Development of a dopaminergic system in sea urchin embryos and larvae

The swimming of marine larvae depends on ciliary propulsion and is critically important for feeding, escape responses (Strathmann, 2007; Stephens, 2008) and photosensitive vertical migration (Ooka et al., 2010). In sea urchin embryos, motile cilia are evident from the blastula stage, and rotatory movement by embryos can be observed in the fertilization envelope. Serotonin plays a role in the regulation of the beating of larval cilia (Yaguchi and Katow, 2003; Katow et al., 2009). However, its role in the regulation of the cilia of blastulae is unknown. The swimming activity of hatched blastulae is insensitive to para-chlorophenylalanine (pCPA), a potent inhibitor of tryptophan 5-hydroxylase, and the serotonergic nervous system (5HT-NS) is not yet formed at this stage (Yaguchi and Katow, 2003). Nevertheless, the swimming behavior of blastulae is organized to a considerable degree. For instance, sea urchin blastulae can swim by orienting their long apical tuft cilia in the target direction (Soliman, 1983a; Stephens, 2008). This suggests that swimming in sea urchin blastulae might be regulated by a distinctive mechanism that is not related to the 5HT-NS.

The beating of cilia is known to be regulated by catecholamines, in particular dopamine (DA) in invertebrates (Soliman, 1983b; Voronezhskaya et al., 1999; Martin et al., 2008) and vertebrates (Maruyama et al., 1983; Tomé et al., 2007). To date, expression of catecholamines has been demonstrated in basic deuterostomes, including asteroidae (Huet and Fraquinet, 1981), brittlestars (De Bremaeker et al., 2000) and holothurian species (Inoue et al., 2002), but not in hemichordates. For instance, embryos of Lytechinus pictus express DA from the 18 h post-fertilization (hpf) swimming blastula stage, as detected by high-performance liquid chromatography with electrochemical detection. This is followed by a period when DA is undetectable from the 28 hpf mesenchyme blastula stage to the 42 hpf late gastrula stage, and trace amounts can be found in 49 hpf late prism larvae (Anitole-Misleh and Brown, 2004). Furthermore, in Strongylocentrotus droebachiensis, DA is only present at the larval stages, as demonstrated by immunohistochemistry (Bisgrove and Burke, 1987). It is possible that catecholamines are involved in morphogenesis during the early development of sea urchins, Psammechinus miliaris (Gustafson and Toneby, 1970; Toneby, 1980; Anitole-Misleh and Brown, 2004). However, the time lag between the onset of embryonic swimming and the immunohistological appearance of dopaminergic cells in larvae suggests the presence of preneural dopaminergic regulation of the beating of the cilia of blastulae.

DA has been shown to signal to the cytoplasm of embryonic cells through DA receptors, which are a group of metabotropic G-protein-coupled receptors (GPCRs) (Girault and Greengard, 2004). Microarray analysis of the expression of DA receptor D1 (Sp-DRD1) and DA receptor D2 (Sp-DRD2) in Strongylocentrotus purpuratus has revealed that the mRNAs are expressed from the two-cell stage to the early prism stage. This suggests that DA functions during early development in this species (Sea Urchin Genome Project, http://www.spbase.org/SpBase/search/anno_search_result.php).

Despite previous biochemical observations that suggest a role for DA in sea urchin embryogenesis (Gustafson and Toneby, 1970; Toneby, 1980; Capasso et al., 1987; Carginale et al., 1995; Anitole-Misleh and Brown, 2004) and embryonic swimming activity (Soliman, 1983a), few morphological studies exist to support this. Therefore, in this study we investigated the localization of DA and its receptor Hp-DRD1, a D1 homolog found in the sea urchin species Hemicentrotus pulcherrimus (Suyemitsu, 2007) (Hp-DRD1; accession number AB510005, see supplementary material Fig. S1), from the pre-hatched blastula stage to metamorphosis using whole-mount immunohistochemistry. To determine whether DA might play a role in swimming behavior, the expression and sub-cellular localization of DA and Hp-DRD1 were examined by immunostaining for DA and Hp-DRD1, as well as for markers of the serotonergic system. In addition, pharmacological and gene knockdown analysis of the loss of function of DA and Hp-DRD1 was performed using S-(–)-carbidopa (carbidopa), which is an inhibitor of aromatic-l-amino acid decarboxylase, and morpholino antisense oligonucleotides (MASO) directed against Hp-DRD1.

Sea urchins, Hemicentrotus pulcherrimus A. Agassiz, were collected in the vicinity of the Research Center for Marine Biology, Tohoku University, Japan. Gametes were obtained by intracoelomic injection of 0.5 mol l⁻¹ KCl. Eggs were inseminated and incubated in filtered seawater (FSW) on a gyratory shaker in an incubator at 18°C until the appropriate developmental stages were reached.

