Planarian fibroblast growth factor receptor homologs expressed in stem cells and cephalic ganglions

Authors

  • Kazuya Ogawa,

    1. Laboratory of Regeneration Biology, Department of Life Science, Faculty of Science, Himeji Institute of Technology, Harima Science Garden City, Akou, Hyogo 678-1297,
    Search for more papers by this author
  • Chiyoko Kobayashi,

    1. Evolutionary Regeneration Biology Group, RIKEN Center for Developmental Biology, Kobe, Hyogo 650-0047 and
    Search for more papers by this author
  • Tetsutaro Hayashi,

    1. Evolutionary Regeneration Biology Group, RIKEN Center for Developmental Biology, Kobe, Hyogo 650-0047 and
    Search for more papers by this author
  • Hidefumi Orii,

    1. Laboratory of Regeneration Biology, Department of Life Science, Faculty of Science, Himeji Institute of Technology, Harima Science Garden City, Akou, Hyogo 678-1297,
    Search for more papers by this author
  • Kenji Watanabe,

    1. Laboratory of Regeneration Biology, Department of Life Science, Faculty of Science, Himeji Institute of Technology, Harima Science Garden City, Akou, Hyogo 678-1297,
    Search for more papers by this author
  • Kiyokazu Agata

    Corresponding author
    1. Evolutionary Regeneration Biology Group, RIKEN Center for Developmental Biology, Kobe, Hyogo 650-0047 and
    2. Laboratory of Regeneration and Developmental Biology, Department of Biology, Faculty of Science, University of Okayama, 3-1-1 Tsushima-naka, Okayama, 700-8530, Japan
    Search for more papers by this author

* Author to whom all correspondence should be addressed. Email: agata@cdb.riken.go.jp

Abstract

The strong regenerative capacity of planarians is considered to reside in the totipotent somatic stem cell called the ‘neoblast’. However, the signal systems regulating the differentiation/growth/migration of stem cells remain unclear. The fibroblast growth factor (FGF)/FGF receptor (FGFR) system is thought to mediate various developmental events in both vertebrates and invertebrates. We examined the molecular structures and expression of DjFGFR1 and DjFGFR2, two planarian genes closely related to other animal FGFR genes. DjFGFR1 and DjFGFR2 proteins contain three and two immunoglobulin-like domains, respectively, in the extracellular region and a split tyrosine kinase domain in the intracellular region. Expression of DjFGFR1 and DjFGFR2 was observed in the cephalic ganglion and mesenchymal space in intact planarians. In regenerating planarians, accumulation of DjFGFR1-expressing cells was observed in the blastema and in fragments regenerating either a pharynx or a brain. In X-ray-irradiated planarians, which had lost regenerative capacity, the number of DjFGFR1-expressing cells in the mesenchymal space decreased markedly. These results suggest that the DjFGFR1 protein may be involved in the signal systems controlling such aspects of planarian regeneration as differentiation/growth/migration of stem cells.

Introduction

Planarians are well known for their strong regenerative capacity. This capacity is considered to reside in the totipotent somatic stem cell called the ‘neoblast’. Over the past several decades, the neoblast has been defined based on its morphologic characteristics. Many electronmicroscopic studies have shown that the neoblast has the morphologic characteristics typical of undifferentiated cells. It is small in size, round to ovoid or teardrop in shape and possesses a high nucleus/cytoplasm ratio. Furthermore, it contains chromatoid bodies in its cytoplasm, which are characteristic of totipotent germ line cells (Pederson 1959; Morita 1967; Hori 1982). Recently, our group demonstrated that a planarian vasa-related gene is specifically expressed in cells containing the chromatoid body (Shibata et al. 1999). The totipotency of the neoblast is supported by X-ray irradiation experiments. Lack of regenerative ability after X-ray irradiation is caused by progressive disappearance and lack of turnover of neoblasts, whereas organisms with an unirradiated region can form blastemas (Wolff & Dubois 1948). Such X-ray-irradiated planarians can be rescued by injection of a neoblast-enriched fraction that is proposed by serial filtration and Ficoll density centrifugation gradient (Baguña et al. 1989). However, the signal systems regulating the differentiation/growth/migration of neoblasts remain unclear because the totipotency of neoblasts has not been examined experimentally.

