Expression patterns of engrailed and dpp in the gastropod Lymnaea stagnalis (original) (raw)

Introduction

The Mollusca is a diverse metazoan phylum, comprising of eight extant classes, and its diversity is only second to that of Arthropoda in the living animals (Lydeard and Lindberg 2003). It is placed among Lophotrochozoa, the sister group of Ecdysozoa in Protostomia (Halanych et al. 1995; Aguinaldo et al. 1997). The possession of highly diversified calcareous shells is a major feature characterizing polyplacophoran and conchiferan molluscs (Salvini-Plawen and Steiner 1996; Brusca and Brusca 2002). Those shells provide an excellent fossil record, which suggests a molluscan origin in the late Precambrian before the “Cambrian explosion” (Runnegar and Pojeta 1985; Runnegar 1996; Brusca and Brusca 2002), and allow us to trace their evolutionary histories since then. But how those molluscan shells first evolved and how they diversified are poorly known, although the clues may be found in the molecular and genetic mechanisms of the shell formation.

Many shell matrix proteins have been identified in molluscs to infer the possible mechanisms of shell formation (see for example, Marin and Luquet 2004; Sarashina and Endo 2006 for reviews). A substantial part of the mechanisms, however, still remains to be resolved. As a complementary approach, upstream genes to those shell matrix protein genes, more specifically the transcription factor and signal molecule genes that may be or could be involved in molluscan shell formation, have also been identified in the recent years.

In the snail Ilyanassa obsoleta, the transcription factor Engrailed was detected in the shell gland progenitor and progeny, and the expression of engrailed mRNA was also observed in the shell plate (Moshel et al. 1998). Subsequently, a close association of the engrailed gene expression with embryonic regions responsible for shell secretion was observed for chiton (Jacobs et al. 2000), clam (Jacobs et al. 2000), tusk shell (Wanninger and Haszprunar 2001), limpet (Nederbragt et al. 2002), cuttlefishes (Baratte et al. 2007; Shigeno et al. 2008), and nautilus (Shigeno et al. 2008). In the limpet Patella vulgata, engrailed expression was localized in the shape of a semicircular band of the shell-forming cells along the shell margin, and the expression of the signal molecule dpp-BMP2/4 was detected in the shape of a circular band in the cells adjacent to and outside of the shell-forming cells that expressed the engrailed gene (Nederbragt et al. 2002). In the abalone Haliotis asinina, the transcription factors Hox1 and Hox4 are expressed in the mantle margin at the veliger stage (Hinman et al. 2003).

Based on this association of gene expression with the shell-forming region, a role in skeletogenesis was suggested for engrailed and for Hox1 and Hox4, in the chiton and clam (Jacobs et al. 2000) and in the abolone (Hinman et al. 2003), respectively. But a role in the compartmentalization to distinguish shell-forming area and the adjacent ectoderm, rather than skeletogenesis, was suggested for engrailed and dpp in the limpet (Nederbragt et al. 2002) and for engrailed in the cuttlefish (Baratte et al. 2007). In the snail Ilyanassa, it was implied that engrailed participates in establishing the polarity and regionalization of the embryo (Moshel et al. 1998). It is, therefore, still unclear how those transcription factors and signal molecules work in the mechanisms leading to the molluscan shell formation, despite the fact that seven species in five molluscan classes have been examined. In order to make clear the genetic basis of shell formation, it would be desirable to establish a model system that allows detailed genetic and embryological studies.

To this end, the pond snail Lymnaea stagnalis is well-suited because the adult individuals of L. stagnalis are easily raised in the laboratory and reproduce after 5 weeks after hatching throughout the year (Meshcheryakov 1990), chiral strains exist, allowing the examination of dextral and sinistral shell morphologies, the decapsulated embryos can survive until the beginning of gastrulation (Meshcheryakov 1990; Freeman and Lundelius 1992), and embryos subjected to micromanipulation in capsules can develop to juveniles (Arnolds et al. 1983). Among the pulmonate gastropods, a genome project is the proceeding in Biomphalaria glabrata (http://biology.unm.edu/biomphalaria-genome/index.html), potentially providing useful insight into the studies of pulmonate gastropods including L. stagnalis. Developmental studies detailed by Morril (1982) and Meshcheryakov (1990) provide an excellent foundation for the future molecular developmental studies of L. stagnalis. Furthermore, it has already been examined from the wide-ranging aspects including those of the genetics (Hosoiri et al. 2003), neurobiology (Voronezhskaya and Elekes 2003; reviewed in Croll 2000), and shell biomineralization (Endo et al. 2004; Sarashina et al. 2006).

In the present study, we identified the full-length cDNAs of engrailed and dpp orthologues in L. stagnalis and examined their spatiotemporal expression during development by the whole mount in situ hybridization (WMISH). The results indicated conserved expression patterns for engrailed but divergent patterns for dpp, implying that the pathways for shell formation may be diversified even among the gastropods.

Materials and methods

Animals and embryos

Adult individuals of the snail L. stagnalis (strains GSS7-1 and GSS22; Endo et al. 2004) have been bred in a deionized tap water at around 25°C and fed with lettuce leaves. Fertilized eggs in capsules coated by jelly are laid throughout the year.

