Enhancement of foot formation in Hydra by a novel epitheliopeptide, Hym-323

In higher metazoans, both invertebrates and vertebrates, signaling molecules involved in pattern formation are commonly proteins, such as members of the Hedgehog, Wingless or TGFβ families. In contrast, the majority of the signaling molecules regulating patterning processes identified so far in Cnidaria appear to be peptides. For example, members of LWamides isolated from sea anemone, Anthopleura elegantissima (Leitz et al., 1994) and from Hydra magnipapillata (Takahashi et al., 1997) induce metamorphosis of planula larvae of marine hydrozoans, Hydractinia. Also in Hydra, head activator, an undecapeptide increases the rates of head regeneration (Schaller, 1973; Javois and Tombe, 1991), foot regeneration (Javois and Frazier-Edwards, 1991) and bud formation (Schaller, 1973; Hobmayer et al, 1997). More recently, two peptides, pedin and pedibin, which enhance foot regeneration, have been identified in H. vulgaris (Hoffmeister, 1996). Whether the involvement of peptides in pattern formation is a trait specific to cnidarians or a widespread phenomenon among other animal phyla remains to be seen. However, there is an increasing body of evidence to indicate that peptides are signaling molecules involved in other developmental processes such as cell proliferation and/or differentiation (e.g. vasopressin (Naro et al., 1997), vasoactive intestinal peptide (Gressens et al., 1997), substance P (Kishi et al., 1996)) or morphogenesis (e.g. bombesin (Sunday et al., 1993) and parathyroid hormone-related peptide (Weir et al., 1996)).

Owing to its morphological simplicity and strong regenerative capacity, Hydra has been a subject of intensive investigations on pattern formation. Classic transplantation experiments, as well as theoretical analyses, have led to a model in which there are two positional value gradients along the body column, one with a maximum in the head and the other with a maximum in the foot (Webster and Wolpert, 1966; MacWilliams et al., 1970; Wolpert et al., 1971; Gierer and Meinhardt, 1972; Sugiyama, 1982). However, the molecular basis of these gradients is still unknown. In order to gain an understanding of the gradients at a molecular level, an effort has been initiated to systematically identify peptide signaling molecules involved in regulation of the patterning processes in H. magnipapillata (Takahashi et al., 1997). Up to now, two peptide molecules that enhance foot regeneration have been found. They are Hym-346 and Hym-323. Hym-346 is a counterpart of pedibin (Hoffmeister, 1996) in H. magnipapillata (Takahashi et al., 1997); these two peptides share the same amino acid sequence except that the latter has an extra glutamic acid at its C terminus (Takahashi et al., 1997). In addition, exogenously added Hym-346/pedibin alters the positional value gradient, which favors foot formation (Grens et al., 1999).

In the present study, purification and detailed characterization of Hym-323 will be reported. The peptide was produced and localized in epithelial cells. It enhanced foot formation by increasing foot activation potential (or lowering positional value) in the peptide-treated tissue. In this context, both Hym-323 and Hym−346 are very similar in their functions. The significance of this apparent overlapping function of these peptides will be discussed.

Strain 105, the standard wild-type strain of Hydra magnipapillata was used for most of the experiments and cultured as described previously (Sugiyama and Fujisawa, 1977). For some experiments, epithelial Hydra, which lack interstitial cells and their derivatives (except for gland cells) were used. The epithelial Hydra were produced from strain 105 by colchicine treatment (Campbell, 1976) and cultured by force-feeding (Nishimiya-Fujisawa and Sugiyama, 1993).

