Smed-SmB, a member of the LSm protein superfamily, is essential for chromatoid body organization and planarian stem cell proliferation

Planarians (phylum Platyhelminthes, class Turbellaria) are bilaterian, triploblastic free-living flatworms best known for their impressive ability to regenerate lost body parts (Morgan, 1898; Saló et al., 2009; Forsthoefel, 2009). This extreme plasticity is due to the presence of a large population (20-30% of the total number of cells) (Hay and Coward, 1975; Baguñà, 1976; Baguñà and Romero, 1981; Newmark and Sánchez Alvarado, 2000; Hayashi et al., 2006) of somatic pluripotent stem cells known as neoblasts (Dubois, 1949; Wolff, 1962). Apart from germ cells, neoblasts are the only cells capable of proliferating in adult planaria (Morita and Best, 1984; Newmark and Sánchez-Alvarado, 2000). They reside in the thickness of the parenchyma (Baguñá, 1973; Newmark and Sánchez Alvarado, 2000; Orii et al., 2005) and are distributed, with few exceptions, throughout the entire body of the animal. Neoblasts maintain tissue homeostasis (Baguñà et al., 1989), generate the germline (Sato et al., 2006) and replace structures lost to damage or amputation through a complex process of pattern re-establishment (re-patterning) (Saló and Baguñà, 1984; Baguñà et al., 1989; Saló, 2006), which begins with the formation of an undifferentiated, unpigmented tissue called blastema (Dubois, 1949; Morita and Best, 1984; Saló and Baguñà, 1984; Newmark and Sánchez Alvarado, 2000). The blastema, however, does not contain proliferating neoblasts. Instead, they are enriched in the region adjacent to the blastema, called post-blastema (PB) (Saló and Baguñà, 1984; Eisenhoffer et al., 2008).

Neoblasts are actively recruited to the PB through signals elicited by the wound response and probably through other subsequent signalling (apoptosis, Wnt-P1) (Pellettieri et al., 2010; Petersen and Reddien, 2009). In the PB, neoblasts actively proliferate to regenerate the lost body part, mainly in two temporally defined waves: the minor one after eight hours, the major one between 48 and 72 hours (Saló and Baguñà, 1984). A similar proliferative response was also found following feeding, especially after a period of starvation (Baguñá, 1976; Newmark and Sánchez Alvarado, 2000). The molecular mechanisms underlying the regulation of neoblast proliferation and differentiation remain, however, largely unknown.

The LSm (like-Sm) RNA-binding protein superfamily is a set of genes involved in many aspects of the RNA metabolism (for a review, see Lührmann et al., 1990). LSm protein products are highly conserved throughout metazoans and relatively well-conserved in bacteria and archaea (Schumacher et al., 2002). LSm proteins are characterized by the presence of the Sm fold, a chaperone-like domain that allows a variety of RNA-RNA and RNA-protein interactions (for a review, see Wilusz and Wilusz, 2005). Complexes of six or seven proteins (the LSm ring) (Khusial et al., 2005) interact with different small nuclear RNAs (snRNAs) to generate small nuclear ribonucleoproteins (snRNPs). LSm proteins take part in several processes such as Cajal body formation, telomere elongation (Fu and Collins, 2006) and ribosomal assembling (Kiss, 2004). LSm proteins are, however, best known for their role in mRNA processing (He and Parker, 2000). They are also involved in the organization of the major and minor spliceosome complexes, which are responsible for the splicing processes of the pre-mRNA (Mayes et al., 1999; Patel et al., 2002). Although these complexes are mainly found in the nucleus, virtually all metazoan germ cells display ribonucleoprotein (RNP) granules in the cytoplasm (Eddy, 1975), like the germ granules in Caenorhabditis elegans and Drosophila melanogaster (reviewed in Strome and Lehmann, 2007) and the nuages or chromatoid bodies (CBs) in mammalian spermatids (reviewed in Parvinen, 2005).