A DNA sequence that encoded the hydrophilic C-terminal region of Hp-DRD1, namely R³¹⁴SFRNAFKTILFCYKCRGLSMRQADSSDDHTAMSNTSRQERLRGRLGNGYNTTPRPSHSDHRSSIENSAEQNQKMVKDTSLVIANKNANANVQDSKKDSSSPVAVATITHSDA⁴²⁶, was inserted into the vector pGEX-KG (Amersham Biosciences, GE Healthcare UK Ltd, Little Chalfont, UK) and the construct transformed into Escherichia coli BL-21 cells. The resulting glutathione S-transferase (GST)-fusion protein was purified from the supernatant obtained after sonication using a glutathione agarose column (Sigma-Aldrich Japan, Tokyo, Japan), and eluted with 10 mmol l⁻¹ glutathione. The protein was concentrated by dialysis and used to inoculate rabbits with Freund's complete adjuvant (Difco Laboratories, Detroit, MI, USA). The immunospecificity of the antiserum was examined using whole-mount immunohistochemistry with or without 1.3 mg ml⁻¹ GST-fusion protein in the primary antiserum.

Embryos and larvae were fixed with 4% paraformaldehyde dissolved in FSW at the 5 hpf morula, 8 hpf unhatched non-rotatory blastula, 9.5 hpf unhatched rotatory blastula, 15 hpf swimming blastula, 20 hpf early gastrula, 24–29 hpf prism, 48 hpf two-arm pluteus, 67–68 hpf two-arm pluteus, and 34 days post-fertilization (dpf) eight-arm pluteus stages. To remove the fertilization envelope to ensure access of the antibody to unhatched blastulae, eggs were pretreated with 1 mmol l⁻¹ amino triazole for 10 min before insemination (Showman and Froeder, 1979). After insemination, the softened fertilization envelopes were removed by passing the eggs through a 62 μm nylon mesh (Kyoshin Ricoh Inc., Tokyo, Japan). The fertilized eggs were incubated in a 6 cm diameter culture dish in an 18°C incubator without shaking until they reached the 8 hpf blastula or 9.5 hpf unhatched rotatory blastula stages. After fixation they were dehydrated in a series of increasing ethanol concentrations, and stored in 70% ethanol at 4°C until use.

Prior to antibody treatment, embryos were hydrated in phosphate-buffered saline with 0.1% Tween-20 (PBST), and then incubated for 24–48 h at 4°C with rabbit antibodies against DA (catalog number DA-1140, BIOMOL International LP, Plymouth Meeting, PA, USA or catalog number DOP11-S, Alpha Diagnostic International, San Antonio, TX, USA; diluted 1:150, the latter antibody was used only as an alternative to the DA-1140 antibody as stated in the text), rabbit antibodies against serotonin (Sigma-Aldrich Japan; 1:500), monoclonal antibody against γ-tubulin (EXBIO Praha, a.s. Vestec, Czech Republic; 1:1000), monoclonal antibody against acetylated α-tubulin (Sigma-Aldrich Japan; 1:500), rabbit antibodies against 5′-bromo-2′-deoxyuridine (BrdU; Sigma-Aldrich Japan; 1:250), monoclonal antibody Epith-2 against sea urchin epithelium plasma membrane (Kanoh et al., 2001), monoclonal antibody 1E11 against synaptotagmin, rabbit anti-Hp-DRD1 antiserum (1:500), or mouse antibodies against serotonin receptor (5HThpr) (Katow et al., 2004) (1:500), either diluted in PBST or undiluted in the case of the monoclonal antibodies. Anti-BrdU antibodies were applied after 1 h of treatment with 2 mol l⁻¹ HCl to reveal incorporated BrdU. Anti-DRD1 antibodies were applied after treatment with 0.2% Triton X-100 for 2 min. After staining with the primary antibodies, the samples were washed with PBST three times (10 min each), and treated with Alexa Fluor 488- or 568-tagged anti-rabbit IgG or anti-mouse IgG antibodies (Invitrogen, Paisley, UK) for 2 h at ambient temperature. Samples were then washed with PBST three times (10 min each). The immunospecificity of the anti-DA antibody was tested using a mixture of the antibody (diluted 1:250 in PBST) and 200 μg ml⁻¹ DA (Sigma-Aldrich), l-DOPA (l-3,4-dihydroxyphenylalanine; Sigma-Aldrich) or noradrenaline (norepinephrine) (Sigma-Aldrich). l-DOPA and noradrenaline exhibited no immunohistochemical reaction with the antibody. For double-immunohistochemistry, two of the above primary antibodies were applied consecutively in the order indicated in the text. To stain the DNA, 1 μmol l⁻¹ propidium iodide (PI; SERVA Electrophoresis GmbH, Heidelberg, Germany) was applied during a PBST wash after incubation with the secondary antibody. The samples were examined under a Micro-Radiance confocal laser scanning microscope (Bio-Rad Microscience, Hemel Hempstead, UK) at the appropriate optical sections indicated in the text. Images were analyzed with NIH ImageJ software.

