Homologues of the T-box gene Brachyury play important roles in mesoderm differentiation and other aspects of early development in all bilaterians. In the diploblast Hydra, the Brachyuryhomologue HyBra1 acts early in the formation of the hypostome, the location of the organiser in adult Hydra. We now report the isolation and characterisation of a second Brachyury gene, HyBra2. Sequence analysis suggests that HyBra1 and HyBra2 are paralogues, resulting from an ancient lineage-specific gene duplication. We show that both paralogues acquired novel functions, both at the level of their cis-regulation as well as through significant divergence of the coding sequence. Both genes are expressed in the hypostome, but HyBra1 is predominantly endodermal, whereas HyBra2 transcripts are found primarily in the ectoderm. During bud formation, both genes are activated before any sign of evagination, suggesting an early role in head formation. During regeneration, HyBra1 is an immediate-early response gene and is insensitive to protein synthesis inhibition, whereas the onset of expression of HyBra2 is delayed and requires protein synthesis. The functional consequence of HyBra1/2 protein divergence on cell fate decisions was tested in Xenopus. HyBra1 induces mesoderm, like vertebrate Brachyury proteins. By contrast, HyBra2 shows a strong cement-gland and neural-inducing activity. Domain-swapping experiments show that the C-terminal domain of HyBra2 is responsible for this specific phenotype. Our data support the concept of sub- and neofunctionalisation upon gene duplication and show that divergence of cis-regulation and coding sequence in paralogues can lead to dramatic changes in structure and function.

The question of how morphological complexity increased during animal evolution has been revitalised with the recent sequencing of genomes and transcriptomes of simple metazoans, such as cnidarians. These studies have revealed that the vast majority of important developmental genes known in vertebrate systems appeared very early during evolution in the common ancestor of cnidarians and bilaterians 600-700 million years ago(Kusserow et al., 2001; Kortschak et al., 2003; Technau et al., 2005; Miller et al., 2005; Putnam et al., 2007). In contrast to the apparent morphological simplicity of the Cnidarians it appears that there is no significant change in the gene set between basal cnidarians and vertebrates (Technau et al.,2005; Putnam et al.,2007). One factor that adds to the complexity of cnidarian genomes is the relatively high number of independent gene duplications(Technau et al., 2005; Chourrout et al., 2006). Gene duplications are thought to be one of the major driving forces of evolution. Although many of the duplicated genes are predicted to end up as pseudogenes,some contribute significantly to the evolution of developmental processes by establishing novel interactions and functions. In principle, a fully functional copy of a gene reduces the selection pressure significantly,allowing for rapid divergence of one or both copies of the gene(Ohno, 1970; Cooke et al., 1997; Krakauer and Nowak, 1999). Comparisons of paralogous genes and dominance of mutations in yeast and humans suggest that haploinsufficient genes, whose loss-of-function alleles are dominant over wild-type alleles, are more likely to be fixed after gene duplication than haplosufficient (i.e. dominant wild-type) genes(Kondrashov and Koonin, 2004). Interestingly, haploinsufficient genes include predominantly developmental regulators, structural and regulatory proteins and transcription regulators.

Predictions surrounding the fate of duplicate genes over long evolutionary time scales have been made in theoretical studies(Cooke et al., 1997; Force et al., 1999). The Duplication-Degeneration-Complementation model of Force and colleagues focuses on the evolution of cis-regulatory regions after the duplication event,although it does not explicitly exclude a divergence of the coding sequence(Lynch and Force, 2000). The detailed experimental analysis of paralogous genes and the degree of conservation and diversification of functions might therefore provide insight into metazoan evolution. Furthermore, in the case of important developmental pathways, paralogous genes might contribute to the evolution of animal body plans.

One of these important developmental genes encodes the T-box transcription factor Brachyury (Herrmann et al., 1990). Brachyury has a conserved role in mesoderm differentiation, and elongation of the posterior body axis in all vertebrates(for reviews, see Smith, 1997; Smith, 1999; Naiche et al., 2005). Comparison of basal deuterostome and protostome larvae, as well as diploblastic cnidarians, indicates that Brachyury has an ancestral function in defining the blastopore (Arendt et al., 2001; Scholz and Technau, 2003) (for a review, see Technau, 2001), although the precise cellular function can vary in the different animal lineages (e.g. Wilkinson et al., 1990; Gross and McClay, 2001; Lartillot et al., 2002). Interference with Brachyury function in vertebrates inhibits convergent extension, causes apoptosis, and reveals separate requirements in the Fibroblast growth factor (FGF)- and Activin-signalling pathways(Conlon and Smith, 1999).

Although Brachyury is considered an early panmesodermal marker in vertebrates, this `mesodermal' gene also occurs in diploblasts, as first reported in the freshwater cnidarian Hydra(Technau and Bode, 1999). The Hydra Brachyury homologue HyBra1 is expressed in the hypostome, the most apical part of the polyp, between the tentacle ring and the mouth opening (Technau and Bode,1999). The hypostome is the location of the head organiser of the animal (Browne, 1909; Broun and Bode, 2002), which is considered equivalent to the Spemann organiser in amphibians(Spemann und Mangold, 1924). Comparative evolutionary studies in basal cnidarians indicate that the oral end of the Hydra polyp derives from the blastopore(Scholz and Technau, 2003). Under all developmental circumstances, i.e. bud formation, head regeneration,reaggregation and embryogenesis, HyBra1 is expressed very early at the site of future hypostome formation, suggesting an early role in head organiser/head formation in Hydra(Technau and Bode, 1999; Technau et al., 2000; Broun and Bode, 2002).

T-Box genes have never been reported from unicellular organisms. It is therefore likely that this class of developmental control genes evolved with multicellularity. Two T-box genes, tbx2 and Brachyury, have been isolated from sponges (Adell et al.,2003; Manuel et al.,2004). T-box genes fall into a number of subfamilies with Brachyury as the most distinct subfamily(Papaioannou and Silver, 1998; Papaioannou, 2001; Wilson and Conlon, 2002). Most animals appear to have only one Brachyury gene; however, a few organisms, including vertebrates, have two or more paralogous Brachyury genes. Where studied, these paralogues are highly similar in sequence and have mostly overlapping expression domains(Strong et al., 2000; Knezevic et al., 1997; Hayata et al., 1999),suggesting rather recent gene duplication events.

In this paper, we report the isolation of a second Hydra Brachyurygene, which appears to be the result of an ancient gene duplication event. Our phylogenetic analysis suggests that the duplication event most likely occurred after the divergence of cnidarians. We show that HyBra1 and HyBra2 occupy distinct expression domains in the hypostome of the polyp, and show very different dynamics of expression during head regeneration. This indicates an evolution of cis-regulatory elements. To investigate the impact of coding-sequence evolution between the two paralogues, we tested the function of both Hydra genes in a heterologous assay system: the Xenopus animal caps. This assay revealed strikingly different inductive capacities for HyBra1 and HyBra2. Domain-swapping experiments indicate that their different activities are defined largely by their divergent C-terminal domains. Our data show that the Hydra Brachyury paralogues have undergone a mixture of subfunctionalisation and neofunctionalisation and the evolution of HyBra gene function occurred both at the cis-regulatory and the protein levels.

Animals and embryological experiments

All Hydra experiments were carried out with polyps of the Basel and AEP strains of Hydra vulgaris. One-day-starved animals were used for all experiments. Hydra polyps were cultured as previously described (Technau and Holstein,1996). Xenopus embryos were in vitro fertilised,de-jellied and cultured by standard methods(Sive et al., 1989). At the four-cell stage, each cell was microinjected with 2.5 nl volume into the animal hemisphere. RNA doses ranged from 0.1 to 1.0 ng/embryo, and were adjusted to result in roughly comparable protein amounts, as judged by western blot analysis of embryos injected with Myc-tagged protein variants. Animal caps were dissected using Dumont N°5 forceps between stages 8.5 and 9(Nieuwkoop and Faber, 1967)and cultivated at 23°C in 0.5× MBS containing 10 μg/ml Gentamycin until stage 36 (hatching).

Cloning of HyBra1 and HyBra2

Fragments of the T-box were cloned by PCR using degenerate primers (forward 5′-TTYGGNGMNCAYTGGATG-3′ and reverse 5′-RAANSCYTTNGCRAANGG-3′) at low annealing temperatures. Random-primed probes of these HyBra1 and HyBra2 T-box fragments were used to screen a Lambda-Zap cDNA library (Stratagene) using standard protocols (Sambrook et al.,1989). A full-length clone of HyBra2 was obtained from the screen of the cDNA library. To obtain the 3′ end of HyBra13′ RACE experiments were performed as described(Frohman et al., 1988). GenBank accession numbers of HyBra1 and HyBra2 full-length clones are (AY366371, AY366372), respectively.