A recent report showed that mammalian cell lines can also be induced to form CB-like structures following the expression of a cytoplasmic form of the prion protein PrP (Beaudoin et al., 2009). Two other studies proposed that LSm proteins regulate cell proliferation in the cytoplasm of mammalian cells through the minor spliceosome (König et al., 2007) and, if perturbed, can interfere with the localization and subcellular distribution of the P-granules of C. elegans (Barbee et al., 2002). Planarian CBs are electron-dense RNP particles characteristic for stem cells (Hay and Coward, 1975) and neurons (Yoshida-Kashikawa et al., 2007). Until now, only two proteins have been described as components of the planarian CB: DjCBC-1 (Yoshida-Kashikawa et al., 2007) and Spoultd-1 (Solana et al., 2009). Their function, not yet fully understood, is related to RNA processing and post-transcriptional modifications (Yoshida-Kashikawa et al., 2007).

In this study, we characterize the first member of the large LSm protein superfamily, Smed-SmB, described in the planarian species Schmidtea mediterranea. Smed-SmB is the ortholog of mammalian SmB/B′/N, and is expressed by stem cells and neurons. Double-stranded RNA (dsRNA)-mediated silencing in intact and regenerating animals resulted in a quick reduction in the number of dividing cells, which compromised both the ability to ensure normal tissue turnover and regeneration after amputation. A massive degenerative process progressed from anterior to posterior in both regenerating and intact animals and caused the animals to die within two to four weeks, respectively.

Our data show that inhibition of Smed-SmB impairs the organization and function of the CB, to which the protein also localizes. Furthermore, Smed-SmB RNAi affects the splicing efficiency of Smed-CycB, which indicates a failure to process pre-mRNAs necessary for cell cycle progression. The degeneration of the CBs and the impairment of the splicing efficiency seem to have a central role in the resulting proliferative failure we observed.

All planarians used in the present study were generated from a single individual of the asexual strain of Schmidtea mediterranea (clonal line BCN-10), collected from a fountain in Montjuic (Barcelona, Spain). Animals were maintained in a 1:1 mixture of tap water treated with AquaSafe (Tetra Aqua) and MilliQ water (Millipore Iberica, Madrid, Spain) and fed twice a week with bovine liver, as previously described (Molina et al., 2007). Animals used for all the presented experiments were starved for one week. To prepare irradiated controls, planarians were γ-irradiated at 75 Gy (1.56 Gy/minute) with a Gammacell 1000 (Atomic Energy of Canada Limited) (Saló and Baguñà, 1985). The animals were collected at different time points (at 12 hours, daily from days 1-11 and at days 14, 15 and 21) according to downstream applications.

The peptide sequence of Smed-SmB was identified among putative neoblast-specific proteins (E.F.-T., G. Rodriguez-Esteban, E.S. and J. Abril, unpublished data), and confirmed by BLAST in a local database generated from the S. mediterranea genome traces (sequenced by Washington University Genome Sequencing Center, USA). Primers were designed to amplify a 237 bp fragment by RT-PCR (forward primer, 5′-CAACACTTCAAGATGGTCG; reverse primer, 5′-CCATTAACGCCTACTGAA) from total RNA preparation (TRIzol). 5′- and 3′-ends were obtained by RNA ligase-mediated rapid amplification of cDNA ends (RLM-RACE)-PCR (5′ RACE primer, 5′-GACUGGAGCAGGAGGACACUGACAUGGAGUGAAGGAGUAGAAA; 3′ RACE primer, 5′-GCTGTCAACGATAGGCTACGTAACGGCATGACAGTG[T]₁₈). The complete sequence was confirmed by BLAST to the planarian genome draft (accession number: GU562964).

Planarian cell dissociation protocol was modified from González-Estévez (González-Estévez et al., 2007). Briefly, animals were treated with 2% L-cysteine hydrochloride monohydrate (Merck, K23484539), pH 7.0, to remove the mucus, transferred in CMF buffer and mechanically dissociated by gentle rocking. The suspension was then serially filtered through 70 and 30 μm nylon meshes, stained for 2 hours with a mixture of 7.5 μg/ml Hoechst 33342 (Sigma, B2261) and 0.05 μg/ml Calcein AM (Invitrogen, Karlsruhe, Germany), pelleted and resuspended for FACS (FACSARIA, BD, Heidelberg, Germany). Eventually, propidium iodide was added at concentration of 1 μg/ml.