DAPI staining

Embryos were washed with phosphate-buffered saline (PBS, 22.3 mM NaH2PO4, 77.4 mM Na2HPO4, 100 mM NaCl) three times for 10 min at room temperature (RT) and stained with 0.2 μg/mL of 4′, 6-diamino-2-phenylindole (DAPI; Dojindo) in PBS for 10 min in the dark at RT. Embryos were observed under the fluorescence microscope (BX51, Olympus) after washing with PBS three times for 10 min at RT. Pictures were taken as digital images (VB6000, Keyence).

Scanning electron microscopy

After the removal of coating jelly by rolling the capsules on a sheet of low lint tissue paper, the embryos of L. stagnalis were dechorionated using a tungsten needle and tweezers in 10% Holtfreter’s solution (Freeman and Lundelius 1992) under a binocular. Embryos were then fixed with 2.5% glutaraldehyde in PBS at RT for an hour. After several washes with PBS, specimens were dehydrated through the ethanol series of 30%, 50%, 80%, and 100% ethanol in PBS for 15 min each. The ethanol was replaced with _t_-butylalcohol three times for 15 min at RT. Specimens in _t_-butylalcohol were freeze-dried and sputter-coated with gold (JFC-1200, JEOL). Specimens were examined with the JSM-5500LV (JEOL) scanning electron microscopy (SEM) at 15 kV.

Amplification of engrailed and dpp gene fragments

Genomic DNA of L. stagnalis was isolated as described previously (Iijima et al. 2006). The degenerate primers were designed basically referring to Nederbragt et al. (2002).

Polymerase chain reaction (PCR) was performed, with EX-Taq DNA polymerase (TAKARA), using the genomic DNA as template and a pair of degenerate primers (0.5 pM/L for each), engrailed-F and engrailed-R for engrailed and dpp-F and dpp-R for dpp. The cycling condition was the following: an initial step (94°C for 1 min), followed by a step of 30 cycles (94°C for 30 s; 50°C for 30 s; 72°C for 1 min), and a last step (72°C for 5 min; GeneAmp PCR System 9700, ABI). The second PCR was performed using the first PCR product as the template.

The PCR products were verified by a gel electrophoresis and purified using a Gel Extraction Kit (QIAGEN). The purified PCR products were ligated into the pGEM-T vector using a DNA ligation kit (Promega). The recombinant plasmids were then used to transform the competent Escherichia coli DH5α (TOYOBO). Cloned PCR products were sequenced by a BigDye Terminator v3.1 Cycle Sequencing Kit (ABI) with ABI Prism 3100-Avant DNA Sequencer (ABI).

Total RNA was isolated from some embryonic stages using ISOGEN (Nippon Gene). The total RNA was applied as template for cDNA synthesis by reverse transcription using ReverTraAce (TOYOBO) with the oligo-dT anchor primer, PN18-R1. Three-prime rapid amplification of cDNA ends (3′-RACE) was performed on the cDNA as template with the specific primers and PN-R1. Specific primers were designed based on the sequences of PCR products we amplified, and the primer sequences are shown below. Additional specific primers were designed also for 5′ RACE.

Complementary DNA for 5′ RACE was synthesized from an adult mantle mRNA, using SMART RACE cDNA Amplification kit (Clontech). The primer pairs universal primer mix (UPM) in the kit and LstR1 (first PCR) and nested universal primer (NUP) in the kit and LstenR2 (second PCR) were used to amplify the 5′-region of the engrailed cDNA, and the primer pairs UPM and LstdppR1 (first PCR) and NUP and LstdppR2 (second PCR) were used to amplify the 5′-region of the dpp cDNA. Full-length cDNAs of engrailed and dpp were obtained with the cDNA used for 5′RACE as template and pairs of the specific primers, LstF0 and LstR0 for engrailed and LstdppF0 and LstdppR0 for dpp.

Primer sequences used in the present study are as follows:

Corresponding amino acid sequences are in parentheses.

Sequences and phylogenetic analysis

Nucleotide and deduced amino acid sequences were analyzed using the basic local alignment search tool (Altschul et al. 1997) to identify the putative homology of the already reported sequences of engrailed or dpp. To detect the putative signal peptide and its cleavage site for Dpp, SignalP 3.0 software analysis (Emanuelsson et al. 2007) was performed on http://www.cbs.dtu.dk/services/SignalP/.

Sequences were aligned using the sequence alignment editor Se-Al v2.2a11 (http://evolve.zoo.ox.ac.uk/) or the ClustalW program (version 1.83). Distance trees were constructed using the ClustalW program with the neighbor-joining method under the default settings (Saitou and Nei 1987). We also performed the maximum likelihood (ML) estimation of phylogeny using the program ProtML (Adachi and Hasegawa 1992) implemented in Molphy version 2.3b3 (Adachi and Hasegawa 1996) using the JTT model (Jones et al. 1992), and the local rearrangement search mode was adopted for searching the topology. In the phylogenetic analyses, the amino acid sequences were used.