It is well known that the fibroblast growth factor (FGF)/FGF receptor (FGFR) signal system is involved in a variety of signaling cascades regulating embryogenesis and organogenesis. Genetic studies in the mouse have demonstrated that FGF signaling is required for cell proliferation/survival at the time of embryo implantation and for cell migration during gastrulation (Feldman et al. 1995; Ciruna et al. 1997; Arman et al. 1998; Sun et al. 1999). At later stages of embryogenesis, FGFs regulate the formation of the brain, limbs, lungs, teeth and many other organs (Goldfarb 1996; Szebenyi & Fallon 1999). Fibroblast growth factor has been found to play a critical role in mesoderm induction during Xenopus development (Slack et al. 1987; Kimelman et al. 1988). Disruption of FGF signaling by expression of a dominant-negative mutant of Xenopus FGFR has been shown to inhibit the formation of mesodermal tissue (Amaya et al. 1991). Fibroblast growth factor and FGFR genes have also been identified in Drosophila and are required for the migration of mesoderm and glial cells and for the control of branching events in the tracheal system (Skaer 1997). In Caenorhabditis elegans, FGF signaling is required for the migration of sex myoblasts (Chen & Stern 1998).

At least four distinct FGFRs constituting a family of receptor-tyrosine kinases have been identified (FGFR1, FGFR2, FGFR3 and FGFR4) in vertebrates, with many additional structural variants resulting from alternative splicing. The basic structure is an extracellular region, with more than two immunoglobulin-like domains, a transmembrane region and tyrosine kinase domains, which resembles the basic structure of the platelet-derived growth factor receptor (PDGFR) and colony stimulating factor-1 receptor (CSF-1R). However, unlike PDGFR and CSF-1R, FGFRs possess additional characteristics, such as a relatively long juxtamembrane region, kinase catalytic sequences split by a short stretch of amino acids and a short carboxyl terminal tail. In invertebrates, FGFRs have been found in Drosophila, C. elegans and sea urchins (Glazer & Shilo 1991; Klambt et al. 1992; DeVore et al. 1995; McCoon et al. 1996). Their overall sequence identity with the vertebrate genes is only approximately 40%.

In previous experiments (Ogawa et al. 1998), we cloned polymerase chain reaction (PCR) fragments from cDNA of planarians using kinase domain-specific degenerate primers. Two of the 12 PCR fragments (DjPTK1 and DjPTK3) isolated were found to encode polypeptides highly homologous in amino acid sequence to both invertebrate and vertebrate FGFRs. For clarification of the above findings and further extension of our analysis, we determined the complete nucleotide sequences of the coding regions of the DjPTK1 and DjPTK3 genes and examined their expression patterns in intact and regenerating planarians. DjPTK1 and DjPTK3 were found to be planarian FGFR homologs, having two or three Ig domains with split tyrosine kinase domains, and were therefore renamed DjFGFR2 and DjFGFR1, respectively. Expression of both DjFGFR1 and DjFGFR2 was observed in the brain; DjFGFR1 was also expressed in stem cells in the mesenchymal space.

Materials and Methods

Organisms

Asexual-state planarians Dugesia japonica (GI), derived from the Irima river in Gifu, Japan, and established in our laboratory were analyzed. All worms used in experiments had undergone 1 week of starvation.

Screening of DjFGFR1 cDNA

The longest cDNA of DjFGFR1 was obtained by the stepwise dilution method, as described by Watanabe et al. (1997), from a cDNA library constructed from poly A+ RNA of whole planarians in λZAPII vector (Stratagene, La Jolla, CA, USA). The sense primer TATTCTCGTGGGTGAACACTTTG-3′ and antisense primer TTGTGTATTTCTTTTCAAGAAGTGC (each ritten in the 5′ to 3′ orientation), which correspond to the sequence of DjPTK3, were used for PCR screening of the cDNA library (Ogawa et al. 1998). The positive cDNA clone with the longest insert was recloned into pBluescript according to the manufacturer's protocol (Stratagene) and sequenced by using a Thermo Sequence cycle sequencing kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and an automatic DNA sequencer (DSQ 1000 L; Shimazu, Tokyo, Japan).

Cloning of the DjFGFR1 and DjFGFR2 coding sequences by 5′RACE

Because screening of the cDNA library yielded partial DjFGFR1/DjPTK3 and DjFGFR2/Dj PTK1 clones (Ogawa et al. 1998), the 5′-extremities of the open reading frame (ORF) were isolated by the Rapid amplification of cDNA ends (RACE)–PCR method (Innis et al. 1990). For 5′RACE, regenerating planarian cDNA was G-tailed using terminal deoxynucleotidyl transferase (Invitrogen, Carlsbad, CA, USA). A first PCR of 30 cycles of 1 min at 94°C, 1 min at 55°C, 3 min at 72°C was performed using (dC)17 primer as the forward primer and ATCAAGAATTTTATTTTCGTTTACATGG as the reverse primer, which corresponds to the sequence of DjFGFR1, or AAATTAAGGCCATGACTCGATCCG (each are written in the 5′ to 3′ orientation), which corresponds to the sequence of DjFGFR2. Two microliters of the PCR products were then used as templates for a second amplification of 40 cycles of (1 min at 94°C, 1 min at 55°C, 3 min at 72°C) using (dC)17 primer and nested reverse primers AGAGTCGACTGCTGTAACATGATCCG, which corresponds to the sequence of DjFGFR1 and TTCCATTCTTGTTGAAAACAAAACACG (each are written in the 5′-to 3′ orientation), which coressponds to the sequence of DjFGFR2. Both products were cloned into TA vector pGM-T Easy (Promega, Madison, WI, USA) and sequenced. Sequence alignments and phylogenetic tree analyses were performed using the ClustalW program (DNA Data Bank of Japan (DDBJ), Mishima, Shizuoka, Japan).