Extraction and nontargeted purification of Hydra peptides was carried out as described previously (Takahashi et al., 2000) with small modifications of the original method (Takahashi et al., 1997). The purification of peptides up to the second HPLC was exactly the same as that reported previously (Takahashi et al., 2000). Briefly, about 150 g of frozen Hydra tissue was boiled in 1.5 l of 5% acetic acid for 5 minutes and homogenized. The homogenate was centrifuged at 16,000 g for 40 minutes at 4°C and concentrated by rotary evaporation. To this solution, 1/10 volume of 1 N HCl was added, mixed and centrifuged as described above. The supernatant was applied to C-18 cartridges (Mega Bond-Elut, Varian) and washed with 10% methanol. The retained material was eluted with 60% methanol. The eluent (RM60) was then fractionated in a C-18 reverse-phase HPLC column (ODP-50, Asahipak; 6 mm×250 mm) with a linear gradient of 0-50% acetonitrile (ACN) in 0.1% trifluoroacetic acid (TFA) (pH 2.2) at a flow rate of 1 ml/minute (Fig. 1A). The fractions eluted between 5 and 37% ACN were divided into 15 groups. Group 8 eluted at approximately 20% ACN was first selected and peptides present in this group were systematically purified.

Hym-323, which was present in group 8, was purified using HPLC in the following steps. Cation-exchange HPLC (SP-5PW, Tosoh; 7.5 mm×75 mm) was carried out with a 70 minute linear gradient of 0-M NaCl in 10 mM phosphate buffer (PB) (pH 7.2) at a flow rate of 0.5 ml/minute (Fig. 1B). The fractions eluted at about 0.3 M NaCl were subsequently subjected to C-18 reverse-phase HPLC (ODS-80TM, Tosoh; 4.6 mm×150 mm) with a 75 minute linear gradient of 15-30% ACN in 0.1% TFA, at a flow rate of 0.5 ml/minute (Fig. 1C). The fractions eluted at about 20% ACN were subjected to the next HPLC (ODS-80TM) with an isocratic elution of 19.5% ACN at a flow rate of 0.3 ml/minute (Fig. 1D). Finally, the fraction was purified as a single peak (Hym-323) on the same column with an isocratic elution of 19.5% ACN in 0.1% TFA (Fig. 1E).

The peptide sequence was determined with an automated gas-phase sequencer (PPSQ-10, Shimadzu). Peptide synthesis was carried out by using a standard solid-phase method (PSSM-8, Shimadzu), followed by TFA-anisol cleavage and HPLC purification. The structures of synthetic peptides, including those used in examining antibody specificity, were confirmed with amino acid sequencing and HPLC analyses. The identity of synthetic and native peptides were confirmed by co-chromatography in both cation- and anion-exchange HPLC.

A cDNA encoding Hym-323 (KWVQGKPTGEVKQIKF; Takahashi et al., 1997) was obtained with the following three steps.

1. Total RNA was extracted from polyps starved for 48 hours using an AGPC method (Chomczynski and Sacchi, 1987) and used as the template for first strand cDNA synthesis, together with the oligo-Dt primer and AMV reverse transcriptase (Roche Diagnostics GmbH). Then, PCR was carried out using the first strand cDNA, the degenerate oligonucleotide primer corresponding to amino acids 1-6 of Hym-323 (5′-AA(G/A)TGGGTNCA(G/A)GGNAA-3′), the complementary primer corresponding to amino acids 16-10 (5′-AA(T/C)T-T(G/A/T)AT(T/C)TG(T/C)TTNAC-3′) and Taq DNA polymerase (Roche Diagnostics GmbH). The reaction was run for 30 cycles of 94°C for 15 seconds, 55°C for 30 seconds and 68°C for 30 seconds. PCR products were fractionated on a 1.5% agarose gel and the region slightly above the primers where no band was seen was isolated. DNA was purified using the Mermaid Kit (BIO 101) and cloned into the pCR2.1 plasmid (Invitrogen). One clone was sequenced and it contained the entire sequence of the Hym-323 peptide.