Probe synthesis and whole mount in situ hybridization

Digoxigenin (DIG)-labeled RNA probes were synthesized for engrailed using DIG RNA Labeling Kit [SP6/T7] (Roche). Fluorescein (FLU)-labeled RNA probes were synthesized for dpp using Fluorescein RNA Labeling Mix (Roche). One-kilobase fragments corresponding to the upstream regions adjacent to EH4 domain of Engrailed and transforming growth factor (TGF) domain of Dpp were amplified by PCR. Those PCR fragments were subcloned into pGEM-T vector to serve as a probe template after the linearization by digestion using the restriction enzymes (TOYOBO): for engrailed, _Pst_I and _Nco_I for antisense and sense probes, respectively; for dpp, _Not_I and _Nco_I for antisense and sense probes, respectively.

Embryos were fixed with 4% paraformaldehyde in 0.1 M 3-(_N_-morpholino)propanesulfonic acid (MOPS; pH 7.4; SIGMA) at 4°C overnight, dehydrated through ethanol series, and stored in 80% ethanol in MOPST (0.1% Tween 20 in 0.1 M MOPS) at −20°C.

Whole mount in situ hybridization was performed as described previously (Nederbragt et al. 2002). Hybridization was performed using RNA probes at a concentration of 2.5 ng/μL at 55°C overnight. After blocking the reaction in a Blocking Solution [0.5% Blocking reagent (Roche) in TN buffer (100 mM Tris-HCl, pH 7.5, 150 mM NaCl)] at RT for an hour, anti-DIG antibody reaction was performed for engrailed in the Blocking Solution containing a 1:2,000 dilution of the anti-Digoxigenin-AP, Fab fragments (Roche) at 4°C overnight. Similarly, anti-FLU antibody reaction was performed for dpp using the anti-Fluorescein-AP, Fab fragments (Roche). Embryos were stained in a staining solution prepared by dissolving a nitro-blue tetrazolium chloride/5-bromo-4-chloro-3′-indolyphosphate p-toluidine salt (NBT/BCIP) ready-to-use tablet (Roche) in the distilled water (final composition: 0.4 mg/mL NBT, 0.19 mg/mL BCIP, 100 mM Tris buffer, pH 9.5, 50 mM MgSO4). Staining reaction was stopped by washing with phosphate buffered saline with Tween 20 (PBST) when reached an appropriate signal level. Some specimens were observed after clearing in 50% glycerol of PBST.

Results

Normal development of L. stagnalis

The development of L. stagnalis that we observed is essentially the same as that described by Morrill (1982) and Meshcheryakov (1990). At around 25°C, early cleavages occur approximately every 1 h. Embryonic and larval directions are assigned (Fig. 1C) according to the scheme of Nederbragt et al. (2002). A spherical blastula develops at 24-h post-first cleavage (hpc). A flattened blastula begins to gastrulate at the vegetal pole at around 30 hpc to become the gastrulating blastula (Fig. 1A). Large cells known as head vesicles (Meshcheryakov 1990) are seen in an anterior region within the animal pole side both under SEM (Fig. 1B) and in DAPI staining (Fig. 2D). A gastrula [the late gastrula in Meshcheryakov (1990)] at around 36 hpc rotating in the capsule becomes more rounded with the shell gland in the dorsal side, and the shell gland begins to invaginate. We divided the gastrula stage into early and late stages. At the late gastrula stage, the shell gland begins to invaginate. DAPI treatment showed weakly stained nuclei in the center of the shell gland (Fig. 2E). Blastoporal stomodaeum in the ventral side opposite to the shell gland becomes narrower [compare mouth in Fig. 1C with blastopore in Fig. 1A]. Until this stage, embryos are opaque. A trochophore [a middle trochophore in Meshcheryakov (1990)] swells to be more translucent at around 60 hpc. The trochophore stage was also divided into early and late stages. At the late trochophore stage, larvae start to become slightly elongated along the antero-posterior (blastopore shell gland) axis. The shell gland invaginates deeply, and a thin transparent shell forms over the shell gland (Fig. 1D). Similarly to the gastrula, weakly stained nuclei were seen in the center of the shell gland of the trochophore (Fig. 2F). At around 10-days post-first cleavage, a larva becomes a juvenile with a developed shell, occupying the entire capsule, and hatches.

Fig. 1

Fig. 1

The alternative text for this image may have been generated using AI.

Full size image

SEM images of L. stagnalis embryos at the gastrulating blastula and trochophore stages. A Gastrulating blastula viewed from the vegetal pole from which the gastrulation begins. B, B’ Gastrulating blastula viewed from the animal pole. A solid line surround large-sized cells that may become head vesicles (B’). C Schematic image of the larval directions superimposed on the trochophore larva, viewed from the left side. D, D’ Early trochophore viewed from the shell gland side. Shell is surrounded by a broken line (D’). BP blastopore, CP cephalic plate, F foot, HV head vesicle, M mouth, P prototroch, S shell. Scale bars, 20 μm

Fig. 2

Fig. 2

The alternative text for this image may have been generated using AI.