DNA constructs and in vitro transcription

For RNA microinjection experiments, the Xbra cDNA was subcloned into the pRNA3 vector; Hybra1 and Hybra2 coding regions were cloned by PCR into BamHI/XbaI sites of pCS2+ and StuI/XbaI sites of pCS2+MT6 respectively. The chimeric constructs XTH2A and H2TXA were generated by subsequent subcloning of the respective PCR products. For both chimerae, the conserved amino acid motif AKAFL-DAKER was used as a junction. Capped in vitro transcripts were generated from SfiI-linearised plasmids and T3 RNA polymerase (HyBra chimerae and Xbra), or by NotI linearisation and Sp6 RNA polymerase (wt HyBra paralogues).

Phylogenetic analysis

ClustalW was used to align amino acid sequences of the T-box proteins. Maximum likelihood program PUZZLE (Schmidt et al., 2002) and MrBayes(Huelsenbeck and Ronquist,2001) were used to perform a phylogenetic tree analysis. In the Puzzle analysis 1000 replica were computed using JTT as a substitution model,assuming a strong rate heterogeneity in the gamma distribution. In MrBayes 3.1, 1,7 Mio generations were run using the WAG model of substitution and eight gamma rate categories.

RNA in situ hybridisation

Hydra in situ hybridisation was carried out as previously described (Grens et al., 1995)with minor modifications. We omitted the elevated temperature step [80°C,and for detection we used NBT/BCIP instead of BMpurple (Roche)]. Xenopus double in situ hybridisations were performed as described(Wittenberg et al., 1999) with a primary Fast Red stain for the fluorescein-labelled probe terminated by heat inactivation for 30 minutes at 65°C in 0.1 M EDTA-TBS buffer, followed by a BM-Purple (Roche) stain for the digoxygenin-labelled probe.

Protein synthesis assay

To assay for protein synthesis, 50 μCi [35S]methionine was injected into the gastric cavity of adult polyps, and incubated for 1 hour. Thereafter, five polyps were dissolved in 100 μl lysis buffer (50 mM Tris-HCl, pH 8.0, 2% SDS, 100 mM β-mercaptoethanol). After lysis, an equal volume of 1 mg/ml BSA was added and 100 μl lysate was spotted and air-dried on a glass fibre filter (Whatman GF/A 2.4 cm; #1820024). Then the filter was rinsed twice with 3 ml ice-cold 20% trichloroacetic acid (TCA) for 5 minutes, followed by two washes in 12.5% TCA to remove unincorporated labeled amino acids. Filters were then washed twice with 95% ethanol followed by one wash with acetone, air dried, placed in scintillation vials with 1 ml scintillation fluid and the total number of counts determined. As a negative control, five unlabeled animals were lysed and then 50 μCi[35S]methionine was added to the lysate.

Inhibition of translation by cycloheximide

Polyps were incubated in 5 or 50 μg/ml cycloheximide (Sigma) in Hydra medium (HM) for 15 minutes, 1 hour or 2 hours. After treatment,polyps were washed five times in HM to remove traces of cycloheximide. To inhibit translation during regeneration, animals were pretreated with cycloheximide for 15 minutes and subsequently decapitated at t0beneath the tentacles. Regenerates were further incubated in cycloheximide for 2 hours, washed five times and allowed to regenerate for the times indicated.

Fig. 1.

Comparison of protein and DNA sequence conservation of HyBra1 and HyBra2. (A) Alignment of T-domains of HyBra1 and HyBra2 with Brachyury proteins from mouse (M-T), Xenopus (X-bra), Amphioxus(Am-bra-1), the sea urchin Hemicentrotus pulcherrimus (Hp-T), the ascidian Halocynthia roretzi (As-T), Platynereis dumerilii(Pd-bra) and Drosophila melanogaster (Dm-Trg). Amino acid identities are black, partial identities (>60%) in grey. Amino acids marked by grey dots under the sequence are involved in DNA-binding, grey bars indicate amino acids involved in dimerisation(Müller and Herrmann,1997). Arrows indicate conserved intron sites in HyBra1and mouse-T. (B) Alignment of the core repression module R1(Kispert et al., 1994) within the activation domain. HyBra2 and Drosophila Trg do not have this motif. (C) Schematic structures of Xbra, HyBra1 and HyBra2 protein illustrate the degree of amino acid identity in the T-box and the R1 domain(blue box) in HyBra1 and HyBra2 compared with XBra.(D) Comparison of exon-intron structure in Xbra, HyBra1 and HyBra2. Filled boxes mark the ORF, white boxes the UTR. Red boxes indicate conserved exon-intron sites, whereas black boxes indicate non-conserved regions. Black numbers represent the length of introns.

Fig. 1.

Comparison of protein and DNA sequence conservation of HyBra1 and HyBra2. (A) Alignment of T-domains of HyBra1 and HyBra2 with Brachyury proteins from mouse (M-T), Xenopus (X-bra), Amphioxus(Am-bra-1), the sea urchin Hemicentrotus pulcherrimus (Hp-T), the ascidian Halocynthia roretzi (As-T), Platynereis dumerilii(Pd-bra) and Drosophila melanogaster (Dm-Trg). Amino acid identities are black, partial identities (>60%) in grey. Amino acids marked by grey dots under the sequence are involved in DNA-binding, grey bars indicate amino acids involved in dimerisation(Müller and Herrmann,1997). Arrows indicate conserved intron sites in HyBra1and mouse-T. (B) Alignment of the core repression module R1(Kispert et al., 1994) within the activation domain. HyBra2 and Drosophila Trg do not have this motif. (C) Schematic structures of Xbra, HyBra1 and HyBra2 protein illustrate the degree of amino acid identity in the T-box and the R1 domain(blue box) in HyBra1 and HyBra2 compared with XBra.(D) Comparison of exon-intron structure in Xbra, HyBra1 and HyBra2. Filled boxes mark the ORF, white boxes the UTR. Red boxes indicate conserved exon-intron sites, whereas black boxes indicate non-conserved regions. Black numbers represent the length of introns.

Alsterpaullone treatment

Alsterpaullone treatment was performed as described(Broun et al., 2005) with some minor modifications. Polyps were treated for 24 hours in 5 μM alsterpaullone (Sigma), then washed extensively and further incubated in HM for 15 hours and 40 hours, respectively. Specimens were then fixed and subjected to in situ hybridisation. DMSO at the corresponding concentration was used as a control with no effect.

Isolation and characterisation of two complete Brachyuryorthologues from Hydra

A partial HyBra1 clone has already been isolated(Technau and Bode, 1999). Using 3′ RACE we have now obtained the 1845-bp-long full-length clone with an open reading frame of 364 amino acids. By degenerate PCR and cDNA library screening, we isolated an additional full-length T-box gene. The clone consists of 1423 bp, encoding a protein of 388 amino acids(Fig. 1C). Phylogenetic analysis revealed that this gene also clearly belongs to the Brachyury subfamily of T-box genes and we therefore named it HyBra2 (Fig. 2). The T-domains of the two HyBra genes show about 70% identity(Fig. 1A). When compared with the T-domain of vertebrate Brachyury proteins, HyBra1 and HyBra2 show about 80% and 70% identities, respectively (Fig. 1A,C). By contrast, the remaining domains are far less conserved,with an overall amino acid identity of <20% (data not shown). Interestingly, two Brachyury protein domains, one positioned N-terminally of the T-box and the other one named R1, which is embedded within the activation domain, have been gained and lost during evolution(Marcellini et al., 2003; Kispert et al., 1995). Whereas the N-terminal domain is absent in non-bilaterian Brachyury proteins, and was therefore probably acquired at the base of the bilaterians(Marcellini, 2006), the ancient R1 fragment in the C-terminal domain is already present in cnidarians,where it is found in HyBra1 and Nematostella Brachyury(Scholz and Technau, 2003),but not in HyBra2. Strikingly, both domains have been lost in Tunicates and arthropods but are conserved in all other bilaterian paralogues(Marcellini, 2006).

Sequence comparison further revealed that HyBra2 carries a spliced leader (SL-B) at the 5′ end (Stover and Steele, 2001), whereas no spliced leader is present in the isolated HyBra1 cDNA (data not shown). Genome walking analysis of the publicly available shotgun sequences of the Hydra genome project(Craig J. Venter Institute, MD, USA) showed that the two HyBraparalogues are not closely linked in the genome. Interestingly, when comparing HyBra1 and Xbra, five out of six intron positions in the coding sequence are conserved, four of which reside in the T-box and are conserved throughout the animal kingdom. These intron sites are also conserved in HyBra2 except the last one linking the T-box to the C-terminal domain (Fig. 1D). In summary,the structural similarities in the T-box, the conserved R1 motif and the arrangement of introns indicate that the ancestral Brachyury gene structure of HyBra1 has been conserved throughout evolution and the HyBra2 paralogue is derivative in various aspects.

Fig. 2.