The aromatic-l-amino acid decarboxylase (AADC) inhibitor (2S)-3-(3,4-dihydroxyphenyl)-2-hydrazino-2-methylpropanoic acid [S-(–)-carbidopa, carbidopa; Sigma-Aldrich Japan] was applied either from soon after insemination until the 20 hpf early gastrula stage (before the appearance of the 5HT-NS) or from the 24 hpf prism larva stage until the 48 hpf pluteus stage (after the appearance of the 5HT-NS). The swimming activity of larvae that had been incubated with 4 ml of 0, 10, 15, 20 or 25 μmol l⁻¹ carbidopa in FSW from the 24 hpf prism larva stage was assayed at 18°C for 2 h in 5 ml plastic bioassay tubes of 1.1 cm diameter, 7.5 cm height, and were mounted perpendicularly on a tube stand. The time period was set so that larvae could re-establish a stable vertical swimming pattern from initial tabulations due to the transfer of the suspension medium to the bioassay tubes but not progress to alter the stage of development. Firstly, 400 μl of larval suspension at the bottom of the tube was collected gently with a Pasteur pipette (bottom fraction). The remaining 3.6 ml of the larval suspension was then collected (vertical swimming fraction). The number of larvae in the bottom and the vertical swimming fractions was counted under a dissection microscope. Each experiment was repeated three to four times using three to four different females with a total of 3512 embryos as controls, and 868, 821, 720 and 3999 embryos for the 10, 15, 20 and 25 μmol l⁻¹ carbidopa treatments, respectively. The embryos that were treated with or without 25 μmol l⁻¹ carbidopa soon after insemination were incubated in 4 ml of FSW in 5 ml plastic tubes and their swimming activity was examined as described above. Each experiment was repeated four times using four different females, and 2740 embryos were used as controls and 2602 embryos for the 25 μmol l⁻¹ carbidopa treatment. The above results were analyzed by a paired two-tailed t-test using GraphPad Software QuickCalcs Online Calculators for Scientists (http://www.graphpad.com/quickcalcs/ttest1.cfm). Significance was set at P<0.05. DA rescue was carried out in the presence of 25 or 50 μmol l⁻¹ of DA (Sigma-Aldrich Japan) in 25 μmol l⁻¹ carbidopa. Oxidation of DA in sea water was prevented by adding anti-oxidant, 50 μmol l⁻¹ of l(+)-ascorbic acid (Braubach et al., 2006). Each experiment was repeated three times using three different females, and a total of 1106 embryos were used for controls, 1176 embryos for the 25 μmol l⁻¹ carbidopa treatment, 2912 embryos for treatment with a mixture of 25 μmol l⁻¹ carbidopa and 25 μmol l⁻¹ DA, and 4704 embryos for treatment with a mixture of 25 μmol l⁻¹ carbidopa and 50 μmol l⁻¹ DA. The above results were analyzed by a paired two-tailed t-test as stated above. An aliquot of embryos treated with a mixture of 25 μmol l⁻¹ carbidopa and 25 μmol l⁻¹ DA was examined with whole-mount immunohistochemistry for DA as stated above.

Expression of Hp-DRD1 mRNA was examined using 1 μg of cDNA template prepared from total RNA extracted from unfertilized eggs, fertilized eggs, 6 hpf morulae, 14 hpf swimming and 16 hpf mesenchyme blastulae, 20 hpf early gastrulae, 26 hpf prism larvae, and 48 hpf and 72 hpf two-arm plutei by PCR. The following primers were used: forward AACAAACGGAACAGACGAAGG; reverse GGACATCGATTGCGTTTCCAT, with an annealing temperature of 70°C (for the samples from unfertilized eggs, an annealing temperature of 60°C was used) for 30 cycles. Amplimers were purified with the Wizard SV Gel and PCR Clean-up System (Promega, Madison, WI, USA) and ligated into the vector pBluescript II (Stratagene, La Jolla, CA, USA). Clones were sequenced using a BigDye Terminator Cycle Sequencing Kit and an ABI PRISM 310 DNA Genetic Analyzer (PE Applied Biosystems, Tokyo, Japan), and found to encode a 1.3 kbp fragment of Hp-DRD1 DNA delimited by the forward and reverse primers (data not shown). Potential contamination with genomic DNA was analyzed by performing PCR without the addition of reverse transcriptase at each developmental stage. For the positive control, the ubiquitin gene from H. pulcherrimus (Hp-ubi) was amplified from 1 μg of H. pulcherrimus cDNA using the following primers: forward DGAGCTGCHATGTATTTGCCAGAYG; reverse TTTGATGGAATAACAAATAACYGATTGCTT.

Unfertilized eggs were attached to plastic dishes coated with 1% protamine sulfate, inseminated, and microinjected with a 1% egg volume of 500 μmol l⁻¹ or 750 μmol l⁻¹Hp-DRD1-MASO (CCGTTTGTTAAATTATTCGTCATGG) or 750 μmol l⁻¹ standard control MASO (Gene Tools, LLC, Philomath, OR, USA). A total of 745 eggs from three different females were used in experiments performed in triplicate. The injected eggs were incubated overnight at 18°C until they hatched, and then transferred to fresh FSW in plastic dishes until the 20 hpf early gastrula stage. At 20 hpf they were transferred to fresh FSW in 1.5 ml polystyrene disposable micro-cuvettes for spectroscopy (10 mm wide, 4 mm deep, 45 mm high; Kartell, Milan, Italy), and incubated for 2 h at 18°C. Firstly, 200 μl of embryo suspension were collected gently from the bottom of the cuvette using a Pasteur pipette, and then the remaining suspension in the cuvettes was collected. These embryos were counted under a dissection microscope. The total number of embryos examined for the bioassay was 186 controls, 201 for the 500 μmol l⁻¹ MASO treatment, and 177 for the 750 μmol l⁻¹ MASO treatment. The results obtained were analyzed by a paired two-tailed t-test using GraphPad Software QuickCalcs Online Calculators for Scientists. Significance was set at P<0.05.

