Foot formation in Hydra: A novel gene, anklet, is involved in basal disk formation

Mechanisms of Development 123 (2006) 352–361 www.elsevier.com/locate/modo

Foot formation in Hydra: A novel gene, anklet, is involved in basal disk formation Yasuko Amimoto 1, Rie Kodama 1, Yoshitaka Kobayakawa * Department of Biology, Faculty of Science, Kyushu University, Ropponmatsu 4-2-1, Chuo-ku, Fukuoka 810-8560, Japan Received 8 December 2005; received in revised form 15 February 2006; accepted 6 March 2006 Available online 27 March 2006

Abstract We isolated a novel gene by a differential-display RT-PCR method comparing basal disk tissue and peduncle tissue in a species of Hydra, Pelmatohydra robusta, and we referred to it as anklet. The putative anklet product has a signal sequence in its N-terminus, and it has one MAC/PF domain and one EGF domain. In normal hydra, the expression of anklet was restricted in the periphery of the basal disk and the lowest region of the peduncle. In foot-regenerating animals, anklet was first expressed in the newly differentiated basal disk gland cells at the regenerating basal end, and then expression became restricted at the periphery of the regenerated basal disk and in the lowest region of the peduncle. This spatially specific expression pattern suggested that the product of the anklet gene plays a role in basal disk formation. We therefore examined the role played by the protein product of the anklet gene by suppressing the transcription level of anklet using an RNA-mediated interference (RNAi) method. Suppression of the level of expression of the anklet gene led to a decrease in basal disk size in normal hydra, and to a delay in basal disk regeneration in footamputated animals. These results suggested that anklet is involved in the formation and maintenance of the basal disk in hydra. q 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Hydra; Anklet; Size regulation; Basal disk formation; RNAi

1. Introduction Size regulation mechanisms are indispensable for the development of all structures (e.g. tissues and organs) in animals, and such mechanisms are also necessary to maintain consistent forms and/or patterns. Therefore, the investigation of size regulation mechanisms in certain tissues or organs in animals is among the most compelling subjects in developmental biology. For the study of such size regulation mechanisms in animal development, basal disk formation and regeneration in hydra are considered as suitable systems. Hydra have one axis from the head to the foot, i.e. the apical–basal axis. Under steady-state conditions, the ectodermal epithelial cells proliferate primarily in the gastric column region and then they are transported to each end, i.e. to the head and foot areas (Campbell, 1967a, b; Du¨bel et al., 1987). The ectodermal epithelial cells then differentiate according to their position along this apical–basal polarity. At the apical end of

* Corresponding author. Tel.: C81 92 726 4760; fax: C81 92 726 4644. E-mail address: [email protected] (Y. Kobayakawa). 1 These authors contributed equally to the work described.

0925-4773/$ - see front matter q 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2006.03.002

the animal, there is a hypostome with a mouth opening. Several tentacles surrounding the hypostome form a tentacle ring. Most of the ectodermal epithelial cells that migrate from the gastric column to the apical end will terminally differentiate into battery cells that cover the surface of the tentacles. At the basal end of the animal, the ectodermal epithelial cells terminally differentiate specifically into mucus-secreting cells, i.e. basal disk gland cells. Using the basal disk, the hydra attaches on a substrate such as waterweed. The basal disk of a hydra is shaped in the form of a simple disk of constant size, and it consists of two layers of epithelial cells, i.e. the endodermal epithelium and the ectodermal epithelium. The basal disk gland cells, which are the ectodermal epithelial cells in the basal disk, differentiate constantly from the ectodermal epitheliomuscular cells in the lowest end of the peduncle (Bode et al., 1986; Du¨bel et al., 1987). On the other hand, older basal disk gland cells are discarded continuously from the central part of the basal disk, the aboral pore. Therefore, the basal disk can maintain its size (Bode and Bode, 1984a; Kobayakawa and Kodama, 2002). When a hydra undergoes amputation of its basal disk, the animal immediately initiates basal disk regeneration. Within several days, a basal disk of the original size will be