Wholemount in situ hybridization (ISH) protocol was modified from Agata (Agata et al., 1998). Briefly, after fixation and rehydration, planarians were permeabilized in 0.1% Triton X-100 in PBS (PBST×0.1) added with 20 μg/ml proteinase K and quenched with 2% glycine. Triethanolamine (Sigma, T9534) treatment was performed as previously described (Nogi and Levin, 2005) before hybridization, which was carried out at 55°C for 18 to 30 hours, using digoxigenin-labelled RNA probes prepared using an in vitro labelling kit (Roche, Mannheim, Germany) at a final concentration of 0.07 ng/μl. ISH on dissociated cells was performed as previously described (González-Estévez et al., 2003).

Wholemount immunohistochemistry (IHC) was performed as described (Cebriá and Newmark, 2005). Briefly, after fixation in Carnoy and rehydration, animals were blocked in 1% BSA in PBST×0.1 for 2 hours at room temperature. The antibody against proliferating cell nuclear antigen (α-PCNA), kindly provided by Dr Hidefumi Orii (Himeji Institute of Technology, Hyogo, Japan), was used 1:1000 in blocking solution for 20 hours under constant agitation at 4°C. Animals were then labelled for 14-16 hours with FITC-conjugated donkey anti-rabbit (Jackson ImmunoResearch, 1:200). PCNA-positive cells were counted using a Leica SP2 confocal laser-scanning microscope (Leica Microsystems, Barcelona, Spain).

In vitro dsRNA preparation of Smed-SmB 5′-region (500 bp) and injection procedure were previously described (Oviedo and Levin, 2007). A total of three injections, distributed over 3 consecutive days (one injection per day with three pulses of 32 nl of 400 ng/μl dsRNA solution) were performed using a Drummond Scientific Nanoject injector (Broomall, PA, USA) targeting the gastrovascular system. The control group was injected with water. Some animals were pre-pharyngeally amputated 1 or 3 days after the last injection (t-cut).

Total RNA was extracted using TRIzol (Invitrogen, Barcelona, Spain) and reverse-transcribed to cDNA with the High-Capacity cDNA Reverse Transcription Kit (AppliedBiosystems, Darmstadt, Germany). Transcript levels were determined on the ABI PRISM SDS 7900HT (AppliedBiosystems, Darmstadt, Germany) using custom-designed oligonucleotides for the 5′-nuclease assays (see Table S1 in the supplementary material) and normalized to the endogenous control SmEf2 that was chosen among other ubiquitously expressed genes for its stable expression throughout the regenerative process, as well as for not being affected by the RNAi treatment performed (see Fig. S1, Fig. S2 in the supplementary material). Relative quantification of gene expression was calculated using the ΔΔCt method. Three technical replicates were used for each real-time PCR reaction; a reverse transcriptase blank and a no-template blank served as negative controls.

For Smed-CycB RT-PCR of spliced and unspliced forms, the equivalent of 5 ng of reverse-transcribed RNA were amplified using oligonucleotides spanning the second intron (forward primer, 5′-ATGCCGCCGAAACTTTATACCTG; reverse primer, 5′-AACTTCTTCGACTTTTGCTGCAA), according to mk4.001494.00.01. Expected amplicon sizes for the spliced and unspliced forms of the transcript were 136 bp and 219 bp, respectively.

Post-blastema fragments of Smed-SmB dsRNA- and water-injected planarians were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, for 2 hours at room temperature. Specimens were post-fixed in 1% OsO₄ and embedded after dehydration in epon. Ultrathin sections of 70 nm were cut (Leica-UC6 ultramicrotome, Vienna, Austria) and counterstained with uranyl acetate and lead citrate. Specimens were observed at 80 kV on a FEI-Tecnai 12 electron microscope (FEI, Eindhoven, Netherlands). Pictures were taken using imaging plates (Ditabis, Pforzheim, Germany).

For statistical analysis, neoblasts in the PB area were inspected and number, size and morphology of their chromatoid bodies were accounted.

