Hydra Peptide Project 1993–2007


  • Toshitaka Fujisawa

    Corresponding authorSearch for more papers by this author
    • Present address: Institute of Zoology, University of Heidelberg, Im Neuenheimer Feld, 230, 69120 Heidelberg, Germany

*Email: toshifujisawa43@yahoo.co.jp


A systematic screening of peptide signaling molecules (<5000 da) in Hydra magnipapillata (the Hydra Peptide Project) was launched in 1993 and at least the first phase of the project ended in 2007. From the project a number of interesting suggestions and results have been obtained. First, a simple metazoan-like Hydra appears to contain a few hundred peptide signaling molecules: half of them neuropeptides and the rest epitheliopeptides that are produced by epithelial cells. Second, epitheliopeptides were identified for the first time in Hydra. Some exhibit morphogen-like activities, which accord with the notion that epithelial cells are primarily responsible for patterning in Hydra. A family of epitheliopeptides was involved in regulating neuron differentiation possibly through neuron–epithelial cell interaction. Third, many novel neuropeptides were identified. Most of them act directly on muscle cells inducing contraction or relaxation. Some were involved in cell differentiation and morphogenesis. During the course of this study, a number of important technical innovations (e.g. genetic manipulations in transgenic Hydra, high-throughput purification techniques, etc.) and expressed sequence tag (EST) and genome databases were introduced in Hydra research. They have already helped to identify and characterize novel peptides and will contribute even more to the Hydra Peptide Project in the near future.


Hydra is a freshwater member of Cnidaria, one of the most basal metazoan phyla. It has a simple body plan with a limited number of cell types (Fujisawa 2003). The cylindrical body is made up of two epithelial layers, ectoderm and endoderm, together with extracellular matrix called mesoglea in between. Both ectodermal and endodermal epithelial cells in the body column continually proliferate, displacing the tissue layers both apically and basally (Campbell 1973). At both ends of the body, these cells stop proliferation and differentiate into tentacle- or basal disk-specific cells, which are continually displaced and eventually sloughed off at the tip. Thus, ectodermal epithelial cells and endodermal epithelial cells in the body column are, respectively, considered to be epithelial stem cells (Dübel et al. 1987). The third stem cell type is interstitial stem cell. The interstitial stem cells also continually proliferate and differentiate into nerve cells and nematocytes. They are capable of differentiating into germ cells (Bosch & David 1987) and gland cells (Schmidt et al. 1986; Bode et al. 1987), although the extent to which the differentiation takes place under normal growth is not known because germ cells and gland cells are also in constant mitosis. These three stem cell lineages do not interchange in Hydra. This feature appears to be specific to Hydra. In other hydrozoans, for example Podocoryne carnea, transdifferentiation occurs from epithelial muscle cells to other cell types including interstitial cells and nematocytes (Schmid & Alder 1984), or in Hydractinia echinata, totipotent interstitial cells are present (Müller et al. 2004). With the three stem cell lineages, a whole Hydra tissue turns over perpetually. In other words, Hydra tissue maintains “quasi-embryonic state” and this is the trait responsible for asexual growth of budding and a strong regenerative capacity.

In higher metazoans, peptides are generally used to maintain homeostasis of a body or tissue as hormones or to transmit signals between neurons or from neurons to muscles as neurotransmitters or neuromodulators. In patterning and morphogenesis signaling proteins play most pivotal roles but little is known about the involvement of peptides. As expressed sequence tag (EST) and genome databases from various animals have become available, it is now apparent that signaling proteins involved in the axis formation and subsequent events in development are conserved in most of the animals including cnidarians. Why should peptides also be involved in developmental pathways? This issue will be discussed later. Since it was suggested that diffusible morphogens were responsible for regulating pattern formation or morphogenesis in Hydra (Crick 1970; Gierer & Meinhardt 1972), and since the first morphogenetic molecule reported in Hydra was an undecapeptide called head activator (Schaller & Bodenmueller 1981), we set out to look for small peptides (<5000 Da) with morphogenetic activity. Although there was no a priori assurance that morphogens were peptides, we were optimistic to obtain peptides that were involved in patterning and morphogenesis. In addition, we expected to identify many interesting peptides with a wide repertoire of functions other than morphogenetic activity. In conventional biochemistry, signaling molecules are isolated one at a time using a specific assay system. The procedure usually requires an assay for each fraction in every step of purification. It is not only laborious and time consuming, but also loses precious samples in assays. The latter is particularly crucial when the peptide content per tissue is low. In the Hydra Peptide Project, we developed a novel systematic approach to identify signaling peptides, which used native peptide samples only once for an assay after the final purification (Takahashi et al. 1997; see next section).