Phylogenetic analysis of Brachyury proteins on the basis of their conserved T-box domain. A group of Tbx6 genes was used as an outgroup for the Brachyury family. Analysis was performed with MrBayes 3.1.2 using the WAG substitution model. Sequences were obtained from GenBank. Aa, Aedes aegypti; Am, Amphioxus (Branchiostoma floridae); As, Halocynthia roretzi; Bf, Branchiostoma floridae; Cj, Clypeaster japonicus; Gg, Gallus gallus; Gm, Grillus maculatus; He, Hydractinia echinata; Hp, Hemicentrotus pulcherrimus; Hs, Homo Sapiens; Lf, Lampetra fluviatilis; Lv, Lytechinus variegatus; Ml, Mnemiopsis leidyi; Mm, Mus musculus; Nem, Nematostella vectensis;Ol, Oikopleura longicauda; Pat, Patella vulgata; Pc, Podocoryne carnea; Pd, Platynereis dumerilii; Pf, Ptychodera flava; Pl, Paracentrotus lividus; Pp, Pleurobrachia pileus; Sr, Scypha raphanus; Ta, Trichoplax adhaerens; Tc, Tribolium castaneum; X, Xenopus laevis; Zf, Zebrafish.

Fig. 2.

Phylogenetic analysis of Brachyury proteins on the basis of their conserved T-box domain. A group of Tbx6 genes was used as an outgroup for the Brachyury family. Analysis was performed with MrBayes 3.1.2 using the WAG substitution model. Sequences were obtained from GenBank. Aa, Aedes aegypti; Am, Amphioxus (Branchiostoma floridae); As, Halocynthia roretzi; Bf, Branchiostoma floridae; Cj, Clypeaster japonicus; Gg, Gallus gallus; Gm, Grillus maculatus; He, Hydractinia echinata; Hp, Hemicentrotus pulcherrimus; Hs, Homo Sapiens; Lf, Lampetra fluviatilis; Lv, Lytechinus variegatus; Ml, Mnemiopsis leidyi; Mm, Mus musculus; Nem, Nematostella vectensis;Ol, Oikopleura longicauda; Pat, Patella vulgata; Pc, Podocoryne carnea; Pd, Platynereis dumerilii; Pf, Ptychodera flava; Pl, Paracentrotus lividus; Pp, Pleurobrachia pileus; Sr, Scypha raphanus; Ta, Trichoplax adhaerens; Tc, Tribolium castaneum; X, Xenopus laevis; Zf, Zebrafish.

HyBra1 and HyBra2 are ancient lineage-specific paralogues

Since HyBra1 is more similar to vertebrate Brachyurygenes than to its cognate partner gene HyBra2, one possibility was that the duplication event occurred before the split of Cnidaria and Bilateria, with maintenance of one or the other (or both) among the Bilateria. So far, paralogous Brachyury genes have been reported for humans, Xenopus, chick and Amphioxus. To test the possibility that these paralogues are descendants of either HyBra1 or HyBra2,respectively, we compared all paralogous Brachyury proteins in a phylogenetic analysis. We also included three other identified Brachyury proteins from different Cnidaria and one from the sponge Suberites(Adell et al., 2003)(Fig. 2). In this more extensive analysis, the cnidarian Brachyury proteins are divided in two separate clusters. HyBra1 clusters with Brachyury homologues from Nematostella (Scholz and Technau,2003) and from a marine relative, the hydrozoan Hydractinia echinata, whereas HyBra2 forms a cluster with a Brachyury from the hydrozoan Podocoryne carnea(Spring et al., 2002). However, none of them were grouped with any of the paralogues from the chordates, nor did the bilaterian Brachyury proteins fall into two distinct(bra1 and bra2) subclasses, although lower Deuterostomes tended to cluster with HyBra2, whereas vertebrate sequences are all closer to HyBra1. However,both HyBra1 and HyBra2 are more similar to deuterostome than protostome Brachyury proteins, reflecting the divergence in the different lineages. Since the basal cnidarian Nematostella vectensis has only a Bra1-type gene (Scholz and Technau, 2003), this analysis is consistent with the view that HyBra1 and HyBra2 represent two ancient paralogues that evolved by a duplication event. Most likely, this event occurred early in cnidarian evolution, after the divergence of the hydrozoan lineage approximately 550 million years ago. It is also clear, that according to this scenario, HyBra2 has undergone a more rapid rate of sequence evolution.

Expression pattern of HyBra1 and HyBra2 in adult polyps and during bud formation

We previously showed that HyBra1 is expressed in the hypostome,the apical tip of Hydra (Technau and Bode, 1999). During various developmental conditions, such as head regeneration, bud formation (the asexual form of Hydrareproduction), the formation of hydra from aggregates of cells, and embryogenesis, it is expressed very early during head formation(Technau and Bode, 1999; Technau et al., 2000). In situ hybridisation showed that HyBra2 was expressed in the hypostome of Hydra polyps in a very similar pattern to that of HyBra1(Fig. 3A,C). However, vibratome sections of adult polyps revealed that HyBra2 was expressed primarily in the ectodermal layer whereas HyBra1 was expressed predominantly in the endoderm (Fig. 3B,D). A comparison of various bud stages showed that HyBra2 was expressed just as early as HyBra1 as a spot in the ectoderm at stage 1(Fig. 3E,H). In subsequent stages of bud formation, HyBra2, like HyBra1, remained restricted to the apical tip of the bud, which is the future hypostome(Fig. 3F,I). Early stages, as well as vibratome sections of later bud stages, confirmed that HyBra1was primarily expressed in the endoderm(Fig. 3G), whereas HyBra2 was predominantly expressed in the ectoderm(Fig. 3J). This suggests that HyBra2, like HyBra1, plays an early role in head formation during bud formation, but acts mainly in the other germ layer.

HyBra1 and HyBra2 expression is differently regulated during regeneration

We next analysed HyBra1 and HyBra2 expression during head regeneration in animals cut at apical and mid-gastric positions of the body column. As shown previously (Technau and Bode, 1999), HyBra1 appears in the regenerating head of animals bisected at the apical end of the body column within 1-3 hours of head removal (Fig. 4A; Fig. 5A-C). In sharp contrast, HyBra2 was first expressed at apical levels after roughly 8-10 hours(Fig. 4A; Fig. 5G-I). There was a substantial variability from animal to animal in the onset of HyBra2expression.

Interestingly, HyBra1 initially appears several hours later in animals bisected in the middle of the body column(Fig. 4B) compared with animals bisected at the apical end (Fig. 4A) (Technau and Bode,1999). This delay is consistent with the delay in head formation at this level and a lower level of the head activation gradient(Bode and Bode, 1984). By contrast, there was only a slight delay in the onset of HyBra2expression in regenerating heads following mid-column bisection compared with those bisected near the head (Fig. 4B). Furthermore, the apical half of these bisected animals transiently, but strongly expressed HyBra1 at the site of foot regeneration, whereas HyBra2 is expressed very weakly, and only in a few individuals (Fig. 4C; H.B. and U.T., unpublished). Hence, HyBra1 is among the earliest genes expressed during head regeneration, and is transiently upregulated during foot regeneration, whereas HyBra2 appears somewhat later and is expressed only where a head is being formed.

Fig. 3.

Expression of HyBra1 and HyBra2. HyBra1and HyBra2 mRNA expression was assessed in adult polyps (A-D)and during budding (E-J). (A,B) In adult hydras HyBra1 is predominantly expressed in the endodermal layer of the hypostome. (C,D) HyBra2 expression was primarily detected in the ectoderm of the hypostome. (B,D) Vibratome cross-sections of the hypostome. (E-J) Both genes are expressed very early during budding, but HyBra1 RNA is predominantly in the endoderm, whereas HyBra2 expression is restricted to the ectoderm (arrows in E and H). (J) Vibratome section of stage 5 bud. (K-M) Switch of HyBra2 expression from the endoderm to the ectoderm during head regeneration. Expression is in the endoderm at 18 hours (K), in both layers at 28 hours (L) and in the ectoderm by 48 hours (M). Arrows indicate the mesoglea, which separates ectoderm from endoderm.

Fig. 3.

Expression of HyBra1 and HyBra2. HyBra1and HyBra2 mRNA expression was assessed in adult polyps (A-D)and during budding (E-J). (A,B) In adult hydras HyBra1 is predominantly expressed in the endodermal layer of the hypostome. (C,D) HyBra2 expression was primarily detected in the ectoderm of the hypostome. (B,D) Vibratome cross-sections of the hypostome. (E-J) Both genes are expressed very early during budding, but HyBra1 RNA is predominantly in the endoderm, whereas HyBra2 expression is restricted to the ectoderm (arrows in E and H). (J) Vibratome section of stage 5 bud. (K-M) Switch of HyBra2 expression from the endoderm to the ectoderm during head regeneration. Expression is in the endoderm at 18 hours (K), in both layers at 28 hours (L) and in the ectoderm by 48 hours (M). Arrows indicate the mesoglea, which separates ectoderm from endoderm.