2. In the second step, nested PCR was carried out. The first PCR was carried out with following steps, 95°C for 1 minute, 40 cycles of 94°C for 30 seconds, 56°C for 30 seconds and 68°C for 3 minutes, and finally 72°C for 5 minutes using the oligonucleotide primer corresponding to amino acids 3-8 of Hym-323 (5′-TNCA(A/G)GGNAAACCAACAG-3′), the M-13 Forward (−40) primer and the Hydra cDNA library constructed in Uni-ZAPII (Stratagene) (Yum et al., 1998a) as the template. For the second PCR, another primer corresponding to amino acids 4-9 of Hym-323 (5′-CA(A/G)GGNAAACCAACAGGA-3′), the T7 promoter primer and 1/50 of the first PCR mixture was used as the template. The amplified DNA was isolated, cloned into the pCR2.1 plasmid (Invitrogen), sequenced as described above and finally a DNA fragment of 174bp containing the Hym-323 precursor was obtained.

3. Finally, the 174 bp fragment was used as a probe to screen the same cDNA library described above. 5×10⁵ clones were screened, which yielded one that contained a full-length cDNA.

Total RNA (5 μg/lane) extracted as described above was fractionated on a formaldehyde-agarose gel and northern blot analysis was carried out according to the standard procedures (Sambrook et al., 1989).

Digoxigenin (DIG)-labeled antisense and sense riboprobes were prepared according to the manufacturer’s procedure (Roche Diagnostics GmbH) using the cDNA clone described above containing the full-length Hym-323 precursor sequence as a template. Whole-mount in situ hybridization was performed by the method described by Grens et al. (Grens et al., 1996). The color reaction was carried out at 37°C for 1 hour.

A cysteine residue was added to the N-terminal end of the C-terminal nonapeptide portion of Hym-323 (CTGEVKQIKF) and the resulting peptide was conjugated to bovine thyroglobulin using succinimidyl m-maleimidobenzoate (MBS; Peerters et al., 1989). The conjugate (100 μg peptide/animal) containing complete Freund’s adjuvant (Difco) was initially injected into two Japanese White rabbits. Thereafter, they were injected six times at intervals of 2-3 weeks with the conjugate (50 μg peptide/animal) together with incomplete Freund’s adjuvant (Difco). The antibody titer was determined by using an enzyme-linked immunosorbant assay (ELISA) kit (ELISAmate, KPL), according to the manufacturer’s method. The serum was preadsorbed with 1 mg/ml of bovine thyroglobulin overnight at 4°C and then centrifuged at 15,000 g. The supernatant was stored at −80°C to be used as an antiserum.

The specificity of the antiserum was examined by competitive ELISA as described previously (Yum et al., 1998b). The antiserum was diluted 10,000-fold. The competitors used were Hym-323, KPTGEVKQIKF, TGEVKQIKF, EVKQIKF, KQIKF, QIKF, IKF, KF-hydrochloride (Sigma) and a peptide with a randomized sequence of Hym-323 (PEFKGQVKQTKWKVGI). The first six peptides all strongly competed with the immobilized Hym-323 peptide to a similar degree. The remaining peptides exhibited essentially no competition. Thus, the antiserum appears to recognize specifically the C-terminal tetrapeptide.

Polyps with several buds were cut transversely immediately below the tentacle ring and directly above the budding region. The resulting cylindrical body columns were treated with a procaine solution (Bode et al, 1987), causing a partial separation of the ectoderm and endoderm. Subsequently, the two layers were completely separated using a pair of forceps.

To examine the tissue localization of Hym-323, indirect immunofluorescence staining on whole mounts of Hydra was performed using the anti-Hym-323 antiserum (1/1,000 dilution). Animals were fixed overnight at 4°C in Zamboni’s fixative (Zamboni and de Martino, 1967). Binding of the antiserum was visualized with biotinylated anti-rabbit Igs (Amersham, 1/100 dilution) and fluorescein isothiocyanate (FITC)-conjugated to streptavidin (Amersham, 1/25 dilution).

Pre-adsorption of the antiserum with peptides was done by incubating diluted antiserum (1/1,000) with three different concentrations of a peptide(100, 10 and 1 μg/ml) at 4°C for 24 hours. At the highest concentration, the immunostaining was abolished, supporting the specificity of the antiserum to Hym-323.