The remaining 181 embryos from the total of 745 were fixed with 4% paraformaldehyde and the expression of Hp-DRD1 and DA was examined with appropriate antibodies under a confocal laser scanning microscope as stated above. To quantify the influence of Hp-DRD1-MASO injection, the number of dopaminergic granules (DAGs) was counted in four control MASO-injected embryos and seven Hp-DRD1 MASO-injected ones. Another 11 control MASO-injected embryos and nine Hp-DRD1-MASO-injected ones were used for counting the number of DRD1 granules. The laser beam-excited positive signals of DA and DRD1 were clear only on the side of the embryo that was closer to the laser discharge device. Thus, the number of DAGs or DRD1 granules that was counted on this side represented one-half of an embryo (half number of positive granules, HNPGs). These calculations were statistically analyzed by a paired two-tailed t-test using GraphPad Software QuickCalcs Online Calculators for Scientists. Significance was set at P<0.05.

DA was not detected before the exhibition of rotatory movement by embryos in the fertilization envelope at the 8 hpf blastula stage (Fig. 1A). At the 9.5 hpf rotatory blastula stage, immediately before they hatched, DA was detected in association with 1 μm diameter DAGs present throughout the apical surface of the ectoderm (Fig. 1B), with DAGs detected on embryonic and larval surfaces until metamorphosis (Fig. 1C–K). DAGs were not detected when the anti-DA antibody was pre-absorbed with DA (DA-absorbed antibody, Fig. 1F), and the anti-DA antibody did not cross-react with l-DOPA or noradrenaline (data not shown). At the 48 hpf pluteus stage, the number of DAGs increased significantly, and formed a broad band around the circumoral ciliary band region (Fig. 1G). The DAGs at the circumoral ciliary band region converged to become a distinctive sharp-edged circumoral stripe (DAG stripe) by the 67 hpf pluteus stage (Fig. 1H). By the 34 dpf eight-arm pluteus stage, the DAGs converged to a narrower stripe at each of the eight arms and at the anterior and posterior epaulets (Fig. 1I). These stripes did not react with the DA-absorbed antibody (Fig. 1J). However, the signal in the middle of the larval arms and posterior area of the larval body was non-specific, because pre-absorbed antibody bound to these areas (Fig. 1J, arrowheads). The DAG stripe at the epaulets remained throughout metamorphosis (Fig. 1K), and the pattern of DAG expression resembled the pattern of distribution of cilia on the surface of embryos and larvae. The pad of the primary podia of the adult rudiment expressed DA (Fig. 1K, small arrows). During metamorphosis, absorption of the dorsal (ab-anal) ectoderm by the adult rudiment occurred earlier (Fig. 1K, ab-anal) than that of the ventral (anal) ectoderm (Fig. 1K, anal). The DAGs were also seen in 24 hpf prism larvae (Fig. 1L) and 68 hpf plutei (Fig. 1M) that were stained with the alternative rabbit anti-DA antibody (DOP11-S). The antibody also recognized DAGs in the ciliary band of 68 hpf plutei (Fig. 1N), indicating that the DAGs were recognized by two antibodies. This excluded the possibility of immunoreaction by antibodies from a particular manufacturer.

To examine how closely the pattern of distribution of the DAG stripes resembled that of the cilia, the cytological localization of the DAGs was examined by immunostaining with anti-DA antibody and an Epith-2 monoclonal antibody, which is specific to the lateral plasma membrane of epithelial cells (Kanoh et al., 2001). This staining revealed that each DAG was positioned apically (Fig. 2A, inset) in the middle of a single ectodermal cell (Fig. 2A).

The sub-cellular localization of DAGs was analyzed further by immunostaining with anti-DA antibody and anti-γ-tubulin antibody as a marker of the basal body of the cilium (Muresan et al., 1993; Hagiwara et al., 2000). These antibodies were co-localized at the apical side of the ectoderm. Anti-γ-tubulin antibody was associated with granules of approximately 0.7 μm in diameter (Fig. 2B). Staining of serial 0.5 μm optical sections revealed granules near the apical surface of the ectoderm that stained positive for both antibodies (Fig. 2C0), which were replaced (Fig. 2C0.5–1.5) by staining for γ-tubulin alone 2 μm from the surface (Fig. 2C2). These findings suggest that DAGs are located on the apical side of γ-tubulin, and therefore are associated with the basal bodies of ectodermal cilia.

The above spatial relationship was tested in 34 dpf plutei by double-immunostaining for DA (Fig. 3A,D,G; DA) and acetylated-α-tubulin (Fig. 3B,E,H; AT). Unlike γ-tubulin, acetylated α-tubulin is a component of various microtubules, which include those located in cilia (Steffen et al., 1994). Accordingly, the anti-acetylated-α-tubulin antibody stained microtubules in the larval ectoderm (Fig. 3B, arrows), cilia at the ciliary bands above the mouth (Fig. 3B rectangle e, Fig. 3E), and epaulets (Fig. 3B rectangle h, Fig. 3H). The merged image indicated clearly that DAGs were localized at the base of each cilium (Fig. 3C,F,I; DA/AT), which suggested that there may be a structural basis for the involvement of DA in ciliary beating, and thus in the swimming activity of embryos and larvae.