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

regenerated, and then the animal will maintain the size of this new basal disk. Histochemical and immunohistochemical methods have enabled investigators to distinguish easily between the differentiated basal disk gland cells and the ectodermal epithelial cells in the peduncle (Hoffmeister and Schaller, 1985; Amano et al., 1997). As mentioned above, basal disk formation in hydra is among the best systems to investigate pattern formation and size regulation mechanisms in animal development. Due to its simple structure and remarkable regeneration activities, the hydra has been the subject of intensive investigations of pattern formation in the field of developmental biology. Extensive studies of experimental morphology and theoretical analyses have been carried out using this animal. Such studies have led to putative models for pattern formation mechanisms in the hydra. According to some of these models, there are two positional value gradients along the axis of the body, one with a maximum in the head and the other with a maximum in the foot. Two inhibitory gradients along the axis of the body, i.e. a head inhibitor gradient and a foot inhibitor gradient, have also been suggested (see Bode and Bode (1984b) and Wolpert et al. (2002) for a review). Models of hydra pattern formation mechanisms have improved with time (e.g. Meinhardt, 1993); quite recently, Berking (2003) proposed a novel model, and this model accounts for the entire bud formation process using only one gradient of positional values from the head to the foot, and without relying on a longrange effective foot inhibitor. Studies of pattern formation in hydra have primarily been focused on head formation or head regeneration, and the features of foot formation have received considerably less attention than those of head formation. However, in comparison to investigative systems involving head formation, foot formation in hydra provides a simpler and more suitable system for the investigation of pattern formation and cell differentiation mechanisms. The foot of a hydra has only one simple structural extremity, the basal disk, whereas the head of the hydra has two extremities, the hypostome and the tip of the tentacle. Although the hydra has provided a suitable model for morphological study, studies of genetic methods have not been carried out using this organism. However, a number of genes that have been suggested as playing a morphogenetic role in hydra (e.g. see Hobmayer et al., 2000; Steele, 2002 for a review) have been isolated, and certain interesting peptides and genes have been reported as candidates that play some role in foot formation (Hoffmeister, 1996; Grens et al., 1996, 1999; Kumpfmu¨ller et al., 1999; Bridge et al., 2000; Harafuji et al., 2001). In addition, several studies have demonstrated the usefulness of the RNAmediated interference (RNAi) method for the suppression of the expression of certain genes to examine their roles in hydra pattern formation (Lohmann et al., 1999; Lohmann and Bosch, 2000; Smith et al., 2000; Cardenas and Salgado, 2003; Takahashi et al., 2005). The results of such studies have suggested that it is now becoming possible to investigate pattern formation mechanisms in hydra at the molecular level. In this paper, we describe a novel gene, anklet, which is expressed at the periphery of the basal disk and in the lowest

353

region of the peduncle, immediately above the basal disk. The expression patterns of anklet in normal hydra as well as in basal disk-regenerating hydra suggested that anklet is involved in basal disk formation and maintenance. In the present study, we used an RNAi approach to identify the role played by anklet in basal disk formation. 2. Results 2.1. Structure of anklet In order to elucidate the mechanism of foot formation in hydra at the molecular level, we attempted to detect genes that are specifically expressed in the foot region, i.e. in the peduncle and the basal disk. Using a differential display RT-PCR method to examine tissue samples obtained from the peduncle and the basal disk, we isolated a cDNA fragment of interest; then, using a 3 0 -RACE and 5 0 -RACE method, we were able to determine the complete cDNA sequence of this gene (Fig. 1A). In accord with the pattern of expression of this gene, we designated this gene as anklet. The cDNA of anklet is thought to encode a single putative protein with a sequence of 618 amino acids (Fig. 1A). Database searches (SMART2 and BLAST) suggested that the predicted anklet protein has a signal sequence in its N-terminus, and that it has one MAC/PF (Membrane attack complex/Perforin) domain and one EGF domain (Fig. 1B). We confirmed the translation of anklet using an in vitro translation system, the Reticulocyte Lysate System (Promega) (Fig. 1C). Hydrophobicity analysis suggested that the protein encoded by anklet is soluble. A similarity search of the databases revealed that certain predicted proteins have a structure similar to that of anklet, i.e. there is one MAC/PF domain and one EGF domain in the putative protein counterparts in rats, mice, and humans. Unfortunately, the respective functions of these proteins have not yet been investigated. To date, there have been reports of BMP/retinoic acid-inducible, neural-specific protein-2 in rats and mice; such reports have indicated that these proteins perform some function(s) in the nervous system. In addition, AvTX-60A, a novel protein with one MAC/PF domain and one EGF domain, was recently identified in the nematocysts of the sea anemone Actineria villosa; moreover, the toxicity of AvTX-60 was confirmed in a previous study (Oshiro et al., 2004). 2.2. The anklet gene is expressed at the boundary between the peduncle and the basal disk Whole-mount in situ hybridisation using DIG-labeled RNA probes revealed that anklet exhibits a spatially specific expression pattern, namely, anklet expression was detected at the boundary between the peduncle and the basal disk. This study also revealed that anklet was expressed in ectodermal epithelial cells. In the basal disk, anklet-expressing cells were found only at the periphery, but not at the center. In the lowest peduncle region that shares a border with the basal disk, high levels of expression of anklet were observed; moreover, with increasing distance from the basal disk, anklet expression levels decreased (Fig. 2A).

354

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

Fig. 1. Analysis of the cDNA sequence and the deduced amino acid sequence of anklet. (A) cDNA sequence of anklet with the predicted amino acid sequence of an open reading frame. One asterisk (*), two asterisks (**), and three asterisks (***) indicate the signal sequence, the MAP/PF domain, and the EGF domain, respectively. (B) Domain structure of the predicted protein which is encoded by anklet. (C) Western blot analysis of GST-fused in vitro translated anklet protein.