Fig. 4.

Kinetics of HyBra1 and HyBra2 expression during regeneration. Expression of HyBra1 (dashed line) and HyBra2 (solid line) during head regeneration following bisection at(A) the upper end of the body column, and (B) in the middle of the body column. (C) Expression of the two genes during foot regeneration following bisection in the mid-body column. Data points are means(± s.d.) of 40-80 animals resulting from 2-5 independent experiments. HyBra1 data modified after Technau and Bode(Technau and Bode, 1999).

Fig. 4.

Kinetics of HyBra1 and HyBra2 expression during regeneration. Expression of HyBra1 (dashed line) and HyBra2 (solid line) during head regeneration following bisection at(A) the upper end of the body column, and (B) in the middle of the body column. (C) Expression of the two genes during foot regeneration following bisection in the mid-body column. Data points are means(± s.d.) of 40-80 animals resulting from 2-5 independent experiments. HyBra1 data modified after Technau and Bode(Technau and Bode, 1999).

Another striking difference between the two genes occurs in the tissue layer in which they are expressed during head regeneration. At about 10 hours after bisection, the initial expression of HyBra2 occurred in the endodermal layer (Fig. 3K). Ectodermal expression first became detectable at the most apical tip after about 24 hours, and then slowly spread over the hypostome(Fig. 3L,M). Endodermal expression persisted for 2-3 days of regeneration and eventually vanished. This shows that HyBra2 expression is differently regulated during head regeneration compared with expression in budding and adult HyBra2 expression. It indicates that the two processes, although sharing many common features, differ at the molecular level.

HyBra1 is an immediate-early response gene during head regeneration

HyBra1 is one of the earliest genes expressed during head regeneration and bud formation (Technau and Bode, 1999). This raised the question whether initiation of HyBra1 expression depends on the synthesis of another factor to activate its expression. To address this question we analysed the expression of HyBra1 and HyBra2 during regeneration in the presence of the translational inhibitor cycloheximide (CHX)(Cascio and Gurdon, 1987). We first established that a treatment with 5-50 μg/ml CHX for up to 2 hours leads to slight but reversible phenotypic effects. Under these conditions, the reappearance of tentacles after head removal is delayed by 12-18 hours (data not shown). Treatment for up to 6 hours was also tolerated by the animals, but had more severe effects. We then measured the inhibition of translation by incorporation of [35S]methionine. We found that after treatment with either 5 or 50 μg/ml CHX for 15 minutes, translation was inhibited by 80-90%. Thereafter, the rate of protein synthesis slowly recovered over the next 2 days (data not shown). Treatment for 2 hours resulted in the 80-90%level of inhibition being maintained for 6 hours, followed by a slow recovery(data not shown).

Fig. 5.

Expression of HyBra1 and HyBra2 during regeneration in the presence or absence of cycloheximide. Animals were pretreated for 15 minutes in 5 μg/ml cycloheximide, decapitated and allowed to regenerate for the times indicated. The first two hours of regeneration were carried out in the presence of the cycloheximide. (A-C) HyBra1 expression under control conditions. (D-F) The onset of HyBra1 is not affected by cycloheximide. HyBra2 expression in cycloheximide-treated animals (J-L) starts later than in control animals (G-I).

Fig. 5.

Expression of HyBra1 and HyBra2 during regeneration in the presence or absence of cycloheximide. Animals were pretreated for 15 minutes in 5 μg/ml cycloheximide, decapitated and allowed to regenerate for the times indicated. The first two hours of regeneration were carried out in the presence of the cycloheximide. (A-C) HyBra1 expression under control conditions. (D-F) The onset of HyBra1 is not affected by cycloheximide. HyBra2 expression in cycloheximide-treated animals (J-L) starts later than in control animals (G-I).

Fig. 6.

Ectopic expression of HyBra1 and HyBra2 in response to alsterpaullone treatment. Animals were exposed to 5 μM alsterpaullone(AP) for 24 hours and subsequently transferred to Hydra media (HM) for 15 hours and 40 hours respectively. HyBra1 (A-D), HyBra2(E-H) and Wnt (I-L) are affected by the GSK-3 inhibitor alsterpaullone. Ectopic HyBra1 and HyBra2 expression starts 24 hours after AP treatment in single cells (B,F, high-magnification inset)whereas normal hypostome expression declines rapidly and is shifted to the tentacle bases (compare A,B with E,F); elevated levels of expression are seen over the uniform endodermal expression (compare A,E with B,F). In later stages, the single cell expression in the body column begins to cluster(arrows in C and G) and subsequently peaks in a stochastic pattern of spots of 10-15 epithelial cells (arrows in G and H). HyWnt hypostome expression is lost 15 hours after completion of AP treatment and clusters along the body column in ectopic expression domains are narrower than those of HyBra1 and HyBra2.

Fig. 6.

Ectopic expression of HyBra1 and HyBra2 in response to alsterpaullone treatment. Animals were exposed to 5 μM alsterpaullone(AP) for 24 hours and subsequently transferred to Hydra media (HM) for 15 hours and 40 hours respectively. HyBra1 (A-D), HyBra2(E-H) and Wnt (I-L) are affected by the GSK-3 inhibitor alsterpaullone. Ectopic HyBra1 and HyBra2 expression starts 24 hours after AP treatment in single cells (B,F, high-magnification inset)whereas normal hypostome expression declines rapidly and is shifted to the tentacle bases (compare A,B with E,F); elevated levels of expression are seen over the uniform endodermal expression (compare A,E with B,F). In later stages, the single cell expression in the body column begins to cluster(arrows in C and G) and subsequently peaks in a stochastic pattern of spots of 10-15 epithelial cells (arrows in G and H). HyWnt hypostome expression is lost 15 hours after completion of AP treatment and clusters along the body column in ectopic expression domains are narrower than those of HyBra1 and HyBra2.

We then performed a regeneration experiment in the presence of CHX. After a pre-treatment of 15 minutes, polyps were decapitated at apical levels and allowed to regenerate in the presence of CHX for 2 hours. Thereafter, they were allowed to regenerate in Hydra medium. Periodically, after end of the CHX treatment, polyps were fixed and analysed for HyBra1 and HyBra2 expression by in situ hybridisation. We found that HyBra1 was expressed as early and strongly as in untreated controls(2-3 hours; Fig. 5A-F). By contrast, the onset of HyBra2 expression was delayed compared with the controls by about 8-10 hours (Fig. 5G-L). This shows that transcriptional activation of HyBra2, but not of HyBra1 requires protein synthesis during head regeneration.

One possible candidate protein to regulate HyBra1 expression isβ-catenin (Hobmayer et al.,2000). It is expressed ubiquitously, transcriptionally upregulated during regeneration and stabilised in the nucleus at the apical end of Hydra (Hobmayer et al.,2000; Broun et al.,2005). Indeed, we obtained support for such a role ofβ-catenin by treatment with the Glycogen synthase kinase 3 β(GSK3β) inhibitor alsterpaullone(Broun et al., 2005). Inhibition of GSK3β results in stabilisation and nuclear translocation ofβ-catenin, where it can act, in concert with T-cell factor (TCF), as a transcriptional regulator (Leost et al.,2000; Bain et al., 2003). Treatment with alsterpaullone for 24 hours followed by incubation in Hydra medium for 15-40 hours leads to an ectopic upregulation of both HyBra1 and HyWnt genes in the body column (Fig. 6),confirming earlier results (Broun et al.,2005). Like HyBra1, HyBra2 is also upregulated broadly in response to alsterpaullone treatment (Fig. 6E-H). HyWnt and both HyBra genes are downregulated at the same time in the hypostome(Fig. 6; data not shown). Interestingly, 15 hours after treatment ectopic upregulation of HyBra1 and HyBra2, but not of HyWnt, in the body column is detectable in single cells, which appear to be interstitial cells(Fig. 6B,C,F,G). After 40 hours of alsterpaullone treatment, both HyBra1 and HyBra2 are strongly upregulated throughout the endoderm of the gastric cavity, with peaks of expression of regular spacing consisting of 10-15 epithelial cells also visible in the ectoderm (Fig. 6D,H). Ectopic HyWnt expression is found in similar spots but does not expand to ubiquitous upregulation in the endoderm(Fig. 6L)(Broun et al., 2005). This suggests that HyBra1 and HyBra2 are direct or indirect targets of the Wnt-β-catenin pathway.