For immuno-EM, fragments of planarians were fixed with either 4% paraformaldehyde or a mixture of 2% paraformaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4. The specimen was then cryoprotected in 2.3 M sucrose and frozen in liquid nitrogen. Ultrathin cryosectioning and immunogold-labelling were performed as described (Slot and Geuze, 2007), except a permeabilization treatment with 1 mg/ml saponin in PBS, pH 7.4, for 10 minutes prior to blocking to improve the accessibility of the antibodies to their epitopes.

Animals were snap-frozen in liquid nitrogen and ground with a disposable potter. After direct lyses in 50 μl 2× Laemmli buffer, samples were loaded on a 10% acrylamide minigel and run under denaturing SDS-PAGE conditions for 3 hours under constant current. Proteins were transferred on a PVDF Immobilon membrane (Millipore, Schwalbach, Germany) and processed for immunodetection with ECL plus chemistry (GE-Healthcare, Solingen, Germany). Antibodies used were: mouse anti-SmB monoclonal antibody (Sigma S0698) diluted 1:1000; rabbit anti-Gapdh polyclonal antibody (Abcam, ab36840) diluted 1:5000; and rabbit anti-Smed-SmB (residues 46-60) affinity-purified polyclonal antibodies SmB310 and 311, both diluted 1:250,000. Secondary HRP-conjugated antibodies were used at 1:50,000, according to the host species of the primary antibodies. Band intensities were quantified using the Chemi-Doc XRS platform equipped with Quantity One software (BioRad, Munich, Germany).

One-way ANOVA with Dunnett's post-test, Student's t-test and Fisher's exact test were performed using GraphPad Prism 4.03 (GraphPad software, La Jolla, CA, USA).

Despite being represented throughout the planarian body and experimentally well accessible, neoblasts are still largely defined by their morphology and by their sensitiveness to γ-irradiation. Therefore, γ-irradiated planaria are instrumental in the discovery of new neoblast-specific markers. Using a proteomic approach, we were able to identify the first planarian member of the LSm superfamily as differentially expressed in wild-type versus γ-irradiated regenerating animals (E.F.-T., G. Rodriguez-Esteban, E.S. and J. Abril, unpublished data). ClustalW alignment showed the conservation of the two Sm fold domains at the N-terminus of the protein (see Fig. S3A in the supplementary material), whereas the bootstrap-based UPGMA method was used to calculate the phylogenetic tree that confirmed Smed-SmB similarity with the SmB proteins of different metazoans (see Fig. S3B in the supplementary material).

Neoblasts are evenly dispersed throughout the parenchyma, a scarcely differentiated tissue that lies under the epidermal and muscle layers, and surrounds all the organs (Baguñá, 1973; Saló, 2006). Wholemount in situ hybridization (ISH) against Smed-SmB transcripts showed the typical neoblast distribution in adult intact planarians (Fig. 1A, 1-3). Fluorescent ISH on dissociated cells revealed that Smed-SmB signal could be detected in small (5-10 μm) round cells with a high nucleus-to-cytoplasm (N/C) ratio (Fig. 1B-D), consistent with the morphological definition of neoblasts (Baguñà, 1981). Additionally, wholemount ISH against Smed-SmB on regenerating animals 48-72 hours after amputation produced a strong signal in the post-blastema (PB; Fig. 1E), where proliferating neoblasts consequently accumulate to the major proliferation peak induced by amputation (Baguñà, 1976).

Following γ-irradiation, Smed-SmB expression was significantly diminished, although irradiation-tolerant Smed-SmB-positive cells persisted in the cephalic region (Fig. 2A), as reported for other neoblast markers like Bruli (Guo et al., 2006) and Pumilio (Salvetti et al., 2005). This observation was further substantiated as the relative expression of the transcript varied depending on the portion of the body considered. Fourteen days following γ-irradiation, animals were cut into head, trunk (pharynx and surrounding tissues) and tail fragments (Fig. 2B). The expression of Smed-SmB mRNA in the cephalic portion of irradiated animals was similar to the non-irradiated control (P=0.5667), whereas in the trunk and tail fragments it was 29±3% (P<0.0001) and 33±3% (P=0.0005), respectively (Fig. 2C). Western blot confirmed a similar trend for the expression of Smed-SmB protein (Fig. 2D,E).

