Receptor cloning

Degenerate oligonucleotide primers were used to amplify 5-HT receptor sequences from H. trivolvis genomic DNA as described previously (Sugamori et al., 1993). In the first round of polymerase chain reaction (PCR), primers transmembrane 3 (TM3) and TM7 (Table 1) were used to amplify from a genomic DNA template (concentration 10^–4 μg μl^–1). MgCl₂ concentration ranged from 1 to 5 mmol l^–1, with a buffer containing 20 mmol l^–1 Tris-HCl (pH 8.3 at 20°C), 25 mmol l^–1 KCl, 100 μg ml^–1 gelatin, 50 μmol l^–1 each dNTP, Taq DNA Polymerase and Tli DNA Polymerase (Promega, Madison, WI, USA). Touchdown PCR was carried out with an annealing temperature range of 60 to 51°C.

PCR products (10 μl) from the first reaction were used as the template for a second round of PCR using nested primers TM3 and TM6 (Table 1). The re-amplification was performed using the same program as above, over a MgCl₂ concentration range of 1 to 9 mmol l^–1.

PCR products were then cloned into pGEM-T. Individual colonies were picked, boiled, screened by PCR, and the DNA sequenced. PCR primers specific for two distinct receptors were designed based on the initial PCR products, and used to screen sections of a H. trivolvis CNS lambda-ZAP cDNA library consisting of 44 fractions (made by Erno Vreugdenhill, 1993, and kindly provided by Garry Hauser and Andy Bulloch, University of Calgary). Primers WJG1157 and WJG1158 were used to identify library sections containing 5-HT_1Hel, while primers WJG1183 and WJG1184 were used to identify those sections containing 5-HT_7Hel (Table 1). Out of the cDNA library sections producing PCR bands of the expected size, one was chosen for screening for each putative receptor clone: CNS fraction no. 1 for 5-HT_1Hel and CNS fraction no. 4 for 5-HT_7Hel.

Inserts from plasmids containing the primary cloned PCR products were excised and labeled with ³²P-αdCTP using the Prime-A-Gene labeling system (Promega, Madison, WI, USA). Labeled probes were used for plaque screening of the lambda-ZAP library sections.

Inserts of plaque-purified lambda clones were excised as plasmids (Stratagene, La Jolla, CA, USA) and sequenced on both strands using a primer walking strategy.

Phylogenetic analysis

A large-scale phylogenetic analysis of G protein-coupled biogenic amine receptors was performed on a set of 768 proteins collected from GenBank (National Center for Biotechnology Information (NCBI, Bethesda, MD, USA). Sequences were aligned using MUSCLE v3.6 (Edgar, 2004). The gapped regions corresponding to poorly conserved N-terminal and C-terminal domains and intracellular and extracellular loops were removed, yielding a final data set of 214 characters. A phylogenetic tree was inferred from this dataset using MrBayes v3.1.2 (Ronquist and Huelsenbeck, 2003). A mixture of 10 different amino acid substitution models was evaluated, and rate variability was modeled as a gamma function with a fraction of sites invariant. Two independent runs consisting of four chains each were carried through a total of 2 000 000 cycles of search, with every 100th cycle being kept. The first 15 001 stored trees from each run were discarded and the remaining 10 000 trees were used to construct the consensus. The two H. trivolvis receptors were robustly partitioned into a clade consisting of the type 1, type 5 and type 7 receptors. A representative subset of the vertebrate and invertebrate type 1, type 5 and type 7 receptors and two cnidarian biogenic amine receptors were selected and aligned and the alignment was trimmed to yield a dataset consisting of 252 characters. This dataset was used to infer another phylogenetic tree as described above.

In situ hybridization

Templates based on the 3′ untranslated region (UTR) for 5-HT_1Hel and the 5′ UTR for 5-HT_7Hel were produced by PCR using specific primers, with nucleotide sequences for either the T7 (sense direction, for control probe production) or SP6 (antisense direction, for experimental probe production) promoter regions on the 5′ end of the appropriate primer (Table 1). PCR products were gel purified and yields quantified, and products were sequenced using dye termination.