Full size image

DAPI images. A–C Bright field images. D–F’ Fluorescence images. Images A–C correspond to D–F and D’–F’, respectively. A, D, D’ Gastrulating embryo. B, E, E’ Gastrula viewed from the shell gland. C, F, F’ Early trochophore viewed from the shell gland side. Solid lines surround large cells that would become head vesicle cells (D’, E’). Broken lines represent roughly the shell gland area, opposite to the stomodaeum (E’, F’). That the size of cells within the broken line is larger than that of the cells out of the solid line can be indirectly confirmed by the density of nuclei stained by DAPI. The center region of the shell gland is where the density of the cells is relatively lower, shown as a black hole (E, E’, F, and F’). Scale bar, 100 μm

Isolation and identification of engrailed and dpp genes of L. stagnalis

For the engrailed orthologue of L. stagnalis, a 1,100-bp fragment with an intron region was first obtained by PCR with the pair of degenerate primers engrailed-F and engrailed-R, designed to the conserved regions within the homeobox, using the genomic DNA as template. Subsequently, 3′ and 5′ RACE were performed using cDNA prepared from embryos and adult mantle tissues, respectively, and finally, the full-length cDNA of 3,129 bp was obtained with the primer set of LstenF0 and LstenR0. The full-length cDNA contained the complete open reading frame (ORF) of 799 amino acids (2,397 bp) with an in-frame stop codon at the position of 221 bp upstream of the putative translation initiation site. All five conserved domains for Engrailed (Ekker et al. 1992; Hui et al. 1992; Logan et al. 1992; Manzanares et al. 1993) were present in the deduced amino acid sequences from L. stagnalis (Fig. 3). Thus, we conclude that this is an engrailed orthologue of L. stagnalis, and it is designated as Lsten (Accession number, AB331395).

Fig. 3

Fig. 3

The alternative text for this image may have been generated using AI.

Full size image

Alignment of conserved Engrailed domains (EH1–5) of deduced amino acid sequences with those from other species. An amino acid residue identical to that of Lsten is shown as a dashed line. The insertion between EH2 and EH3 is shown, if present. A dot indicates a gap. Afr, Artemia franciscana (crustacea arthropod), Bfl, Branchiostoma floridae (cephalochordate protochordate, chordate), Bmo, Bombyx mori (lepidoptera insect, arthropod), Cap, Capitella sp (polychaete annelid), Cel Caenorhabditis elegans (nematode), Cin Ciona intestinalis (urochordate protochordate), Dme Drosophila malanogaster (diptera insect, arthropod), Lst Lymnaea stagnalis (gastropod mollusc), Mmu Mus musculus (mammal vertebrate, chordate), Pvu Patella vulgata (gastropod mollusc), Sko, Saccoglossus kowalevskii (enteropneust hemichordate), Spu Strongylocentrotus purpuratus (echinoid echinoderm). Accession numbers are: Afr-en X70939, Bfl-en BFU82487, Bmo-en M64335, Cap-en AY578983, Cel-en P34326, Cin-en Q81AC7, Dme-en M10017, Mmu-en1 AK140408, Mmu-en2 AK157766, Pvu-en AF440096, Sko-en AY313158, Spu-en XP_794753.1

Degenerate PCR on genomic DNA template amplified 240-bp fragment of a putative dpp orthologue of L. stagnalis using the pair of primers dpp-F and dpp-R, designed from the conserved regions within the TGF-β domain. Subsequent 3′ and 5′RACE extended the PCR fragment in both the 3′ and 5′ directions, and finally, the full-length cDNA of 4,012 bp was obtained using the pair of primers LstdppF0 and LstdppR0. The full-length cDNA contained the complete ORF of 589 amino acids (1,767 bp) with an in-frame stop codon at the position of 1,908 bp upstream of the putative translation initiation site. A preproprotein of bone morphogenetic proteins (BMP) subfamily consists of three regions: the signal peptide, the prodomain that is proteolytically cleaved from the mature domain immediately after the site RXXR (R = arginine, X = any amino acids), and the mature ligand domain that contains seven cysteine residues at the conserved sites and form a dimer to yield an active ligand (Lelong et al. 2000; reviewed in Hogan 1996). The amino acid sequence deduced from the obtained full-length cDNA apparently contains all three regions: a putative signal peptide with its cleavage site between 45 (A) and 46 (S) amino acid residues, a prodomain with a potential proteolytic cleavage sites (RTRR: T = threonine) at the amino acid position 453, and a ligand domain with seven conserved cysteine residues in the carboxyl-terminal (Fig. 4). Thus, it appears highly likely that the cDNA is an L. stagnalis orthologue of BMP subfamily genes. In the maximum likelihood tree constructed for BMP subfamily proteins, the deduced protein was included in the group of the dpp-BMP2/4 proteins with a bootstrap value of 99% (Fig. 5), although a relatively long branch for this protein implies that the protein experienced a higher extent of amino acid substitutions. The neighbor-joining tree showed almost the same topology as the ML tree (data not shown). Altogether, we conclude that the cDNA that we obtained is an L. stagnalis orthologue of dpp-BMP2/4 and designated it as Lstdpp (Accession number, AB331396).

