Planarians are known to have strong regenerative ability. A small fragment from almost any part of their body can regenerate to form a complete animal, and planarians utilize this ability to propagate asexually. Grown planarians divide by fission into two pieces transversely at the anterior or posterior end of the pharynx, with each fragment able to subsequently regenerate the whole body. Furthermore, planarians can easily be converted from an asexual state to a sexual state in response to changes in their food source, or by cooling the water temperature of their surroundings (Kenk,1937; Kobayashi et al.,1999,2002). After conversion of the animals to a sexual state, germline cells are formed and hermaphroditic sexual organs are developed. It has been suggested that the presence of pluripotent stem cells, also known as neoblasts in planarians, permits this flexibility (Baguñà,1981). Previous studies have shown that neoblasts are able to give rise to all cell types according to positional information, including the germline cells in the planarian (Baguñà et al.,1989; Newmark and Sánchez-Alvarado,2000; Agata et al.,2003; Reddien et al.,2005; Sato et al.,2006, Handberg-Thorsager and Saló,2007; Wang et al.,2007). Neoblasts are the only cells that have the ability to proliferate in the planarian body (Morita and Best,1984; Newmark and Sánchez-Alvarado,2000), and these cells proliferate to give rise to new differentiated cells, even in intact animals (Baguñà et al.,1990).
Under the light microscope, neoblasts appear as small round or ovoid cells scattered throughout the mesenchymal space (Orii et al.,2005; Sánchez-Alvarado and Kang,2005; Hayashi et al.,2006). Numerous electron microscopy studies have shown that neoblasts possess a relatively large nucleus and limited cytoplasm, which contains a high concentration of ribosomes, but little or no endoplasmic reticulum and only a few mitochondria (Hay and Coward,1975; Morita et al.,1976; Hori,1982). One of the most remarkable features of neoblasts is the presence of an electron-dense, membrane-lacking chromatoid body in their cytoplasm (Hay and Coward,1975; Hori,1982). A close relationship is usually observed between chromatoid bodies and pore regions in the nuclear envelope in regenerative cells (Hori,1982; Auladell et al.,1993). The chromatoid bodies continue to decrease in size during the cytodifferentiation of regenerating cells, though they do not disappear entirely during the regenerative process (Hori,1982,1997). It has been reported that following regeneration, chromatoid bodies remain only in germline cells (Teshirogi,1962; Teshirogi and Ishida,1987). The presence of RNA in the chromatoid body has been shown by cytochemical studies (Hori,1982) and RNase treatment experiments (Auladell et al.,1993). Thus, all of these features of chromatoid bodies resemble those of germline-specific structures called germ granules or polar granules, which are observed in almost all animals (Eddy,1975).
These findings have also led us to speculate that chromatoid bodies contain mRNAs required for cell proliferation and differentiation, and regulate these mRNAs at the post-transcriptional level. We assumed that RNA-binding proteins involved in various steps of the post-transcriptional regulation process may be components of the chromatoid bodies as well as germ granules. Based on the similarities between planarian chromatoid bodies and germ granules, we tried to isolate a planarian homolog of vasa, which is known to be a major component of germ granules, and succeeded in identifying DjvlgA (Dugesia japonica vasa-like gene A; Shibata et al.,1999). After this success, many genes encoding homologs of germ granule components of other animals, especially RNA binding proteins, were isolated from planarians, and these homolog genes were found to be expressed in neoblasts or germline stem cells; for example, Djpum is a pumilio homolog; Smedwi1, 2, and Djpiwi-1 are piwi homologs; bruli is a bruno-like gene; and Djnos or Smednos are nanos-related genes (Reddien et al.,2005; Salvetti et al.,2005; Guo et al.,2006; Rossi et al.,2006; Sato et al.,2006; Handberg-Thorsager and Saló,2007; Wang et al.,2007). However, it has not yet been shown whether any of these proteins are localized in the chromatoid body, although electron microscopic mRNA hybridization analysis clearly revealed that Djnos mRNA was localized in the chromatoid bodies in the germline stem cells (Sato et al.,2006). Recently, it has also been demonstrated by RNAi experiments that some of these genes are required for neoblast maintenance and regulation (Reddien et al.,2005; Salvetti et al.,2005; Guo et al.,2006).
We obtained 99 genes encoding RNA-binding proteins from a planarian EST database (Mineta et al.,2003). Their expression patterns were analyzed in the present study by in situ hybridization and then classified into four types. This analysis revealed that many of these proteins were expressed not only in the stem cells in the mesenchymal space, but also in the brain neurons of planarians. Amongst these 99 genes, we focused on that corresponding to 1352HH cDNA, which encodes a putative member of a conserved DEAD box helicase subfamily that also includes Xenopus Xp54, Drosophila Me31B, and C. elegans CGH-1, which have all been reported to be components of RNP complexes (Ladomery et al.,1997; Nakamura et al.,2001; Navarro et al.,2001). Me31B has also been shown to be involved in translational silencing of some maternal mRNAs during transport from nurse cells to oocytes in Drosophila (Nakamura et al.,2001).