Structure-function relationship in animal cap assays for mesoderm induction

We were interested whether the high degree of sequence divergence between the two HyBra paralogues reflects two different functions. To test this, we used a well-defined in vivo system, the Xenopus animal cap. Untreated animal caps differentiate into epidermal tissue, but can be forced to form mesoderm if injected at the two- to four-cell stage with Xenopus bra mRNA (Smith et al.,1991). Injection of mRNAs encoding other T-box genes such as VegT or eomesodermin, induces endoderm in addition to mesoderm, indicative of the endogenous functions of these proteins(Marcellini et al., 2003; Conlon et al., 2001; Zhang and King, 1996; Zhang et al., 1998). Other factors, such as otx2 and ath-3 have been shown to induce differentiation of neural tissue in animal caps(Gammill and Sive, 1997; Takebayashi et al., 1997). Hence, animal caps can display a variety of very distinct and specific differentiation fates in response to an exogenous transcription factor provided by mRNA microinjection. Notably, Brachyury mRNAs from a wide range of bilaterian species induce mesoderm under these assay conditions(Marcellini et al., 2003). Thus, this assay can be used to rapidly identify functionally crucial domains or motifs generated by sequence divergence during animal evolution.

Fig. 7.

Distinct morphological phenotypes induced in Xenopus animal caps by Hydra Brachyury proteins. (A) Schematic view of the microinjected constructs (AD, activation domain; C-T, C-terminal domain).(B) Sketch of the animal cap experiment (st, stage). (C-E)Control specimens: (C) wt sibling embryo at NF36, (D) uninjected and (E) GFP-injected animal caps with default epidermal morphology. Animal caps injected with Xbra (F), HyBra1 (G) or H2TXA (I) with elongated morphology, indicating mesoderm formation. (H) HyBra2-injected animal cap with two darkly pigmented cement gland areas. (J) XTH2A-injected explant with single cement gland area. (K) Statistical overview of morphological characteristics [uninduced, elongated and cement gland celltype-specific (cg cells)] induced by different Brachyury paralogues (3-10 independent injections). RNA doses were titrated from 0.1 to 1.0 ng/embryo, resulting in consistent, qualitatively different phenotypes represented by the displayed specimen.

Fig. 7.

Distinct morphological phenotypes induced in Xenopus animal caps by Hydra Brachyury proteins. (A) Schematic view of the microinjected constructs (AD, activation domain; C-T, C-terminal domain).(B) Sketch of the animal cap experiment (st, stage). (C-E)Control specimens: (C) wt sibling embryo at NF36, (D) uninjected and (E) GFP-injected animal caps with default epidermal morphology. Animal caps injected with Xbra (F), HyBra1 (G) or H2TXA (I) with elongated morphology, indicating mesoderm formation. (H) HyBra2-injected animal cap with two darkly pigmented cement gland areas. (J) XTH2A-injected explant with single cement gland area. (K) Statistical overview of morphological characteristics [uninduced, elongated and cement gland celltype-specific (cg cells)] induced by different Brachyury paralogues (3-10 independent injections). RNA doses were titrated from 0.1 to 1.0 ng/embryo, resulting in consistent, qualitatively different phenotypes represented by the displayed specimen.

We first injected Xbra, Hybra1 and Hybra2 mRNA into four-cell-stage embryos, dissected animal caps and cultured them until the hatching stage, when we scored morphological(Fig. 7) and molecular phenotypes (Fig. 8). Uninjected or GFP-preloaded explants were round and lacked any sign of non-ectodermal differentiation (Fig. 7D,E; Fig. 8A,B). The large majority of HyBra1- or Xbra-injected caps were elongated (Fig. 7G,F) and expressed the striated muscle marker cardiac actin(Fig. 8C,E). Under these conditions, Xbra and HyBra1 demonstrate comparable mesoderm inducing activities. By contrast, HyBra2 never induced elongation or mesodermal marker gene expression, even when very high amounts of mRNA were injected, suggesting that HyBra2 lacks a common mesoderm-inducing activity (Fig. 7K; Fig. 8M).

Intriguingly, it turned out that HyBra2 is a highly potent cement gland inducer, even when very low RNA concentrations were injected(Fig. 7H,K; Fig. 8G,M). The Xenopus cement gland, which is temporarily formed during the tadpole stage, consists of palisade-like epithelial cells. These cells have a distinct morphological structure and are therefore easy to identify. Of all Hybra2-injected caps, 86% contained one or several cement gland-like areas characterised by their regular arrangement of strongly pigmented cells(Fig. 7H,K). Additional evidence that these structures are cement glands comes from two observations. First, like their endogenous counterparts at the anterior end of sibling embryos (Fig. 7C), their surface was highly adhesive, indicating mucus secretion (data not shown). Second, 84% of Hybra2-injected caps expressed Xcg-1(Sive et al., 1989), a cement-gland-specific marker (Fig. 8G,M). Since there is evidence that cement glands, although epithelial in nature, develop as a consequence of neural induction(Sasai and De Robertis, 1997),we analysed the animal caps for expression of the pan-neural marker N-CAM (see Fig. S1 in the supplementary material) andβ -tubulin, a marker for differentiated neurons(Fig. 8). Although these neural mRNAs were hardly detectable in Hybra1- and Xbra-injected caps, both markers were strongly expressed in a significant fraction of the Hybra2-injected explant (Fig. 8F,H,M and see Fig. S1 in the supplementary material). The absence of muscle and endoderm suggests that this induction results from the direct conversion of ectoderm into neural tissue in response to HyBra2. Together, the Xenopus animal cap experiments reveal that the activities of HyBra1 and vertebrate Xbra are very similar,but that HyBra2 has significantly diverged and acquired cement gland and neural-inducing properties.

This unexpectedly distinct, qualitative difference in activity motivated us to look into the domains of the proteins in more detail. We therefore constructed two chimeras, swapping the T-Box and C-terminal domains of Xbra and HyBra2, as shown in Fig. 7A. Surprisingly, animal caps injected with the H2TXAchimera contained not a single cement gland, whereas 40% of them were elongated (Fig. 7I,K). Although the latter suggests the induction of mesoderm, muscle actin was only rarely induced (Fig. 8I,M). Moreover,no Xcg-1 or β-tubulin expression was detected(Fig. 8I,J,M). The converse fusion, XTH2A, induced cement glands and neuralβ -tubulin approximately to the same extent as HyBra2,but neither induced elongation or muscle actin expression(Fig. 7J,K; Fig. 8K,L,M). Thus, the XTH2A variant phenocopies the activity of HyBra2 and is unable to induce mesoderm.

Fig. 8.

Molecular analysis of the differentiated animal cap phenotypes induced by HyBra proteins. Marker analysis by double(A,C,E,G,I-L) or single(B,D,F,H) RNA in situ hybridisation; probes are listed within panels, black letters indicate BM-Purple stain, red letters indicate Fast-Red stain (n≥3 independent repeats). (A,B) GFP-injected control explants showed no staining of tested marker genes. Injections of Xbra (C) and HyBra1 (E) induced muscle actin, but not xcg-1 expression, whereas HyBra2 (G) triggers the converse gene expression pattern. Furthermore, mRNA of the differentiated neuronal markerβ -tubulin was often present in HyBra2- (H), but never in HyBra1- (F), and only rarely in Xbra-injected explants(D). (I-L) Double in situ staining of animal caps injected with the Hybra2/Xbra chimerae. (I,J) H2TXA-injected animal caps show no xcg-1, β-tubulin induction and very limited muscle actin expression. (K,L) By contrast, XTH2A-injected caps clearly show strong induction of xcg-1 and β-tubulin, but no muscle actin mRNA. (M) Statistical overview of induced markers(xcg-1, muscle actin, β-tubulin) by injection of particular mRNAs (2-6 independent single and/or double in situ hybridisation experiments).

Fig. 8.

Molecular analysis of the differentiated animal cap phenotypes induced by HyBra proteins. Marker analysis by double(A,C,E,G,I-L) or single(B,D,F,H) RNA in situ hybridisation; probes are listed within panels, black letters indicate BM-Purple stain, red letters indicate Fast-Red stain (n≥3 independent repeats). (A,B) GFP-injected control explants showed no staining of tested marker genes. Injections of Xbra (C) and HyBra1 (E) induced muscle actin, but not xcg-1 expression, whereas HyBra2 (G) triggers the converse gene expression pattern. Furthermore, mRNA of the differentiated neuronal markerβ -tubulin was often present in HyBra2- (H), but never in HyBra1- (F), and only rarely in Xbra-injected explants(D). (I-L) Double in situ staining of animal caps injected with the Hybra2/Xbra chimerae. (I,J) H2TXA-injected animal caps show no xcg-1, β-tubulin induction and very limited muscle actin expression. (K,L) By contrast, XTH2A-injected caps clearly show strong induction of xcg-1 and β-tubulin, but no muscle actin mRNA. (M) Statistical overview of induced markers(xcg-1, muscle actin, β-tubulin) by injection of particular mRNAs (2-6 independent single and/or double in situ hybridisation experiments).

