vleminckx, K., Stemmler, M. P., van Roy, F. and Kemler, R. (
2000
). The p300/CBP acetyltransferases function as transcriptional coactivators of β-catenin in vertebrates.
EMBO J.
19
,
1839
-1850.
Hoppler, S., Brown, J. D. and Moon, R. T.(
1996
). Expressiion of a dominant-negative Wnt blocks induction of MyoD in Xenopus embryos.
Genes Dev.
10
,
2805
-2817.
Huelsken, J., Vogel, R., Brinkmann, V., Eerdman, B., Birchmeier,C. and Birchmeier, W. (
2000
). Requirement for β-catenin in anterior-posterior axis formation in mice.
J. Cell Biol.
148
,
567
-578.
Imai, Y., Gates, M. A., Melby, A. E., Kimelman, D., Schier, A. F. and Talbot, W. S. (
2001
). The homeobox genes voxand vent are redundant repressors of dorsal fates in zebrafish.
Development
128
,
2407
-2420.
Inoue, A., Takahashi, M., Hatta, K., Hotta, Y. and Okamoto,H. (
1994
). Developmental regulation of islet-1 mRNA expression during neuronal differentiation in embryonic zebrafish.
Dev. Dyn.
199
,
1
-11.
Kawahara, A., Wilm, T., Solnica-Krezel, L. and Dawid, I. B.(
2000a
). Antagonistic role of vega1 and bozozok/dharma homeobox genes in organizer formation.
Proc. Natl. Acad. Sci. USA
97
,
12121
-12126.
Kawahara, A., Wilm, T., Solnica-Krezel, L. and Dawid, I. B.(
2000b
). Functional interaction of vega2 and goosecoid homeobox genes in zebrafish.
Genesis
28
,
58
-67.
Kelly, C., Chin, A. J., Leatherman, J. L., Kozlowski, D. J. and Weinberg, E. S. (
2000
). Maternally controlledβ-catenin-mediated signaling is required for organizer formation in the zebrafish.
Development
127
,
3899
-3911.
Kelly, G. M., Erezyilmaz, D. F. and Moon, R. T.(
1995a
). Induction of a secondary embryonic axis in zebrafish occurs following the overexpression of β-catenin.
Mech. Dev.
53
,
261
-273.
Kelly, G. M., Greenstein, P., Erezyilmaz, D. F. and Moon, R. T. (
1995b
). Zebrafish wnt8 and wnt8b share a common activity but are involved in distinct developmental pathways.
Development
121
,
1787
-1799.
Kim, C.-H., Oda, T., Itoh, M., Jiang, D., Artinger, K. B.,Chandrasekharappa, S. C., Driever, W. and Chitins, A. B.(
2000
). Repressor activity of Headless/Tcf3 is essential for vertebrate head formation.
Nature
407
,
913
-916.
Kim, S.-H., Shin, J., Park, H.-C., Yeo, S.-Y., Hong, S.-K., Han,S., Rhee, M., Kim, C.-H., Chitnis, A. B. and Huh, T.-L.(
2002
). Specification of an anterior neuroectoderm patterning by Frizzled8a-mediated Wnt8b signaling during late gastrulation in zebrafish.
Development
129
,
4443
-4455.
Koos, D. S. and Ho, R. K. (
1998
). The nieuwkoid/dharma homeobox gene is essential for bmp2b repression in the zebrafish pregastrula.
Dev. Biol.
215
,
190
-207.
Kudoh, T., Concha, M. L., Houart, C., Dawid, I. B. and Wilson,S. W. (
2004
). Combinatorial Fgf and Bmp signaling patterns the gastrula ectoderm into prospective neural and epidermal domains.
Development
131
,
3581
-3592.
Larabell, C. A., Torres, M., Rowning, R. A., Yost, C., Miller,J. R., Wu, M., Kimelman, D. and Moon, R. T. (
1997
). Establishment of the dorsoventral axis in Xenopus embryos is presaged by early asymmetries in beta-catenin that are modulated by the Wnt signaling pathway.
J. Cell Biol.
136
,
1123
-1136.