RNA probes were transcribed and digoxigenin (DIG) labeled using the DIG-RNA labeling kit (SP6/T7; Roche Diagnostics, Laval, QC, Canada). Labeled probe was quantified by dot blot comparison to standards.

H. trivolvis animals of an inbred lab-raised albino strain were maintained and embryos collected as described before (Goldberg et al., 1994). Embryos were staged as described previously (Diefenbach et al., 1998; Goldberg et al., 1988). Embryos were collected from their egg capsules, rinsed with 0.01 mol l^–1 phosphate buffered saline (PBS) and immediately fixed in freshly prepared and filtered 4% paraformaldehyde (pH 7.5) in PBS at 4°C for 2 h to overnight. All solutions from this point on were RNase free. Embryos were rinsed twice in PBT (PBS with 0.1% Triton X-100) then dehydrated in an ascending ethanol series in PBT (25%, 50%, 75%) for 10 min each at room temperature, followed by two washes in 100% ethanol. Embryos were incubated at– 20°C for at least 2 h. Embryos were rehydrated in a descending (75%, 50%, 25%) ethanol series in PBT at room temperature for 10 min per wash, and then rinsed twice in PBT. Embryos were digested in either 20 μg ml^–1 proteinase K for 20 min at 37°C or 0.6 mg ml^–1 trypsin (Sigma-Aldrich, Oakville, ON, Canada) for 5–10 min at 37°C. The digestion was stopped by replacing the enzyme solution with 1 mmol l^–1 phenylmethylsulfonyl fluoride in PBT for 10 min at room temperature. Embryos were refixed in 0.2% gluteraldehyde/4% paraformaldehyde for 20 min at 4°C. They were washed twice in PBT to remove all fixative, and then incubated in prewarmed (55°C) prehybridization solution [40% deionized formamide, 10% dextran sulfate, 1× Denhardt's solution, 4× saline sodium citrate (SSC), 10 mmol l^–1 dithiothreitol (DTT), 1 mg ml^–1 yeast tRNA, 1 mg ml^–1 denatured and sheared salmon sperm DNA] for at least 2 h at 55°C with gentle shaking.

Hybridization with 100 ng probe ml^–1 hybridization solution was carried out at 55°C overnight on a shaking platform. Washes, antibody incubation and antibody washes were performed as described previously (Nieto et al., 1996). Briefly, embryos were washed thrice for 10 min at 55°C in 2× SSC, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and then thrice for 10 min at 55°C in 0.2× SSC, 0.1% CHAPS. Embryos were rinsed in KTBT (50 mmol l^–1 Tris-HCl pH 7.5, 150 mmol l^–1 NaCl, 10 mmol l^–1 KCl, 1% Triton X-100) at room temperature for 10 min and preblocked in 20% fetal calf serum (FCS) in KTBT for at least 2 h at 4°C. Embryos were incubated overnight at 4°C on a rocking platform in 1:2000 diluted anti-DIG alkaline phosphatase-conjugated Fab fragments (Roche Diagnostics, Laval, QC, Canada) in 20% FCS in KTBT. Embryos were washed five times for 1 h in KTBT at room temperature and incubated overnight in KTBT at 4°C.

Chromogenic precipitate formation was carried out in 1 mmol l^–1 levamisole using either the NBT/BCIP color system (for purple precipitate) or the Fast Red/HNPP (for red fluorescent precipitate; Roche Diagnostics). Purple-stained embryos were visualized using brightfield microscopy, while fluorescent embryos were cleared with glycerol and visualized using confocal microscopy.

The in situ hybridization data presented are representative of the results obtained in three repeat experiments using the purple precipitate and two repeat experiments using the Fast Red precipitate. Sense controls were included in every experiment. In each experiment, at least 10 embryos were included from each embryonic stage tested.