Fig. 4

Fig. 4

The alternative text for this image may have been generated using AI.

Full size image

Alignment of conserved TGF-b domain of deduced amino acid sequences with those from other species. An amino acid residue identical to that of Lstdpp is shown as a dashed line. Dots indicate gaps. BMP58, univin, and 60A belong to different subclasses within TGF-β superfamily, respectively. Afuni of the ophiuroid echinoderm belongs to a univin subclass. Af Amphiura filiformis (ophiuroid echinoderm), Ate Achaearanea tepidariorum (arachnid chelicerate, arthropod), Bma Brugia malayi (nematode), Bpa Brugia pahangi (nematode), Cgi Crassostrea gigas (bivalve mollusc), Hpu Hemicentrotus pulcherrimus (echinoid echinoderm), Iob Ilyanassa obsoleta (gastopod mollusc), Lva Lytechinus variegatus (echinoid echinoderm), Pdu Platynereis dumerilii (polychaete annelid), Pfl Ptychodera flava (enteropneust hemichordate), Pfu Pinctada fucata (bivalve mollusc), Pli Paracentrotus lividus (echinoid echinoderm), Tca Tribolium castaneum (coleoptera insect, arthropod), Tgr Tripneustes gratilla (echinoid echinoderm). Other abbreviations are in the legend of Fig. 3. Accession numbers are: Afuni AY954372, Ate-dpp AB096072, Bfl-BMP24 AF068750, Bma-dpp AF012878, Bpa-dpp AF010495, Cel-dpp O02424, Cgi-dpp AJ130967, Dme-dpp M30116, Hpu-BMP AB088684, Iob-dpp AF499914, Lva-BMP24 AF119712, Mmu-BMP2A P21274, Mmu-BMP4 P21275, Pdu-dpp AM114782, Pfl-BMP24 AB028219, Pfu-dpp AB176952, Pli-BMP24 DQ536194, Pli-univin DQ536195, Pvu-dpp AF440097, Spu-BMP24 AF119713, Spu-BMP58 Z48313, Spu-univin U10533, Tca-dpp U63132, Tgr-BMP24 AF133305, 60A M77012

Fig. 5

Fig. 5

The alternative text for this image may have been generated using AI.

Full size image

Phylogenetic tree of the genes cloned in the present study and those from databases for _dpp_-BMP2/4 genes, representing a maximum likelihood tree (see “Material and methods”). Bootstrap values lower than 50% are not included. Drosophila 60A is used as an outgroup. Abbreviations and accession numbers are stated in the previous figure legends

We identified only one orthologue for either engrailed or dpp gene in the present study after screening a number of clones. But we cannot exclude the possibility that other paralogous genes exist in L. stagnalis.

Expression of Lsten and Lstdpp in early development of L. stagnalis

The expression patterns of Lsten and Lstdpp were examined through the embryonic developmental stages, gastrulating blastula, early and late gastrula, and late trochophore stages. At the gastrulating stage, no expression of Lsten was observed (Fig. 6A). At the early gastrula stage, Lsten is expressed in the dorsal ectoderm, showing a semicircular band of stained cells (Fig. 6B). This cell band corresponds to a part of the peripheral area of the presumptive shell gland (Morrill 1982). At four spots along the posterior end, Lsten is also expressed (Fig. 6B inset). Each spot corresponds to a cell or a group of a few cells constituting a part of the posterior ectoderm. At the late gastrula stage, the expression of Lsten was detected in the peripheral ectoderm of the shell gland (Fig. 6C,D). The four spots at the posterior end continue to express Lsten just behind the presumptive foot (Fig. 6D). At the late trochophore stage, Lsten continues to be expressed in four spots behind the foot (Fig. 6E inset) and in the peripheral ectoderm of the shell gland at the same stage, forming a double circle (Fig. 6E). In addition, bilateral expression pattern of Lsten was observed as two small spots at the base of the foot (Fig. 6F). Expression patterns of Lsten in the peripheral areas of the presumptive shell gland or invaginating shell gland suggest that Lsten play an important role in the shell gland formation or in the shell formation that follows. We also observed Lsten expression behind and at the base of the foot, suggesting that Lsten has other roles than shell formation at these developmental stages. We will discuss the expression patterns involved in shell formation. Throughout all the examined stages, the sense probe of Lsten did not show any signals (Fig. 6H–K).

Fig. 6

Fig. 6

The alternative text for this image may have been generated using AI.