In situ hybridization

A digoxigenin-labeled RNA probe was prepared according to the manufacturer's protocol (Boehringer, Mannheim, Germany), with the DjFGFR1 or DjFGFR2 full-length cDNA as a template. Whole-mount in situ hybridization was performed as described by Umesono et al. (1999). For fixation, the relaxant solution was 1% HNO3, 2.25% formalin and 50 µm MgSO4 in modified Holtfleter solution. Fixed samples were embedded in paraffin and serially sectioned at 4 µm. In situ hybridization of sections was performed as described by Kobayashi et al. (1998). Cell nuclei were labeled with Hoechst No. 33342 (Sigma, St Louis, MO, USA).

X-Ray irradiation

Worms placed on wet filter paper on ice were irradiated with 12 R X-rays using an X-ray generator (SOFTEX B-4; SOFTEX, Tokyo, Japan). Under these conditions, irradiated worms could not regenerate after amputation and died within 20 days (Ito et al. 2001; Kato et al. 2001). After worms had recovered for 7 days, they were cut transversely in the pre- or post-pharyngeal region, allowed to regenerate at approximately 22°C and fixed for in situ hybridization of tissue sections after 1 day. Intact planarians were allowed to recover for 10 days after X-ray irradiation.

Synthesis of double-strand RNA

Double-strand RNA (dsRNA) was basically synthesized as described previously(Sánchez Alvarado & Newmark 1999). pBluescriptII SK+ containing the appropriate cDNA inserts was linearized for in vitro transcription. Clone DjFGFR1, 2246 bp in length (nucleic acid sequence position 197–2442; Fig. 1), was digested with SacI or KpnI to synthesize antisense (T7) or sense (T3) RNA, respectively; clone DjFGFR2, 2304 bp in length (nucleic acid sequence position 724–3028, Fig. 2), was digested with BamI or XhoI to synthesize antisense (T7) or sense (T3) RNA, respectively. The RNA were denatured for 20 min at 65°C and annealed for 40 min at 37°C. After ethanol precipitation, dsRNA was resuspended in diethyl pyrocarbonate (DEPC)-treated H2O. Electrophoretic mobilities of dsRNA and single-strand (ss) RNA were assessed in 1.5% Agarose gels.

Figure 1.

Nucleotide and deduced amino acid sequences of DjFGFR1. The box at the N-terminus indicates the predicted signal peptide. Conserved Cys residues in the Ig domain are circled. Seven potential N-glycosylation sites are shaded. The shaded box indicates the predicted transmembrane region. The tyrosine kinase catalytic sequence is boxed. The conserved Tyr residue in the C-terminal tail is indicated by a triangle. The full-length cDNA sequence of DjFGFR1 has been deposited in DDBJ/EMBL/GenBank databases (http://www.ddbj.nig.ac.jp) with the accession number AB074425.

Figure 2.

Nucleotide and deduced amino acid sequences of DjFGFR2. The box at the N-terminus indicates the predicted signal peptide. Conserved Cys residues in the Ig domain are circled. Six potential N-glycosylation sites are shaded. The shaded box indicates the predicted transmembrane region. The tyrosine kinase catalytic sequence is boxed. The conserved Tyr residue in the C-terminal tail is indicated by a triangle. The full-length cDNA sequence of DjFGFR2 has been deposited in DDBJ/EMBL/GenBank databases (http://www.ddbj.nig.ac.jp) with the accession number AB074426.

Microinjection and amputation

Intact planarians were injected with dsRNA three times (32 nL/injection) for 3 consecutive days using a Drummond Scientific Nanoject injector (Broomall, PA, USA). Control animals were injected with H2O. Between 2 and 3 h after the last injections, animals were amputated at different levels along the anteroposterior axis using sterile surgical blades (Keisei Medical Industrial, Tokyo, Japan). Three kinds of regenerating fragments were obtained after amputation: (i) head fragments capable of regenerating a new central and tail regions; (ii) trunk fragments able to regenerate new head and tail regions; and (iii) tail fragments capable of regenerating a new head and central regions. During the regeneration process, samples were maintained at 21°C. Brain regeneration was analyzed according to Cebriàet al. (2002).