The swimming activity of larvae is regulated by the 5HT-NS (Yaguchi and Katow, 2003; Katow et al., 2007). Given that the larval 5HT-NS expresses synaptotagmin (anti-synaptotagmin antibody-positive nervous system, Syn-NS) (Katow et al., 2009), the distribution of the 5HT-NS in larvae was examined using an anti-synaptotagmin antibody. The spatial distribution patterns of the Syn-NS and DAGs were examined in 34 dpf plutei by immunohistochemistry, which revealed that the Syn-NS and DAGs showed distinctly separate patterns of distribution (Fig. 4A–C). In the larval arms, DAGs were aligned parallel to the apical surface of the ectoderm. They were separate from the Syn-NS fibers, but were associated closely with regularly arranged perpendicular Syn-NS fibers (Fig. 4D, arrowheads), which suggested that the DAG stripe was not connected structurally to the Syn-NS. The DAG stripes continued along the larval arms (Fig. 4E). Although the primary Syn-NS branch was aligned parallel to the DAG stripe, with a space between the most anterior edge of the DAG stripe at the epaulets, the minor Syn-NS fibers were localized closer to the DAG stripe (Fig. 4F, arrow).

To examine the changes over time of the expression of the DA receptor, the expression patterns of both the Hp-DRD1 mRNA and the associated protein were examined. Analysis by RT-PCR showed that the Hp-DRD1 transcript was present in unfertilized eggs (Fig. 5A, lane 1) and that its relative concentration was significantly lower in fertilized eggs (Fig. 5A, lane 2). However, the transcript was then detected continuously and clearly from the 6 hpf morula stage (Fig. 5A, lane 3) to the 72 hpf pluteus stage (Fig. 5A, lane 9).

Hp-DRD1 protein was not detected immunohistochemically in 8 hpf unhatched non-rotatory blastulae (Fig. 5B). However, by the time the embryos had acquired rotatory movement in the fertilization envelope, before hatching at 9.5 hpf, DRD1 granules were expressed on the embryonic surface in an irregular pattern (Fig. 5C). After the 20 hpf early gastrula stage, the size of the DRD1 granules decreased, and the granules were arranged throughout the embryonic surface (Fig. 5D). Immunofluorescence was not detected when the antibody was pre-absorbed with the Hp-DRD1-GST-fusion protein (Fig. 5E), which indicated that the positive immunoreaction was specific to Hp-DRD1. The distribution pattern was remarkably similar to that of the DAGs (Fig. 1). Like the DAGs, the DRD1 granules converged at the circumoral ectodermal region by the 48 hpf pluteus stage (Fig. 5F). The DRD1 granules were localized in stripes on the larval arms and epaulets by the 34 dpf pluteus stage (Fig. 5G, arrows and double-arrows, respectively, inset labeled a). These DRD1 granules were aligned in five rows of stripes (Fig. 5G, inset labeled b), which were reminiscent of the five-row striped pattern of DAGs seen in the larval arms (Fig. 4D, inset, numbers 1–5). Each of these DRD1 granules overlapped partially with γ-tubulin on the apical side (Fig. 6A–C). DRD1 granules were located at the apical side of the cell (Fig. 6D), and overlapped with γ-tubulin 2.9 μm below the cell surface (Fig. 6E). At 4 μm from the apical surface, only γ-tubulin was observed to be present (Fig. 6F). Although the distances of the DRD1 granules and γ-tubulin from the apical surface of the cell of 34 dpf larval arms were not the same as those of DAGs and γ-tubulin in 24 hpf prism larvae (Fig. 2), the DRD1 granules and DAGs were located very close to γ-tubulin, which suggests that DAGs and DRD1 granules are strongly co-localized at the base of the cilium. The exact relationship between DAGs and DRD1 granules was not examined by immunohistochemical double-staining owing to the fact that the Triton X-100 treatment necessary for the Hp-DRD1 immunohistochemistry results in the loss of the antigenicity of DA.

Immunostaining of DA and synaptotagmin (Fig. 4) suggested the existence of an interaction between the Hp-DRD1 stripes and the 5HT-NS. This could provide the morphological basis of the swimming activity of larvae involving both the 5HT-NS and the DA/Hp-DRD1 system. In 34 dpf eight-arm plutei, an Hp-DRD1 stripe at the anterior and at the posterior epaulet was located between two major branches of the serotonin receptor cell network (SRCN) (Fig. 7A–D for anterior epaulet). Its anterior edge was close to a minor branch of the SRCN (Fig. 7C,D, arrow), whereas two major branches were distinctly separate from the Hp-DRD1 stripe. The SRCN, on the other hand, had a pattern similar to that of the 5HT-NS (Fig. 7E–H). In particular, the SRCN, 5HT-NS (Fig. 7H) and Hp-DRD1 stripes (Fig. 5G) exhibited highly similar patterns at the base of the larval arms. At the epaulet region, these three components exhibited patterns similar to those observed at the base of the larval arms (Fig. 7I).

The dynamic reorganization of the pattern of DA/DRD1 granule (DA/DRD1-G) distribution on the larval surface that was observed during formation of the ciliary band was initiated at around the 48 hpf pluteus stage (Fig. 1G, Fig. 5F). This implicates several morphogenetic mechanisms, such as local changes in cell shape associated with convolution of the ectodermal cell sheet and/or local proliferation of ectodermal cells, in this process.