Full size image

Expression pattern of Lsten during L. stagnalis development. A–F WMISH with antisense probe for Lsten. H–K WMISH with sense probe. Sense probe does not show positive signals throughout the development examined. Anterior side is to the top of the figures. A, H Gastrulating embryos viewed from the side of the blastopore. B Early gastrula viewed from the shell gland, opposite to the blastoporal stomodaeum. C, D, I Late gastrulae. C, I Shell gland views. D Right side view. E, F, G, J, K Late trochophore larvae. E, J Shell gland views. F, G, K Left side views. G The same specimens as in F seen through polarized light, showing a crystallized shell on the shell gland opening (boxes in F and G). Expression of Lsten is not observed in the gastrulating blastula (A). The early gastrula (B) shows Lsten expression in a semicircular, peripheral band (black arrowheads) of the presumptive shell gland, and four spots along the posterior end (inset in B, white arrowheads). The late gastrula (C, D) expresses Lsten circularly in the peripheral zone of shell gland (D, black arrowheads). The late trochophore larva (E, F) expresses Lsten in the peripheral zone of shell gland as a double circle (black arrowheads). Four spots in the posterior end continue to be expressed (D, E, F, white arrowheads). Additional bilateral expression of Lsten is observed in the base of the foot (F, double arrowheads). An anterior, L left, P posterior, R right, Sh shell gland opening, St stomodaeum. Scale bar, 100 μm

At the gastrulating blastula stage, expression of Lstdpp is not observed (Fig. 7A). At the early gastrula stage, Lstdpp is expressed in the right-hand side ectoderm within the presumptive shell gland (Fig. 7B). Weak signals of Lstdpp were also observed in the tip of the invaginating stomodaeum (data not shown). Bilateral expression patterns were detected in the dorso-lateral ectoderm and in the antero-lateral ectoderm (Fig. 7B). At the late gastrula stage, the right-hand side of the invaginating shell gland ectoderm expressed Lstdpp strongly (Fig. 7C). Gene expression was also seen in the invaginating stomodaeum (Fig. 7D). Relatively weak signals of Lstdpp remained to be observed bilaterally in the dorso-lateral and antero-lateral ectoderm (Fig. 7C,D). At the late trochophore stage, Lstdpp continued to be strongly expressed in the right-hand side ectoderm of the invaginated shell gland and in the posterior part of the stomodaeum including the radular sac depression (Fig. 7E,F). There were also continuous bilateral patterns of Lstdpp expression in the dorso-lateral and antero-lateral ectoderm (Fig. 7E,F). Expression of Lstdpp continued in a right-sided manner in the presumptive and invaginated shell gland ectoderm, suggesting that Lstdpp is involved not only in the shell formation but also in the specification of the left–right asymmetry within the shell gland. We also observed Lstdpp expression in the dorso-lateral and antero-lateral ectoderm, suggesting that Lstdpp has other roles than shell formation at these developmental stages. We will discuss the expression patterns involved in the shell formation. The sense probe of Lstdpp did not show any signals (Fig. 7G–J).

Fig. 7

Fig. 7

The alternative text for this image may have been generated using AI.

Full size image

Expression pattern of Lstdpp during L. stagnalis development. A–F WMISH with antisense probe for Lstdpp. G–J WMISH with sense probe. Sense probe does not show positive signals throughout the development examined. Anterior side is to the top of the figures. A, G Gastrulating embryos viewed from the side of the blastopore. B Early gastrula viewed from the shell gland, opposite to the blastoporal stomodaeum. C, D, H Late gastrulae. C, H Shell gland views. D Right side view. E, F, I, J Late trochophore larvae. E, I Shell gland views. F, J Left side views. Expression of Lstdpp is not observed in the gastrulating blastula (A). Lstdpp is expressed in the right-hand side area within the presumptive shell gland at the early gastrula stage (B, black arrow). Right-hand side ectoderm of invaginating shell gland expresses Lstdpp strongly at the late gastrula stage (C, D, black arrows). Invaginating stomodaeum also expresses Lstdpp (D, white arrow). At the late trochophore stage (E, F), the right-hand side ectoderm of invaginated shell gland (E, F, black arrows) and the posterior part of the stomodaeum with radular sac (F, white arrow) continue to express Lstdpp. At this stage, the view with the polarized light (bottom-right inset in F) shows the beginning of the crystallization around the shell gland opening (box in F). Bilateral expression patterns of Lstdpp are observed in the dorso-lateral (B, C, E, black arrowheads) and antero-lateral ectoderm (B, D, F, asterisks). Bottom inset in B represents the magnified image of the bilateral expression in antero-lateral ectoderm (asterisks), viewed from the anterior side. Bottom inset in D represents the magnified image of the right side expression of bilateral expression in the antero-lateral ectoderm (asterisk), viewed from the right side. Bottom-left inset in F represents the magnified image of the left side of the antero-lateral expression of Lstdpp, focusing on the left side ectoderm. Abbreviations are as in the previous figure. Scale bar, 100 μm. Scale bars in insets, 25 μm

To further elucidate the relationship between the expression patterns of Lsten and Lstdpp in the shell gland region, we performed the whole mount in situ hybridization using both Lsten and Lstdpp probes simultaneously, with a single coloring procedure. At the gastrulating blastula stage, expression patterns of the two genes are neither overlapped nor adjacent to each other (Fig. 8A). Similarly, both at the gastrula and trochophore stages, no overlapping or adjoining expression was observed between Lsten and Lstdpp (Fig. 8B–I). This overall pattern is different from that in the other gastropod, Patella vulgata, in which engrailed and dpp are expressed in an adjoining manner in the peripheral zone of the shell gland (Nederbragt et al. 2002). Moreover, the spatial relationships of expression patterns of engrailed and dpp are different between L. stagnalis and P. vulgata at one more important point. While the engrailed gene is expressed in the peripheral most area of the shell gland ectoderm in both L. stagnalis and P. vulgata, and the dpp gene of P. vulgata is expressed outside the _engrailed_-expressing circle while the dpp gene of L. stagnalis is expressed inside this circle. Concerning the expression patterns of Lsten and Lstdpp in the shell gland area, overview images are shown in Fig. 9.