Fluoresence-activated cell sorting analysis

Single-cell suspensions from dsRNA-injected planarians were prepared by trypsin digestion and gently pipetting in modified Holtfreter solution. After filtration, single-cell suspensions were incubated with Hoechst No. 33342 (Sigma) for 2 h at 20°C. Samples were sorted with fluoresence-activated cell sorting (FACS) Vantage SE (Becton Dickinson, San Jose, CA, USA). Data analysis was performed using CELL Quest software ver. 3.3 (Becton Dickinson).

Results

Cloning and characterization of DjFGFR1 cDNA

To characterize DjPTK3, the longest cDNA for each gene was isolated from a cDNA library containing 3.5 × 106 independent clones in λZAPII vector by stepwise dilution screening (Watanabe et al. 1997). The longest cDNA for DjPTK3 contained a 2197 bp insert with an ORF encoding 716 amino acids. Subsequently, the missing 5′ region was obtained by the 5′RACE method and a full-length sequence encoding 854 amino acids of DjPTK3 was thereby obtained. The predicated amino acid sequence of DjPTK3 contained a signal peptide, an entire extracellular region, a transmembrane region and an entire intracellular region (Figs 1,3).

Figure 3.

Diagram showing structures of DjFGFR1, DjFGFR2, Drosophila fibroblast growth factor receptor (DFR1) and human fibroblast growth factor receptor (FLG).

The intracellular region of DjPTK3 contains a relatively long juxtamembrane domain (amino acid positions 407–545), a split tyrosine kinase catalytic sequence (545–818) and a short carboxyl-terminal tail (818–854; Fig. 1). All tyrosine kinase subdomains (I to XI; Hanks et al. 1988) were conserved. Except for the kinase insert region, the overall sequence homology between the kinase domain of DjPTK3 and that of FGFRs from other animals was estimated to be approximately 45%. The carboxyl-terminal tail of DjPTK3 includes a tripeptide sequence, Tyr-Leu-Glu, which may provide a tyrosine phosphorylation site supposedly required for binding of SH2 of phosphoinositide-specific phospholipase C-γ (Mohammadi et al. 1991). To clarify whether DjPTK3 belonged to the FGFR or PDGFR family, we performed phylogenetic analyses using the kinase domain, except for the insertion region. These analyses suggested that DjPTK3 could be included within the FGFR family (Fig. 4). Thus, we renamed DjPTK3 as DjFGFR1.

Figure 4.

Phylogenetic relationship among the fibroblast growth factor receptor (FGFR) and platelet-derived growth factor receptor (PDGFR) families. A closer view of the vertebrate FGFR and PDGFRa sub-trees is shown on the right. The tree was drawn by the NJ method. The numbers show bootstrap values.

The extracellular domain of FGFRs is, in general, characterized by a series of Ig domains, each containing two Cys residues, possibly connected by a disulfide-bound β-sheet structure (Williams & Barclay 1988). The consensus sequence for the Ig domains is Vx(I/L)xC(8x-12x)W (20x-50x)DxGxYxCx (where x is an arbitary amino acid). The DjFGFR1 protein contains three Ig domains (most FGFRs have two or three Ig domains). However, there was no short stretch of acidic amino acids in the DjFGFR1 extracellular domain. One and two of seven potential N-glycosylation sites of DjFGFR1 were conserved in DFR1 and flg, respectively (Ruta et al. 1989; Shishido et al. 1993).

Cloning and characterization of DjFGFR2 cDNA

In previous experiments (Ogawa et al. 1998), the longest cDNA found for DjPTK1 did not contain either a putative signal peptide sequence or a probable initial methionine. Therefore, we performed 5′RACE and thereby obtained a full-length sequence encoding 887 amino acids of DjPTK1. The predicated amino acid sequence of DjPTK1 contained a signal peptide, an entire extracellular region, a transmembrane region and an entire intracellular region (Figs 2,3).

As noted for DjFGFR1, the intracellular region of DjPTK1 contained a split tyrosine kinase catalytic domain, highly homologous to those of other animal FGFRs (Fig. 4). Thus, DjPTK1 was renamed DjFGFR2. The Tyr-Leu-Glu sequence was found near the carboxyl terminus of DjFGFR2. The extracellular region included two Ig-domains, but no acidic domains. The locations of two N-glycosylation sites were found to be conserved between DjFGFR1 and DjFGFR2. The similarity between the DjFGFR1 and DjFGFR2 kinase domains was 50%.