In this study, we show that the 1352HH gene product is a component of chromatoid bodies in X-ray-sensitive neoblasts. We named this gene Djcbc-1 (Dugesia japonica chromatoid-body-component 1). Djcbc-1 is also expressed in the X-ray-insensitive neurons of the brain. At the ultrastructural level, we demonstrated that the DjCBC-1 protein is localized to electron-dense structures within these cells. Furthermore, we observed strong expression of DjCBC-1 in sexual-state germline cells. These results suggest that post-transcriptional regulation, in which DjCBC-1 may be involved, is a common mechanism for gene expression control in both somatic stem and germline cells, and also in brain neurons in planarians as well as other organisms.
Expression Pattern of Planarian Genes Encoding RNA Binding Proteins
We obtained 99 genes that we thought would encode RNA-binding proteins from the planarian EST project (Mineta et al.,2003). We then analyzed the expression pattern of 60 of these genes, which were confirmed by BLAST analysis to have RNA binding motifs (Table 1) in asexual D. japonica using whole-mount in situ hybridization. To investigate whether these genes are actually expressed in neoblasts in the mesenchymal space, we compared their expression patterns in normal and X-ray-irradiated planarians. It has been shown that X-ray irradiation specifically eliminates neoblasts, but leaves differentiated cells unaffected. Additionally, we also analyzed the expression of these genes during regeneration. If they are expressed in the neoblasts, it should be possible to observe up-regulation of their signals in the post-blastema region, that is, the region of the original body remaining near the stump.
Table 1. Classification of Planarian EST Clones Including RNA Binding Motif
–, indicates no obvious motifs in the EST sequence.
Polyadenylate-binding protein protein (D. melanogaster)
SF1 protein, pre-mRNA splicing (D. melanogaster)
Y-Box factor protein (A. californica)
Translation elF4E protein (A. californica)
Tudor protein (X. laevis)
Cap binding protein 80 protein (D. melanogaster)
snRNP70K protein (D. melanogaster)
Heterogeneous nuclear ribonucleoprotein H1 protein (M. musculus)
Poly(rC)-binding protein 3 (Alpha-CP3) (M. musculus)
Heterogeneous nuclear ribonucleoprotein K, isoform b protein (H. sapiens)
DjVLGA protein (D. japonica)
Splicing factor 3B subunit 1 (X. laevis)
Putative RNA helicase (DEAD box) protein (D. rerio)
Splicing factor protein (H. sapiens)
DEAD (Asp-Glu-Ala-Asp)box polypeptide 1 protein (M. musculus)
Small nuclear ribonucleoprotein (D. melanogaster)
Splicesome-associated factor 61 protein (D. rerio)
Sam68-like mammalian protein 1 protein (H. sapiens)
U5 snRNP-specific protein protein (H. sapiens)
RNA binding motif protein 5 protein (H. sapiens)
Heterogeneous nuclear ribonucleoprotein F protein (H. sapiens)
DEAD/H box 56 RNA helicase/noh61 protein (D. rerio)
PRP3 pre-mRNA processing-factor 3 homolog protein (M. musculus)
Growth regulated nuclear 68 protein (H. sapiens) PRH14 [D. japonica]
DEAD-box protein protein (Ddx1) (D. melanogaster)
TLS-associated protein TASR-2 protein (H. sapiens)
DEAD(Asp-Glu-Ala-Asp) box polypeptide 48 protein (D. rerio)
ATP-dependent RNA helicase protein (F. nucleatum) PRH4 [D. japonica]
IGF-II mRNA-binding protein 2 protein (M. musculus)
PASILLA, putative RNA binding protein (D. melanogaster)
Probable pre-mRNA splicing protein PRP2 protein (N. crassa)
Fragile X mental retardation protein 1 (X. laevis)
mpc2 protein protein (D. rerio)
Bruno-like 5, RNA binding protein protein (H. sapiens)
hnRNP K protein (X. laevis)
RNA-dependent ATPase, putative protein (S. cerevisiae)
Heterogeneous nuclear ribonucleoprotein L protein (H. sapiens)
Polyadenylate binding protein (H. sapiens)
Dead box protein 15 (H. sapiens)
Ras-GTPase-activating protein SH3-domain-binding protein protein (D. rerio)
WM6 protein (D. melanogaster)
MAP2 RNA trans-acting protein MARTA1 protein (R. norvegicus)
Nucleolar RNA helicase
II/Gu protein (M. musculus)
Fibrillarin protein (P. falciparum)
Eukaryotic translation initiation factor 4A, isoform 1A protein (D. rerio)
Fib protein (D. melanogaster)
Heterogenous nuclear ribonucleoprotein A2/B1 protein (M. musculus)
Mammalian A1, A2/B1 hnRNP homologue protein (S. americana)