To examine possible changes in cell shape, the ciliary bands of the larval arms were analyzed by immunostaining with anti-Hp-DRD1 antiserum and Epith-2 (Fig. 8A–C). The Epith-2 staining pattern indicated that the region contained a larger number of smaller cells (Fig. 8B, solid line rectangle) than was found in the non-ciliary band of ectoderm that surrounded the ciliary band (Fig. 8B, dashed-line rectangle). The merged image shows a single DA/DRD1-G per cell in the ciliary band (Fig. 8C, solid-line rectangle and inset). Thus, the possibility of the local appearance of cells with multiple DA/DRD1-Gs was ruled out. The cell number seemed too large for simple changes in cell shape, and seemed to suggest an increase in the number of cells that expressed DA/DRD1-G. The merged image also indicated that the non-ciliary band of ectoderm that surrounded the ciliary band did not express DA/DRD1-G (Fig. 8C, dashed-line rectangle). These observations suggest that local proliferation of cells and the loss of DA/DRD1-Gs from the ectodermal cells in the region adjacent to the ciliary band are responsible for the distribution pattern of DA/DRD1-Gs.

To examine the local proliferation of cells, plutei were incubated with 0.1 μmol l⁻¹ BrdU for 7 h from either the 41 hpf to the 48 hpf stage or the 44 hpf to the 51 hpf stage. These are the periods during which the ciliary band emerges (Fig. 1H). In 48 hpf plutei, BrdU-positive cells comprised a broad circumoral band (Fig. 8D) in which nuclei also accumulated (Fig. 8E,F). In 51 hpf plutei, the circumoral band of BrdU-positive cells was narrower than that in younger plutei (Fig. 8G), and the nuclei accumulated in the circumoral region (Fig. 8H,I). These observations indicate that formation of the ciliary band is associated with: (1) local changes in the shape of ectodermal cells, (2) withdrawal of DA/DRD1-Gs from nearby ectodermal cells that are not part of the ciliary band, and (3) active proliferation of ectodermal cells at the ciliary band.

The immunohistochemical studies described herein suggest that DA/DRD1-Gs are the morphological basis of a dopaminergic system that regulates swimming in embryos and larvae. In support of this, we found that synthesis of DA was inhibited by carbidopa, a potent inhibitor of AADC (Gilbert et al., 2000), in larvae that were treated from the 24 hpf prism stage, the stage when the 5HT-NS emerges, to the 48 hpf pluteus stage. The swimming activity of these larvae was not affected significantly at concentrations of carbidopa ≤15 μmol l⁻¹. However, upon treatment with 20–25 μmol l⁻¹ carbidopa, the swimming activity decreased significantly in a dose-dependent manner (P<0.05; Fig. 9A). Larvae that had sunk to the bottom of the bioassay tubes showed spasmodic, weak movement.

In plutei treated with 25 μmol l⁻¹ carbidopa, DAGs associated with the ciliary band were scarcely detected (Fig. 9B). Carbidopa also caused a dose-dependent delay or inhibition of morphogenesis in some cases, which resulted in features such as the slightly under-developed larval arms (Fig. 9B). Carbidopa was lethal at a concentration of 50 μmol l⁻¹ (data not shown).

The target enzyme of carbidopa is also involved in the synthesis of serotonin. Thus, it was possible that the effect of carbidopa on larval swimming was not due to effects on DA but was instead due to the inhibition of serotonin synthesis. However, immunohistochemical analysis indicated clearly that the 5HT-NS was not affected visibly by carbidopa even at a concentration of 25 μmol l⁻¹ (Fig. 9C,D), and therefore the reduced level of swimming by larvae could not have been due to a decreased level of serotonin synthesis or morphogenetic delay, but rather to the perturbation of DA synthesis.

The 5HT-NS of the larvae of sea urchins has long been regarded as the primary apparatus that regulates the beating of cilia during larval swimming (Yaguchi and Katow, 2003; Katow et al., 2004; Katow et al., 2007). However, the mechanisms that underlie the period of serotonin-independent swimming that occurs from soon after hatching until immediately before serotonergic ganglion development at the 24 hpf prism larva stage have been an ongoing mystery (Yaguchi and Katow, 2003). In the present study, we detected DA/DRD1-Gs at 9.5 h before hatching, which is close to the time when the blastulae acquire rotatory movement in the fertilization envelope (Fig. 1B). This strongly suggests that the DA/DRD1-G system could be an important contributor to ciliary beating in early embryos.

Embryos treated with 25 μmol l⁻¹ carbidopa soon after insemination developed normally until the 15 hpf swimming blastula stage and had apparently normal swimming activity. However, a few hours later, at the 20 hpf early gastrula stage, their swimming activity decreased rapidly, and 90% of them had sunk to the bottom of the bioassay tubes (Fig. 10A), and showed weak spasmodic movement. Expression of DAGs in these embryos was found to have decreased considerably (Fig. 10B,C), whereas DRD1 granules were not affected (Fig. 10D,E). Longer periods of incubation with carbidopa resulted in the death of most embryos. The effects of this reagent were considerably more severe in early embryos than in 24 hpf prism larvae (Fig. 9A). However, in the presence of exogenous DA, the swimming activity was restored in a dose-dependent manner; P<0.05 between control and 25 μmol l⁻¹ DA, and P>0.05 between control and 50 μmol l⁻¹ DA. Expression of DA granules was considerably restored in the embryos that were treated with a mixture of 25 μmol l⁻¹ carbidopa and 25 μmol l⁻¹ DA (Fig. 10G), which indicated that the swimming activity and the formation of DA granules of early embryos are heavily dependent on DA.