Lekven, A. C., Thorpe, C. J., Waxman, J. S. and Moon, R. T.(
2001
). Zebrafish wnt8 encodes two Wnt8 proteins on a bicistronic transcript and is required for mesoderm and neurectoderm patterning.
Dev. Cell
1
,
103
-114.
Miller-Bertoglio, V. E., Fisher, S., Sanchez, A., Mullins, M. C. and Halpern, M. E. (
1997
). Differential regulation of chordin expression domains in mutant zebrafish.
Dev. Biol.
192
,
537
-550.
Momoi, A., Yoda, H., Steinbesser, H., Fagotto, F., Kondoh, H.,Kudo, A., Driever, W. and Furatani-Seiki, M. (
2003
). Analysis of Wnt8 for neural posteriorizing Frizzled 8c and Frizzled 9 as functional factor by identifying receptors for Wnt8.
Mech. Dev.
120
,
477
-489.
Morita, T., Nitta. H., Kiyama, Y., Mori, H. and Mishina, M.(
1995
). Differential expression of 2 zebrafish Emx homoprotein messenger-RNAs in the developing brain.
Neurosci. Lett.
198
,
131
-134.
Nieuwkoop, P. D. and Nigtevecht, G. V. (
1954
). Neural activation and transformation in explants of competent ectoderm under the influence of fragments of aanterior notochord in urodeles.
J. Embryol. Exp. Morphol.
2
,
175
-193.
Nieuwkoop, P. D., Botterenbrod, E. C., Kremer, A., Bloesma, F. F. S. N., Hoessels, E. L. M. J., Meyer, G. and Verheyen, F. J.(
1952
). Activation and organization of the central nervous system in amphibians.
J. Exp. Zool.
120
,
1
-108.
Oxtoby, E. and Jowett, T. (
1993
). Cloning of the zebrafish Krox-20 gene (krx-20) and its expression during hindbrain development.
Nucleic Acids Res.
21
,
1087
-1095.
Pelegri, F. and Maischen, H.-M. (
1998
). Function of zebrafish β-catenin and TCF-3 in dorsoventral patterning.
Mech. Dev.
77
,
63
-74.
Piedra, J., Martínez, D., Castaño, J., Miravet,S., Duñach, M. and de Herreros, A. G. (
2001
). Regulation of β-catenin structure and activity by tyrosine phosphorylation.
J. Biol. Chem.
276
,
20436
-20443.
Postlethwait, J. H., Yan, Y. L., Gates, M. A., Horne, S.,Amores, A., Brownlie, A., Donovan, A., Egan, E. S., Force, A., Gong, Z. Y. et al. (
1998
). Vertebrate genome evolution and the zebrafish gene map.
Nat. Genet.
18
,
345
-349.
Prince, V. E., Joly, L., Ekker, M. and Ho, R. K.(
1998
). Zebrafish hox genes: genomic organization and modified collinear expression patterns in the trunk.
Development
125
,
407
-420.
Ramel, M. C. and Lekven, A. C. (
2004
). Repression of the vertebrate organizer by Wnt8 is mediated by Vent and Vox.
Development
131
,
3991
-4000.
Rebagliati, M. R., Toyama, R., Fricke, C., Haffter, P. and Dawid, I. B. (
1998
). Zebrafish nodal-related genes are implicated in axial patterning and establishing left-right asymmetry.
Dev. Biol.
199
,
261
-272.
Reversade, B., Kuroda, H., Lee, H., Mays, A. and De Robertis, E. M. (
2005
). Depletion of Bmp2, Bmp4, Bmp7 and Spemann organizer signals induces massive brain formation in Xenopus embryos.
Development
132
,
3381
-3392.
Rowning, B. A., Wells, J., Wu, M., Gerhart, J. C., Moon, R. T. and Larabell, C. A. (
1997
). Microtubule-mediated transport of organelles and localization of beta-catenin to the future dorsal side of Xenopus eggs.
Proc. Natl. Acad. Sci. USA
94
,
1224
-1229.
Schier, A. F. and Talbot, W. S. (
2005
). Molecular genetics of axis formation in zebrafish.
Annu. Rev,Genet.
15
,
561
-613.
Schneider, S., Steinbesser, H., Warga, R. M. and Hausen, P.(
1996
). β-catenin translocation into nuclei demarcates the dorsalizing centers in frog and fish embryos.