Systematic approach to identify peptide signaling molecules

A newly developed method (Takahashi et al. 1997) consisted of the following five steps (Fig. 1A). (i) Peptides extracted from Hydra tissue were purified to homogeneity in a systematic manner using high performance liquid chromatography (HPLC) without involving any biological assays. (ii) To select the peptides of potential interest, the differential display-polymerase chain reaction (DD-PCR) (Liang & Pardee 1992) was used to compare the effect of a peptide on gene expression between treated and untreated control Hydra. (iii) Those that affected gene expression were subjected to peptide sequencing by an automated sequencer and/or a tandem mass spectrometer. (iv) Based on the sequence, the peptides were chemically synthesized. (v) The synthetic peptides obtained in a large quantity were used for various biological assays to examine their functions.

Figure 1.

Systematic identification of peptide signaling molecules. (a) Flowchart of the approach. The inserted gel pattern shows the result of differential display-polymerase chain reaction (DD-PCR) from treated and untreated Hydra. Dimethylsulfoxide (DMSO) was used to permearize cells, but was later proved to be unnecessary. Short arrows indicate different band patterns. (b) Purification steps of Hym-54 by series of high performance liquid chromatography (HPLC).

Figure 1(B) shows five steps of purification of a peptide Hym-54 as an example. The peptide extract was separated into 15 groups in the first step of reverse-phase HPLC (Fig. 1B-A). Group 8 eluting at 22% of acetonitrile was subjected to the second step of cation-exchange HPLC (Fig. 1B-B). A pooled fraction eluting at approximately 0.15 m of NaCl was then subjected to three successive steps of reverse-phase HPLC (Fig. 1B-CDE), resulting in a homogenous peak of a peptide, Hym-54. DD-PCR analysis showed that Hym-54 downregulated several messages as shown in Figure 1(A). Hym-54 was then subjected to amino acid sequencing and the tandem mass analysis, which determined its structure to be GPMTGLWamide. Synthetic Hym-54 was used to examine its function by various assays described in Figure 1(A). It induced detachment of buds from the parental body by contracting sphincter or constrictor muscles in the basal disk (Takahashi et al. 1997). Furthermore, the peptide induced bud detachment in epithelial polyps (Takahashi et al. 1997), which are essentially made of epithelial muscle cells and contain no cells in the interstitial cell lineage including neurons (Campbell 1976; Sugiyama & Fujisawa 1978a). This indicates that the peptide acts directly on muscle cells. GLWamides are produced and localized in nerve cells (Fig. 2; Takahashi et al. 2003), so they seem to act as neurotransmitters or neuromodulators at the neuro-muscular junction (see details below).

Figure 2.

Fluorescein isothiocyanate-immunostaining (FITC) of neurons in the Hydra head by using an anti-GLWamide antibody prepared from a rabbit.

The approach might at a first glance seem formidable, but the fact that many novel peptides were identified indicated that the approach was remarkably effective. However, the DD-PCR screening was later omitted for the following reasons. First, about 45% of purified peptides turned out to be DD-PCR positive. Second, since the Hydra EST and more recent genome databases have become available, it is much faster to sequence every purified peptide and to look for the precursor protein from which it is derived in the databases (see below).

Novel peptides identified by the project

By the approach described above, 817 peptides were purified and 527 among them were sequenced. Based on the results of the DD-PCR assay and the structural characteristics deduced from the precursor proteins, 55 peptides were chemically synthesized and we examined their biological activities. Expression patterns of their encoding genes and/or localization of peptides by immunostaining using antipeptide antibodies were examined. From these results, peptides were classified into two groups. One is a group of epitheliopeptides that are derived from the epithelial cells and the other is a group of neuropeptides. A rough estimation indicates that Hydra tissue contains about 500 peptide signaling molecules and 50% are epitheliopeptides and the rest neuropeptides (Takahashi et al. 1997). The number initially seemed extremely high in a simple organism like Hydra. Later the genome information became available and a rough estimation became possible for the numbers of peptide encoding genes and receptor (GPCR, G-protein coupled receptor with seven transmembrane domains) genes in Drosophila melanogaster and Caenorhabditis elegans. Table 1 compares the estimated numbers of neuropeptides, neuropeptide genes and GPCRs among three invertebrates, Hydra, Drosophila and C. elegans. As can be seen, all of the numbers in these organisms are not vastly different.

Table 1.  Comparison of the numbers of neuropeptides, their genes and receptors among Hydra, Drosophila and Caenorhabditis elegans
NumberHydraDrosophila (Hewes & Taghert 2001)Caenorhabditis elegans (Nathoo et al. 2001)
  • Estimated by the author assuming that each gene encodes an average of five peptide species. GPCR, G-protein coupled receptor.