Fig. 8

Fig. 8

The alternative text for this image may have been generated using AI.

Full size image

WMISH of double staining with a single coloration using Lsten and Lstdpp probes. Anterior side is to the top of the figures, except for E and H, in which the stomodaeum is to the left of the figures. A Early gastrula dorsal view. B, C, D, E Late gastrulae. B, C Shell gland views. D Left side view. E Posterior view. F, G, H, I Late trochophore larvae. F, G Shell gland views. H Anterior view. I Left side view. At the early gastrula stage (A), Lsten and Lstdpp are detected with a non-adjoining expression pattern. B and C show the same specimens, showing different focal plans. The former (B) shows Lsten expression in the peripheral zone of shell gland, while the latter (C) shows Lstdpp expression in the right-hand side ectoderm of the invaginating shell gland. At the late gastrula and late trochophore stages (D–I), it is indicated that, around the shell gland, Lsten (black arrowheads) and Lstdpp (black arrows) are expressed in separate locations. White arrows show Lstdpp expression in invaginating stomodaeum (D, E, H, I). Abbreviations are as in Fig. 6. Scale bar, 100 μm

Fig. 9

Fig. 9

The alternative text for this image may have been generated using AI.

Full size image

Overview of expression patterns of Lsten and Lstdpp around the shell gland at the early gastrula and the late trochophore stages. A A superimposed image of the early gastrula viewed from the dorsal side. Black line and white lines delineate approximately the expression area of Lsten and Lstdpp, respectively. B, C, D Superimposed images of the late trochophore stage. B An anterior view, stomodaeum and shell gland opening are to the left and right, respectively, of the figure. Broken line outlines roughly the invaginating shell gland. Black solid lines represent roughly the inner and outer boundaries of Lsten expression. The right-hand side ectoderm of the invaginating shell gland express Lstdpp (white line). C Shell gland view representing the inner and outer boundaries (black circles) of Lsten expression. D Shell gland view showing the Lstdpp expression area (a white line). Broken lines outline roughly the invaginating shell gland. In the shell gland ectoderm, Lstdpp is expressed only in the right-hand side. The same specimens is seen from the same direction with different focuses (C, D). Abbreviations are as in Fig. 6. Scale bar, 100 μm

Discussion

Terminology and homology of the shell-forming region

In gastropods, bivalves, and scaphopods, invagination of dorsal epithelium occurs to form the ‘shell gland’ that later evaginates to begin the secretion of shell matrix [see Kniprath (1981) for the details]. There are some different designations for the progenitor and progeny of the dorsal invagination region during the development: the shell plate (Moshel et al. 1998), shell gland (Dictus and Damen 1997; Moshel et al. 1998), and shell field (Dictus and Damen 1997; Moshel et al. 1998; Wanninger and Haszprunar 2001; Kniprath 1981). The term “shell gland” has been used to refer not only to the invaginated depression itself but also to the region developed from the depression. The latter usage resulted from an erroneous interpretation of its functional morphology (Kniprath 1981). Hereafter, following Moshel et al. (1998), we distinguish three different terms for the cell region concerned with shell formation, depending on the different developmental stages: first, the shell plate consisting of the thickened epithelium in the dorsal region before invagination; second, the shell gland forming invagination before developing shell formation; third, the shell field, resulting from the shell gland evagination and flattening, secreting the shell matrix to develop the shell. It is most likely that the molluscan shells are homologous, and thus, it could be that all the diverse shell fields that produce shells are also homologous (Kniprath 1981).

Involvement of engrailed genes in the shell formation of molluscs

Molluscan animals with developed shells form diverse shells that are the characteristic of each class (e.g., Brusca and Brusca 2002). However, the molecular mechanisms to yield the diverse shell morphologies have not been elucidated. One commonality that has been found for shell-forming cells of the several molluscan species representing five different classes is the expression of the engrailed gene as described below.