The results described above suggest that it is primarily the part located C-terminally of the T-box that accounts for the qualitatively distinct protein activities of HyBra1 and HyBra2, rather than the divergence of the T-Box domains. However, the minuscule mesoderm-forming activity contained within the H2TXA chimera indicates that the two domains cannot be freely exchanged, but also have to cooperate to a certain extent. From this we conclude that the specific activity, which in the Xenopus animal cap assay induces cement gland and neural cell fates, is confined to the C-terminal domain of HyBra2 and probably involves more than constitutive transcriptional activation or repression.

The diploblastic metazoan hydra has two ancient paralogues of Brachyury

Among non-bilaterian animals, the first T-box gene described was the Brachyury homologue HyBra1 of the diplobast Hydra(Technau and Bode, 1999). We now report the identification of a second Brachyury gene from this basal metazoan, termed HyBra2. The phylogenetic analysis strongly suggests that the duplication event occurred at the base of the Hydrozoan lineage, at least 550 million years ago(Chen et al., 2002). Hydrozoa have retained two Brachyury genes, Bra1 and Bra2. Whether the duplicated Brachyury genes in Hydractinia(Bra2-type) and Podocoryne (Bra1-type) have been preserved as in Hydra or have been lost is not yet clear. Our sequence analysis further suggests that bilaterian, in particular all vertebrate Brachyury proteins, share more features with HyBra1 than with HyBra2. This suggests that a HyBra1-like molecule was present in the common ancestor of cnidarians and bilaterians and that it has retained more of its ancestral structural features than HyBra2. Interestingly, HyBra1 shares a small conserved motif in the activation domain with the vertebrate Brachyurys, corresponding to the core of the repression module 1 (R1) as described previously(Kispert et al., 1995). The conservation of this motif suggests that it might play a role as a conserved interface for protein-protein interaction.

The role of Brachyury during head formation in Hydra

HyBra1 and HyBra2 are expressed in endoderm and ectoderm,respectively, of the hypostome, which is considered equivalent to the Spemann organiser (Browne, 1909; Broun and Bode, 2002). Interestingly, although HyBra1 and HyBra2 mRNA expression is activated at the same stage during budding, the kinetics of expression of HyBra1 and HyBra2 differs significantly during regeneration: HyBra1 is an immediate-early gene, in the sense that its onset of expression does not require protein synthesis, whereas HyBra2expression does require protein synthesis and occurs significantly later. This points to some crucial differences in the molecular regulatory network between regeneration (for reviews, see Holstein et al., 2003; Galliot et al.,2006; Bosch, 2007)and budding, despite the fact that overall the same genes are used. Head formation is thought to be suppressed in the body column of a normal polyp by an inhibitor that forms a concentration gradient from the head to the budding region (MacWilliams, 1983). While the nature of this postulated inhibitor is still unclear, we propose that decapitation rapidly removes the head inhibitor from the tissue, which leads to a local post-translational modification of a transcriptional regulator present in the tissue that activates HyBra1 (and other immediate-early response genes). Members of the MAPK and CREB pathway are implicated in early head regeneration and possibly involved in the regulation of HyBra1 (Cardenas et al.,2000; Cardenas and Salgado,2003; Manuel et al.,2006; Kaloulis et al.,2004). Another candidate molecule to activate HyBra1expression upon post-translational modification is β-catenin.β -catenin is expressed throughout the animal, but upregulated at the apical tip during regeneration(Hobmayer et al., 2000) and stabilised in the apical region in nuclei(Broun et al., 2005). In accordance with a role in regulating HyBra expression, HyBra1 and HyBra2 expression is ectopically upregulated in animals where β-catenin has been ectopically stabilised by alsterpaullone treatment.

Hence, inhibition of early head-specific genes in the body column might be lifted in response to regeneration signals, preparing the tissue for rapid regeneration, irrespective of its future fate. By contrast, HyBra2does not belong to the immediate-early-response genes during head regeneration and is not detectable during foot regeneration. After the immediate-early response, additional factors, which stabilise the polarity of the tissue and tell the regenerating end whether to differentiate a head or a foot, must be involved.

The evolutionary divergence of the Hydra Brachyuryparalogues occurred at diverse levels

It has been repeatedly proposed that divergence and evolution of cis-regulatory regions may drive such increasing complexity as observed in the lineage of the vertebrates (Levine and Tijan, 2003; Davidson and Erwin, 2006). The duplication-degeneration-complementation (DDC)model provides a theoretical framework for the fate of paralogues upon duplication (Force et al.,1999; Lynch and Force,2000). Essentially, paralogues that are retained over long evolutionary time scales underwent either subfunctionalisation or neofunctionalisation. Subfunctionalisation mainly emphasises the divergence and subdivision of the cis-regulatory elements of the parental gene. Indeed,there are cases where the gene expression domains of both duplicates together make up for the expression domain of a single orthologue in a species, which has branched off before the duplication event(De Martino et al., 2000). In the case of HyBra1 and HyBra2 the acquired cis-regulatory elements resulted in a paralogue-specific spatial and temporal regulation of the two Hydra Brachyury genes. If the combined expression of both paralogues reflects the expression of the ancestral gene, this would mean that Brachyury is marking a domain in Hydra, rather than a particular germ layer.

In addition to subfunctionalisation, divergence of coding sequences in paralogues can lead to neofunctionalisation. In the case of ancient duplications, the paralogues may have diverged significantly and possibly taken up novel functions. Therefore, ancient lineage-specific paralogues in particular provide the opportunity to study the levels and target sites of the diverging mutations. They also provide us with evolutionary variants of proteins, which can then be studied for structure-function relationships and help to rapidly identify crucial motifs for specific subfunctions. In the absence of functional assays in Hydra with specific read-outs for distinct cellular fates, we used a heterologous system - the Xenopusanimal caps - as an assay system to test whether both Hydraparalogues have functionally diverged. This assay system is ideal, because this naive tissue, which otherwise develops normally into undifferentiated epidermis, but can be pushed to differentiate endoderm, mesoderm or neural ectoderm, depending on the input. It is well established that Brachyury induces mesoderm, both on the molecular level and in terms of tissue properties such as convergent extension(Smith et al., 1991). Although Hydra does not contain mesoderm, HyBra1 is able to induce mesoderm just like the Xenopus Brachyury gene(Marcellini et al., 2003)(this study). In contrast, HyBra2 behaves differently to other bilaterian Brachyury genes in this assay by strongly inducing formation of cement glands and neural tissue. Domain-swapping experiments demonstrate that the HyBra2 C-terminal domain is necessary and sufficient(when fused to a Brachyury T-box) to induce neural tissue formation. Interestingly, Hybra2 is not the only Brachyury protein with neuralising activity. The Xbra paralogue Xbra3 activates both mesodermal and neural differentiation (Strong et al.,2000). In animal caps, however, it induced only posterior neural tissue and cement glands were not observed (Hartmann et al., 2002). It is therefore not clear, whether this is a secondary effect of the neuralising effect of embryonic FGF (eFGF), which forms a feedback loop with Brachyury(Schulte-Merker and Smith,1995). In addition, specific truncations of the Xbraactivation domain can convert the mesoderm inducer into a neuralising factor(Rao, 1994). However,C-terminal truncations of HyBra2 either did not alter the neuralising phenotype or had no effect at all (data not shown). Further, the HyBra2 C-terminal domain is even longer than that of HyBra1, which is a mesoderm inducer. We conclude that the similar phenotypes obtained with the HyBra2 and with truncated version of Xbra in the study by Rao(Rao, 1994) must be of different origin, yet might reveal a common mechanism.

Brachyury has previously been shown to be a transcriptional activator, so one possibility to explain the radical difference in the inductive behaviour of HyBra2 in animal caps is that the C-terminal activation domain evolved into a repression domain. However, this does not seem to be the case, as an H2T-EnR variant (HyBra2 T-box fused to the Engrailed repressor domain) did not mimic the neuralising activity of HyBra2 (data not shown). Furthermore, repression of Brachyury function by injection of a comparable XbraT-EnR construct caused mesodermal defects in convergent extension and notochord formation and induction of anterior endoderm (Conlon et al., 1996; Conlon and Smith,1999); however, the induction of cement glands had not been reported in these studies. Therefore, it is unlikely that Hybra2 simply represents an antimorphic variant of mesoderm-inducing Brachyury proteins. Future studies aimed to identify proteins interacting with variants of Brachyury might be key for our understanding of how transcriptional activity and specificity is modulated.

We acknowledge the publicly available genome sequence of Hydra magnipappillata by the Craig J. Venter Institute with funding by the NHGRI, which helped to determine the exon-intron structures of HyBragenes. We thank Jim Smith for sending a Xbra clone. We thank Fabian Rentzsch and Rein Aasland for critically reading the ms. This work was funded by a grant of the DFG (Te311-1/3) and the Sars core budget to U.T., by the National Science Foundation (IBN-0130375 and IBN-9904757) to H.R.B. and the DGF grant to R.R. (SFB646/A6).