Animals were first fixed with 99.5% ethanol for 30 minutes at room temperature. Immunoreactive peptides usually disappear with this treatment, presumably being washed away by the ethanol. The animals were subsequently fixed with Zamboni’s fixative overnight at 4°C. The rest of the procedure for immunohistochemistry was same as described above.

All biological assays were carried out at 18±1°C using a synthetic peptide at a concentration of 10⁻⁶ M unless otherwise mentioned.

Head regeneration was assayed by monitoring morphological changes, particularly tentacle formation under a dissecting microscope (Sugiyama and Fujisawa, 1977). Young polyps detached from the parents during previous 24 hours (stage I polyps; Fujisawa, 1987) were collected and starved for 24 hours. These polyps were bisected transversely in the middle of the body column and the lower halves were allowed to regenerate in the presence or absence of Hym-323. When two tentacles or more appeared, it was judged as head regeneration.

Foot regeneration of normal Hydra was examined with two different methods. In the first method, a monoclonal antibody (mAb) AE03, which specifically recognizes basal disk ectodermal epithelial cells (a kind gift of Y. Kobayakawa; Amano et al., 1997) was used. The regenerating feet were detected on whole mounts by indirect immunofluorescence staining using mAb AE03 according to the method described by Koizumi et al. (Koizumi et al., 1988). In the second method, the foot-specific peroxidase assay was used to monitor foot formation in a group of regenerates (Hoffmeister and Schaller, 1985).

Foot regeneration of epithelial Hydra was examined using mAb AE03 as described above.

To examine the effect of the peptide on cell proliferation and differentiation of neurons and nematocytes, the standard BrdU (2 mM) labeling method was used as described previously (Takahashi et al., 1997; Takahashi et al., 2000).

Lateral transplantation to measure changes in the foot- or head-forming potential in peptide treated polyps was performed as described by Grens et al. (Grens et al., 1999) with the following modification: donor polyps were treated with Hym-323 for 3 to 7 days.

Hym-323 is one of the peptides that was isolated, purified and sequenced as described previously (Takahashi et al., 1997).

The purification procedure, which was partially described previously (Takahashi et al., 2000) is shown in Fig. 1. After five steps of purification, a single peak was obtained that was subsequently designated as Hym-323. The structure of Hym-323 was determined to be KWVQGKPTGEVKQIKF.

A cDNA encoding Hym-323 was cloned in three steps: PCR was used to obtain the exact sequence corresponding to middle four amino acids of Hym-323. Then, nested PCR was used to obtain a longer fragment of the gene, and, finally, the longer fragment was used to screen a cDNA library to obtain a full-length clone. Fig. 2 shows the cDNA sequence and the deduced amino acid sequence of the precursor protein. Since the size of the transcript is about 450 bases long as determined by northern blot analysis (data not shown), the deduced amino acid sequence in Fig. 2 appears to represent a full-length precursor protein. The precursor protein contains no typical signal peptide sequence at its N terminus (von Heijne, 1983), although there is a hydrophobic region (underlined). It contains a single copy of Hym-323 at its C terminus (in bold letters). Since the Hym-323 sequence is preceded by a threonine residue at the N terminus, and threonine residues are known to be a precursor protein cleavage site in cnidarians (Grimmelikhuijzen et al., 1996), this threonine residue could act as a cleavage site to produce mature Hym-323. Both the cDNA and precursor sequences showed no significant similarity to known genes or proteins.

Expression of the gene encoding Hym-323 was investigated with both northern blot analysis and in situ hybridization. Fig. 3 shows the results of northern blot analysis. Total RNA was extracted from various parts of the Hydra body: head (H), gastric region (G), peduncle/basaldisk (PB) (Fig. 3A), the ectodermal tissue layer (Ecto) and the endodermal tissue layer (Endo) (see Materials and Methods). RNA was also prepared from epithelial (Epi) Hydra. As shown in Fig. 3B, the message with the same size was detected ubiquitously. The results indicate that the Hym-323 gene is expressed (1) throughout the body, (2) both in the ectoderm and endoderm, and (3) at least in epithelial cells. The general trend of this expression pattern was confirmed with whole-mount in situ hybridization (Fig. 4). The Hym-323 gene is expressed in the epithelial cells of both the ectoderm and endoderm throughout the body column (Fig. 4A). As shown in Fig. 4B, the gene is expressed in the basal portion of the ectodermal epithelial cells. In the extremities, the gene is not expressed in the basal disk or the tentacles. However, it is expressed in the endoderm, but not the ectoderm of the hypostome.