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

355

Fig. 2. Whole-mount in situ hybridisation of anklet in an adult hydra, and in a bud. Immunofluorescence of AE03-positive, differentiated basal disk gland cells in the forming basal disk of a bud. (A) Expression of anklet was observed at the periphery of the basal disk, and in the lowest region of the peduncle, just above the basal disk. A gradient pattern of expression was observed. (B) In a forming bud basal disk, anklet expression was first detected at the base of the bud closest to the basal (aboral) end of the parental animal. Bars indicate 100 mm. (C) AE03-positive, differentiated basal disk gland cells also initially appeared in an asymmetrical manner at the base of the bud closest to the basal end of the parental animal.

Upon the initiation of basal disk formation in buds, anklet expression was not symmetrical around the bud axis. Initially, anklet expression was detected at the base of the bud closest to the basal (aboral) end of the parental animal (Fig. 2B). Then, the anklet expression pattern assumed a ring-like shape around the bud axis, as is typical for these animals. We also ascertained that basal disk gland cell differentiation in the developing buds also occurred in an asymmetrical manner, as determined by immunohistochemical staining using the monoclonal antibody AE03, which specifically recognizes basal disk gland cells. Basal disk gland cells appeared first at

the base of the bud closest to the basal end of the parental animal (Fig. 2C). This interesting asymmetry in the bud’s basal disk formation will be discussed latter. 2.3. Expression of anklet in newly differentiated basal disk gland cells in foot-regenerating hydra Changes in the anklet expression pattern during foot regeneration were examined by whole-mount in situ hybridisation. The feet of the animals were amputated at the middle of the peduncle and were then allowed to regenerate for certain

Fig. 3. Expression pattern of anklet in basal disk-regenerating animals with an amputated foot at the middle peduncle. (A) Eight hours after the amputation of the basal disk. (B) Twelve hours after amputation. (C and C 0 ) 24 h after amputation. In some cases, the entire regenerating basal disk area expressed anklet (C); and in some cases, no anklet expression was detected in the central region of the regenerating basal disk (C 0 ). (D) Thirty-six hours after amputation. The expression pattern of anklet appeared similar to that in normal hydra (Fig. 2A). The bar indicates 100 mm. (E) Changes in the spatial expression pattern of anklet during foot regeneration were schematized.

356

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

amounts of time (8,12, 24, and 36 h). Sample animals were then fixed in order to detect anklet expression. Here, no expression of anklet was detected at the regenerating basal end of the samples that had been allowed to regenerate for 12 h (Fig. 3B). However, in the 24-h regenerated animals, anklet expression was initially detected. In some of the 24-h regenerated animals, anklet expression appeared as a field that covered the regenerating basal ends (Fig. 3C). However, other animals from the 24-h regeneration group had already lost anklet expression in the central region of the regenerated basal disk (Fig. 3C 0 ). In the 36-h regeneration group, anklet was expressed at the boundary between the peduncle and the regenerated basal disk, as would be expected in normal animals (Fig. 3D). Changes in the expression pattern of anklet during foot regeneration has been schematized in Fig. 3E. At first, anklet expression was detected in the entire regenerated basal disk region. As regeneration proceeded, anklet expression vanished in the central region of the regenerated basal disk. When basal disk regeneration was complete, expression was restricted to the boundary between the peduncle and the basal disk, as was observed in the normal controls. 2.4. Depression of anklet transcripts in intact hydra as determined by RNA-mediated interference (RNAi) The expression pattern of anklet in steady-state hydra and in the foot-regenerating hydra suggested that anklet is involved in

the basal disk meintanance and regeneration. Therefore, we attempted to characterize the involvement of anklet in the formation and maintenance of the basal disk by inhibiting anklet transcription using an RNAi protocol described by Lohmann et al. (1999). We synthesized a dsRNA corresponding to the full-length cDNA of anklet. This dsRNA was introduced into the entire hydra by electroporation. Three days after the electroporation of the anklet dsRNA, we ascertained by whole-mount in situ hybridisation whether or not the expression of anklet had been suppressed. We found that anklet expression was indeed suppressed in the animals that had been subjected to electroporation with the anklet dsRNA (Fig. 4A), whereas no such suppression was observed in the control animals that had been subjected to electroporation without any dsRNA (Fig. 4B) or with luciferase dsRNA (Fig. 4C); these results are consistent with those reported by Lohmann et al. (1999). However, not all animals that had been electroporated with anklet dsRNA exhibited suppressed anklet expression. In some of the experiments, some sample animals did express certain levels of anklet (marked with an asterisk in Fig. 4A), although they had been subjected to electroporation with anklet dsRNA. In the experiment shown in Fig. 4A, seven of nine samples exhibited suppressed expression of anklet (four representative samples are shown in Fig. 4A). On the average, 78% of samples, that had been electroporated with anklet dsRNA, exhibited suppressed anklet expression (five experiments, SDZ20%, nZ5–9).

Fig. 4. (A) The expression of anklet was examined by in situ hybridisation in anklet dsRNA-introduced animals. The expression of anklet in control animals that had been subjected to electroporation in DW without dsRNA (B), and anklet expression in luciferase dsRNA-introduced animals (C) was also examined by in situ hybridisation. The introduction of the dsRNA of anklet suppressed the expression of anklet, but this was not observed with electroporation only (B) or with the introduction of luciferase dsRNA (C) in control animals. However, some samples did not exhibit suppressed anklet expression (asterisk in A) with the introduction of the dsRNA of anklet. The bar indicates 100 mm.