The expression patterns of these 60 genes were classified into four types, A, B, C, and D (Table 1). Type A genes were expressed mainly in neoblasts distributed throughout the mesenchymal space, but not in the brain (Fig. 1A). These cells were eliminated by X-ray irradiation (Fig. 1A'). Also, positive cells were observed among dorsally distributed cell clusters, which were thought to be germline-committed cells expressing nanos (Sato et al.,2006; Handberg-Thorsager and Saló,2007; Wang et al.,2007). To check whether these dorsal cell clusters are germline stem cells or not, we attempted co-expression analysis by fluorescent in situ hybridization and immunofluorescent experiments (Fig. 2). Djpiwi4, isolated from the EST database, is a D. japonica homolog of the Smedwi1 gene in S. mediterranea (Reddien et al.,2005), which is known to be a neoblast-specific gene. Expression of Djpiwi4 was observed in mesenchymal cells from the eye to the tail region, like the expression of Smedwi1 (see Supplemental Fig. S1A, which can be viewed at www.interscience.wiley.com/jpages/1058-8388/suppmat; Guo et al.,2006). This expression disappeared upon X-ray irradiation, suggesting that these cells were neoblasts (Supplemental Fig. S1B). Reactivity with anti-PIWI4 antibody was observed in cells in a wider region (Supplemental Fig S1C). In front of the eye, cells positive for DjPIWI4 protein but not for Djpiwi4 mRNA were observed (Supplemental Fig. S1D), whereas almost all of the cells expressing DjPIWI4 protein also expressed Djpiwi4 mRNA in the pre- and post-pharyngeal mesenchymal region (Supplemental Fig. S1E and F). Only Djpiwi4-expressing cells incorporated BrdU, and were positive for staining with antibody against a mitotic cell marker, phospho-histone H3 (data not shown), as is also true of Smedwi1-expressing cells. All of these findings about the features of Djpiwi4 were in close accord with the reported features of Smedwi1 and its protein in S. mediterranea (Guo et al.,2006), and indicate that Djpiwi4 is expressed in all of the neoblasts and its protein is expressed in the neoblasts and their early-stage differentiated progeny in D. japonica, like Smedwi1 in S. mediterranea (Guo et al.,2006). Dorso-lateral cell clusters expressing Djnos were also positive for immunostaining with anti-DjPIWI4 antibody (Fig. 2A–C). 5895HH-positive dorsal cell clusters were also positive for immunostaining with anti-PIWI4 antibody (Fig. 2D–F), as were somatic neoblasts located in a more inner region of the mesenchyme (Fig. 2G). Furthermore, these 5895HH-positive dorsal cell clusters did not accumulate in the blastema region (Fig. 1A”), suggesting that these cells were germline stem cells (Sato et al.,2006). The 5895HH gene, a representative clone of type A genes, contains the tudor domain also seen in Drosophila tudor and its homologs, which are known to be specifically expressed in germline cells (Chuma et al.,2003; Ikema et al.,2002; Golumbeski et al.,1991). From these observations, we speculate that Type A genes are specific to both somatic stem cells and germline stem cells.
Many of the analyzed genes were categorized as type B genes (Fig. 1B). They were expressed not only in the cells found throughout the mesenchymal space from head to tail and dorsal cell clusters, but also in brain neurons. The positive cells in the mesenchymal space were specifically eliminated in X-ray-irradiated planarians (Fig. 1B') and expression of these genes was strongly detected in the post-blastema region during regeneration (Fig. 1B”). Also, dorsal cell clusters expressing type B genes, especially the 1352HH gene, showed sensitivity to X-rays and expressed DjPIWI4 protein (Figs. 1B, 2H–J), as did somatic neoblasts (Fig. 2K). In contrast, the positive cells in the brain region were not eliminated by X-ray-irradiation (Fig. 1B'), suggesting that these cells were fully differentiated brain neurons. More than half of the clones analyzed (39 out of 60 clones) were classified as this type, suggesting that the majority of RNA-binding proteins are expressed in both stem cells and brain neurons in planarians. All types of RNA-binding proteins were found in this class.
Four clones showed a brain neuron-specific expression pattern, and were categorized as Type C genes. The expression of these genes was not affected by X-ray-irradiation (Fig. 1C,C'). One of these genes, 2659HH, encodes a protein containing 3 RRMs, and belongs to the Etr (elav-type ribonucleoprotein) family, which includes a neural marker in Xenopus (Knecht et al.,1995).
Type D genes were expressed in all regions of the body (Fig. 1D). Neither X-ray-irradiation nor amputation had a marked effect on their expression pattern (Fig. 1D',D”). Double detection analysis of a Type D gene, 0417HH, and DjPIWI4 protein showed that signals for the 0417HH gene were observed not only in cells expressing DjPIWI4 in dorsal cell clusters and in somatic neoblasts, but also in DjPIWI4-negative cells (Fig. 2L–O). These genes were considered to be ubiquitously expressed in all cell types, and so-called housekeeping genes, such as eIF4A (4001HH) and polyA-binding protein (0312HH), were included in this class.
In conclusion, many RNA-binding protein genes expressed in stem cells were identified, and the findings were in accord with those of DNA chip analysis (Rossi et al.,2007).