Immunohistochemistry

Immunofluorescence localization of 5-HT_1Hel and 5-HT_7Hel receptors was performed on histological sections from the CNS of mature snails and whole-mounts of cultured identified neurons. The intact CNS or subsets of CNS ganglia were dissected out of mature snails as previously described (Young et al., 1999), fixed in 4% paraformaldehyde in PBS for 24 h at 4°C, and washed three times in PBS for 60 min at room temperature. The tissues were dehydrated through a series of ethanol into toluene, embedded in paraffin wax, cut into 10 μm sections, mounted on slides and incubated for 24 h at 37°C. The sections were washed three times with toluene to remove the paraffin, rehydrated though a series of ethanol and rinsed three times with PBS over 30 min. Primary antibodies to 5-HT_1Hel and 5-HT_7Hel were raised in rabbits against peptides derived from intracellular loop sequences of the respective receptor proteins, as described previously (Doran and Goldberg, 2004). The antibodies were diluted 1:500 in blocking medium (4% horse serum, 0.1% NaN₃, 0.1% Triton X-100 in 0.1 mol l^–1 PBS) and applied to sections for 24 h at 4°C under gentle agitation. The sections were washed six times in PBS over 3 h and then exposed to 1:400 diluted goat anti-rabbit IgG conjugated to Alexa 488 (Invitrogen, Carlsbad, CA, USA) for 3 h at 4°C under gentle agitation. The slides were washed six times in PBS over 3 h, mounted in 80% glycerol in PBS, and stored for 3 days at 4°C before viewing. The immunoreactivity data presented are representative of the results obtained from four different experiments on a total of 12 isolated CNS.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Phylogenetic reconstruction illustrating the placement of H. trivolvis 5-HT receptor proteins as orthologs of 5-HT₁ and 5-HT₇ receptors in other animals. The tree illustrates the relationship between the two H. trivolvis receptors and other receptors of the 5-HT₁, 5-HT₅ and 5-HT₇ families. This figure represents a subtree of a larger phylogenetic tree generated from 768 biogenic amine receptors, aligned and trimmed to yield a matrix of 214 amino acid characters. For simplicity, clades encompassing a large number of vertebrate receptors are represented diagrammatically as triangles. Each sequence is represented by the name of the species from which it was obtained and a GenPept identification number that is uniquely assigned to each sequence. The scale bar represents an amount of evolutionary change corresponding to an expected 0.1 changes per site.

Immunolocalization of 5-HT_1Hel and 5-HT_7Hel on the identified cerebral ganglion neuron C1 in culture was done according to the protocol used to detect 5-HT_1Hel and 5-HT_7Hel immunoreactivity in cultured embryonic ciliary cells (Doran and Goldberg, 2004). Neuron C1 was isolated and cultured according to the methods of Price and Goldberg (Price and Goldberg, 1993), with one exception. The brain-conditioned medium was prepared using H. trivolvis brains that were first washed for 24 h in defined medium, and then incubated in defined medium for 96 h. In control experiments, pre-immune serum from the rabbit used to generate the primary antibody replaced the primary antibody. In addition, control experiments were also performed where the primary antibody was excluded, or the primary antibody was pre-absorbed either with blocking peptides from the sequence used to generate the antibodies (negative control), or with Keyhole limpet hemocyanin, a carrier protein used to generate the antibody (positive control). Sections or whole-mounts were viewed using an Axiovert 135 fluorescence microscope (Zeiss, ON, Canada) equipped with differential interference contrast (DIC) optics. Images were collected with a Retiga Ex digital charged-coupled device camera (Q-Imaging, Burnaby, BC, Canada) linked to a Pentium 4 PC computer running Northern Eclipse software (Empix Imaging Inc., Mississauga, ON, Canada). The immunoreactivity data are representative of the results obtained from three repeat experiments on a total of 14 isolated C1 neurons.