In polyplacophoran and conchiferan molluscs other than monoplacophorans, engrailed gene is expressed in or around the shell-forming cells, suggesting its involvement directly or indirectly in the shell formation. In the chiton (Lepidochitona caverna, Polyplacophora), expression of engrailed is detected in the border regions between the shell plates and in the girdle region where spicules form (Jacobs et al. 2000). In the clam (Transenella tantilla, Bivalve), Engrailed protein detected using 4D9 antibody raised against a portion of the fly Engrailed protein (Patel et al. 1989) is localized around the margin of the shell and along the hinge (Jacobs et al. 2000). In the tusk shell (Antalis entalis, Scaphopoda), Engrailed is localized in the marginal cells of the embryonic shell field (Wanninger and Haszprunar 2001). In the snail (Gastropoda), Ilyanassa obsoleta, the expression of the Engrailed is located in the shell gland progenitor and progeny, and the expression of the engrailed mRNA is also seen in the shell plate (Moshel et al. 1998). In the limpet (Gastropoda), Patella vulgata, engrailed expression is localized in the shape of a semicircular band of the shell-forming cells along with the shell margin (Nederbragt et al. 2002). In the cuttlefish (Sepia officinalis, Cephalopoda), Engrailed is located in both the mantle and the whole shell sac area at the earlier stages, restricted to both the shell sac and mantle edges at the subsequent stage (Baratte et al. 2007). In another cuttlefish (Idiosepius paradoxus, Cephalopoda) and the nautilus (Nautilus pompilius, Cephalopoda), the similar expression pattern of Engrailed proteins are observed in the marginal cells of the shell field (Shigeno et al. 2008).

In the present study, the engrailed expression patterns in the gastropod snail L. stagnalis are found to be similar to those of other molluscs. It is likely that engrailed gene has been incorporated in the genetical pathway for shell formation among molluscan animals as a synapomorphy, but it has not been determined whether it functions directly in the skeletogenesis (Jacobs et al. 2000) or in the regional specification of the specific tissue (Nederbragt et al. 2002). In L. stagnalis, engrailed begins to be expressed remarkably earlier than the formation of the shell gland (Fig. 6). Among all the shell gland-forming molluscs hitherto examined, including the scaphopod, the bivalve, and the gastropods, the timing of engrailed expression is generally earlier than that of the shell gland invagination. Thus, the engrailed gene could be involved in the specification of the shell gland progenitor or the shell plate. At later stages, the temporally continuous expression of engrailed genes around the shell fields implies their roles in the skeletogenesis or the maintenance of secretion of the shell matrix proteins, rather than the regional specification of the shell-forming ectoderm.

Possible diversification of the mechanisms involved in the shell formation in molluscs

Expression of the signaling molecule, dpp-BMP2/4, was detected in the shape of a circular band in the cells adjacent to and outside of the shell-forming cells that expressed the engrailed in the limpet P. vulgata (Nederbragt et al. 2002). In contrast, our observation in L. stagnalis indicates that dpp is localized in the shell gland ectoderm with the preferential expression in the right-hand side, in the inside of the shell gland margin that expresses engrailed gene. The difference in dpp expression patterns between L. stagnalis and P. vulgata implies that the roles of the dpp genes are different at least among gastropods. Different expression patterns of dpp genes between L. stagnalis and P. vulgata could reflect the modifications of the framework in the gene regulatory network.

The expression pattern of Lstdpp is quite curious. During almost all of the developmental stages examined, Lstdpp expression in the shell gland ectoderm exhibited left–right asymmetry or chirality, whereas other expression sites showed bilateral arrangements (Figs. 7 and 9). This observation suggests that the dpp gene is involved in the signaling pathway importantly for the chirality in the shell gland ectoderm and possibly for the coiling direction of the adult shell, i.e., dextrality or sinistrality. The mechanisms to form the chirality in the gastropods have been studied in many fields, e.g., developmental biology (Freeman and Lundelius 1982; Shibazaki et al. 2004), genetics (Hosoiri et al. 2003), population genetics (Ueshima and Asami 2003), and molecular biology (Harada et al. 2004), but the details remain to be uncovered. According to Meshcheryakov (1990) and our observation, the shell gland opening of L. stagnalis migrates from the sagittal plane in the dorsal side at the earlier stages (until about late gastrula stage) to the left at the later stages (after early trochophore stage; compare Fig. 8B,F). The invaginated shell gland in the larval body also leans toward the left side, extending dominantly its right-hand side ectoderm where dpp is strongly expressed (e.g., Figs. 8H and 9B). It is likely that the dpp gene, one member of BMP subclasses of TGF-β superfamily, is involved in the morphogenesis of the shell gland ectoderm. The shell gland, however, evaginates at the later stages (Kniprath 1981). It is, therefore, required to examine the dpp expression patterns in the juveniles or adults that developed the coiled shells. Such examination is also required in the sinistral strains at the several developmental stages. Those experiments could provide insights into the involvement of dpp gene in the shell formation and the chirality of the shell morphology. BMP4 signaling pathway is shown to be involved in the left–right asymmetry in the chick (Monsoro-Burq and Le Douarin 2000), posing interesting questions for further comparative studies.

In order to fully understand the roles of dpp and engrailed genes in shell formation, it should be required to carry out functional analyses of those genes. The functional analyses for biomineralization have been examined in deuterostomes, e.g., mouse Cbfa1 (Komori et al. 1997) or echinoid P16 (Cheers and Ettensohn 2005), but not in lophotrochozoans. For this purpose, L. stagnalis is arguably well-suited, and our present study provides a basis for such future studies.

References

Download references