Adell, T., Grebenjuk, V. A., Wiens, M. and Müller, W. E. (
2003
). Isolation and characterization of two T-box genes from sponges, the phylogenetically oldest metazoan taxon.
Dev. Genes Evol.
213
,
421
-434.
Arendt, D., Technau, U. and Wittbrodt, J.(
2001
). Evolution of the bilaterian larval foregut.
Nature
409
,
81
-85.
Bain, J., McLauchlan, H., Elliot, M. and Cohen, P.(
2000
). The specificities of protein kinase inhibitors: an update.
Biochem. J.
371
,
199
-204.
Bode, P. M. and Bode, H. R. (
1984
). Patterning in Hydra. In
Pattern Formation, a Primer in Developmental Biology
(ed. G. M. Malacinski and S. V. Bryant), pp.
213
-241. New York: MacMillan.
Bosch, T. C. G. (
2007
). Why polyps regenerate and we don't: towards a cellular and molecular framework for Hydraregeneration.
Dev. Biol.
15
,
421
-433.
Broun, M. and Bode, H. R. (
2002
). Characterization of the head organizer in
Hydra. Development
129
,
875
-884.
Broun, M., Gee, L., Reinhardt, B. and Bode, H. R.(
2005
). Formation of the head organizer in hydra involves the canonical Wnt pathway.
Development
132
,
2907
-2916.
Browne, E. N. (
1909
). The production of new Hydrants in Hydra by the insertion of small grafts.
J. Exp. Zool
.
8
,
1
-23.
Cardenas, M. M. and Salgado, L. M. (
2003
). STK,the src homologue, is responsible for the initial commitment to develop head structures in Hydra.
Dev. Biol.
264
,
495
-505.
Cardenas, M. M., Fabila, Y. V., Yum, S., Cerbon, J., Bohmer, F. D., Wtzker, R., Fujisawa, T., Bosch, T. C. G. and Salgado, L. M.(
2000
). Selective protein kinase inhibitors block head-specific differentiation in hydra.
Cell. Signal.
12
,
846
-857.
Cascio, S. and Gurdon, J. B. (
1987
). The initiation of new gene transcription during Xenopus gastrulation requires immediately preceding protein synthesis.
Development
100
,
297
-305.
Chen, J.-Y., Oliveri, P., Gao, F., Dornbos, S. Q., Li, C.-W.,Bottjer, D. J. and Davidson, E. H. (
2002
). Precambrian animal life: probable developmental and adult cnidarian forms from southwest China.
Dev. Biol.
248
,
182
-196.
Chourrout, D., Delsuc, F., Chourrout, P., Edvardsen, R. B.,Rentzsch, F., Renfer, E., Jensen, M. F., Zhu, B., de Jong, P., Steele, R. E. et al. (
2006
). Minimal ProtoHox cluster inferred from bilaterian and cnidarian Hox complements.
Nature
442
,
684
-687.
Conlon, F. L. and Smith, J. C. (
1999
). Interference with brachyury function inhibits convergent extension,causes apoptosis, and reveals separate requirements in the FGF and Activin signalling pathways.
Dev. Biol.
213
,
85
-100.
Conlon, F. L., Sedgwick, S. G., Weston, K. M. and Smith, J. C. (
1996
). Inhibition of Xbra transcription activation causes defects in mesodermal patterning and reveals autoregulation of Xbra in dorsal mesoderm.
Development
122
,
2427
-2435.
Conlon, F. L., Fairclough, L., Price, B. M., Casey, E. S. and Smith, J. C. (
2001
). Determinants of T box protein specificity.
Development
128
,
3749
-3758.
Cooke, J., Nowak, M. A., Boerlijst, M. and Maynard-Smith, J.(
1997
). Evolutionary origins and maintenance of redundant gene expression during metazoan development.
Trends Genet.
13
,
360
-364.
Davidson, E. H. and Erwin, D. H. (
2006
). Gene regulatory networks and the evolution of animal body plans.
Science
311
,
796
-800.
De Martino, S., Yan, Y. L., Jowett, T., Postlethwait, J. H.,Varga, Z. M., Ashworth, A. and Austin, C. A. (
2000
). Expression of sox11 gene duplicates in zebrafish suggests the reciprocal loss of ancestral gene expression patterns in development.
Dev. Dyn.
217
,
279
-292.
Force, A., Lynch, M., Picket, F. B., Amores, A., Yan, Y. and Postlethwait, J. (
1999
). Preservation of duplicate genes by complementary, degenerative mutations.
Genetics
151
,
1531
-1545.
Frohman, M. A., Dush, M. K. and Martin, G. R.(
1988
). Rapid production of full-length cDNAs from rare transcripts: amplification using a single gene-specific oligonucleotide primer.
Proc. Natl. Acad. Sci. USA
85
,
8998
-9002.
Galliot, B., Miljkovic-Licina, M., de Rosa, R. and Chera, S.(
2006
). Hydra, a niche for cell and developmental plasticity.
Semin. Cell Dev. Biol.
17
,
492
-502.
Gammill, L. S. and Sive, H. (
1997
). Identification of otx2 target genes and restrictions in ectodermal competence during Xenopus cement gland formation.
Development
124
,
471
-481.
Grens, A., Mason, E., Marsh, J. L. and Bode, H. R.(
1995
). Evolutionary conservation of a cell fate specification gene: the Hydra achaete-scute homolog has proneural activity in Drosophila.
Development
121
,
4027
-4035.
Gross, J. M. and McClay, D. R. (
2001
). The role of Brachyury (T) during gastrulation movements in the sea urchin Lytechinus variegatus.
Dev. Biol
.
239
,
132
-147.
Hartman, D., Haldin, C. E., Stott, D. and Jones, E. A.(
2002
). Xbra3 elicits the production of neural tissue by a non-BMP-dependent mechanism in Xenopus sp.
Mech. Dev.
118
,
65
-75.
Hayata, T., Eisaki, A., Kuroda, H. and Asashima, M.(
1999
). Expression of Brachyury-like T-box transcription factor, Xbra3 in Xenopus embryo.
Dev. Genes Evol.
209
,
560
-563.
Herrmann, B. G., Labeit, S., Poustka, A., King, T. R. and Lehrach, H. (
1990
). Cloning of the T gene required in mesoderm formation in the mouse.
Nature
343
,
617
-622.
Hobmayer, B., Rentzsch, F., Kuhn, K., Happel, C. M., Cramer von Laue, C., Snyder, P., Rothbächer, U. and Holstein, T. W.(
2000
). WNT signalling molecules act in axis formation in the diploblastic metazoan
Hydra. Nature
407
,
186
-189.
Holstein, T. W., Hobmayer, E. and Technau, U.(
2003
). Cnidarians: an evolutionarily conserved model system for regeneration?
Dev. Dyn.
226
,
257
-267.
Huelsenbeck, J. P. and Ronquist, F. (
2001
). MRBAYES: Bayesian inference of phylogenetic trees.
Bioinformatics
17
,
754
-755.
Kaloulis, K., Chera, S., Hassel, M., Gauchat, D. and Galliot,B. (
2004
). Reactivation of developmental programs: the cAMP-response element-binding protein pathway is involved in hydra head regeneration.
Proc. Natl. Acad. Sci. USA
101
,
2363
-2368.
Kispert, A., Herrmann, B. G., Leptin, M. and Reuter, R.(
1994
). Homologs of the mouse Brachyury gene are involved in the specification of posterior terminal structures in Drosophila, Tribolium and Locusta.
Genes Dev
.
8
,
2137
-2150.
Kispert, A., Koschorz, B. and Herrmann, B. G.(
1995
). The T protein encoded by Brachyury is tissue-specific transcription factor.
EMBO J
.
14
,
4763
-4772.
Knezevic, V., DeSanto, R. and Mackem, S.(
1997
). Two novel chick T-box genes related to mouse Brachyury are expressed in different, non-overlapping mesodermal domains during gastrulation.
Development
124
,
411
-419.
Kondrashov, F. A. and Koonin, E. V. (
2004
). A common framework for understanding the origin of genetic dominance and evolutionary fages of gene duplications.
Trends Genet.
20
,
287
-291.
Kortschak, R. D., Samuel, G., Saint, R. and Miller, D. J.(
2003
). EST analysis of the cnidarian Acropora milleporareveals extensive gene loss and rapid sequence divergence in the model invertebrates.
Curr. Biol.
14
,
R106
-R108.
Krakauer, D. C. and Nowak, M. A. (
1999
). Evolutionary preservation of redundant duplicated genes.
Semin. Cell Dev. Biol.
10
,
555
-559.