Expression of DjFGFR1 and DjFGFR2 in intact planarians

To investigate the expression patterns of the DjFGFR1 and DjFGFR2 genes in intact planarians, we performed whole-mount in situ hybridization. Expression of both DjFGFR1 and DjFGFR2 was observed in the brain and mesenchymal space (Fig. 5A,B). The intensity of the positive signals of DjFGFR1 in the mesenchymal space was stronger than that of the DjFGFR2 signals. The localization of the cells expressing DjFGFR1 and DjFGFR2 was analyzed more precisely by in situ hybridization. The expression of both DjFGFR1 and DjFGFR2 was observed in the brain, especially intensely in the posterior region of the cephalic ganglion (Fig. 5C,D; Agata et al. 1998). Expression of DjFGFR2 could not be detected in the mesenchymal space of sections with our in situ hybridization procedures (data not shown), although we think that DjFGFR2 may be expressed at a low level in the cells of the mesenchymal space. In contrast, expression of DjFGFR1 was clearly observed in a certain population of cells in the mesenchymal space (Fig. 5E). Higher magnification views revealed that these cells were small and round or ovoid or teardrop shaped (Fig. 5F). Figure 5 shows counter-nuclear staining of Fig. 5F using Hoechst No. 33342. The positive cells had a small cytoplasmic space. These observations suggest that DjFGFR1-expressing cells in the mesenchymal space may have characteristics identical to those of stem cells.

Figure 5.

Whole-mount in situ hybridization of DjFGFR1 and DjFGFR2 in intact planarians. (A) Expression of DjFGFR1. (B) Expression of DjFGFR2. Both DjFGFR1 and DjFGFR2 are expressed in the cephalic ganglion and mesenchymal space. (C) Expression of Djsyt (Tazaki et al. 1999) is seen in the cephalic ganglion (cg). (D–G) Spatial expression pattern of DjFGFR1 in intact planarians. (D) A transverse section of a head region. (E) Higher magnification view of (D). Expression of DjFGFR1 is seen in the cephalic ganglion. (F) A transverse section of the mesenchymal space. (G) Higher magnification view of (F). (H) A counter-nuclear staining of (G) using Hoechst No. 33342. The DjFGFR1-expessing cells have a large nucleus and are small in size.

Expression of DjFGFR1 in regenerating planarians

To investigate whether the expression of DjFGFR1 is influenced by regeneration, we performed whole-mount in situ hybridization of planarians regenerating from the head or from pieces of the body containing the pharynx or from tail pieces (Fig. 6). In regenerating planarians, 2 days after cutting, expression of DjFGFR1 had become strong in the blastema. Three days after cutting, the accumulation of DjFGFR1-expressing cells was observed in not only the blastema, but also in the regenerating pharynx and brain (Fig. 6B). Five days after cutting, accumulation of positive cells was no longer observed in the blastema, but still remained in the regenerating pharynx and brain. Regeneration is almost completed 7 days after amputation. The expression pattern of DjFGFR1 returned to the original one in each regenerate at this stage. To investigate more precisely the expression patterns of DjFGFR1 in regenerating planarians, we analyzed the distribution of DjFGFR1-expressing cells in the sections by in situ hybridization. Expression of DjFGFR1 was not observed in stump region just after cutting (Fig. 7A). In regenerating planarians half a day after cutting, a few DjFGFR1-expressing cells already appeared in the blastema just on the epidermis (Fig. 7B). Two days after cutting, accumulation of DjFGFR1-expressing cells was observed in the blastema (Fig. 7D). Three days after cutting, in both the regenerating head and tail pieces, the expression of DjFGFR1 had become strong, with a higher level of expression in regions nearer the stump (Fig. 7F,G). These positive cells were concentrated in the blastema and regenerating pharynx. These results clearly indicate that DjFGFR1-positive cells play roles in both blastema formation and regenerating organs, suggesting that DjFGFR1 is specifically expressed in the stem cells and regenerating cells. Expression of DjFGFR2 during regeneration was observed in the regenerating brain, but not in the blastema or regenerating pharynx (data not shown).

Figure 6.

Expression of DjFGFR1 during regeneration. (A) Whole-mount in situ hybridization of DjFGFR1 in regenerating planarians. Upper row: expression pattern of DjFGFR1 during the process of tail regeneration from head pieces; middle row: expression pattern of DjFGFR1 during the process of bidirectional regeneration from body pieces containing pharynx; lower row: expression pattern of DjFGFR1 during the process of head regeneration from tail pieces. (B) Schematic drawing showing the regenerating planarians 3 days after cutting. Blue indicates the regenerating brain. Green indicates the regenerating pharynx. DjFGFR1-positive cells are indicated in red.

Figure 7.

Spatial expression patterns of DjFGFR1. Anterior to the left. Dorsal to the top. (A–E) In the process of tail regeneration from head pieces. (A) Just after cutting; (B) half a day after cutting; (C) 1 day after cutting; (D) 1.5 days after cutting; (E) 2 days after cutting. Expression of DjFGFR1 was upregulated in the blastema. (F,G) Expression of DjFGFR1 in regenerating planarians 3 days after cutting. (F) A transverse section of a head piece. (G) A transverse section of a tail piece. Expression of DjFGFR1 has become strong in proportion to the nearness to the stump.