Finally, immunohistochemical localization of the Hym-323 peptide was examined on whole mounts of Hydra using an anti-Hym-323 antiserum. The immunoreactivity was detected throughout the body column in both the ectodermal and endodermal epithelial cells, including tentacles and the hypostome (Fig. 5A). This is a sharp contrast to in situ hybridization, where no message was detected in the tentacles and the ectoderm of the hypostome. The basal disk was free of immunoreactivity as was free of the message in in situ hybridization (Fig. 5B). The immunostaining disappeared when the Hydra tissue was fixed with 99.5% ethanol, which generally washes away neuropeptides from the tissue (data not shown).

All of these results indicate that Hym-323 is expressed in epithelial cells.

In order to examine the function of Hym-323, we subjected it to a series of biological assays as described in Materials and Methods. Since Hym-323 shares identical amino acid residues in five out of ten positions in its C-terminal region with head activator (pEPPGGSKVILF; Schaller and Bodenmueller, 1981), the enhancement of cell proliferation, neuron differentiation and head regeneration (the characters attributed to head activator) were carefully examined. To examine if the peptide affects cell proliferation, animals were treated with the peptide for 48 hours and pulse-labeled with BrdU for the last 1 hour. Thereafter, the samples were macerated into a single cell suspension (David, 1973) and fixed to score labeling index of epithelial cells, large interstitial cells, which include multipotent stem cells and early committed cells of the interstitial cell lineage, nematoblasts and gland cells. All of these cells are constantly in mitotic cycle (David and Campbell, 1972; Campbell and David, 1974) although the mitotic index of gland cells drops sharply upon starvation (T. F., unpublished observations). As shown in Table 1, the peptide had no significant effect on the cell proliferation rate of these cell populations. To examine if the peptide affects differentiation of neurons and nematocytes, animals were pulse-labeled with BrdU for 1 hour in the presence of peptide and then chased for 47 hours. At 48 hours, samples were macerated, and the labeling index of neurons and the ratio of 4s to epithelial cells were determined as previously described (Takahashi et al., 2000). (4s are clusters of four nematoblasts that are the first cell type recognizable as being unique to the nematocyte pathway (David and Gierer, 1974)). As Table 2 shows, the labeling index and the ratio were similar both in the peptide-treated and control samples, suggesting that the peptide had no effect on either neuron or nematocyte differentiation. The peptide exhibited no effect on head regeneration either. Young polyps were bisected and the lower halves were allowed to regenerate in the presence or absence of 10⁻⁶ M of Hym-323 (see Materials and Methods). Head regeneration was quantitated by calculating the fraction of regenerates that produced more than two tentacles. As shown in Fig. 6, no difference in kinetics of head regeneration was detected between treatment and control. There was no difference either in the number of tentacles formed (data not shown).

The effect of Hym-323 on foot regeneration was examined with two different methods. In the first method, polyps that had a protrusion of the first bud (stage II polyps) were treated with 10⁻⁶ M of Hym-323 for 24 hours. The middle third of the peduncle (the region immediately below the bud protrusion and above the basal disk) was isolated and allowed to regenerate for 40 hours at 15°C. The regenerating feet were detected on whole mounts by indirect immunofluorescence staining using mAb AE03 (see Materials and Methods). The regeneration time (40 hours) and the temperature (15°C) were selected to obtain the maximum difference between control and treatment after repeating the experiments with different regeneration times and temperatures. As shown in Fig. 7, the fraction of pieces with AE03-positive feet increased significantly over untreated control regenerates. In the second method, 20 stage I polyps were bisected transversely in the middle of the body column and the upper halves were allowed to regenerate in the presence of 10⁻⁶ M of Hym-323 for 0, 24 and 48 hours. At the end of regeneration, 20 polyps per sample were subjected to the peroxidase assay according to Hoffmeister and Schaller (1985). The results are shown in Fig. 8A. The peroxidase activity was significantly higher in the peptide-treated samples than in untreated controls at both 24 hours and 48 hours. This result confirms the effect of Hym-323 shown in Fig. 7.