Neuropeptide genes1004090
GPCRs for neuropeptides>10045130


Proline-tryptophan (PW) peptides involved in nerve cell differentiation. Table 2 shows novel epitheliopeptides identified by the Project. The first group of peptides is a PW family (Takahashi et al. 1997). Four members that shared a PW sequence at the C-terminus were initially identified. These peptides all inhibited neuron differentiation and counteracted to the enhancing activity of neuron differentiation by a neuropeptide Hym-355 (see below; Table 3). However, prolonged treatment with the peptide nullified its own effect, suggesting that the feedback mechanism is involved in the regulation. Immunostaining using a polyclonal antibody against Hym-33H indicated that the peptides were localized in the ectodermal epithelial cells throughout the body column except the basal disk (Takahashi et al. unpubl. data, 2007). Thus, we speculate that the density of neurons (neurons/epithelial cells) is maintained constant by interaction between neurons and epithelial cells and that PW peptides are a negative signal emitted from epithelial cells if the neuron density is high enough. If the density is low, PW peptides are not emitted and the effect of Hym-355 predominates. Cloning the PW peptide-encoding gene had been unsuccessful for a long period of time because the short sequences of the peptides made it difficult to design proper primers. However, the information from the Hydra EST Project enabled us to clone a full-length cDNA. The precursor encodes not only all four sequences but also four additional putative peptides. The expression pattern of the gene was similar to the immunostaining result.

Table 2.  Epitheliopeptides identified by the Hydra Peptide and Express Sequence Tag (EST) Projects
  • The peptide sequence was deduced from the precursor sequence. ND, not determined. PW, proline-tryptophan.

PW family
Hym-33HAALPW Takahashi et al. (1997)
Hym-35EPSAAIPWInhibition of neuronTakahashi et al. (1997)
Hym-37SPGLPWdifferentiationTakahashi et al. (1997)
Hym-310DPSALPW Takahashi et al. (1997)
EPSAALPWNDTakahashi et al. (1997)
IPALPWNDTakahashi et al. (1997)
TRTLPWNDTakahashi et al. (1997)
Hym-1397TPALHWUnknownFujisawa, unpubl. data, 2006
Hym-301KPPRRCYLNGYCSPamideEnhancement of tentacle formationTakahashi et al. (2005)
Hym-323KWVQGKPTGEVKQIKFEnhancement of foot formationHarafuji et al. (2001)
Pedibin/Hym-346AGEDVSHELEEKEKALANHSEnhancement of foot formationHoffmeister (1996); Grens et al. (1996)
Hym-342EVETFDKSKLKKTETVEKNPUnknownTakahashi et al. unpubl. data, 1997
Hym-330/thypedinEELRPEVLPDVS(E)Enhancement of foot formation and buddingTakahashi et al. unpubl. data, 1997 year; Hermann et al. 2005
Table 3.  Neuropeptides identified by the Hydra Peptide and Express Sequence Tag (EST) Projects and by others
  • The fourth residue proline is hydroxylated.

  • Four Hym-176 related genes were found in the Hydra ESTs.

  • §

    One Hym-355 related gene was found in the Hydra ESTs.

  • ¶The peptide sequence was deduced from the precursor sequence. < Q, C-terminal pyroglutamate; ND, not determined.

GLWamide family
 Hym-53NPYPGLWa Takahashi et al. (1997)
 Hym-54GPMTGLWaEnhancement of bud detachmentTakahashi et al. (1997)
 Hym-248EPLPIGLWaInduction of metamorphosis of planulaTakahashi et al. (1997), Takahashi et al. (2003)
 Hym-249KPIPGLWalarvae from HydractiniaTakahashi et al. (1997)
 Hym-331GPPPGLWaHym-248 also induces body elongationTakahashi et al. (1997)
 Hym-338GPPhPGLWa Takahashi et al. (1997)
 Hym-370KPNAYKGKLPIGLWa Takahashi et al. (2003)
 Hym-1071KPPWRGGM(O)WUnknownTakahashi et al. unpubl. data, 2006
NPENRLPLGLWaNDLeviev et al. (1997)
< QPPIGMWaNDLeviev et al. (1997)
Hym-176 gene family
 Hym-176APFIFPGPKVaContraction of peduncleYum et al. 1998a.
YPFYNQNPKVaUnknownHayakawa et al. unpubl. data, 2003
NPKNKNFMIFVGPKVaUnknownHayakawa et al. unpubl. data, 2003
 Hym-357KPAFLFKGYKPaContraction of whole bodyYum et al. 1998b.
 Hym-690KPLYLFKGYKPaContraction of whole bodyFujisawa unpubl. data, 2006
NPFIFKGHKHaUnknownHayakawa et al. unpubl. data, 2003
Hym-355 gene family (§)
 Hym-355FPQSFLPRGaEnhancement of neuron differentiationTakahashi et al. (2000)
DFQDYAPRGaUnknownHayakawa et al. unpubl. data, 2003
DARPRAaUnknownTakahashi et al. (2000)
ENRPRPaUnknownHayakawa et al. unpubl. data, 2003
FRamide family
 Hym-65 (FRamide 1)IPTGTLIFRaBody elongationHayakawa et al. 2007
 Hym-1533 (FRamide-2)APGSLLFRaBody contractionHayakawa et al. 2007
RFamide family
 RFamide I< QWLGGRFaUnknownMoosler et al. (1996)
 RFamide II< QWFNGRFaUnknownMoosler et al. (1996)
 RFamide III/IV(KP)HLRGRFaEnhancement of body pumpingMoosler et al. (1996), Shimizu & Fujisawa (2003)
 RFamide V< QLMSGRFaNDDarmer et al. (1998)
 RFamide VI< QLMRGRFaNDDarmer et al. (1998)
 RFamide VII< QLLRGRFaNDDarmer et al. (1998)
 RFamide VIII/IX(KP)HYRGRFaNDDarmer et al. (1998)
 RFamide XKPHLIGRFaNDHayakawa et al. unpubl. data, 2003
 RFamide XI< QLMTGRFaNDHayakawa et al. unpubl. data, 2003