Loss of DjFGFR1-expressing cells in X-ray-irradiated planarians

To investigate whether the DjFGFR1-expressing cells were neoblasts, we analyzed the expression of DjFGFR1 in X-ray-irradiated planarians. It is known that X-ray-irradiated planarians lack neoblasts and cannot regenerate (Wolff & Dubois 1948; Baguña et al. 1989), presumably because neoblasts or undifferentiated cells may be more sensitive to X-rays than differentiated cells. As expected, the number of DjFGFR1-expressing cells in the mesenchymal space was markedly reduced by X-ray irradiation (Fig. 8B). However, DjFGFR1-expressing cells in the brain were not affected (Fig. 8B), suggesting that DjFGFR1-expressing cells in the mesenchymal space may be stem cells that are eliminated specifically X-ray irradiation. We also analyzed the distribution pattern of DjFGFR1-expressing cells in regenerating planarians from tail pieces 10 days after irradiation (Fig. 8C,D). Although accumulation of DjFGFR1-expressing cells was observed in the growing blastema and regenerating pharynx of non-irradiated planarians (Fig. 8C), DjFGFR1-expressing cells were not detected and no growth of a blastema or regeneration of a pharynx was observed in irradiated planarians (Fig. 8D). These results strongly suggest that loss of regenerative activity in X-ray-irradiated planarians is caused by disappearance of DjFGFR1-expressing cells in the mesenchymal space and that DjFGFR1 may be expressed in the stem cells that participate in regeneration.

Figure 8.

Expression of DjFGFR1 in X-ray-irradiated planarians. Anterior to the left. Dorsal to the top. (A) Expression of DjFGFR1 in a normal planarian. (B) Expression of DjFGFR1 in a planarian 10 days after X-ray irradiation. DjFGFR1-expressing cells decreased markedly in X-ray-irradiated planarians. However, positive cells remained in the cephalic ganglion. (C) Expression of DjFGFR1 in regenerating planarians 3 days after cutting. (D) A planarian was cut 7 days after X-ray irradiation and fixed 3 days thereafter.

Loss of function of DjFGFR1 and DjFGFR2 by RNA interference

Recently, the RNA interference (RNAi) method has been used to analyze gene function in planarians (Sánchez Alvarado & Newmark 1999; Pineda et al. 2000). To examine the function of DjFGFR1 and DjFGFR2, we applied this technique to regenerating planarians. Unexpectedly, all dsRNA-injected animals could regenerate normally (both the blastema and pharynx were formed in these animals) and no difference was found when compared with control animals (Table 1; Fig. 9A,B). Immunostaining with an antibody against planarian synaptotagmin (DjSYT; Tazaki et al. 1999) revealed that the central nervous system was normally regenerated in all dsRNA-injected animals (Fig. 9C–E). To investigate whether dsRNA-injected animals show any difference compared with control animals, we performed FACS analysis. The data obtained from these experiments show that no overt effect could be detected in dsRNA-injected animals (Fig. 10). However, it should be noted that the number of Hoechst No. 33342-sensitive cells in animals injected with both DjFGFR1 and DjFGFR2 was slightly higher (Fig. 10, inset). We could not identify the cell type of the affected cells. In summary, these results indicate that planarian regeneration was not influenced by loss-of-function of DjFGFR1 and DjFGFR2 after RNAi experiments, although some cellular events may be affected.

Table 1.  Summary of RNA interference experiments
 BlastemaPharynxBrain
DjFGFR1(–)
 110/1010/1010/10
 210/1010/1010/10
 310/1010/1010/10
 410/1010/1010/10
DjFGFR2(–)
 1
 2
 310/1010/1010/10
 410/1010/1010/10
DjFGFR1(–)/DjFGFR2(–)
 1
 2
 310/1010/1010/10
 410/1010/1010/10
Water
 110/1010/1010/10
 210/1010/1010/10
 310/1010/1010/10
 410/1010/1010/10
Figure 9.

Morphologic analyses of double-strand (ds) RNA-injected planarians. (A,B) Histlogic sections of regenerants from head pieces, stained with hematoxylin and eosin. (A) DjFGFR1 dsRNA-injected planarian; (B) water-injected planarian. Anterior to the left. Dorsal to the top. DjFGFR1 dsRNA-injected planarian regenerates normally, including both the blastema (arrows) and the pharynx (arrowheads). (C–F) Immunostaining of regenerating brains from tail pieces stained with DjSYT antibody. (C) Control (watar-injected animals); (D) DjFGFR1 dsRNA-injected animals; (E) DjFGFR2 dsRNA-injected animals; (F) both DjFGFR dsRNA-injected animals. All dsRNA-injected animals were normally stained with DjSYT antibody.

Figure 10.