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

357

Fig. 5. In animals that exhibited RNAi-suppressed anklet expression, basal disk size was smaller than that of control animals. (A) The basal disks in anklet expression-suppressed animals. Six days after the introduction of dsRNA. The basal disks were visualized by immunofluorescence with a basal disk gland cellspecific monoclonal antibody, AE03. (B) The basal disks of animals that had been subjected to electroporation in DW without dsRNA. The bar indicates 100 mm. (C) AE03-positive area extracted and measured by NIH-image software.

The suppression of anklet expression was also observed among animals examined 6 days after electroporation, which is consistent with the results reported by Lohmann et al. (1999). However, 10 days after electroporation, most of the anklet dsRNA-introduced polyps expressed anklet at levels similar to those observed in normal controls. These results suggest that the suppression of anklet expression using this RNAi approach was maintained for at least 1 week following the introduction of the dsRNA, but that this effect did not last for a period of 10 days thereafter. Although this RNAi approach using the full-length dsRNA of anklet did not lead to the complete suppression of gene expression, as described above, our results suggest that the RNAi did suppress gene-specific expression to a certain extent, and this suppression was maintained for at least several days after the introduction of the dsRNA into the animals. We therefore, chose to examine basal disk phenotype changes in those animals into which the dsRNA of anklet had been introduced, in order to

help elucidate the role played by anklet in basal disk formation and maintenance. 2.5. Basal disk size decreased in anklet-depressed hydra We then examined the effects of anklet-depression by RNAi on the basal disk phenotype. Using the monoclonal antibody AE03, which specifically recognizes basal disk gland cells, we identifiied the basal disk area and measured this area on whole mount samples using NIH-image software (Fig. 5C). The dsRNA of anklet was introduced into intact animals by electroporation, and the samples were cultured under typical culture conditions for 6 days. Then, the samples from the dsRNA-introduced animals and the control animals (electroporated in DW without anklet dsRNA) were fixed, and the size of the basal disks was measured. The basal disks in the RNAitreated, anklet-depressed animals (Fig. 5A) were smaller than those of the control animals (Fig. 5B). In a representative

358

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

experiment, the average basal disk area measured 27,800 mm2 (SD: 8000) in anklet dsRNA-introduced animals, and this area was 44,600 mm2 (SD: 9000) in animals that had been subjected to electroporation in DW without anklet dsRNA (nZ6, P! 0.007 Student’s t-test). This reduction in the size of the basal disk in anklet dsRNAintroduced animals depended on the extent of the decrease in the number of basal disk gland cells. In other words, the size of each basal disk gland cell in the anklet expression-suppressed animals (Fig. 5A) was roughly the same as that of the control animals (Fig. 5B). 2.6. Anklet-depressed hydra and delayed basal disk regeneration after foot amputation In order to confirm the involvement of anklet in basal disk formation more directly, we examined whether or not the suppression of anklet expression would affect basal disk regeneration in the feet of amputated animals. The anklet dsRNA was introduced into the animals by electroporation and the animals were allowed to remain 4 days under the standard

Fig. 6. Delay of basal disk regeneration in anklet expression-suppressed animals. Twenty-four hours after the amputation of the basal disk, dsRNAintroduced animals were only able to form basal disks smaller than those of control animals into which luciferase dsRNA had been introduced (*P!0.02 using a t-test, nZ9). However, after 72 h, the basal disk size in the dsRNAintroduced animals did not significantly differ from that of the control animals. Each bar indicates the average relative size of the basal disk (where 100% indicates the average size of the basal disk in normal hydra), with SD.

culture conditions. Then, the feet of the anklet dsRNA-introduced animals and control animals (into which luciferase dsRNA had been introduced) were amputated at the middle peduncle. Some of the foot-amputated animals were allowed to regenerate feet for 24 h and the others were allowed to regenerate for 72 h. After the regeneration period, anklet dsRNA-introduced animals and control animals were fixed and basal disk gland cells were detected by monoclonal antibody AE03. The sizes of the regenerated basal disks were then compared. After 24 h of regeneration, anklet dsRNA-introduced animals were only able to form basal disks smaller than those of control animals (Fig. 6). However, after 72 h of regeneration, the regenerated basal disk size in the anklet dsRNA-introduced animals did not significantly differ from those of control animals. This finding might have been the result of the long time lapse, which led to a reduction in the effect of the introduced anklet dsRNA. These results suggest that the rate of basal disk regeneration was lower in those animals in whom anklet function had been suppressed by RNAi. 3. Discussion 3.1. Involvement of anklet in basal disk formation and maintenance In steady-state hydra, anklet was found to be expressed at the periphery of the basal disk and in the lowest region of the peduncle where ectodermal epithelial cells differentiate into basal disk gland cells (Fig. 2A, B). In foot-regenerating hydra, changes in the anklet expression pattern correlated with the location at which the ectodermal epithelial cells differentiated into basal disk gland cells (Fig. 3). These observations suggested that anklet might be involved in basal disk formation and maintenance. When the expression of anklet was depressed in an animal by the RNAi method, the basal disk size became smaller (Fig. 5). This reduction in basal disk size is expected to have depended on the decreased number of basal disk gland cells, as the size of each cell size had not changed. In the case of anklet expression-depressed hydra, the rate of differentiation into basal disk gland cells decreased; however, it is expected that the rate of defluxion of old basal disk gland cells remained unaffected because anklet was not originally expressed in the central region of the basal disk. Therefore, the basal disk size was reduced in anklet expression-depressed animals. The delay in basal disk regeneration after foot amputation in the anklet expression-depressed animals (Fig. 6) could also be accounted for in the same manner. Moreover, these results suggest that anklet plays a role in the promotion of the differentiation of peduncle ectodermal epithelial cells into basal disk gland cells. The predicted structure of protein which is encoded by anklet suggested some information about the function of anklet gene. The membrane attack complex/perforin (MAC/PF) is a large domain that can be integrated in the cell membrane. Complement molecules that are functional in the vertebrate immune system are known to contain this type of domain. This domain functions as a component in the formation of the membrane attack complex, which in turn results in the lysis of