Cloning of 5-HT receptors

We used nested degenerate PCR, as described by Sugamori et al. (Sugamori et al., 1993) to amplify fragments of 5-HT receptor genomic DNA. PCR products were cloned, and those with sequences similar to those of known 5-HT receptors were radiolabeled and used as probes to screen H. trivolvis CNS cDNA libraries. Two full-length putative 5-HT receptor cDNA clones were isolated in this manner. The sequences of these clones were deposited in GenBank [5-HT_1Hel is under accession numbers AY395746.1 (nucleic acid) and AAQ95277.1 (protein), 5-HT_7Hel is under accession numbers AY395747.1 (nucleic acid) and AAQ84306.1 (protein)].

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Alignment of the seven predicted transmembrane helices of the two H. trivolvis receptors (Hel1, Hel2) with representative vertebrate (Homo sapiens; Hum1D, Hum7) and invertebrate (Drosophila melanogaster; Dmel1, Dmel7) 5-HT₁ and 5-HT₇ receptors. The conserved residues that have been shown to be involved in 5-HT binding to the receptor are marked with asterisks. The arrows along transmembrane (TM) helices 3 to 7 indicate the portion of the sequence that forms the vestibule leading from the extracellular space to the 5-HT binding site.

Phylogenetic analysis

The amino acid multiple sequence alignment of the two H. trivolvis 5-HT receptor proteins with 766 biogenic amine receptors from a variety of organisms yielded a data set of 214 characters after removal of gapped regions (supplementary material Fig. S1). A consensus phylogenetic tree constructed using MrBayes v3.1.2 (Ronquist and Huelsenbeck, 2003) robustly placed one H. trivolvis receptor in a clade of invertebrate type 1 5-HT receptors and the other in a clade of invertebrate type 7 5-HT receptors (Fig. 1 and supplementary material Figs S2 and S3). Thus we have designated one receptor as 5-HT_1Hel and the other as 5-HT_7Hel, based on phylogenetic relationships and the nomenclature suggested by Tierney (Tierney, 2001). The large scale of the phylogenetic analysis supports a robust placement of the 5-HT₅ family of receptors as the sister group of the 5-HT₁ family, with the 5-HT₇ receptor family being the sister group of the 5-HT₁/5-HT₅ clade. To confirm that the phylogenetic partitioning was not due to an artifact arising from aligning a large number of sequences with widely varying sequence, we performed a similar analysis using selected examples of Type 1, 5 and 7 receptors and two cnidarian receptor sequences as an out group. This alignment of 26 sequences yielded 252 aligned positions, and a phylogenetic tree (not shown) that was completely compatible with the type 1,5,7 clade from the larger analysis (Fig. 1).

Unlike the vertebrate 5-HT₁ receptor clade, which contains five paralogs that apparently arose in the common ancestor of vertebrates, the invertebrate 5-HT₁ receptor clade generally contains only single homologs from each species that is represented. In the cases where there are more than one paralogous sequence in an organism (e.g. the two 5-HT₇ receptors from Dugesia japonica and the three 5-HT₁ receptors from Drosophila melanogaster) the paralogy appears to have arisen late in evolution.

Molecular characterization

Full-length 5-HT_1Hel cDNA is over 4000 base pairs (bp) in length, excluding the poly-adenylated tail. The 5′ UTR is 144 bp long, the 3′ UTR is over 2000 bp and the open reading frame (ORF) is 1509 bp long, predicting a protein of 503 amino acids (aa) in length. The putative 5-HT_1Hel receptor has seven hydrophobic helical domains that are strongly predicted by the Transmembrane Hidden Markov Model program (TMHMM) (Sonnhammer et al., 1998), with an extracellular N-terminus that is weakly predicted to contain a non-cleaved signal peptide sequence. The seven transmembrane helices align with the characteristic motifs within transmembrane regions of other seven-pass G protein-coupled receptors (Fig. 2). Interestingly, the species homolog 5-HT_1Lym from the related pond snail Lymnaea stagnalis (Sugamori et al., 1993) shares a similar protein structure profile, being 508 aa long, having nearly identically sized amino- and carboxy-termini (94 aa and 19 aa, respectively, for 5-HT_Lym compared with 95 aa and 19 aa, respectively, for 5-HT_1Hel) and comparably sized third intracellular loops (150 aa for 5-HT_Lym, 154 aa for 5-HT_1Hel).