Kusserow, A., Pang, K., Sturm, C., Hrouda, M., Lentfer, J.,Schmidt, H. A., Technau, U., von Haeseler, A., Hobmayer, B., Martindale, M. Q. et al. (
2001
). Unexpected complexity of the Wnt gene family in a sea anemone.
Nature
433
,
156
-160.
Lartillot, N., Lespinet, O., Vervoort, M. and Adoutte, A.(
2002
). Expression pattern of Brachyury in the mollusc Patella vulgata suggests a conserved role in the establishment of the AP axis in Bilateria.
Development
129
,
1411
-1421.
Leost, M., Schultz, C., Link, A., Wu, Y. Z., Biernat, J.,Mandelkow, E. M., Bibb, J. A., Snyder, G. L., Greengard, P., Zaharevitz, D. W. et al. (
2000
). Paullones are potent inhibitors of glycogen synthase kinase-3beta and cyclin-dependent kinase 5/p25.
Eur. J. Biochem.
267
,
5983
-5994.
Levine, M. and Tjian, R. (
2003
). Transcription regulation and animal diversity.
Nature
424
,
147
-151.
Lynch, M. and Force, A. (
2000
). The probability of duplicate gene preservation by subfunctionalization.
Genetics
154
,
459
-473.
MacWilliams, H. K. (
1983
). Hydratransplantation phenomena and the mechanism of Hydra head regeneration. I. Properties of the head inhibition.
Dev. Biol
.
96
,
217
-238.
Manuel, G. C., Reynoso, R., Gee, L., Salgado, L. M. and Bode, H. R. (
2006
). PI3K and ERK-1-2 regulate early stages during head regeneration in hydra.
Dev. Growth Differ.
48
,
352
-361.
Manuel, M., Le Parco, Y. and Borchiellini, C.(
2004
). Comparative analysis of Brachyury T-domains, with the characterization of two new sponge sequences, from a hexactinellid and a calcisponge.
Gene
340
,
291
-301.
Marcellini, S. (
2006
). When Brachyury meets Smad1: the evolution of bilateral symmetry during gastrulation.
BioEssays
28
,
413
-420.
Marcellini, S., Technau, U., Smith, J. C. and Lemaire, P.(
2003
). Evolution of Brachyury proteins: identification of a novel regulatory domain conserved within Bilateria.
Dev. Biol
.
260
,
352
-361.
Miller, D. J., Ball, E. E. and Technau, U.(
2005
). Cnidarians and ancestral genetic complexity in the animal kingdom.
Trends Genet.
21
,
536
-539.
Müller, C. W. and Herrmann, B. G. (
1997
). Crystallographic structure of the T domain-DNA complex of the Brachyury transcription factor.
Nature
389
,
884
-888.
Naiche, L. A., Harrelson, Z., Kelly, R. G. and Papaioannou, V. E. (
2005
). T-box genes in vertebrate development.
Annu. Rev. Genet.
39
,
219
-239.
Nieuwkoop, P. D. and Faber, J. (
1967
).
Normal Table of Xenopus laevis.
Amsterdam: North Holland.
Ohno, S. (
1970
).
Evolution by Gene Duplication.
Heidelberg: Springer Verlag.
Papaioannou, V. E. (
2001
). T-box genes in development: from Hydra to humans.
Int. Rev. Cytol.
207
,
1
-70.
Papaioannou, V. E. and Silver, L. M. (
1998
). The T-box gene family.
BioEssays
20
,
9
-19.
Putnam, N. H., Srivastava, M., Hellsten, U., Dirks, B., Chapman,J., Salamov, A., Terry, A., Shapiro, H., Lindquist, E., Kapitonov, V. V. et al. (
2007
). Sea anemone genome reveals ancestral eumetazoan gene repertoire and genomic organization.
Science
317
,
86
-94.
Rao, Y. (
1994
). Conversion of a mesodermalizing molecule, the Xenopus Brachyury gene, into a neuralizing factor.
Genes Dev
.
8
,
939
-947.
Sambrook, J., Fritsch, E. F. and Maniatis, T.(
1989
).
Molecular Cloning: A Laboratory Manual
. New York: Cold Spring Harbor Laboratory Press.
Sasai, Y. and De Robertis, E. M. (
1997
). Ectodermal patterning in vertebrate embryos.
Dev. Biol.
182
,
5
-20.
Schmidt, H. A., Strimmer, K., Vingron, M. and von Haeseler,A. (
2002
). TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing.
Bioinformatics
18
,
502
-504.
Scholz, C. B. and Technau, U. (
2003
). The ancestral role of Brachyury: expression of NemBra1 in the basal cnidarian Nematostella vectensis (Anthozoa).
Dev. Genes Evol.
212
,
563
-570.
Schulte-Merker, S. and Smith, J. C. (
1995
). Mesoderm formation in response to Brachyury requires FGF signaling.
Curr. Biol.
5
,
62
-67.
Sive, H. L., Hattori, K. and Weintraub, H.(
1989
). Progressive determination during formation of the anteroposterior axis in
Xenopus laevis. Cell
14
,
171
-180.
Smith, J. (
1997
). Brachyury and the T-box genes.
Curr. Opin. Genet. Dev
.
7
,
474
-480.
Smith, J. (
1999
). T-box genes: what they do and how they do it.
Trends Genet.
15
,
154
-158.
Smith, J. C., Price, B. M. J., Green, J. B. A., Weigel, D. and Herrmann, B. G. (
1991
). Expression of Xenopushomolog of Brachyury (T) is an immediate-early response to mesoderm induction.
Cell
67
,
79
-87.
Spemann, H. and Mangold, H. (
1924
). Über Induktion von Embryonalanlagen durch Implantation artfremder Organisatoren.
Wilhelm Roux Arch. Entw. Mech. Org
.
123
,
389
-517.
Spring, J., Yanze, N., Josch, C., Middel, A. M., Winninger, B. and Schmid, V. (
2002
). Conservation of Brachyury,Mef2, and Snail in the myogenic lineage of jellyfish: a connection to the mesoderm of bilateria.
Dev. Biol
.
244
,
372
-384.
Stover, N. A. and Steele, R. E. (
2001
). Trans-spliced leader addition on mRNAs in a Cnidarian.
Proc. Natl. Acad. Sci. USA
98
,
5693
-5698.
Strong, C. F., Barnett, M. W., Hartman, D., Jones, E. A. and Stott, D. (
2000
). Xbra3 induces mesoderm and neural tissue in Xenopus laevis.
Dev. Biol
.
222
,
405
-419.
Takebayashi, K., Takahashi, S., Yokota, C., Tsuda, H.,Nakanishi, S., Asashima, M. and Kageyama, R. (
1997
). Conversion of ectoderm into a neural fate by ATH-3, a vertebrate basic helix-loop-helix gene homologous to Drosophila proneural gene atonal.
EMBO J.
15
,
384
-395.
Technau, U. (
2001
). Brachyury, the blastopore and the evolution of the mesoderm.
BioEssays
23
,
788
-794.
Technau, U. and Holstein, T. W. (
1996
). Phenotypic maturation of neurons and continuous precursor migration in the formation of the peduncle nerve net in Hydra.
Dev. Biol
.
177
,
599
-615.
Technau, U. and Bode, H. R. (
1999
). HyBra1, a Brachyury homolog, acts during head formation in Hydra.
Development
126
,
999
-1010.
Technau, U., Cramer von Laue, C., Rentzsch, F., Luft, S.,Hobmayer, B., Bode, H. R. and Holstein, T. W. (
2000
). Parameters of self-organization in Hydra aggregates.
Proc. Natl. Acad. Sci. USA
97
,
12127
-12131.
Technau, U., Rudd, S., Maxwell, P., Gordon, P. M., Saina, M.,Grasso, L. C., Hayward, D. C., Sensen, C. W., Saint, R., Holstein, T. W. et al. (
2005
). Maintenance of ancestral complexity and non-metazoan genes in two basal cnidarians.
Trends Genet.
21
,
633
-639.
Wilkinson, D. G., Bhatt, S. and Herrmann, B. G.(
1990
). Expression pattern of the mouse T gene and its role in mesoderm formation.
Nature
343
,
597
-598.
Wilson, V. and Conlon, F. L. (
2002
). The T-box family.
Genome Biol.
3
,REVIEWS3008.
Wittenberger, T., Steinbach, O. C., Authaler, A., Kopan, R. and Rupp, R. A. W. (
1999
). MyoD stimulates Delta-1 transcription and triggers Notch signaling in the Xenopus gastrula.
EMBO J.
18
,
1915
-1922.
Zhang, J. and King, M. L. (
1996
). Xenopus VegT RNA is localized to the vegetal cortex during oogenesis and encodes a novel T-box transcription factor involved in mesodermal patterning.
Development
122
,
4119
-4129.
Zhang, J., Houston, D. W., King, M. L., Payne, C., Wylie, C. and Heasman, J. (
1998
). The role of maternal VegT in establishing the primary germ layers in Xenopus embryos.
Cell
94
,
515
-524.