Morphogenetic peptides.  Morphogenesis of Hydra is primarily controlled by epithelial cells (Sugiyama & Fujisawa 1978b) and morphogens are thought to be diffusible small molecules (Crick 1970; Gierer & Meinhardt 1972). We therefore expected to find some epitheliopeptides involved in morphogenesis of Hydra. Hym-301 is the first epitheliopeptide identified in Hydra that has an amidated residue in the C-terminus (Takahashi et al. 2005). It has an intramolecular disulfide linkage (Table 2). The expression of the Hym-301-encoding gene is restricted to the ectodermal epithelial cells of the hypostome except its tip and the subtentacle region (Takahashi et al. 2005). A similar pattern was also obtained by immunostaining using an anti-Hym-301 antibody. Several functional analyses indicate that the peptide regulates tentacle formation (Takahashi et al. 2005). Furthermore, overexpression of Hym-301 under the control of actin promoter produced ectopic tentacles in the body of transgenic Hydra (Khalturin and Bosch, pers. comm., 2007) confirming our above conclusion. Recently, we identified three more Hym-301 related genes in the Hydra ESTs (Berger et al. unpubl. data, 2007). All of them were expressed differentially along the body axis in the ectodermal epithelial cells. They encode a copy of different peptides with a possible disulfide linkage within each molecule. All of the Hym-301 related genes appeared to be downstream of the Wnt pathway, which is envisaged as a principal signaling cascade involved in the axis formation in Hydra (Hobmayer et al. 2000). The observation suggests that both protein signaling molecules and peptide signaling molecules are involved in morphogenesis in Hydra. The recent discovery of small peptides involved in actin-based morphogenesis in D. melanogaser (Kondo et al. 2007) indicates possible morphogenetic functions of peptides in higher metazoans in general.

Two other morphogenetic peptides that regulate foot formation have been identified. They are Pedibin/Hym-346 and Hym-323 (Hoffmeister 1996; Grens et al. 1996; Harafuji et al. 2001; see Fujisawa 2003 for review). Treatment of Hydra with either peptide alters positional information, which is one of the important characteristics of a morphogen (Grens et al. 1996; Harafuji et al. 2001). The pedibin/Hym-346-encoding gene is expressed in the endodermal epithelial cells not only in the foot but also at the tentacle base (Herrmann et al. 2005). Therefore, the gene might also be involved in tentacle formation. On the contrary, the peptide Hym-323 is localized in both ectodermal and endodermal epithelial cells except the basal disk (Harafuji et al. 2001). During foot regeneration, the peptide disappears when the basal disk cells are formed. These results indicate that Hym-323 is required for foot formation but not for its maintenance (Harafuji et al. 2001; Fujisawa 2003). Another peptide, pedin/Hym-330 is reported to induce foot formation (Hoffmeister 1996). However, we did not find this effect in our study, instead the peptide enhanced bud outgrowth (Herrmann et al. 2005). Pedibin/Hym-346 and Hym-323 appear to be involved in the same pathway because both affect the same target gene Farm1 (foot specific astacin metalloprotease-1) (Kumpfmueller et al. 1999). A putative peptide Heady is reported to be an inducer of the apical fate (Lohmann & Bosch 2000). Whether the peptide truly exists in Hydra tissue or not remains to be tested.