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

exogenous cells or infected bacteria. Killer T-cells and natural killer cells produce the complement molecules, which are then secreted to attack exogenous cells or infected bacteria. On the other hand, the EGF domain is distributed widely in signal molecules such as Notch-Delta. Among the reported proteins that contain one MAC/PF domain and one EGF domain, only the toxicity of AvTX-60, which was found in the nematocysts of the sea anemone, Actineria villosa (Oshiro et al., 2004), was confirmed. Based on this information, it was suggested that anklet protein exerts cytotoxic activity. As discussed above, however, anklet is expressed in differentiating and newly differentiated basal disk gland cells. Therefore, it is conceivable that anklet protein plays some role in the early stages of basal disk gland cell differentiation. However, at the boundary between the peduncle and the basal disk, many drastic phenotypic changes occur in both the epithelial cells and the nerve cells. Ectodermal epithelial cells differentiate into basal disk gland cells and begin to secret mucus just after passing through that boundary. RFamide-positive neurons and Hym-176-positive neurons are densely distributed in the lower peduncle (Yum et al., 1998a, b; Grimmelikhuijzen et al., 2002). It is expected that these neurons are displaced with the ectodermal epithelial cells into the basal disk, because they extend nerve fibers between the ectodermal epithelium and form connections with the epithelial cells. However, no RFamide-positive neurons and no Hym-176-positive neurons have been observed in the basal disk. This drastic disappearance of specific nerve cells at the boundary is suggestive of a phenotypic change in the nerve cells; however, it is also possible that cell death occurs at the boundary between the peduncle and the basal disk. These findings suggest that basal disk formation and maintenance is not merely reflective of basal disk gland cell differentiation from peduncle ectodermal epithelial cells, but that it is instead a more complicated phenomenon. In this context, it is not unreasonable to assume that anklet plays a cytotoxic role and promotes basal disk formation at the boundary. Some peptides and genes have been reported as candidates that affect pattern formation in the foot region and basal disk gland cell differentiation (Hoffmeister, 1996; Grens et al., 1996, 1999; Takahashi et al., 1997; Bridge et al., 2000; Harafuji et al., 2001; Steele, 2002 for review). An astacin matrix metalloprotease has been reported as a target gene for the hydra foot activator peptides Hym-346 and Hym-323 (Kumpfmu¨ller et al., 1999). Recently, Thomsen et al. (2004) suggested that CnNK-2 and pedibin orchestrate foot-specific differentiation, as demonstrated by in vitro electrophoretic mobility shift assay. It will be intriguing to investigate how anklet responds to these molecules in the course of the basal disk formation process.

359

appearance of an asymmetrical structure in hydra is an intriguing phenomenon with respect to investigation of the origins of bilateral symmetry. In this study, we demonstrated that during basal disk formation in the bud, the expression of anklet was initially asymmetrical. Expression was observed only on one side of the developing basal disk, i.e. on the side closer to the aboral end of the parental animal (Fig. 2B). This asymmetry of early anklet expression is consistent with the asymmetrical appearance of the basal disk gland cells in the early stages of basal disk formation in the bud (Fig. 2C). There is yet another indication of asymmetry in the developing buds in Pelmatohydra robusta. In this species of Hydra, the tentacles begin to form at the side of bud closer to the parent animal’s oral end. At first, two tentacles appear on that side of the developing bud’s head. Therefore, at this stage, the buds of Pelmatohydra robusta exhibit bilateral symmetry. Asymmetrical foot formation in the bud has been reported and thoroughly discussed by Bridge et al. (2000), who clearly presented the strikingly asymmetrical expression of shin guard and manacle in the initial stage of development of the feet in buds in Hydra vulgaris. Bridge et al. also elegantly demonstrated that the manacle expression pattern was inverted in the developing bud foot of a polyp with a flipped bud field. The extreme edges of the bud field develop into the basal disk of the bud. Based on these results, they suggested that the positional value gradient in the parental animal plays a role in bud formation. Since the side of the bud’s foot nearest the parent’s foot has a lower positional value than does the other side of the bud’s foot, the basal disk tissue would be expected to form there first. They also revealed that this asymmetrical development of a bud’s foot occurs in permanently footless animals. These observations and accounts of the findings are quite consistent with Berking’s model of budding (2003), with only one gradient of positional values from head to foot and a local increase in the positional value at the center of the bud field in the absence of long-range effective inhibitors. The asymmetrical expression pattern of anklet in initial stage of foot formation in developing buds in Pelmatohydra robusta is consistent with both Bridge’s suggestions and Berking’s model. Asymmetrical tentacle development in the bud of Pelmatohydra robusta can also be interpreted as a type of dependence on the parental positional value. It is thus reasonable to conclude that the side of the bud’s head nearest the parent’s head has a higher positional value than does the other side. This higher positional value may induce the earlier development of tentacles on the side of the bud that is closer to the parent animal’s head. 4. Methods 4.1. Animal culture