Structure of Clone 1352HH and Production of Antibody Against the Gene Product
Next, we tried to produce antibodies against the proteins encoded by clones whose expression in the neoblasts was confirmed, to investigate whether their products were components of the chromatoid body. Unfortunately, most of the fusion proteins were not synthesized in bacterial hosts. Moreover, for the fusion proteins that were synthesized and purified, most of the antisera produced against these proteins could not be used for immunodetection on EM specimens. However, we did succeed in producing an antibody against the product of the 1352HH gene, which was classified into Type B in the expression analysis described above, and we proceeded to further analyze the expression and possible function of this gene. We isolated a clone containing a full-length insert of the 1352HH gene from a cDNA library. This gene encodes a DEAD box type RNA helicase with 503 amino acids (Fig. 3A), including all eight motifs conserved in DEAD box proteins (Linder et al.,1989). However, this gene differed from planarian DEAD box family genes previously identified in our laboratory (Shibata et al.,1999). Multiple alignment showed that the product of this gene belongs to the RCK/p54/Me31B subfamily of the DEAD box RNA helicase family (Fig. 3B). It showed about 70% similarity to Me31B (Drosophila), RCK (human), CGH-1 (C. elegans), and ste13 (S. pombe). In contrast, it showed less than 30% identity to DjVLGA and DjVLGB, two vasa-like genes in the planarian, which also belong to the DEAD box helicase family (Shibata et al.,1999).
DjCBC-1 Protein Is One of the Components of Chromatoid Bodies
We made an affinity-purified antibody against 1352HH protein that recognized a 57-kDa protein band in planarian samples and did not react with the recombinant DjVLGA and DjVLGB proteins (Fig. 4A). Using this antibody, we performed immunofluorescent staining. Paralleling its mRNA expression, 1352HH protein was detected in the X-ray-sensitive cells of the mesenchymal space (Fig. 4C,E) and in X-ray-resistant cells in brain (Fig. 4B,D).
We initially focused on the protein's expression in the mesenchymal space. To confirm that positive cells in the mesenchyme were neoblasts, we performed a BrdU labeling experiment (Fig. 4F–H). Two days after labeling, cells labeled with BrdU were also 1352HH positive (Fig. 4H). Additionally, we confirmed that 1352HH-positive mesenchymal cells also expressed DjPIWI4 protein (Fig. 4I–K). Some 1352HH-positive cells were negative for staining by anti-DjPIWI4 (Fig. 4K, arrow), suggesting that 1352HH protein was expressed not only in neoblasts, but also in late-stage differentiating cells, like DjvlgA (Shibata et al.,1999; Newmark and Sánchez-Alvarado,2002).
We then performed immunoelectron microscopic analysis to investigate whether the 1352HH protein is a component of chromatoid bodies, which are known to contain RNA (Hori,1982; Auladell et al.,1993), as some 1352HH orthologs are known to form RNP complexes in other organisms (Ladomery et al.,1997; Navarro et al.,2001; Nakamura et al.,2001). In neoblasts, gold particles corresponding to 1352HH signals were observed on or around the chromatoid bodies (Fig. 5A and B), and therefore we named this gene Djcbc-1 (Dugesia japonica chromatoid-body-component 1). DjCBC-1 signals were also observed within the nucleus in a speckled pattern (Fig. 5B). Fifty percent of the observed neoblasts contained DjCBC-1-positive chromatoid bodies (Table 2) and the immunopositivity was also heterogeneous amongst the chromatoid bodies within one neoblast (Fig. 5, Table 2). There was no remarkable difference in the morphology between DjCBC-1-positive and -negative chromatoid bodies (Fig. 5B and C). This suggests that there is heterogeneity among chromatoid bodies within one neoblast at the molecular level.
Table 2. Localization of DjCBC Protein in the Chromatoid Bodies
No. of CB
No. of CB with signal
Nucleus with signal
Total number of DjCBC-1-positive NB, 11 (78%); total number of NB containing DjCBC-1-positive CB, 7 (50%); total number of DjCBC-1-positive CB, 15 (68%). Djcbc-1-positive CB/total CB in neoblast No. 1–7.
DjCBC-1 Protein Forms Chromatoid Body-Like Complexes in Brain Neurons
We wondered why DjCBC-1 is expressed in X-ray-resistant, fully differentiated neurons in the brain. To address this question, we examined the expression of DjCBC-1 in the brain in detail. In the planarian brain, there are bundles of axons that are stained with anti-DjSyt protein antibody (Tazaki et al.,1999) and located in the central portion of the brain surrounded by cell bodies (Fig. 6B). In the brain region, DjCBC-1 protein was highly expressed in all of the cell bodies of neurons adjacent to the axon bundles (Fig. 6B and C). The DjCBC-1 signal was also detected relatively weakly in the axon region, but was not detected in the muscle fibers (Fig. 6C and D). Also in the ventral nerve cords, no signal was observed (data not shown). Interestingly, the DjCBC-1 signals were detected as puncta (Fig. 6B and C).