Finally, the concentration-dependent effect of Hym-323 was examined by the peroxidase assay using the same experimental scheme described above, except for the treatment with different concentrations (10⁻⁹–10⁻⁵ M) of Hym-323 for 27 hours. As shown in Fig. 8B, the effect of Hym-323 is concentration dependent, being most significant at 10⁻⁶ M. At a higher concentration (10⁻⁵ M) the effect appeared to lessen. Thus, the choice of 10⁻⁶ M of Hym-323 for biological assays was appropriate.

Hym-323 is distributed evenly throughout the body except for the basal disk (Fig. 5), and yet it enhanced foot formation (Figs 7, 8). Thus, it would be interesting to examine the behavior of Hym-323 during foot regeneration. Stage II polyps were cut transversely at the middle of the body column and the regenerating upper tissues were stained immunohistochemically using an anti-Hym-323 antiserum. The results are shown in Fig. 9. Hym-323 was detected continually at the regenerating tip, even at 36 hours, when foot-specific cells became obvious (Fig. 9B). The staining intensity appears to have become stronger in the foot. At 48 hours, the regenerating foot became bigger and started to lose the immunoreactivity (Fig. 9C). No obvious difference in the immunostaining was detected in other regions than the foot (data not shown). These results indicate that Hym-323 persists in the regenerating tip until the nearly intact foot was formed.

Removal of the interstitial cell lineage from adult Hydra results in animals that consist only of both ectodermal and endodermal epithelial layers (Marcum and Campbell, 1978; Sugiyama and Fujisawa, 1978). Such animals, known as epithelial Hydra, maintain their morphology and gradients indefinitely, as well as being able to regenerate head and foot normally. Thus, the patterning processes, i.e. the gradients, must reside in the epithelial cells. However, a more recent study involving a mutant that is defective in head regeneration (Sugiyama and Wanek, 1993) indicates that cells of the interstitial cell lineage, possibly a subset of neurons, also play a role in these patterning processes.

Therefore, we examined if Hym-323 acts on epithelial cells directly or through other cell types (e.g. nerve cells). To do this, epithelial Hydra were treated with 10⁻⁶ M of Hym-323 for 3 hours and then bisected transversely into two halves. The upper halves were allowed to regenerate in the presence of 10⁻⁶ M of Hym-323. At 30 hours of regeneration, the pieces were examined for foot formation by immunostaining with mAb AE03. As shown in Fig. 10, Hym-323 significantly enhanced foot regeneration of epithelial polyps, strongly suggesting that Hym-323 directly acted on epithelial cells.

Enhancement of the rate of foot regeneration by Hym-323 treatment suggests that the peptide alters a positional value gradient along the head-foot axis favoring foot formation. This was examined by lateral transplantation experiments of the type described previously (Grens et al., 1999). Polyps were treated with 10⁻⁶ M of Hym-323 for 3-7 days and stage II polyps were selected as donors. A piece of tissue from position 2 of the donor was cut out and laterally grafted to the same position of an untreated host polyp of the same stage (Fig. 11). The fate of grafted tissue was followed for 11 days in the absence of the peptide. The implant derived from donors treated with Hym-323 for 3-7 days invariably induced an ectopic foot on the host at a significantly higher level than the untreated implant. Table 3 shows the result of 7-day treatment as a representative. The results suggest that Hym-323 increases the foot-forming potential in the treated tissue, thereby enhancing foot formation. A slight increase in ectopic foot formation in control was probably due to structural irregularity introduced by transplantation (Shimizu and Sawada, 1987).