Is Hym-323 a cryptic peptide?  In general, a precursor that produces bioactive peptides contains a signal peptide at the N-terminus, which directs transportation to endoplasmic reticulum. In trans-Golgi networks the precursor is processed and active peptides are stored in secretory vesicles. However, the deduced precursor from Hym-323 or Hym-346 contains no apparent signal peptide. Furthermore, the Hym-323 precursor has a putative conserved domain of mitochondrial adenosine triphosphate (ATP) synthase epsilon subunit (pfam04627). Although it has not been tested whether the Hym-323 precursor has the enzyme activity, it is likely that the precursor is a mitochondrial enzyme. If this is the case, the question arises whether or not Hym-323 is a degradation product of the protein. There are early reports that the N-terminal fragment of porcine β-hemoglobin has activity to release growth hormone (Schally et al. 1971) and that a fragment of porcine α-hemoglobin stimulates adrenocorticotropic hormone (ACTH) release (Schally et al. 1978). These peptides were, however, considered as artifacts of extraction or isolation procedures. In 1999 Wakasugi and Schimmel reported that human tyrosyl t-RNA synthetase was secreted outside of cells without a signal sequence during an apoptotic process and that it was cleaved by an elastase to produce two distinct peptides with cytokine activities. More recently, fragments of mitochondrial proteins (e.g. cytochrome c oxidase) that are produced during protein maturation or degradation have been reported to activate neutrophils (Ueki et al. 2007). The activity is totally different from that of parental protein and is usually hindered unless the ‘cryptic peptides’ are generated. Although the issue of whether Hym-323 is a non-specific degradation product or a cryptic peptide is not solved, more evidence supporting cryptic peptides is accumulating. If cryptic peptides are indeed generated in vivo, they multiply the functional repertoire of a single gene, thus serving the economy of genetic materials.


It is generally believed that neurotransmission in the Hydra nervous system is mediated by neuropeptides, and not by classic neurotransmitters such as biogenic amines (Grimmelikhuijzen et al. 1992). Although the complexity of Hydra behavior is limited, there are some distinct concerted movements, for example, tentacle waving and elongation/contraction, mouth opening/closing, peristalsis similar to esophageal reflex and intestinal movements, defecation (Shimizu et al. 2004), pumping of the body column (Shimizu & Fujisawa 2003), locomotion by somersaulting (Ewer 1947), and attachment and detachment of basal disk to the substrate or air. These coordinated movements must be regulated by neurons and therefore by neuropeptides. As will be described in this section, some of the neuropeptides have been shown to be involved. Table 3 lists the Hydra neuropeptides, most of which are identified by the Hydra Peptide Project.

GLWamide family.  Eight members of the (G)LWamide family are purified and seven, except Hym-1071, have a GLWamide motif at the C-terminus (Takahashi et al. 1997; Takahashi et al. 2003). A peptide having the C-terminal GLWamide, metamorphosin A was first purified from sea anemone Anthopleura elegantissima, which induced metamorphosis of planula larvae of a marine hydrozoan Hydractinia echinata (Leitz et al. 1994). Immunostaining in H. echinata indicates metamorphosin A to be a neuropeptide (Leitz & Lay 1995). A GLWamide gene from Hydra encodes all of the GLWamides that we have isolated (Leviev et al. 1997) and is expressed in a subset of neurons (Mitgutsch et al. 1999). Immunohistochemistry using an antibody against the GLWamide moiety shows that neurons throughout the Hydra body are recognized, confirming that GLWamides are neuropeptides (Takahashi et al. 2003; Fig. 2). We have shown that all of the Hydra GLWamides induce metamorphosis of Hydractinia larvae (Takahashi et al. 1997, 2003). Furthermore, structure-function analysis indicates that the C-terminal GLWamide moiety of the peptides is necessary and sufficient for the induction of metamorphosis. RFamides, in contrast, are shown to block the effect of GLWamides in metamorphosis, suggesting interplays of neuropeptides (Katsukura et al. 2003). Since Hydra has no planula stage in its life cycle, GLWamides must play different roles in the animal. One of the functions we have discovered is the enhancement of bud detachment from the parental body. In Hydra, budding is a way of asexual reproduction. Buds are formed at the side of the parental body column, first by protrusion, elongation, head formation and finally foot formation. At the final stage of budding, circular muscles, called sphincter (or constrictor) muscles are formed in the ectoderm of basal disk (Campbell, unpubl. data, 1997). The GLWamides induce contraction of the sphincter muscles, thus the buds are squeezed off prematurely from the parent. Only one GLWamide, Hym-248 has an additional function (Takahashi et al. 2003). This peptide induces elongation of the body column and tentacles (see below). Since the ectodermal muscle processes in Hydra run longitudinally and the endodermal muscle processes run circumferentially, the elongation of the body is caused by the contraction of endodermal muscles while ectodermal muscles relax. All of the myoactive effects of GLWamides are reproduced on epithelial Hydra. Since epithelial cells in both ectoderm and endoderm, are muscle cells, the epithelial Hydra is an excellent pure in vivo muscle sample for testing myoactivity. Thus, the result suggests that GLWamides are neurotransmitters or neuromodulators working at the neuromuscular junctions. Since the GLWamide positive neurons are distributed throughout the body, it is possible that the peptides have some other functions yet to be discovered.