We then examined the localization of DjCBC-1 protein in the cell bodies of neurons in the brain at the electron microscopic level. The area observed is shown in Figure 6A as a box. The cells in this area appeared to be nerve cells, as they possessed a type of neurosecretory vesicle in their cytoplasm, as has been previously reported (Fig. 6E, arrows; Oosaki and Ishii,1965). We found that DjCBC-1 signals were localized on the chromatoid body-like structures (Fig. 6E, arrowhead). These electron-dense bodies were 0.1–0.3 μm in diameter and were not bound by a membrane, so they clearly were not dense bodies surrounded by a membrane (Oosaki and Ishii,1965).
DjCBC-1 Protein Is Expressed In Germline Cells in Sexual Planarians
In some animals, Djcbc-1 orthologs have been reported to be expressed mainly in germline cells (Ladomery et al.,1997; Navarro et al.,2001; Nakamura et al.,2001). We therefore investigated whether DjCBC-1 protein is expressed in germline cells in sexual state planarians. Although our strain basically regenerates asexually, some worms convert to a sexual state by chance. These sexual planarians have bilateral lines of testes in the dorsal side and a pair of ovaries in the ventral side (Fig. 7A). It has been reported that chromatoid bodies are found in planarian germline cells (Teshirogi and Ishida,1987). In sexual planarians, DjCBC-1 protein was also expressed in the ovary and testis, as well as in the neoblasts and brain neurons (Fig. 7B–E, data not shown). In the testis, DjCBC-1 protein was expressed in the spermatogonia, spermatocytes, and spermatids (Fig. 7D and D'). In ovaries stained with anti-DjCBC-1, punctate signals were observed in oogonia and early oocytes (Fig. 7C), and were considered to indicate staining of the chromatoid bodies, whereas the signal was uniformly observed in mature oocytes. The expression pattern of DjCBC-1 was thus different from that of DjvlgB mRNA, which was expressed only in spermatocytes in the testis, but similar to that of DjvlgA (Shibata et al.,1999).
Djcbc-1 Is an RCK/Xp54/Me31B Ortholog in Planarians
DEAD box proteins are well conserved and appear to be involved in a wide range of intracellular events in which alteration of the RNA secondary structure is required. In the planarian, 12 fragments of RNA helicase genes (planarian RNA helicase, PRH) containing motifs of DEAD box RNA helicases were previously identified (Shibata et al.,1999). In the planarian EST database, we found a total of 18 genes encoding DEAD box RNA helicases, which included Djcbc-1 (Table 1). Amongst them, 6055HH and 5302HH completely corresponded to PRH4 and PRH14, respectively. Thus, there are at least 28 genes encoding DEAD box RNA helicases in the planarian. Most (13/18) of the DEAD box RNA helicases whose expression was analyzed showed a type B expression pattern (i.e., expression in neoblasts and brain neurons) identical to that of Djcbc-1. We tried to analyze the function of Djcbc-1 using the RNA interference (RNAi) method. The knockdown of the expression level of the mRNA and protein by Djcbc-1 dsRNA was confirmed by real-time PCR and immunostaining (data not shown). However, the dsRNA-injected planarians were able to regenerate normally. The morphology of the dsRNA-injected planarians' central nervous system also showed no difference compared to that of the control animals. We suspect that this is because some genes have a redundant function to that of Djcbc-1, and we speculate that they might be identified amongst other genes showing the same expression pattern as Djcbc-1.
DjCBC-1 Is a Component of RNP Complexes in Planarians
Control of gene expression via post-transcriptional processing of mRNA, the export of mRNA from the nucleus to the cytoplasm, and mRNA localization and translational regulation have been shown to be essential for various aspects of development and cell differentiation (Wickens et al.,2000). RNA-binding proteins involved in these processes form complexes with mRNAs. These complexes can be found as electron-dense granules in various cell types, especially in neurons (Havik et al.,2003; Krichevsky and Kosik,2001; Kanai et al.,2004) and germline cells (Eddy,1975; Ikenishi,1998; Seydoux and Braun,2006), which are required to respond to environmental changes more quickly than other cell types. Recently, another important function of RNA-binding proteins has been discovered. During spermatogenesis in mouse, Piwi-related proteins are localized in chromatoid bodies (Kotaja et al.,2006). Piwi proteins have been shown to be involved in the RNA interference (RNAi) machinery via piRNA, and to regulate the expression of retrotransposons (Aravin et al.,2006; Girard et al.,2006, Brennecke et al.,2007), suggesting that the chromatoid body is a facility containing RNAi machinery. Thus, electron-dense granules may play important roles in translational and also transcriptional control by regulating RNAs.