Fluoresence-activated cell sorting (FACS) profile of DjFGFR1 (right upper) and DjFGFR2 (left lower) and both of DjFGFRs (right lower) double-strand (ds) RNA-injected planarians. Control is shown in the left upper panel. Each profile contains 10 000 cells. The x-axis displays forward scattering property and the y-axis displays Hoechst No. 33342 intensity. The number of Hoechst No. 33342-sensitive cells (within a square) in animals injected with both DjFGFR1 and DjFGFR2 simultaneously was slightly higher. The numbers show the average number of cells within a square in two experiments.

We also tested whether SU5402 treatment blocked planarian regeneration. However, we could not detect any phenotype. SU5402-treated planarians were normally regenerated and formed new brain and pharynx. SU5402 specifically blocks vertebrate FGFR1 kinase activity. We could not detect any consensus binding site to SU5402 within DjFGFR1 nor DjFGFR2 sequences (Mohammadi et al. 1997).

Disscussion

DjFGFR1 and DjFGFR2 are FGFR family members

We have shown that DjFGFR1 and DjFGFR2 encode receptor tyrosine kinases that are structurally very similar to both vertebrate and invertebrate FGFRs. Like other FGFRs, DjFGFR1 and DjFGFR2 are composed of extracellular Ig domains and intracellular split tyrosine kinase domains. While most members of the FGFR family contain an acidic region in the extracellular domains, we could not detect an acidic region in DjFGFR1 or DjFGFR2. Molecular phylogenetic analyses of the extracellular and intracellular domains showed that both are categorized in the FGFR family, but not in a PDGFR family. These analyses also suggest that the DjFGFR1 and DjFGFR2 genes may have diverged after the segregation of planarians. Although both are expressed in the brain, the major cells expressing them are different. DjFGFR1 is mainly expressed in somatic stem cells of asexual-state planarians, but DjFGFR2 is expressed in germ-line cells of sexual-state planarians (Ogawa et al. 1998).

DjFGFR are used in different cell processes

Fibroblast growth factor/FGFR signaling systems are essential for a variety of developmental processes of both vertebrates and invertebrates. During mouse development, FGFR1 and FGFR2 primarily function in the primitive ectoderm at the egg cylinder stage (Orr-Urtreger et al. 1991). Later, they are used more widely in mesoderm- and neuroectoderm-derived tissues (Safran et al. 1990). The expression pattern of DjFGFR1 suggests that the signaling system relevant to DjFGFR1 may be used in several different processes of the planarian cell system. In intact planarians, DjFGFR1 is expressed in the brain-composing cells and neoblast-like cells distributed in the mesenchymal space. X-Ray irradiation experiments clearly showed that the nature of these cells is completely different. DjFGFR1-expressing cells in the mesenchymal space are sensitive to X-ray irradiation, but those in the brain region are not, suggesting that the DjFGFR1-signaling system may be used in different processes. In regenerating planarians, at early stages of regeneration, accumulation of DjFGFR1-expressing cells is observed in the blastema and in both pharynx and brain rudiments. At later stages, DjFGFR1 expression continues in the brain region, where neuronal differentiation actively occurs. These results suggest that the DjFGFR1-signaling system may be involved in growth/migration/differentiation of both undifferentiated cells and neuronal cells, as it is in other organisms.

In the case of DjFGFR2, we have already shown that the signaling system of DjFGFR2 may have an important role in germ cell development and migration during sexualization (Ogawa et al. 1998). While expression of DjFGFR2 is also clearly observed in the brain region the same as for DjFGFR1, expression in the mesenchymal space is obscure. We cannot speculate on its function in regeneration processes at this moment. In conclusion, both DjFGFR1 and DjFGFR2 may function in brain formation. The DjFGFR1-signaling system may be involved mainly in the somatic stem cell system in asexual-state planarians, whereas the DjFGFR2-signaling system may be used in germ-line cells in sexual-state planarians.

Possible function of DjFGFR1 during regeneration

DjFGFR1-expressing cells are detected in the mesenchymal space of intact animals and their accumulation is observed in the blastema and rudiments at early stages of regeneration, suggesting that the DjFGFR1-signaling system may have an important function in the planarian stem cell system. However, we have not yet clearly determined whether DjFGFR1-expressing cells in the mesenchymal space are limited to real stem cells or not. Electron microscopic studies have revealed that the neoblast is small and round in shape, has a high nucleus/cytoplasm ratio and contains chromatoid bodies in the cytoplasm (Morita 1967; Hori 1982). The DjFGFR1-expressing cells are also small in size and round to ovoid in shape and have a high nucleus/cytoplasm ratio. In regenerating planarians, 2 days after cutting, accumulation of DjFGFR1-expressing cells was observed in the blastema and blastema–proximal region. After 3 days, the number of positive cells in the blastema decreased gradually and the expression pattern returned to the original state after 7 days. In addition, DjFGFR1-expressing cells are specifically eliminated by X-ray irradiation and irradiated worms lose their regenerative ability. All these data support the idea that the behavior of DjFGFR1-expressing cells orresponds to that of neoblasts.