In order to understand the function of Smed-SmB, we performed RNAi experiments on regenerating and intact animals. RNAi efficiency was assessed at both the RNA and protein levels. qRT-PCR showed a sudden downregulation of the transcript (more than 90%), which followed the same kinetics in both intact and regenerating animals (Fig. 3A). Smed-SmB protein expression was also clearly influenced by the RNAi, showing different kinetics between intact and regenerating animals. In regenerating animals, Smed-SmB protein was barely detectable as early as 2 days after amputation. A comparable level of protein expression is attained in intact animals after 9 days from the last dsRNA injection (Fig. 3A).

All Smed-SmB dsRNA-treated animals (n=95), which were amputated 1 or 3 days after the third round of injections (t-cut), failed to regenerate and died within two weeks after amputation. Interestingly, in no cases could we observe blastema formation (Fig. 3B). Within three to four weeks following RNAi, intact animals (n=50) also died, with signs of degeneration becoming evident 10 days after the RNAi treatment (Fig. 3C).

To understand whether Smed-SmB has a neoblast-specific function, we observed the changes induced by Smed-SmB RNAi to the planarian cell populations. Largely based on the morphological features of planarian cells, Hoechst 33342/Calcein AM double-staining is able to separate the irradiation-sensitive cells (X1 and X2) from the differentiated cells (Xin) on a FACS plot (Reddien et al., 2005; Hayashi et al., 2006). In a similar way to γ-irradiation, Smed-SmB RNAi induced a dramatic reduction of irradiation-sensitive fractions X1 and X2 in intact animals (Fig. 4A). Irradiation-sensitive X1 and X2 cells were drastically reduced in number after 11 days (18±8.3% and 27±11.3% of the relative control, respectively; P<0.0001), with the decrease beginning after 7 days (P<0.05; Fig. 4B,C). These data nicely correlate with the availability of Smed-SmB protein shown in Fig. 3A. FACS analysis of PB tissue separated from the rest of the body (RoB) of regenerating RNAi animals showed that the reduction of X1 and X2 cells happened much faster in the PB (day 3, P<0.05), where actively proliferating neoblasts are enriched compared with the RoB (day 6, P<0.05; Fig. 4D-G).

We analyzed the expression of neoblast-specific markers in both intact and regenerating animals and we found they were, in both cases, significantly reduced following Smed-SmB RNAi. qRT-PCR of regenerating animals showed that Smed-Bruli (Guo et al., 2006) and Smedwi1 (Reddien et al., 2005; Rossi et al., 2006) expression rapidly disappeared, beginning 3 days after Smed-SmB RNAi (P=0.0002 and P=0.0013, respectively; Fig. 5A). By contrast, the expression of differentiated cell markers, like Smed-Pax6 (Pineda et al., 2002) and Smed-Tmus (Cebriá et al., 1997), was essentially unaltered (P=0.4551 and P=0.0983, respectively; Fig. 5B). The expression of Smedwi2, involved in neoblast differentiation (Reddien et al., 2005), also suffered a significant reduction (P<0.05), although milder and at a later time (day 5) than the other neoblast markers (Fig. 5A).

Given the incapacity shown by Smed-SmB knocked-down animals to form a regenerating blastema, we decided to investigate neoblast proliferative potential. PCNA-positive cells were reduced in number after 10 days from Smed-SmB knockdown in intact animals (P<0.0001; Fig. 5C-E). PCNA is an essential component of the DNA replication machinery that accumulates in cells that are virtually able to proliferate (Bravo et al., 1987; Orii et al., 2005). Conversely, cyclinB is expressed only by actively cycling neoblasts of the X1 subpopulation (Reddien et al., 2005; Eisenhoffer et al., 2008). In wild-type regenerating planarians, its expression increases around day 2-3 of regeneration (P<0.05; Fig. 5F), related to the major peak of mitotic activity (Saló and Baguñà, 1984). As a consequence of Smed-SmB RNAi, Smed-CycB expression was promptly reduced 2 days after amputation (P<0.05), reaching a basal level (more than 90% downregulation; P<0.0001) after 7 days (Fig. 5F).