Recently, it have been reported that neoblasts express many types of RNA metabolism-related genes (Rossi et al.,2007). In this study, we also showed that several genes encoding RNA-binding proteins are expressed in neoblasts in the asexual planarian's mesenchymal space and in planarian brain neurons. Although we could not yet confirm whether all of these genes are expressed in germline cells in the sexual state, at least DjCBC-1 was expressed in completely differentiated germline cells. The presence of electron-dense RNP complexes in neoblasts and germline cells has been the focus of several studies (Hay and Coward,1975; Hori,1982; Teshirogi,1962; Teshirogi and Ishida,1987). We were able to detect such RNP complexes in completely differentiated neurons based on the localization of DjCBC-1. Figure 8 illustrates possible functions of DjCBC-1 in undifferentiated neoblasts, differentiated neurons, and germline cells of the planarian. DjCBC-1 is one of the components of RNP complexes commonly observed in these cell types. Each RNP complex may contain other proteins optimized for the transport or other regulation of the RNA(s) in the respective cell types. In other animals, DjCBC-1 orthologs have been reported to form mRNP complexes in the cytoplasm of germline cells, and some of them have been shown to be involved in translational silencing of maternal mRNAs during transport (Ladomery et al.,1997; Nakamura et al.,2001; Navarro et al.,2001; Weston et al.,2006). This suggests that post-transcriptional regulation, in which DjCBC-1 may be involved, is a common mechanism for gene expression control in both somatic stem and germline cells, and also in brain neurons in planarians as well as other organisms (Weston and Sommerville,2006).
DjCBC-1 in Chromatoid Bodies in Neoblasts
Chromatoid bodies are assumed to play an important role in neoblast differentiation; it is possible that these bodies represent the transient accumulation of mRNAs required for differentiation. In this study, we have shown that DjCBC-1 is localized in the chromatoid bodies of neoblasts (Fig. 5). We also emphasize that DjCBC-1 signals could be observed within the nucleus in a speckled pattern (Fig. 5). Xenopus Xp54 has been shown to have the ability to shuttle between the nucleus and cytoplasm, to bind to nascent RNP transcripts, and to accompany maternal mRNA out of the nucleus and into the cytoplasm as RNP particles (Smillie and Sommerville,2002). Together, these considerations lead us to speculate that DjCBC-1 detected on the chromatoid bodies may transport mRNAs from the nucleus to the cytoplasm, and store mRNAs required for cell differentiation and thus regulate them at the post-transcriptional level until the appropriate time for their translation. Controlling gene expression by regulation at the post-transcriptional level is thought to allow a cell to respond to environmental signals more quickly than regulation at the transcriptional level. This idea is in line with the observations that chromatoid bodies are derived from nuclear material (Hori,1982) and contain RNA (Auladell et al.,1993).
In morphological studies, the chromatoid bodies of neoblasts do not show any difference in electron density, although there is considerable variation in their number, size, and distribution depending on the cell differentiation stage of the neoblasts (Hori,1982; Auladell et al.,1993). However, it has been suggested that there is remarkable diversity amongst chromatoid bodies at the molecular level, based on a study labeling chromatoid bodies with RNase-gold markers (Auladell et al.,1993). Auladell et al. (1993) suggested that this diversity in the chromatoid bodies could be correlated with the complexity of the protein components. We showed here that there is also heterogeneity in the reactivity of the chromatoid bodies with anti-DjCBC-1 (Table 2). We consider that this heterogeneity may reflect different stages of differentiation of neoblasts and/or subgroup(s) of them. This idea is in accord with evidence about the heterogeneity of Drosophila neuronal RNPs (Barbee et al.,2006).
Chromatoid Body-like Structures in Brain Neurons
The presence of chromatoid body-like structures has been reported in various cell types. For example, mRNAs encoding proteins involved in synaptic plasticity assemble in electron-dense granules in mammalian neurons (Krichevsky and Kosik,2001). Several studies have revealed that translational regulation of localized mRNA is important for the synaptic plasticity responsible for learning and memory. Translational machinery stationed near the synapse could engage specific mRNA to affect local synthesis and perhaps influence synaptic function (Brittis et al.,2002). Some of the protein and mRNA components of these bodies have already been identified (Krichevsky and Kosik,2001; Kanai et al.2004; Barbee et al.,2006). In planarians, Djcbc-1 is expressed in brain neurons but not in ventral nerve cords (Fig. 1). We observed electron-dense bodies, in which DjCBC-1 protein was localized, in the cell bodies of these brain neurons (Fig. 6). The expression of Djcbc-1 orthologs in somatic cells and neurons has been reported in other animals. In many cases, these ortholog proteins are localized in granular structures, such as p-bodies or stress granules, and repress the translation or promote the degradation of mRNA (De Valoir et al.,1991; Navarro et al.,2001; Cougot et al.,2004; Wilczynska et al.,2005; Barbee et al.,2006). It is possible that in planarians, DjCBC-1 may have a role in mRNA regulation in neurons, like its orthologs in other animals. In addition to expressing DjCBC-1, neurons in the brain region express some RNA-binding proteins related to the components of neural RNP complexes in other animals. For example, clone 6718HH encodes a member of the FRMP (fragile X mental retardation protein 1) family, whose members are known to be key components of RNP complexes in neurons in other animals (Bardoni et al.,2001; Barbee et al.,2006). This again leads us to speculate that RNP complexes in neurons are also conserved in planarians. We should investigate the target mRNAs of DjCBC-1 as the next step in future experiments.