To examine whether knockdown of the Hp-DRD1 protein affected the swimming activity of larvae, a MASO that was directed against Hp-DRD1 was microinjected into eggs soon after fertilization. The eggs developed normally until around the 18 hpf swimming blastula stage. However, during the following 2–3 h, the swimming activity of the MASO-treated embryos decreased rapidly in a dose-dependent manner (Fig. 11A). In 20 hpf embryos injected with 750 μmol l⁻¹ MASO, swimming activity decreased significantly (P<0.05), whereas only a small decrease was observed between the control and 500 μmol l⁻¹ MASO groups. Expression of DRD1 granules was severely inhibited (Fig. 11B,C); HNPGs for DRD1 in Hp-DRD1-MASO-injected embryos was 77.6±22.8/embryo (N=9) and that of control MASO-injected embryos was 226.8±75.4/embryo (N=11) (P<0.005). However, formation of DAGs was virtually unaffected by MASO treatment (Fig. 11D,E); HNPGs for DA of control MASO-injected embryos was 400.3±94.6/embryo (N=4) and that in Hp-DRD1-MASO-injected embryos was 386±61.8/embryo (N=7) (P=0.9259). Taken together with the depletion of DAGs induced by carbidopa (Fig. 9B), the MASO results suggest that the formation of DAGs is independent of the formation of DRD1 granules.

The high concentration of DA/DRD1-Gs in the ciliary bands of larval arms and epaulets supports the notion that the dopaminergic system is involved in the regulation of the beating of cilia. This pattern of expression is remarkably similar to that identified in a previous whole-mount in situ hybridization study of the expression of Hp-DRD1 mRNA (Suyemitsu, 2007). Convergence of DA/DRD1-Gs from the initial evenly distributed pattern in embryonic stages to the well-organized ciliary band pattern occurs at around the 48 hpf early pluteus stage. This is similar to the findings related to the pattern of ciliary band formation in the doliolaria larvae of Florometra serratissima that were obtained by scanning electron microscopy (Lacalli and West, 1986). Our immunohistochemical results suggest that the major driving mechanism of ciliary band formation includes the local proliferation of ectodermal cells and the loss of DA/DRD1-Gs from the cells around the ciliary band ectoderm. According to the in situ hybridization database of the Sea Urchin Genome Project (http://goblet.molgen.mpg.de/eugene/cgi/eugene.pl), several region-specific genes are transcribed in the ciliary band ectoderm, which includes genes related to cell proliferation such as Sp-early histone H3, a histone H3 homolog (Kase et al., 2006); Sp-Hes, a hes homolog (Castella et al., 2000); and Sp-PaxC, a neurogenetic paxC homolog (Hadrys et al., 2005). These genes might provide the genetic basis for local cell proliferation during formation of the ciliary band, and will be an interesting focus of future study.

As mentioned above, the five-row striped pattern of DA/DRD1-Gs, which was typically seen in the ciliary band of the larval arm, emerged at the 48 hpf pluteus stage. The number of rows remained constant throughout larval development until immediately before metamorphosis at around the 34 dpf eight-arm pluteus stage. At this stage, the number of rows of DA/DRD1-Gs increased to approximately seven at the apical tuft region. The mechanism of formation of the rows of granules remains to be elucidated.

We have shown here that DA/DRD1-Gs are expressed at the same time as the acquisition of rotatory movement before hatching. Our previous observations indicated that embryonic swimming activity is insensitive to the absence of serotonin in embryos treated with pCPA (Yaguchi and Katow, 2003). These observations, together with those from a previous study (Soliman, 1983b), suggest that the swimming activity of early embryos is not regulated by the 5HT-NS. Formation of DAGs was inhibited by the suppression of DA synthesis by carbidopa, which is a potent inhibitor of AADC activity that can affect the synthesis of serotonin and formation of the 5HT-NS. Although AADC is involved in the synthesis of both DA and serotonin (for a review, see Nagatsu, 1991), our findings indicate that inhibition of AADC by carbidopa did not inhibit serotonin synthesis. Thus, the decreased swimming activity of larvae treated with carbidopa is due to the decreased synthesis of DA.