Positional value gradients are thought to play an important role in patterning processes in developing animal embryos as well as in the maintenance of the adult form of Hydra (e.g. Bode and Bode, 1984). Although the molecular nature is still unknown, a large number of experiments carried out at the tissue level indicate that two pairs of morphogen gradients, one from the head to foot and the other from the foot to head (Wolpert et al., 1971; Wolpert et al., 1974; Gierer and Meinhardt, 1972; MacWilliams et al., 1970), are central to the maintenance of the adult Hydra form. In this context, the head and foot can be considered to be organizing centers. Since the cells of the body column are constantly in the mitotic cycle, Hydra tissue constantly changes its axial location, because of the displacement of epithelial sheets in both apical and basal directions (Campbell, 1967). Thus, the maintenance of positional value gradients appears to be a dynamic process and there may be signals that act continuously to maintain these gradients. Hym-323 reported in this study may be one such signal.

Hym-323 is encoded in a small gene, and appears to be the only peptide or protein encoded by this gene (Fig. 2). Although it is common to find genes that encode a number of copies of the same peptide, or several different peptides, in Cnidaria (Leviev and Grimmelikhuijzen, 1995; Gajewski et al., 1996; Leviev et al., 1997; Yum et al., 1998a), as well as in vertebrates and insects (Chretien and Li, 1967; Nakanishi et al., 1979; Kawano et al., 1992), the gene for Hym-323 is not such a gene. Recently, we have cloned even a smaller gene that encodes a single peptide, Hym-346/pedibin (DDBJ Accession Number AB030084), which also enhances foot formation (Hoffmeister, 1996; Grens et al., 1999).

The precursor protein contained no typical signal peptide sequence in the N-terminal region (von Heijne, 1983), although a stretch of a few hydrophobic amino acids was present (Fig. 2). How and where the peptide is processed without the signal peptide sequence remains to be seen. However, a recent study of human tyrosyl t-RNA synthetase indicates that the protein is secreted outside of the cells without a signal sequence during an apoptotic process; and that it is cleaved by an elastase to produce two distinct peptides that possess cytokine activities (Wakasugi and Schimmel, 1999). The Hym-323 precursor might use a similar mechanism. It could be released during foot regeneration and processed enzymatically to produce Hym-323. However, the possibility that Hym-323 is processed within the cell still remains.

The transplantation and regeneration experiments indicate that Hym-323 plays a role in foot formation. These data, coupled with the expression data, provide a more precise idea of that role. The transplantation experiments indicate that tissue treated with the peptide has an enhanced capacity to form a foot (Table 3), indicating that the peptide affects a patterning process involved in foot formation. This is consistent with either raising the level of the foot activation gradient in a two-gradient model (Bode and Bode, 1984) or simply lowering the positional value in a single positional value gradient model (Müller, 1996). The effects of Hym-323 on regeneration indicate that the peptide specifically affects foot regeneration, but not head regeneration. This was measured in terms of an enhanced rate of foot regeneration, as reflected in the earlier appearance of two markers specific for foot differentiation: the antigen recognized by the AE03 antibody and a foot-specific peroxidase (Figs 7, 8). These latter results could reflect the effect of Hym-323 on foot patterning, or they could indicate that Hym-323 affects both foot patterning, as well as the subsequent step of foot differentiation.

The regional distribution of the peptide, as measured with the anti-Hym-323 antibody, provides a more detailed view. Immunofluorescence studies showed that the peptide was found in all regions of the animal except the foot (Fig. 5). Bisection of the body column at any axial level results in the formation of a foot at the basal end of the upper part, indicating that the body column always has the potential to form a foot. The presence of the Hym-323 peptide in the epithelial cells of the body column would reflect this potential. When the tissue is displaced from the peduncle onto the foot, Hym-323 in the epithelial cells vanishes. A similar drop in the level of the peptide occurs in those epithelial cells regenerating a foot following bisection of the body column (Fig. 9). However, Hym-323 did not vanish until a foot of a specific size was developed (48 hours after the bisection; Fig. 9C). As shown in Fig. 4, Hym-323 mRNA is also absent in the foot. This is in a sharp contrast to the body column where the peptide must be synthesized continuously to maintain the level of peptide in the epithelial cells, since these cells are constantly in the mitotic cycle. Thus, a simple interpretation is that upon initiation of foot formation, the stored Hym-323 peptide is released from the epithelial cells, binds to a receptor on these same epithelial cells, and induces differentiation of the basal disk cells of the foot.