DjCBC-1 in Germline Cells
In addition to being able to regenerate, planarians can also convert to a sexual reproductive state. It has been shown that germline cells are also derived from neoblasts (Baguñà et al., 1998; Sato et al.,2006; Handberg-Thorsager and Saló,2007; Wang et al.,2007). We observed that DjCBC-1 protein is strongly expressed in germline cells, as well as in neoblasts and neurons, in sexual state planarians (Fig. 7B–E). In C. elegans, CGH-1 is expressed in germline cells entering meiosis, and is required for oogenesis and spermatogenesis (Navarro et al.,2001). This suggests that this family is essential for meiosis entry and/or germline development. In S. pombe, Ste13, a member of the RCK/Xp54/Me31B family, has been reported to be essential for sexual reproduction but not for growth (Maekawa et al.,1994; Weston and Sommerville,2006). The ste13-mutant phenotype can be suppressed by Drosophila Me31B but not by the vasa gene (Maekawa et al.,1994). The expression pattern of DjvlgB in the testis is different from that of DjCBC-1 (Shibata et al.,1999). These facts suggest that this RCK/subfamily protein may have conserved functions distinct from those of the vasa subfamily, whereas it remains unknown whether Djcbc-1 has a similar function in meiosis entry in planarians.
DjCBC-1 orthologs in other animals are known to form complexes with Y-box proteins, another type of RNA-binding protein, in germline cells (Nakamura et al.,2001; Bouvet and Wolffe,1994). In planarians, two Y-box proteins have been cloned (Salvetti et al.,1998,2002). One of them, DeY1, is detected in the nucleus of spermatogonia and in both the nucleus and the cytoplasm of spermatocytes at the electron microscopic level (Salvetti et al.,2002). Female germline-specific expression of a Y-box gene has not been reported. We identified a new Y-box family gene, 3689HH, showing a type A expression pattern (i.e., expression in neoblasts and germline stem cells). Although we have not yet analyzed the expression pattern of 3689HH in sexual state planarians, it is possible that 3689HH encodes a Y-box protein specifically expressed in female germline cells, and that it functions as a partner of DjCBC-1.
A clonal strain, SSP (diploid, 2n=16), of the planarian Dugesia japonica was used (Ito et al.,2001). The worms were maintained in autoclaved tap water at room temperature (22–24°C). For sexualization, SSP were maintained at 18°C. In this condition, some SSP worms were sexualized, forming testes and ovaries.
In this study, worms were starved for 1 week before they were used for experiments.
Worms placed on wet filter paper on ice were irradiated with 12 R of X-rays by using an X-ray generator (SOFTEX B-4). Four days after irradiation, worms were fixed for in situ hybridization and immunohistochemistry.
Cloning of Full-Length cDNA
The longest cDNA of Djcbc-1 was obtained from a D. japonica head cDNA library by the stepwise dilution method as described previously (Mineta et al.,2003; Watanabe et al.,1997).
In Situ Hybridization
All digoxigenin-labeled RNA probes were prepared using Djcbc-1 cDNApBluescriptSK(−) vectors containing the insert from the EST clone and a digoxigenin-labeling kit according to the manufacturer's instructions (Roche, Indianapolis, IN). Whole-mount in situ hybridization using alkaline phosphatase and NBT/BCIP color development was performed as described by Umesono et al. (1997) and Takano et al. (2007).
Fluorescent whole mount in situ hybridization (FISH) was also performed as described until the wash after the hybridization. After blocking, samples were incubated with 1:100 anti-Dig-POD antibody (Molecular Probes, Eugene, OR) overnight at 4°C. Then samples were washed more than five times for 1 hr with maleate buffer (0.1 M maleic acid, 0.15 M NaCl and 0.1% Triton X 100). To develop fluorescent color, we used TSA Kit no. 2 (Molecular Probes). After the wash, samples were incubated with amplification buffer containing 1:100 Alexa Fluor 488 tyramide for 30 min. Then samples were washed five times for 1 hr with maleate buffer, and used for immunohistochemistry using anti-PIWI4 antibody.
cDNA fragments encoding the C-terminal region of the DjCBC-1 protein and DjPIWI4 were amplified by PCR with primers 5′-TGGATCCTCACAATATTATGCATATGTCC- 3′ and, 5′-TCTGCAGTCAATGTCGCATATTCTGCTG- 3′ for DjCBC-1 and 5′-AGTATTGAAGAACAGCCTATTAC-3′, and 5′-CTATTTTACAAGTAAAATAAACGTTCC-3′ for DjPIWI4 from a planarian head cDNA library, and digested with BamHI and PstI and then subcloned into pQE30 vector (QIAGEN). Recombinant 6×His-DjCBC-1 and DjPIWI4 protein were purified and used as antigens. Polyclonal rabbit antiserum for DjCBC-1 and mouse monoclonal antibody for DjPIWI4 were generated by MBL (Nagano, Japan). The anti- DjCBC-1 antiserum was affinity-purified using the antigen. The affinity-purified antibody was then used in experiments as an antibody for DjCBC-1 protein. Recombinant DjVLGA (a.a. 436-611) and DjVLGB (a.a. 162-410) were produced, purified, and used for Western blot analysis.