DA and serotonin have been thought to have counteracting functions in the beating of cilia and/or the regulation of swimming activity (Braubach et al., 2006; Martin et al., 2008), because DA was shown to suppress the beating of cilia in the gill of the mollusk Mytilus edulis (Catapane et al., 1979) as well as that in the palatine mucosa of frog (Maruyama et al., 1983). However, in sea urchin larvae (Soliman, 1983b), as well as in the motile cilia that line the foot of the mollusk Tritonia diomedea (Woodward and Willows, 2006), both serotonin and DA have been shown to stimulate the beating of cilia. Our present observation that the suppression of DA synthesis results in decreased swimming activity in embryos and larvae and our previous observation that the suppression of serotonin synthesis results in decreased swimming activity in larvae (Yaguchi and Katow, 2003) are supported by previous reports (Soliman, 1983b; Woodward and Willows, 2006). These apparently contradictory results might indicate the presence of a tissue-specific mechanism to regulate the beating of cilia. In this scenario, both serotonin and DA could have opposite effects depending on the concentrations of the compounds to which cilia are exposed (Soliman, 1983b). On the other hand, our immunohistochemical results suggest a synergistic relationship between the 5HT-NS, SRCN and the preneural dopaminergic system. In early plutei, serotonin that is injected into the blastocoel triggers the release of calcium ions in the cytoplasm of ectodermal cells (Katow et al., 2007); therefore, physical contact between the 5HT-NS and SRCN might not be essential.

Although dopaminergic cells have been reported to exist from the four-arm pluteus stage, be found in the lower lip ganglion and the post-oral arm ganglia, and possess axonal connections (Bisgrove and Burke, 1987), our whole-mount immunohistochemistry conducted using two different anti-DA antibodies did not detect these structures. In our study, ectodermal cells that contained DA/DRD1-Gs had neither cytoplasmic DA nor axonal connections in embryos or larvae. These data suggest that DA is responsible for a preneural form of signal transduction, which contradicts a previous report on S. droebachiensis (Bisgrove and Burke, 1987). These discrepancies could be the result of differences between the species.

Although no immunohistochemical observation of a close localization between the DA/DA receptor and the basal body of cilia in organisms other than the present sea urchin embryos and larvae has been made to date, such sub-cellular localization and previously identified morphological and functional relationships may provide the morphological basis of DA involvement in the regulation of the beating of cilia through DA receptors (Catapane et al., 1979; Soliman, 1983b; Maruyama et al., 1983; Braubach et al., 2006; Martin et al., 2008).

The results reported here could represent the first observation of the association of DA with granules that measure several micrometers in diameter. However, the association of Hp-DRD1 with such granules shows remarkable similarity to previous immunohistochemical findings on several subtypes of DA receptors in the human brain (Kumar and Patel, 2007) and multiciliated ependyma in rat, mouse and bovine species (Tomé et al., 2007). The observed DA/DRD1-Gs are far larger than a single GPCR protein (Cherezov et al., 2007), which suggests that each DA/DRD1-G is composed of multiple heteropolymers of components of the DA/DRD1 signaling pathway (Capasso et al., 1988; Carginale et al., 1992; Girault and Greengard, 2004; Arguello and Gogos, 2008).

We observed that DA and Hp-DRD1 appeared in granules on the surface of the ectoderm from the 9.5 hpf unhatched rotatory blastula stage until the end of the larval stages. Thus there was an apparent discrepancy between the continuous presence of Hp-DRD1 mRNA from the unfertilized egg stage and the detection of Hp-DRD1 protein only from the 9.5 hpf unhatched blastula stage. The results obtained by RT-PCR were consistent with the pattern of accumulation of Sp-DRD1 mRNA that was detected by microarray analysis (http://urchin.nidcr.nih.gov/cgi-bin/exp.plx?glean=11320). The DA binding site of DRD1 is located in the transmembrane domain, and it has been suggested that binding of DA to this site is associated with allosteric conformational change (Missale et al., 1998). However, the present study strongly suggests that the formation of DAGs and DRD1 granules occurs mechanically independently. Allosteric conformational change of the DRD1 molecule may not contribute to the acquisition of immunoreactivity. The mechanism underlying this phenomenon is yet to be found.

Although double-immunostaining for DA and Hp-DRD1 was not possible, the spatial relationship between DAGs and DRD1 granules that was detected by DA/γ-tubulin and Hp-DRD1/γ-tubulin double-staining strongly suggests that DAG is located close to DRD1 granules at the basal bodies of cilia. This suggests a close relationship between DA/DRD1-Gs and the beating of cilia, and hence to the swimming of embryos and larvae.

We thank M. Washio for collecting the sea urchins for this study and Dr Y. Nakajima for generously providing us with the 1E11 antibody.

 
• AADC

  aromatic-l-amino acid decarboxylase

 
• AT

  acetylated tubulin

 
• BrdU

  5′-bromo-2′-deoxyuridine

 
• DA

  dopamine.

 
• DA/DRD1-G

  dopamine/DRD1 granule

 
• DAGs

  dopaminergic granules

 
• dpf

  days post-fertilization

 
• FSW

  filtered seawater

 
• 5HT-NS

  serotonergic nervous system

 
• GPCRs

  G-protein-coupled receptors

 
• GST

  glutathione S-transferase

 
• HNPGs

  half number of positive granules

 
• Hp-DRD1

  Hemicentrotus pulcherrimus homolog of the dopamine receptor 1

 
• hpf

  hours post-fertilization

 
• l-DOPA

  l-3,4-dihydroxyphenylalanine

 
• MASO

  morpholino antisense oligonucleotides

 
• PBST

  phosphate-buffered saline with 0.1% Tween-20

 
• pCPA

  para-chlorophenylalanine

 
• PI

  propidium iodide

 
• RT-PCR

  reverse transcriptase polymerase chain reaction

 
• SRCN

  serotonin receptor cell network

 
• Syn-NS

  anti-synaptotagmin antibody-positive nervous system