Using as primer oligonucleotides able to amplify both the spliced and unspliced forms of the Smed-CycB transcript, we observed that the decrease in mature Smed-CycB mRNA expression was accompanied by an increase in the expression of its unspliced form (Fig. 5G). The ratio between the expressions of unspliced and spliced transcripts in Smed-SmB RNAi animals increased up to 5-fold after 3 days of regeneration (Fig. 5H). This suggests that Smed-SmB RNAi does not prevent the expression of cyclinB, but rather, at least in part, prevents the splicing of its transcript.

We found that Smed-SmB is important to guarantee the proliferative ability of the neoblasts. We then had a closer look at the ultrastructural level of stem cells in both intact and regenerating Smed-SmB RNAi animals. We noticed that, after 5 days of regeneration, the cytoplasm of the neoblasts was less-densely packed (Fig. 6A) compared with water-injected animals (Fig. 6B). This change was accompanied by an increase in cellular membranes of the endolysosomal system and the Golgi (Fig. 6A, asterisks) — normally not present in neoblasts (Baguña, 1973). CBs are electron-dense granules not delimited by membranes (Fig. 6B, arrows), which resemble the appearance of heterochromatic nuclear spots (Auladell et al., 1993). A loss of integrity in the CB structure was also observed at higher magnification (Fig. 6A,C, white frame). Several small CB-like structures (Fig. 6C, arrowheads) appeared, probably owing to the fragmentation of pre-existing CBs. Occasionally, we also observed the presence of large autophagic vacuoles (Fig. 6D, asterisks), suggesting that the decrease in the X1 and X2 populations might be regulated by this degenerative pathway (González-Estévez et al., 2007). Interestingly, we also found that the degeneration of the CBs follows the Smed-SmB protein availability (Fig. 3A). In intact animals, we occasionally observed degenerated CBs, but only after 11 days from RNAi treatment onwards. In regenerating animals, however, the CB state reflected the neoblast position in the planarian body. CBs observed in neoblasts of the posterior stripe began to degenerate after 5 days from amputation, whereas those found in neoblasts of the PB began to degenerate as early as 1 day after cutting (P<0.05; Fig. 6E). As CBs degenerated, an increasing number of neoblast-like cells without CBs were observed in the PB. The percentage of these cells increased from 11.8±1.8% at t-cut to 50.5±12.5% after 3 days (P<0.05), with a clear trend (R²=0.923; Fig. 6E). Conversely, the percentage of healthy neoblasts with normal CBs dropped from 69.4±8.5% at t-cut to 13.7±11.1% after 3 days (P<0.05; Fig. 6E).

The evidence we collected at the ultrastructural level support the idea that knockdown of Smed-SmB impairs the functional organization of the CB, as was also observed for the germ granules of C. elegans after RNAi of either SmE or SmB (Barbee et al., 2002).

The dynamics of the event characteristics of the Smed-SmB phenotype showed that the degeneration of the CBs is one of the first outcomes of the phenotype, particularly when considering the neoblasts located in the PB (Fig. 6E; Fig. 7A). As several articles proved the localization and functional role of the LSm proteins in CBs and CB-like structures (Parvinen, 2005; Strome and Lehmann, 2007; Beaudoin et al., 2009), we tried to uncover whether Smed-SmB protein also localizes to the CB. Cryoimmuno-EM using two different affinity-purified rabbit polyclonal antibodies (SmB310 and SmB311), specifically raised against residues 46-60 of the Smed-SmB protein, showed gold particle labelling of both the nucleus — except the nucleolus — and the CB of wild-type neoblasts (Fig. 7B). These were mainly labelled at the edge, probably as consequence of steric hindrance of the antigens in the condensed structure. Results were confirmed by using both the antibodies specifically raised. Nuclear Smed-SmB signal was not found in differentiated cells (data not shown). These results indicate with good confidence the presence of the Smed-SmB protein inside as well as outside the nuclear compartment of planarian neoblasts.