The Hym-176 gene family.  The Hym-176 peptide is localized at high concentrations in neurons located in the lower peduncle region and at low concentrations in neurons in the gastric region (Yum et al. 1998a,b). It induces body contraction at 10−5 M (Yum et al. 1998a, b), but at lower concentrations only peduncle contraction is evoked coinciding with the localization of peduncle-specific neurons (Noro et al. unpubl. data, 2007). The gene encoding Hym-176 contains another peptide Hym-357. This peptide strongly induces both tentacle and body contraction of normal polyps (Fig. 3), but has no effect on epithelial polyps. The result indicates that the peptide does not work directly on muscles but presumably activates other neurons which in turn release neurotransmitters to directly induce muscle contraction. Comparison of the peptide effect between normal and epithelial polyps provides a unique methodology to differentiate the mode of actions of a given peptide. In the presence of Hym-248 (10−6 M), Hym-357 (10−6 M) failed to exert its effect (Fig. 3, bottom panels). These peptides, possibly together with other unknown peptides, may be involved in coordinated movements of tentacles.

Figure 3.

Effects of neuropeptides Hym-357 and Hym-248 on tentacles of Hydra. A head was isolated to show the peptide effect more clearly. Upper panels (A–D). Strong but slow contraction of tentacles by Hym-357 (10−6 M). Lower panels (E–H). Effect of Hym-248 10−6 M on contracted tentacles in the continual presence of Hym-357.

In Hydra ESTs, four genes related to Hym-176 have been obtained (Hayakawa et al. unpubl. data, 2003). Peptides encoded by these genes are variable, although three of them have PKVamides and one of them encodes the Hym-357-like peptide, Hym-690, whose effect is essentially same as Hym-357 (Table 3). The expression patterns of these genes are unique in that all of them are expressed in subpopulations of neurons, some of which form distinct compartments along the body axis with a sharp boundary between them. The feature is being exploited to examine both functional and evolutionary significance of neuron compartments (Noro et al. unpubl. data, 2007).

Peptides involved in neuron differentiation.  Hym-355 is a neuropeptide that positively regulates neuron differentiation as mentioned above. A feedback mechanism involving PW peptides and Hym-355 has been proposed (Takahashi et al. 2000). According to the model, Hym-355 produced by neurons increases the rate of neuron differentiation at the early stage in the pathway. To counteract this positive feedback loop, epithelial cells produce PW peptides that block the neuron precursors to differentiate. However, an additional factor is required to complete the feedback mechanism. When interstitial cells are introduced into an epithelial polyp, the rate of neuron differentiation is higher than in a normal polyp (Fujisawa 1992). If PW peptides are active all the time, new neuron differentiation should not occur. Yet, the interstitial cells do produce neurons under these conditions (Bosch et al. 1991). Thus, there must be interaction between neurons and epithelial cells to measure the local neuron density. We currently favor an idea that the third factor produced by mature neurons through cell–cell interaction or short-range diffusion regulates the release of PW peptides. This three-part mechanism would permit adjustment in the rate of nerve cell production to maintain the nerve cell density constant in the adult Hydra. For example, should the neuron density decline precipitously, the level of PW would become so low as to be ineffective. Subsequently, the autocatalytic part of the feedback loop involving Hym-355 would take over, leading to a rise in nerve cell density.

In Hydra ESTs, a Hym-355-related gene was found and it appeared to encode both Hym-355-like PRGamide and PRPamide. The two genes are expressed in different subpopulations of neurons (Hayakawa et al. unpubl. data, 2003). The function of these peptides as well as PRAamide, another peptide encoded in Hym-355 remains to be discovered.

The last point worth mentioning is that Hym-355 is an entity that has long been considered to be a vasopressin-like peptide in Hydra (Morishita et al. 2003). Immunostaining of Hydra polyps using polyclonal antibodies against mammalian Arg-vasopressin suggested the presence of a related peptide in nerve cells (Grimmelikhuijzen et al. 1982). However, purification and structure determination of immunoreactive peptides from Hydra shows that they are Hym-355 and its shorter form (Morishita et al. 2003). The only feature common in both Arg-vasopressin and Hym-355 is their C-terminal motif of PRGamide. Thus, the vasopressin family conserved from annelids to mammals appears to be absent in one of the most basal metazoans, Hydra.