The full-length 5-HT_7Hel cDNA is 4295 bp in length excluding the poly-adenylated tail. The 5′ UTR is 820 bp in length, the 3′ UTR is 2034 bp long and the ORF is 1437 bp in length, predicting a protein of 479 aa. Hydrophobicity analysis based on TMHMM (Sonnhammer et al., 1998) predicts that the putative 5-HT_7Hel protein has seven distinct transmembrane regions. Both amino- and carboxy-termini are 70 aa in length, and the third intracellular loop is 97 aa long. There is no predicted cleaved signal peptide at the N-terminus, but the N-terminus is identified as extracellular by TMHMM.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Localization of 5-HT_1Hel and 5-HT_7Hel transcripts in whole-mounts of H. trivolvis embryos. In situ hybridization of digoxigenin (DIG)-labeled anti-sense RNA probes was immunohistochemically localized with alkaline phosphatase-conjugated anti-DIG antibodies and reactants that form either a purple (A–C; light microscope) or fluorescent red (D and E; confocal microscope) precipitate. Using anti-sense probes in stage E45 embryos (A, B and D), 5-HT_1Hel RNA expression was localized mainly to the ciliated epithelium of the foot (open arrows), anterior head (filled arrows), tentacles (^*), ciliated mantle (filled arrowheads) and gut (open arrowheads). Using sense probes to 5-HT_1Hel, no signal was observed (C). In stage E30 embryos (E), 5-HT_7Hel RNA expression was localized to the ciliated foot epithelium (open arrows), developing gut (open arrowheads), and ciliated mantle (filled arrowhead). Using sense probes to 5-HT_7Hel, no signal was observed (not shown). Scale bar, 100 μm.

The regions of both receptors that were identified as transmembrane helices aligned well with each other and with homologous regions from other 5-HT receptors (Fig. 2). The conserved residues that are involved in 5-HT binding are conserved in helices 3–7. The sequences of the helices are more conserved on the cytoplasmic side of the binding cavity than they are in the vestibule of the binding cavity that forms the pathway for 5-HT to reach the ring of binding residues. The DRY (aspartate–arginine–tyrosine) sequence motif that is found at the cytoplasmic boundary of the third transmembrane helix of essentially all biogenic amine receptors (D and R are both present in more than 99% of the sequences and Y is present in more than 95% of the sequences) is present in these two receptor proteins.

Localization in embryos and the CNS of adult H. trivolvis

Localization of the 5-HT_1Hel and 5-HT_7Hel gene transcripts was initially examined using in situ hybridization in Helisoma embryos, where 5-HT has been shown to be involved in the regulation of ciliary beating (Diefenbach et al., 1991; Kuang and Goldberg, 2001), neurite outgrowth (Diefenbach et al., 1995; Goldberg and Kater, 1989; Goldberg et al., 1991; Goldberg et al., 1992) and neuronal intracellular calcium concentration (Goldberg et al., 1992). The RNA DIG labeled probe for 5-HT_1Hel was based on the full-length insert corresponding to the coding region between TM3 and TM6. Embryos at stage E45–50, which represents the end of the prototrocal–juvenile transition, displayed consistent expression of 5-HT_1Hel on the ciliated foot, tentacles and ciliated mantle (Fig. 3A,B). Control reactions using the corresponding sense, rather than anti-sense, sequences revealed no expression (Fig. 3C). Furthermore, a second probe that was generated from the 3′ UTR of the 5-HT_1Hel gene revealed a similar expression pattern to that seen with the initial probe, suggesting that the expression did not result from cross-reactivity with other receptor gene sequences. This latter probe was revealed using Fast Red labeling of anti-DIG immunoreactivity and confocal microscopy, confirming that the whole-mount techniques were sufficient to reveal staining in the interior of the embryos (Fig. 3D). This approach revealed additional expression where the neurites of serotonergic neurons innervate foot ciliary cells (Koss et al., 2003), as well as within the gastrointestinal tract.