Western Blot Analysis
Western blot analysis was performed basically as described (Orii et al.,2005). Planarian extract (2.5 μg) and 0.05 μg of each recombinant protein were applied onto a 15 % polyacrylamide gel. After electrophoretic transfer and blocking, the membrane was incubated with anti-DjCBC-1 antibody diluted 1:1,000 in blocking buffer (10% goat serum in phosphate buffered saline supplemented with 0.1 % Triton X-100; TPBS). A 1:5,000 dilution of anti-rabbit IgG peroxidase conjugate (SIGMA, Saint Louis, MO) in blocking solution was used as a secondary antibody, and detection was performed using ECL Western Blotting Detection Reagents (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK).
To prepare sections, animals were fixed with cold relaxant solution (1% HNO3, 4.5% formaldehyde, 50 mM MgSO4) for 18–24 hr. Fixed samples were dehydrated, embedded in paraffin, and sectioned at 7–8-μm thickness. After deparaffination, the sections were treated with 10 mM citric acid buffer at 120°C for 10 min, rinsed with TPBS, and incubated with blocking buffer (10% goat serum in TPBS) for 30 min at RT. They were incubated with the primary antibody, anti-DjCBC-1 antibody diluted 1:500, anti-DjPIWI4 diluted 1:1,000, or anti-DjSyt diluted 1:500 (Tazaki et al.,1999), with blocking buffer at 4°C overnight. After washing, signals were detected using AP-conjugated anti-rabbit IgG secondary antibody (1:2,000) and NBT/BCIP (Roche) or Alexa Fluor 488-conjugated anti-rabbit IgG antibody (diluted 1:1,000) (Molecular Probes). FISH samples were incubated with blocking buffer containing 1:1,000 anti-DjPIWI4 antibody overnight at 4°C. Then samples were washed with TPBS seven times for 1 hr each, and treated with Alexa Fluor 488-conjugated anti-rabbit IgG antibody diluted 1:1,000 (Molecular Probes) in TPBS overnight at 4°C. After the wash, samples were mounted with mounting medium (DakoCytomation, Carpinteria, CA).
BrdU labeling and detection were performed by injection basically as described previously (Newmark and Sánchez-Alvarado,2000; Inoue et al.,2007). Two days after injection, detection of BrdU was performed with Alexa Fluor 594 tyramide using TSA Kit no. 15 (Molecular Probes).
Worms were fixed in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 120 min at 4°C. After rinsing in cacodylate buffer twice for 5 min each, and in TBST [10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1% Tween 20] once for 5 min, the samples were soaked in 50 mM ammonium chloride for 10 min to neutralize residual aldehyde groups. After rinsing in TBST three times for 5 min each, the samples were dehydrated in an ethanol series (70, 80, 90, 95% for 15 min each at 4°C, and 100% ethanol twice for 30 min each at 4°C) and embedded in LR White (Agar Scientific Ltd.). Samples were transferred into gelatin capsules filled with neat resin (Lowicryl HM20, Polysciences), and polymerized for 24 hr at 55°C.
Ultra-thin sections were then cut and collected on nickel grids (no.150 mesh). Sections on the nickel grids were incubated in a drop of blocking solution (0.5% BSA in TBST) for 10 min, and in a drop of primary antibody (1:50) for 1 hr at room temperature. As a negative control, blocking solution alone was used. After rinsing the grids in a drop of TBST six times for 1 min each, the grids were incubated in a drop of secondary antibody solution [15-nm-gold-conjugated anti-rabbit antibody (Pharmacia) diluted with TBST, 1:50] for 1 hr at room temperature. After the grids were rinsed in a drop of TBST six times for 1 min each, the sections on the grids were refixed in a drop of 0.5% glutaraldehyde for 10 min and were rinsed in distilled water several times. The sections were stained with uranyl acetate and lead citrate and observed under an electron microscope.
We thank Tomomi Kudome-Takamatsu, Sayaka Higuchi, and Osamu Nishimura for technical assistance with the sequencing and in situ hybridization, and Yumi Saito for technical assistance with RNAi, and Elizabeth Nakajima for careful reading of the manuscript. We are also grateful to Chiaki Ogasa and Akiko Tanaka of the Protein Research Group, Genomic Sciences Center (GSC), RIKEN, for supplying us with DjVLGA and DjVLDB recombinant proteins. This work was supported by a Grant-in-Aid for Creative Scientific Research to K.A. (17GS0318), a Grant-in-Aid for Scientific Research on Priority areas to K.A., and the Global COE program A06 of Kyoto University.