The FRamide family.  The power of the ESTs was most conspicuously exhibited in identifying peptides of the FRamide family. A native sample of Hym-65 (now called FRamide 1; tentative sequence of IPTGTLIFR) was negative in the DD-PCR assay and was set aside. Blasting all of the peptide sequences, irrespective of whether they were DD-PCR positive or not, against the translated Hydra EST database picked up one precursor that contained a signal peptide at the N-terminus and the Hym-65 sequence, which was flanked by GKK, a possible amidation motif, at its C-terminus (Hayakawa et al. 2007). The native peptide was subjected to mass analysis to find that the C-terminal R was indeed amidated. In the precursor, another presumptive FRamide was encoded. The peptide, Hym-1533 (now called FRamide-2) was later obtained in the Project. FRamide 1 and FRamide 2 exhibit an opposite effect; the former evokes contraction of endodermal muscles thereby elongating the body column and the latter evokes contraction of ectodermal muscles, thereby contracting the body column. There are a few possible explanations for how peptides derived from a single precursor can act in opposite ways. (i) The precursor could be differentially processed depending on the neuron types. A typical example for this is proopiomelanocortin. It is differentially processed in the anterior lobe and intermediate lobe of pituitary gland in mammals and, as a result, different repertoires of peptides are expressed in each region (Eipper & Mains 1980; Rosa et al. 1980). (ii) Different peptides might be differentially sorted to excretory vesicles. In case of the prohormone of the egg-laying hormone of a freshwater snail Lymnaea stagnalis, processed peptides are sorted to different vesicles depending on the neuron type, thus creating unequal distribution of peptides in different neurons (Klumperman et al. 1996). (iii) Like many hormones target cells might express different receptors specific to each peptide. The peptide-specific antibodies might help to solve the problem.

Other peptides identified by the Hydra Peptide and EST Projects.  In a similar way as mentioned above, a novel neuropeptide Hym-121 is identified. The peptide is 18 amino acids long and is the first neuropeptide in Hydra that has no amidation at the C-terminus (Hayakawa et al. unpubl. data, 2007). Hym-121 has a myoactivity on epithelial Hydra thus directly acting on muscle cells. Later, we also found that Hym-1071, a member of the GLWamide family, has no C-terminal amidation (Table 3). In this case, however, GLW is flanked by an amide donor G and a processing site K in the precursor. The Hym-121 encoding gene is found in the Hydra ESTs (Hayakawa et al. unpubl. data, 2006) and is also detected in microarray analysis as a neuron-specific gene (hmp_11958) (Hwang et al. 2007). The precursor protein deduced from the cDNA sequence contains several other presumptive peptide sequences both amidated and nonamidated at the C-termini.

The insulin signaling pathway has been characterized in detail in bilaterian animals, where it has been shown to regulate metabolism, growth and longevity. In Hydra, the insulin receptor homolog, HTK7 has been known for some time and is suggested to regulate growth and patterning (Steele et al. 1996). However, its ligands were not identified. The search in the Hydra ESTs has yielded three genes encoding insulin-like peptides (Steele et al. unpubl. data, 2007). One of the genes is expressed in neurons, whereas the other two are expressed in epithelial cells. The structures deduced from nucleotide sequences suggest that all three have three intramolecular disulfide linkages typical to insulin-like peptides. Two of the genes can rescue the growth defects in Drosophila, which have had cells producing insulin-like peptides ablated (Theofiles et al. 2007). Thus, the functions of the peptides are conserved from Cnidaria to Arthropoda, although in vivo functions in Hydra remain to be discovered.

Evolutionary conservation of peptides

The function of a peptide is generally determined by a motif of a few amino acids that is recognized specifically by its receptor. The rest of the peptide sequence can vary almost randomly. Consequently, it is not easy to study evolution of peptides or peptide encoding genes. There are only several peptide families that are conserved from invertebrates to vertebrates. The best example is a family (or families) of FMRFamide-related peptides (FaRPs or RFamides). FMRFamide was originally identified in the nervous system of the clam Macrocallista nimbosa as a cardioexcitatory peptide (Price & Greenberg 1977). Later a gene encoding multiple copies of RFamides was isolated (Schaefer et al. 1985). Since then a large number of RFamides and their encoding genes have been isolated from Hydra to humans. The most extreme case is found in C. elegans that contains as many as 25 flp (FMRFamide-like peptide) genes and 61possible FaRPs (Li et al. 1999; Kim & Li 2004). Among them 15 FaRPs have been biochemically isolated (Papaioannou et al. 2005). The diversity and complexity of RFamides in many groups of metazoans, however, have made it difficult to draw a simple picture in the context of evolution. In Hydra, four RFamide encoding genes and three peptides are identified (Moosler et al. 1996; Darmer et al. 1998; Hayakawa et al. unpubl. data, 2003). Only RFamide III/IV among three peptides evokes pumping activity of the Hydra body column, which we believe is the ancestral heart movement (Shimizu et al. 2004). The cardioregulatory role of RFamides seems to be well conserved from the diploblastic to triploblastic metazoans.

A search for counterparts of Hydra peptides in higher metazoans has revealed that nematodes, both C. elegans and C. briggsea, have a gene, respectively, encoding three and two presumptive (G)LWamides. Two of the former three peptides were recently identified biochemically (Husson et al. 2005). The gene is expressed in a small number of interneurons in C. elegans (Ishihara et al. unpubl. data, 1998). However, since mutations in the gene exhibit no obvious defect, the function of the gene is still unknown. The GLWamide encoding gene in higher metazoans other than nematodes has not been discovered.