3.2. Asymmetrical expression pattern of anklet in basal disk formation in the bud Hydra have a radially symmetrical body around the basal– apical axis. Cnidarians, among which genus Hydra belongs, are considered a sister group to the bilaterians. Therefore, the

Pelmatohydra robusta was used in all experiments. Animals were maintained at 18–22 8C in a hydra culture solution (1 mM NaCl, 1 mM CaCl2, 0.1 mM KCl, 0.1 mM MgSO4, 1 mM tris–(hydroxymethyl)-aminomethane; pH 7.4, adjusted with HCl). The animals were fed with newly hatched nauplii of Artemia three times a week. Under these conditions, the animals remained asexual and reproduced by budding. The animals were starved for

360

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361

24 h before use. P. robusta is a species belonging to the oligactis group, and they clearly have a thin, long peduncle. It is therefore easy to distinguish the peduncle from the gastric column region in this species.

4.2. Differential display of RT-PCR (DD-RT-PCR) DD-RT-PCR was performed according to the standard protocol. The mRNAs were extracted from three parts of hydra, namely, the gastric column, the peduncle, and the basal disk, using a QuickPrepw Micro mRNA purification kit (Amersham Pharmacia Biotech). These mRNAs were also used for the RACE (rapid amplification of cDNA ends). To synthesize the first strand cDNA, we used the Ready-To-Go T-Primed First-Strand Kit (Amersham Pharmacia Biotech). The synthesized cDNA samples were purified using the QIA quickw PCR Purification Kit (QIAGEN). PCR amplification of cDNA was performed using five kinds of 10mer arbitrary primers (AP). The base sequences of the APs were as follows: AP1, 5 0 -AGCCAGCGAA-3 0 ; AP2, 5 0 -GACCGCTTGT-3 0 ; AP3, 5 0 -AGGTGACCGT-3 0 ; AP4, 5 0 -GGTACTCCAC-3 0 ; and AP5, 5 0 -GTTGCGATCC-3 0 . All of the PCR products were electrophoresed on an extended format 6% denaturing polyacrylamide gel. All bands were detected by silver staining. To extract the cDNAs that were differentially displayed, the silver-stained bands were excised from the gels using a scalpel and forceps. The pieces of the gel were boiled in distilled water in order to extract the cDNA fragments. Those cDNA fragments were then re-amplified by PCR. The PCR products were directly inserted into the plasmid vector pCRII using the TOPO TA Cloningw kit (Invitrogenw). Plasmids, in which PCR products had been inserted, were then transformed into TOP10-competent cells (Invitrogenw). We created DIG-labeled RNA probes from the obtained cDNA fragmentinserted plasmids, and then we confirmed their expression patterns by wholemount in situ hybridisation. To determine the full-length sequence of the obtained cDNA, we carried out RACE.

4.3. RNAi RNA-mediated interference (RNAi) was performed as described in Lohmann et al. (1999), with slight modifications. The DNA templates for RNA synthesis were generated twice by PCR using M13 Reverse and M13 Forward primers and a pCRII plasmid containing the entire coding region of anklet. The sense RNA and the antisense RNA were produced from the purified DNA template using RiboMAX Large Scale RNA Production Systems (Promega). The sense and antisense fractions were combined and heated to 65 8C for 20 min in order to eliminate any secondary structures. The complementary strands were annealed in a solution (750 mM NaCl, 75 mM sodium citrate) by cooling the samples slowly to room temperature. The size and quantity of double-stranded RNA (dsRNA) were examined by agarose gel electrophoresis. As a control, luciferase dsRNA was produced using Luciferase Control DNAs (Promega, Coleoptera luciferase). For the transfection of dsRNA in the hydra polyps, electroporation was carried out with the following modifications. Twenty to thirty polyps were washed twice in chilled DEPC–H2O. Electroporation was carried out by placing polyps in 200 ml DEPC–H2O containing 10 mg dsRNA in chilled electroporation cuvettes with a 0.4-cm gap. To minimize the possibility of degradation, the dsRNA was added into the cuvettes just prior to electroporation. Control transfections were performed without dsRNA or with luciferase dsRNA. Whole polyps were pulsed with a BioRad Gene Pulsere (Bio-Rad) adjusted to 250 V and 25 mF capacitance. After being subjected to the electric pulse, the polyps exhibited a fluffed surface and were damaged. They were then immediately transferred into chilled hydra culture solution, which was supplemented with 20% hyperosmotic dissociation medium (pH 6.9) containing 6 mM CaCl2, 1.2 mM MgSO4, 3.6 mM KCl, 12.5 mM N-tris– [hydroxymethyl]-2-aminoethanesulfonic acid, 6 mM sodium pyruvate, 6 mM sodium citrate, 6.0 mM glucose, and 50 mg/ml rifampicin. To facilitate recovery, the polyps were kept at 10 8C for up to 2 days. Two days after electroporation, the medium was exchanged for standard hydra culture solution, and the polyps were kept at 18–22 8C. Four to six days after electroporation, the polyps were used for the experiments.