In this study we characterized the function of the planarian SmB/B′/N ortholog, Smed-SmB, the first member of the conserved LSm superfamily identified in planarians.

The data we collected suggest an essential role for Smed-SmB in sustaining the proliferative potential of the planarian stem cells during regeneration and in tissue homeostasis.

Smed-SmB transcripts were mainly found in the parenchyma and around the cephalic ganglia. Expression of Smed-SmB by postmitotic cells of the cephalic ganglia explains the partial downregulation we found after γ-irradiation, which only affects actively dividing cells.

The silencing of Smed-SmB, efficiently achieved through dsRNA-induced RNAi, produced a late degenerative phenotype in intact animals (Fig. 3C) and an early non-regenerating phenotype in amputated ones (Fig. 3B). The phenotype was accompanied, in both cases, by downregulation of the neoblast markers (Fig. 5B) and reduction in the number of neoblasts (Fig. 4A-G). These effects were also described following the perturbation of other genes — like Pumilio, Smedwi2 and Bruli — involved in neoblast maintenance or differentiation (Salvetti et al., 2005; Reddien et al., 2005; Guo et al., 2006). Compared with the latter, however, the Smed-SmB phenotype showed some peculiarities.

The degenerative effects produced by Smed-SmB knockdown in intact animals progressed at a slower rate compared with the other genes essential for neoblast function. At the beginning of the fourth week, more than 60% of the intact animals were still alive, albeit with a degenerated anterior part (Fig. 3C). This implies that maintenance of the differentiated cells was virtually unaffected (Fig. 4A-C), and that the delay in degeneration observed in intact animals was due to the prevention of tissue turnover.

Also, in Smed-SmB RNAi regenerating animals, the inability to form a blastema was not affected by the length of time between injection and amputation, as observed in the case of Smed-Bruli and Smedwi2 RNAi (Guo et al., 2006; Reddien et al., 2005). Animals cut at different time points after injection (1 or 3 days) invariably failed to form a blastema, suggesting that neoblasts could proliferate, and possibly differentiate, for a short time after either Smed-Bruli or Smedwi2 RNAi, but they were unable to do so following Smed-SmB RNAi. On the contrary, proliferation was severely hampered, as depicted by a decrease in the expression of both PCNA (Fig. 5C-D) and Smed-CycB (Fig. 5E). Although we cannot exclude an effect of Smed-SmB RNAi on the early neoblast progeny, we concluded that the inability to regenerate a blastema as demonstrated by Smed-SmB RNAi animals resulted from the absence of actively proliferating cells.

Another interesting observation was that the expression of a gene, which was otherwise ubiquitously expressed and involved in basal cell functions, was restricted in planaria to stem cells and neurons. Conversely, SmB is known to be a component of RNP granules, also known as CBs, which are present in the germ cells of virtually all metazoans (Eddy, 1975). Recently, their presence was also observed in CB-like structures induced in mammalian cells by expressing the cytoplasmic form of the Prion protein (Beaudoin et al., 2009). The large RNP complexes found in the germ cells share remarkable similarities with planarian CBs, which are a characteristic feature of neoblasts (Hay and Coward, 1975), germ cells (Coward, 1974) and neurons (Yoshida-Kashikawa et al., 2007) and are expected to function in posttranscriptional gene regulation.

Like the germ cell granules, planarian CBs assemble components of the machinery to process mRNA in the cytoplasm, like DjCBC-1 (Yoshida-Kashikawa et al., 2007) and Spoltud-1 (Solana et al., 2009). They might also contain small non-coding RNAs (miRNAs, snRNAs) and members of the LSm protein superfamily. Following neoblast differentiation, CBs gradually disappear, only to persist in the germline (Coward, 1974) and in neurons (Yoshida-Kashikawa et al., 2007).