There are many aspects that the Hydra Peptide Project has left out. The numbers of peptides and their encoding genes are still small compared with the expected numbers (a few hundred). More efficient ways of identifying peptides are required. The recent development of high-throughput devices aligning HPLC and tandem mass-spectrometers on line (LC-MS/MS) should be of great help in speeding up purification and identification of peptides. Particularly, the exploitation of the Hydra EST and genome databases, which has recently become publicly available, should greatly contribute to the quick identification of peptides. Although preliminary, Takahashi has recently introduced LC-MS/MS and quickly identified a few hitherto-unidentified peptides from Hydra, proving the power of the method. Another technique to be used is the single-cell mass spectrometry (see review, Li et al. 2000). The in situ preparations of Hydra tissues or cells may be subjected to mass analysis. Since sensitivity in detecting peptides in mass spectrometry has greatly increased, it will be a powerful tool to identify peptides that are present in vivo tissue.

As can be seen in Table 3, functions of many of the peptides are still to be discovered. Myoactivity of peptides has been successfully assayed by using epithelial and normal Hydra as described above. However, proper assays to examine whether a peptide could regulate the release of other factors including peptides are to be developed.

In order to know the function(s) of a peptide, information on its receptor is required. In most cases, the receptors that use peptides as ligands are G protein-coupled receptors with seven transmembrane domains (family 1 GPCRs). A survey of the Hydra genome database has uncovered about 700 members belonging to the family 1 GPCRs (Hayakawa and Fujisawa, unpubl. data, 2007). All of them are orphan receptors whose ligands are unknown. We have attempted to deorphanize some of the GPCRs by expressing each GPCR gene in mammalian cells or Xenopus oocytes and assaying with peptides or other substances. Preliminary results suggest that ligands for three of them are determined (Takahashi et al. unpubl. data, 2007). This line of approach should be continued. In the snail Helix aspersa, the presence of FMRFamide-gated Na+ channel was discovered (Lingueglia et al. 1995). This is the first report on a peptide-gated ion channel. Recently, it was shown that Hydra-RFamide I and II are the ligands for a heteromeric Na+ channel (HyNaC2 and HyNaC3) (Golubovic et al. 2007). Whether there exist other ion channels gated by other peptides in Hydra remains to be seen.

Finally some remarks about the recent establishment of infrastructures for Hydra research should be added. As already described, the Hydra EST and genome databases have become publicly available. Needless to say, the information is extremely powerful in carrying out any kinds of genetic research. Also, a technique for producing transgenic Hydra has recently become available (Wittlieb et al. 2006; Khalturin et al. 2007). Functional analysis of genes in Hydra was hampered by the lack of gene manipulation techniques. RNAi has been partially successful but not in many cases. Transgenic Hydra provides a way of over-production of a particular gene product or obtaining dominant-negative effects if the interacting molecules are known. We have also successfully used promoter regions of several peptide genes to drive GFP expression that recapitulated the original patterns (Noro and Fujisawa, unpubl. data, 2007). These techniques will be further exploited for functional analysis of a gene or a lineage tracing of a particular cell type.


For systematic identification of small peptides a bioinformatics approach has a severe limitation. Unlike proteins, the precursors that encode peptides generally have no motifs except for the N-terminal signal peptide. The only other feature may be processing sites that flank a peptide sequence both in its N- and C-terminal regions. Thus, it is absolutely necessary to obtain peptide sequences first to exploit EST and genome databases. The Hydra Peptide Project was the first attempt to analyze peptide signaling molecules in a systematic manner, well ahead of the beginning of the proteomic era. Despite initial skepticism, our project has successfully and efficiently identified many novel peptides, particularly peptides involved in morphogenesis and cell differentiation in Hydra. Although the general importance of peptides, particularly in metazoan morphogenesis and cell differentiation is still un-established, the recent discoveries of short peptides in Drosophila and mammalian cells as described in this article is a step forward in this area of research.


The author expresses sincere gratitude to a number of coworkers who have contributed to the work described in this article. Hydra Peptide Project: Yojiro Muneoka, Tsutomu Sugiyama, Toshio Takahashi, Osamu Koizumi, Yoshitaka Kobayakawa, Osamu Matsushima, Fumihiro Morishita, Charles N. David, Thomas Bosch, Jan Lohmann, Hans Bode, Masayuki Hatta, Hiroshi Shimizu, Seungshic Yum, Takeshi Ishihara, Naoe Harafuji, Eisuke Hayakawa, Yukihiko Noro, Claudia Berger, Akemi Hayashiuchi and Ikuko Goto. Hydra EST project: Takashi Gojobori, Shiho Hayakawa, J. Shan Hwang, Chiemi Nishimiya-Fujisawa, Mamiko Hirose and Sumiko Minobe. Financial support from the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Space Agency, the Sumitomo Foundation, Deutsche Forschungsgemeinschaft (DFG) and NIH are also appreciated.

Conflict of Interest

No conflict of interest has been declared by T. Fujisawa.