4.4. Whole mount in situ hybridisation We used DIG-labeled RNA probes that corresponded to the entire coding region of the anklet cDNA for the whole mount in situ hybridisation; this was done in order to examine the expression pattern of anklet. When they were used as samples for in situ hybridisation, the animals were relaxed for 5 min in 1% urethane solution in hydra culture solution and were then fixed overnight at 4 8C in 4% paraformaldehyde in hydra culture solution. Hybridisation was performed according to the standard procedure at 55 8C overnight. The DIG-labeled RNA probes were detected with alkaline phosphatase-conjugated anti-digoxygenin Fab fragments (Roche Diagnostics). The samples were then incubated in NBT (nitro blue tetrazolium)—BCIP (5-bromo-4-chloro-3-indolyl phosphate) stock solution: NTMT buffer at 1:50 in the dark for the coloring reaction.

4.5. Immunohistochemistry to identify basal disk gland cells A monoclonal antibody, AE03 culture supernatant of the hybridoma cell line AE03 was used to identify the basal disk gland cells. This monoclonal antibody recognizes mucous granules in basal disk gland cells and also the secreted mucus (Amano et al., 1997). Therefore, it can easily detect differentiated basal disk gland cells. An indirect immunofluorescence technique was used on both whole-mount and sectioned samples in order to detect antibody binding. To compare basal disk sizes, we used NIH-image software. Each AE03-positive area was selected by setting an appropriate threshold, and the size of the area was measured (Fig. 5C).

4.6. In vitro translation and Western blotting The coding region of the anklet cDNA was ligated in pGX4T-1 expression vector (Amersham), and GST-fused anklet RNA was synthesized in vitro. For the translation of GST-fused anklet, we used an in vitro translation system, Reticulocyte Lysate System (Promega). The proteins were separated by SDSPAGE on 10% acrylamide gels, and then they were electroblotted onto nitrocellulose membranes (BioTraceNT, NIPPON Genetics Co.). The GSTfused anklet protein was detected by anti-GST antibodies with secondary peroxidase-conjugated antibody (Amersham).

Acknowledgements This work was supported by a grant from Japan Society for the Promotion of Science to Y.K. References Amano, H., Koizumi, O., Kobayakawa, Y., 1997. Morphogenesis of the atrichous isorhiza, a type of nematocyst in Hydra observed with a monoclonal antibody. Dev. Genes Evol. 207, 413–416. Berking, S., 2003. A model for budding in hydra: pattern formation in concentric rings. J. Theol. Biol. 222, 37–52. Bode, P.M., Bode, H.R., 1984a. Formation of pattern in regenerating tissue pieces of hydra attenuata II. Degree of proportion regulation is less in the hypostome and tentacle zone than in the tentacles and basal disc. Dev. Biol. 103, 304–312. Bode, P.M., Bode, H.R., 1984b. Patterning in hydra. In: Malacinski, G., Bryant, S. (Eds.), Primers in Developmental Biology, vol. 1. Macmillan, New York, pp. 213–241. Bode, H.R., Dunne, J., Heimfeld, S., Huang, L., Javois, L., Koizumi, O., et al., 1986. Transdifferentiation occurs continuously in adult hydra. Curr. Top. Dev. Biol. 20, 257–280. Bridge, D.M., Stover, N.A., Steele, R.E., 2000. Expression of a novel receptor tyrosine kinase gene and a paired-like homeobox gene provides evidence of differences in patterning at the oral and aboral ends of Hydra. Dev. Biol. 220, 253–262. Campbell, R.D., 1967a. Tissue dynamics of steady state growth in Hydra littoralis l. Patterns of cell division. Dev. Biol. 15, 487–502.