The Smed-SmB RNAi effects we observed at the CB level showed striking similarities with the knockdown of either SmE or SmB in C. elegans embryos, which alters the sub-cellular distribution of the P-granules and leads to the inaccurate segregation of the germline (Barbee et al., 2002). Our data suggest that Smed-SmB RNAi results in a loss of CB structure and organization, which subsequently translates into a loss of function (Fig. 6C). As generally believed, CBs store the transcripts to provide the cell with a quick response to physiological needs (Kotaja and Sassone-Corsi, 2007). Hence, it seems plausible that the observed degeneration of the CBs triggers the proliferative failure of the neoblasts and the inability of the animals to form a blastema. This is in contrast to the Spoltud-1 phenotype, where the CB is probably unaffected, but the neoblast population is depleted regardless, even though after a much longer time (Solana et al., 2009). The observed increase of unspliced Smed-CycB transcripts suggests that Smed-SmB RNAi additionally leads to a splicing defect of genes essential for neoblast proliferation. The moderate increase in the expression of cyclinB pre-mRNA is probably underestimated owing to its active degradation by nonsense-mediated decay following the introduction of stop codons into the transcript sequence by the unspliced introns (for a review, see Isken and Maquat, 2008). Strikingly, the degeneration of the CBs in neoblasts of the PB occurs in parallel to Smed-SmB protein downregulation. Therefore, we propose that CB degeneration is not the effect of a general splicing defect caused by the loss of Smed-SmB. Such an indirect effect would require the degradation of both the mRNAs and proteins of Smed-SmB splicing targets, a process we would expect to take longer to produce the degeneration and disappearance of the CBs that we observed.

Without the possibility to trace planarian cells and look at their fate, we could only speculate about the fate of Smed-SmB RNAi neoblasts. Nonetheless, we frequently observed lysosomes and autophagic vacuoles (Fig. 6A,D) in Smed-SmB RNAi neoblasts, suggesting that these cells might undergo resorption by autophagy (González-Estévez et al., 2007). It is interesting, however, that neoblasts are able to survive longer if not actively cycling, as in the case of intact RNAi animals. Non-regenerating animals rely on their pool of stem cells to guarantee cellular turnover, and this is the reason why the neoblast population of intact animals is virtually unaffected by Smed-SmB knockdown for up to 7 days after the RNAi treatment (Fig. 4A-D). This capacity to survive the Smed-SmB RNAi for an extended period of time is probably owing to the long half-life of the Smed-SmB protein (Fig. 3A; Fig. 7A).

In this work, we have shown that a member of the LSm superfamily, Smed-SmB, is required for the proliferation and maintenance of the adult stem cells of Schmidtea mediterranea. Smed-SmB is expressed by stem cells and neurons. Knockdown of Smed-SmB expression, however, apparently affects only neoblasts, causing the downregulation of the expression of specific neoblast markers like Smedwi1 and Smed-Bruli. Expression of the proliferation markers PCNA and Smed-CycB was also quickly impaired, resulting in a massive proliferative failure and drastic reduction of viable neoblasts. As a direct consequence, regenerating animals do not form blastema and die within 2 weeks. Smed-SmB knockdown also leads to the death of intact animals, presumably as consequence of a lack of cellular turnover. At the subcellular level, we observed a severe degeneration of the CBs. Within 3 days, the number of neoblasts with ‘healthy’ CBs in the PB was reduced to 1 out of 8. By immunogold-labelling, we could localize the Smed-SmB protein to the nucleus of planarian stem cells and to the CB. In addition, we also found an increased expression of unspliced Smed-CycB transcripts. The latter is indicative of a failure of the splicing machinery, which further impairs the proliferative capacity of the stem cells.

We thank H. Orii for providing DjPCNA antibody, C. Ortmeier and G. Verberk for real-time qRT-PCR, K. Mildner for technical assistance with the electron microscopy sample preparation, J. Müller-Keuker for assistance with the figures and S. Kölsch for proofreading. This work was supported by grant BFU-2005-00422 and BFU2008-01544 from the Ministerio de Educación y Ciencia (Spain) and grants 2005SGR00769 and 2009SGR1018 from AGAUR (Generalitat de Catalunya, Spain) to E.S., grant DAAD D/06/12871 to H.R.S., and E.F.T. received an FPI fellowship from the Ministerio de Ciencia y Cultura, Spain.