However, our recent studies using molecular markers and observations made in an electron microscopic study of Hori clearly indicated that the most cells classically identified as the neoblasts have been committed to certain cell types before migration to the blastema or rudiments (Hori 1997; Agata & Watanabe 1999; Kobayashi et al. 1999). The committed cells are not distinguishable from uncommitted stem cells by classic morphologic observations. Based on these findings, we speculate that DjFGFR1 may be expressed in uncommitted stem cells and their expression may be maintained during migration after commitment. The DjFGFR1-signaling system may have an important role in the maintenance of the undifferentiated state in both uncommitted stem cells and migrating committed cells. In the case of pharynx regeneration, during the early stage of regeneration, the stem cells in the mesenchymal space (defined as the pharynx region as a consequence of the rearrangement of the body proportions after cutting) may be committed as pharynx-forming cells within 2 days. These cells start to transcribe a muscle-specific gene (DjMHC-A; myosin heavy chain gene A) and may migrate to form a pharynx rudiment (Kobayashi et al. 1999). However, they may maintain the undifferentiated state during migration. These features are strikingly similar to those of myoblasts in other animals. For example, fibroblast growth factor receptor-like embryonic kinase (FREK) expression in chick embryos is upregulated in migrating myoblasts (Marcelle et al. 1995). Targeted expression of a dominant-negative FGFR variant in the lungs of mice prevents migration of smooth muscle myoblasts (Peters et al. 1994). In sea urchins, expression of FGFR is observed in migrating myoblasts (McCoon et al. 1998).

To understand the role of FGFR-signaling systems in planarians, we have investigated DjFGFR1 and DjFGFR2 functions by RNAi experiments. Unexpectedly, we could not obtain any clear phenotypes. Even though we injected dsRNA of both FGFRs, those planarians could normally regenerate. However, FACS analyses suggested that the Hoechst No. 33342-sensitive cells of DjFGFR1 dsRNA-injected planarians were enriched compared with those of control and DjFGFR2 dsRNAi-injected animals. Such an effect was much clearer in double-injected animals, suggesting that the redundancy of FGFRs may shade the DjFGFR1(–) phenotype. Although FGFRs can interact with multiple FGFs in other organisms (Mansukhani et al. 1990; Vainikka et al. 1992; Werner et al. 1992; Wang et al. 1994), we have not been able to find the ligand molecules nor other FGFRs in planarians. To clarify the function of the DjFGFR family genes, we need to identify not only their ligand molecules, but also another type of FGFR. In addition, we are trying to produce antibodies against the FGFRs to identify the DjFGFR-expressing cells.

Possible DjFGFR1 and DjFGFR2 functions in cephalic ganglions

In intact planarians, the expression of DjFGFR1 and DjFGFR2 is observed in cephalic ganglions. X-Ray irradiation experiments indicate that DjFGFR1- and DjFGFR2-expressing cells in the cephalic ganglion may have already been committed or differentiated as neurons, suggesting that DjFGFR1 and DjFGFR2 may be involved in neurogenesis and maintenance of cephalic ganglions. It is well known that various members of an immunogloblin superfamily, including FGFR, are essential for neural development and maintenance in many organisms (Doherty & Walsh 1996; Hall et al. 1996). Planarians readily change their body size and reform their body proportions according to their feeding conditions. Thus, it is reasonable that DjFGFR1 and DjFGFR2 are expressed continuously in the brains of intact animals.

Recently, a FGFR-like 1 (FGFL1)-like gene (nou-darake) was identified in planarians (M Nakazawa et al., unpubl. obs.). In planarians with loss of nou-darake function due to RNAi, cephalic ganglions expand to a more posterior region, suggesting that nou-darake may be involved in inhibition of cephalic ganglion formation outside of the brain. Interestingly, nou-darake contains two Ig domains in an extracellular region, like DjFGFR2. However, nou-darake lacks the kinase domain in the intracellular region. Investigation of the function of FGFR-related genes in the planarian brain may provide new insights into brain formation. However, we could not detect any defect on brain regeneration in dsRNA-injected planarians. Now, investigations are under way using other molecular markers.

Acknowledgements

We thank Takahiro Kunisada and Keiji Okamoto for technical support, Kentaro Kato, Norito Shibata and Akira Tazaki for technical advice and Kimihiro Sato, Kenichi Tominaga for their helpful discussion and Francesc Cebriá and Shigeru Kuratani for critical reading of the manuscript. This work was supported by JSPS Research Fellowships for Young Scientists (KO), Special Coordination Funds for Promoting Science and Technology (KA) and a Grant-in-Aid for Scientific Research on Priority Areas (KA, KW).

Ancillary