Y. Amimoto et al. / Mechanisms of Development 123 (2006) 352–361 Campbell, R.D., 1967b. Tissue dynamics of steady state growth in Hydra littoralis ll. Pattern of tissue movement. J. Morphol. 121, 19–28. Cardenas, M.M., Salgado, M., 2003. STK, the src homologue, is responsible for the initial commitment to develop head structures in Hydra. Dev. Biol. 264, 495–505. Du¨bel, S., Hoffmeister, S.A.H., Schaller, H.C., 1987. Differentiation pathways of ectodermal epithelial cells in hydra. Differentiation 35, 181–189. Grens, A., Gee, L., Fisher, D.A., Bode, H.R., 1996. CnNK-2, an NK-2 homeobox gene, has a role in patterning the basal end of the axis in hydra. Dev. Biol. 180, 473–488. Grens, A., Shimizu, H., Hoffmeister, S.A., Bode, H.R., Fujisawa, T., 1999. The novel signal peptides, pedibin and Hym-346, lower positional value thereby enhancing foot formation in hydra. Development 126, 517–524. Grimmelikhuijzen, C.J.P., Williamson, M., Hansen, G.N., 2002. Neuropeptides in cnidarians. Can. J. Zool. 80, 1690–1702. Harafuji, N., Takahashi, T., Hatta, M., Tezuka, H., Morishita, F., Matsushima, O., Fujisawa, T., 2001. Enhancement of foot formation in Hydra by a novel epitheliopeptide, Hym-323. Development 128, 437–446. Hobmayer, B., Rentzsch, F., Kuhn, K., Happel, C.M., Laue, C.C., Snyder, P., et al., 2000. WNT signaling molecules act in axis formation in the diploblastic metazoan Hydra. Nature 407, 186–189. Hoffmeister, S.A.H., 1996. Isolation and characterization of two new morphogenetically active peptides from Hydra vulgaris. Development 122, 1941–1948. Hoffmeister, S., Schaller, H.C., 1985. A new biochemical marker for footspecific cell differentiation in hydra. Roux’s Arch. Dev. Biol. 194, 453–461. Kobayakawa, Y., Kodama, R., 2002. Foot formation in Hydra: commitment of basal disk cells in the lower peduncle. Dev. Growth Differ. 44, 517–526. Kumpfmu¨ller, G., Rybakine, V., Takahashi, T., Fujisawa, T., Bosch, T.C.G., 1999. Identification of an astacin matrix metalloprotease as target gene for Hydra foot activator peptides. Dev. Genes Evol. 209, 601–607. Lohmann, J.U., Bosch, T.C.G., 2000. The novel peptide HEADY specifies apical fate in a simple radially symmetric metazoan. Genes Dev. 14, 2771–2777.

361

Lohmann, J.U., Endl, I., Bosch, T.C.G., 1999. Silencing of developmental genes in Hydra. Dev. Biol. 214, 211–214. Meinhardt, H., 1993. A model of biological pattern formation of hypostome, tentacles and foot in Hydra: how to form structures close to each other, how to form them at a distance. Dev. Biol. 157, 321–333. Oshiro, N., Kobayashi, C., Iwanaga, S., Nozaki, M., Namikoshi, M., Spring, J., Nagai, H., 2004. A new membrane-attack complex/perforin (MACPF) domain lethal toxin from the nematocyst venom of the Okinawan sea anemone Actineria villosa. Toxicon 43, 225–228. Smith, K.M., Gee, L., Bode, H.R., 2000. HyAlx, an aristaless-related gene, is involved in tentacle formation in hydra. Development 127, 4743–4752. Steele, R.E., 2002. Developmental signaling in Hydra: what does it take to build a ‘Simple’ animal? Dev. Biol. 248, 199–219. Takahashi, T., Muneoka, Y., Lohmann, J., de Haro, L.M., Solleder, G., Bosch, T.C.G., et al., 1997. Systematic isolation of peptide signal molecules regulating development in hydra: LWamide and PWfamilies. Proc. Natl Acad. Sci. USA 94 (4), 1241–1246. Takahashi, T., Hatta, M., Yum, S., Gee, L., Ohtani, M., Fujisawa, T., Bode, H.R., 2005. Hym-301, a novel peptide, regulates the number of tentacles formed in hydra. Development 132, 2225–2234. Thomsen, A., Till, A., Wittlieb, J., Beetz, C., Khalturin, K., Bosch, T.C.G., 2004. Control of foot differentiation in Hydra: in vitro evidence that the NK2 homeobox factor CnNK-2 autoregulates its own expression and uses pedibin as target gene. Mech. Dev. 121, 195–204. Wolpert, L., Beddington, R., Jessel, T., Lawrence, P., Meyerowitz, E., Smith, J., 2002. Regeneration in Hydra, Principles of Development, second ed. Oxford University Press, Oxford, pp. 458–465. Yum, S., Takahashi, T., Hatta, M., Fujisawa, T., 1998a. The structure and expression of a preprohormone of a neuropeptide, Hym-176 in Hydra magnipapillata. FEBS Lett. 439, 31–34. Yum, S., Takahashi, T., Koizumi, O., Ariura, Y., Kobayakawa, Y., Mohri, S., Fujisawa, T., 1998b. A novel neuropeptide, Hym-176, induces contraction of the ectodermal muscle in Hydra. Biochem. Biophys. Res. Commun. 248, 584–590.