Mech. Dev.
57
,
191
-198.
Shimizu, T., Bae, Y. K., Muraoka, O. and Hibi, M.(
2005
). Interaction of Wnt and caudal-related genes in zebrafish posterior body formation.
Dev. Biol.
279
,
125
-141.
Shimoda, N., Knapik, E. W., Ziniti, J., Sim, C., Yamada, E.,Kaplan, S., Jackson, D., de Sauvage, F., Jacob, H. and Fishman, M. C.(
1999
). Zebrafish genetic map with 2000 microsatellite markers.
Genomics
58
,
219
-232.
Smith, W. C. and Harland, R. M. (
1991
). Injected Xwnt-8 RNA acts early in Xenopus embryos to promote formation of a vegetal dorsalizing center.
Cell
67
,
753
-765.
Sokol, S. Y. (
1999
). Wnt signaling and dorsoventral axis specification in vertebrates.
Curr. Opin. Genet. Develop.
9
,
405
-410.
Stachel, S. E., Grunwald, D. J. and Myers, P. Z.(
1993
). Lithium perturbation and goosecoid expression identify a dorsal specification pathway in the pregastrula zebrafish.
Development
117
,
1261
-1274.
Takemaru, K.-I. and Moon, R. T. (
2000
). The transcriptional coactivator CBP interacts with β-catenin to activate gene expression.
J. Cell Biol.
149
,
249
-254.
Talbot, W. S., Egan, E. S., Gates, M. A., Walker, C., Ullmann,B., Neuhauss, S. C. F., Kimmel, C. B. and Postlethwait, J. H.(
1998
). Genetic analysis of chromosomal rearrangements in the cyclops region of the zebrafish genome.
Genetics
148
,
373
-380.
Talbot, W. S. and Schier, A. F. (
1999
). Positional cloning of mutated zebrafish genes.
Meth. Cell Biol.
60
,
259
-286.
Tao, Q. H., Yakota, C., Puck, H., Kofron, M., Birsoy, B., Yan,D., Asashima, M., Wylie, C. C., Lin, X. H. and Heasman, J.(
2005
). Maternal Wnt11 activates the canonical wnt signaling pathway required for axis formation in Xenopus embryos.
Cell
120
,
857
-871.
Weidinger, G., Thorpe, C. J., Wuennenberg-Stapleton, K., Ngai,J. and Moon, R. T. (
2005
). The Sp1-related transcription factors sp5 and sp5-like act downstream of Wnt/β-catenin signaling in mesoderm and neuroectoderm patterning.
Curr. Biol.
15
,
489
-500.
Weinberg, E. S., Allende, M. L., Kelly, C. S., Abdelhamid, A.,Murakami, T., Andermann, P., Doerre, O. G., Grunwald, D. J. and Riggleman,B. (
1996
). Developmental regulation of zebrafish MyoD in wild-type, no tail and spadetail embryos.
Development
122
,
271
-280.
Westerfield, M. (
1993
).
The Zebrafish Book: A Guide for the Laboratory Use of Zebrafish (Brachydanio rerio)
. Eugene, OR: University of Oregon Press.
Woo, K. and Fraser, S. E. (
1997
). Specification of the zebrafish nervous system by nonaxial signals.
Science
277
,
254
-257.
Woods, I. G., Kelly, P. D., Chu, F., Ngo Hazelett, P., Yan,Y.-L., Huang, H., Postlethwait, J. H. and Talbot, W. S.(
2000
). A comparative map of the zebrafish genome.
Genome Res.
10
,
1903
-1914.
Woods, I. G., Wilson, C., Friedlander, B., Chang, P., Reyes, D. K., Nix, R., Kelly, P. D., Chu, F., Postlethwait, J. H. and Talbot, W. S.(
2005
). The zebrafish gene map defines ancestral vertebrate chromosomes.
Genome Res.
15
,
1307
-1314.
Wylie, C., Kofron, M., Payne, C., Anderson, R., Hosobuchi, M.,Joseph, E. and Heasman, J. (
1996
). Maternal beta-catenin establishes a `dorsal signal' in early Xenopus embryos.
Development
122
,
2987
-2996.

Supplementary information