This viewpoint is supported by the data on the cell types affected by Hym-323. Hydra consists of three cell lineages:the ectodermal epithelial cell lineage, the endodermal epithelial cell lineage and the interstitial cell lineage. Treatment of animals with the peptide had no effect on the proliferation or differentiation of cells of the interstitial cell lineage. Conversely, in animals devoid of the interstitial cell lineage, Hym-323 increased the rate of the appearance of the AE03 antigen during foot regeneration, indicating the peptide directly affected the ectodermal epithelial cells. Hence, it is likely that during displacement of the epithelial cells onto the foot or during foot regeneration, the released Hym-323 peptide directly affects the epithelial cells that released the peptide. The possibility also remains that the peptide acts internally, being released from a stored state to an active state within the epithelial cells.

A final point concerns the difference between the regional distributions of the peptide and the production of the mRNA encoding the peptide. The gene is expressed in the body column and part of the head, but not in the differentiated extremities, which are the tip of the hypostome, the tentacles and the foot. Thus, it is expressed in the parts of the animal that are still undergoing cell division and for the most part capable of foot formation. The fact that the peptide is found in the tentacles and the hypostome is a reflection of the stability of the possibly stored peptide as tissue is displaced from the upper body column onto the tentacles or from the lower hypostome into the tip of the hypostome. Since tissue in the lower body column is continuously displaced onto the foot, the loss of the peptide from the epithelial cells in the foot emphasizes the difference between the two ends of the animal with respect to the behavior of the peptide. No changes occur in the apical end to the stored peptide, while the peptide is actively removed from a stored state in the foot, most likely to participate in foot formation.

Previously, foot activator, though its structure is unknown (Grimmelikhuijzen, 1977), and two other peptides, pedin/Hym-330 and pedibin/Hym-346 (Hoffmeister, 1996: Grens, et al., 1999), have been reported to enhance foot formation. Foot activator and pedin have been reported to have mitogenic activity and stimulate neuron differentiation (Hoffmeister, 1989; Hoffmeister, 1996). In contrast, neither Hym-330 nor Hym-346 exhibited these activities in our assays (data not shown; Grens et al., 1999). At the moment, it is not clear why three (even four) different peptides (Hym-323, pedin/Hym-330, pedibin/Hym-346 or foot activator) with apparently a similar function are present in Hydra tissue. These peptides may function in the same pathway that leads to the foot formation, being complementary to each other or competing with each other for the same target. Alternatively, they may be involved in totally independent pathways. There is some evidence that two of the peptides act in the same pathway. Recently, it was reported that both Hym-346 and Hym-323 downregulate the expression of a novel metalloprotease gene, foot activator-responsive metalloprotease 1 (Farm1; Kumpfmueller et al., 1999). Furthermore, the co-treatment with Hym-323 and Hym-346 did not enhance the rate of foot regeneration any more than did treatment with either of these peptides alone (data not shown). One possibility is that Hym-323 and Hym-346 act at different stages in the same pathway. This point is currently under investigation.

Authors thank Dr Hans R. Bode, University of California, Irvine for valuable comments, discussions and careful reading of the manuscript. This work was supported by grants from the Ministry of Education, Science, Sports and Culture of Japan, the Sumitomo Foundation and Japan Space Forum (Ground Research Announcement for Space Utilization) to T. F., and by the Sasakawa Scientific Research Grant to N. H. T. T. was a postdoctoral fellow supported by the Japan Society for Promotion of Science.