In situ hybridization experiments on the 5-HT_7Hel receptor gene revealed a similar expression pattern to that described for 5-HT_1Hel (Fig. 3E). Confocal optical sectioning of Fast Red–labeled embryos showed expression of 5-HT_7Hel in the region where the pedal cilia are innervated by the ENC1 neurons, as well as in the gastrointestinal tract and at the ventral mantle (Fig. 3E). Expression was also observed in the primordial tentacles, whereas no expression was observed when a sense probe was used (data not shown).

The expression of 5-HT_1Hel and 5-HT_7Hel receptors was also characterized in histological sections of the CNS from mature snails. Using antibodies raised against peptides from within distinct intracellular loop regions of the receptor proteins (Doran and Goldberg, 2004), immunohistochemistry revealed widespread expression of both receptors throughout the CNS (Figs 4 and 5). Whereas neuritic staining of both receptor proteins occurred in the neuropile, connectives and peripheral nerves in every ganglion of the CNS, neuronal somata appeared unstained in all cases. The absence of cell body staining precluded an analysis of the expression profile in known identified neurons. Interestingly, when cerebral ganglion neuron C1 was isolated and grown in cell culture, 5-HT_1Hel (Fig. 5C–E) and 5-HT_7Hel (data not shown) immunoreactivity was observed primarily in the cell body. The widespread expression of the 5-HT_1Hel and 5-HT_7Hel receptors in both embryos and mature snails suggests that these receptors may mediate some of the well-established developmental and physiological actions of 5-HT.

Phylogenetic relationships of the 5-HT_1Hel and 5-HT_7Hel receptors

Phylogenetic analysis clearly identifies the two new receptors as homologs of type 1 and type 7 5-HT receptors from other organisms. Not unexpectedly, 5-HT_1Hel grouped entirely within a molluscan clade of 5-HT₁ receptor genes, with the 5-HT_1Lym receptor being the closest ortholog. The non-molluscan 5-HT₁ receptor sequences in our analysis were from arthropod and nematode species. In every case, sequences from the same phylum were more similar to each other than to those from different phyla. Furthermore, arthropod and nematode sequences were more closely related to each other than they were to the molluscan sequences, as expected from the ecdysozoan relationship between arthopods and nemotodes. The position of the 5-HT_1Hel gene in the phylogentic tree produced in our analysis supports the hypothesis that a common ancestor of the ecdysozoans and lophotrochozoans contained 5-HT₁ receptors.

Interestingly, the phylogenetic relationships for 5-HT₇ genes were quite different, with 5-HT₇ genes from various insect species grouping together as expected, but the two nematode genes were more related to 5-HT_7Hel and planarian genes than to the insect genes. 5-HT₇ sequences from additional arthropods, nematodes, planarians and molluscs are required to confirm this surprising separation of insect and nematode genes.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

In situ and cultured neurons display different cellular localization of 5-HT_1Hel and 5-HT_7Hel receptors. In situ, 5-HT_1Hel-like immunoreactivity (A) and 5-HT_7Hel-like immunoreactivity (B) was restricted to neurites in neuropile (filled arrows) and neurite tracts within connectives and nerves (arrowhead; also see Fig. 4). Cell bodies (open arrows) were always unstained as revealed by overlaying the differential interference contrast pattern on the fluorescence signal. In contrast, after isolation and culture of cerebral neuron C1, 5-HT_1Hel-like immunoreactivity occurred primarily in the cell body (C and D). Preabsorption of antibody with Keyhole limpet hemocyanin (D) and replacement of antiserum with pre-immune serum (E) provided positive and negative controls for antibody specificity, respectively. Scale bar, 50μ m.

Expression and function of 5-HT_1Hel and 5-HT_7Hel receptors in H. trivolvis

Invertebrates have long been known to express a wide variety of 5-HT receptors, based on the variety of cellular responses, signal transduction elements and pharmacological profiles associated with the actions of 5-HT (Peroutka, 1994; Tierney, 2001). However, only after the primary structure of many different 5-HT receptor proteins was revealed through molecular cloning did it become evident that most, if not all, invertebrate 5-HT receptors belong to one of the seven major families of mammalian 5-HT receptors. It is not yet clear whether all invertebrate 5-HT receptors in a particular family couple to the same signal transduction elements (see below). Furthermore, the pharmacological profile of specific invertebrate receptors is less likely to match that of their mammalian homologs, as the changes in molecular structure during 600 million years of evolution beyond the divergence of invertebrates and vertebrates would probably result in altered receptor binding characteristics (Tierney, 2001).

Several invertebrate 5-HT₁ receptor genes have been cloned to date, including at least seven from molluscan species. Of particular interest is the 5-HT_1Lym receptor cloned from Lymnaea stagnalis (Sugamori et al., 1993), which provided the template for the original primers used in the present study. While a phylogenetic analysis suggested that this receptor also fell into the 5-HT₁ family of receptors, binding analysis indicated a mixed pharmacology relative to the 5-HT₁ profile of vertebrate receptors, typical for invertebrate 5-HT receptors (Goldberg et al., 1994). In contrast to 5-HT₁ receptors, the only invertebrate 5-HT₇ receptor genes cloned to date are from arthropod, roundworm and possibly flatworm species (Hobson et al., 2006; Pietrantonio et al., 2001; Witz et al., 1990). Therefore, the present study is the first to describe the molecular structure of a 5-HT₇ receptor from a molluscan species.

The molecularly identified invertebrate 5-HT receptors generally couple to their respective effector systems in a similar way to their vertebrate homologs (reviewed by Tierney, 2001). Similar to the vertebrate 5-HT₁ receptors, activation of invertebrate 5-HT₁ receptors decreases cyclic AMP levels through G_αi protein-mediated inhibition of adenylate cyclase. Likewise, both vertebrate and invertebrate 5-HT₇ receptors couple to G_αs proteins, causing activation of adenylate cyclase and cyclic AMP production. There are, however, some invertebrate 5-HT receptors that do not belong in any of the seven 5-HT receptor families, even though they couple normally to an effector pathway. For example, the 5-HT_AAp and 5-HT_BAp receptors cloned from Aplysia do not group in any of the known families, even though they activate the enzyme phospholipase C (PLC), a coupling characteristic of 5-HT₂ receptors (Li et al., 1995). Likewise, the MOD-1 5-HT receptor of C. elegans is a ligand-gated chloride channel that does not appear to be a member of the 5-HT3 family of ligand-gated 5-HT receptors (Ranganathan et al., 2000).

Although confirmation of the coupling characteristics of 5-HT_1Hel and 5-HT_7Hel receptors awaits functional analyses performed both in an expression system and in situ, it is reasonable to hypothesize that these receptors act in the same manner as their vertebrate and invertebrate homologs. Evaluation of various 5-HT-mediated responses in H. trivolvis therefore allows for a tentative association of the receptors cloned in this study with particular response pathways. One such pathway is the 5-HT-induced inhibition of neurite outgrowth in buccal ganglion neuron B19 (Haydon et al., 1984; Price and Goldberg, 1993). When regenerating in cell culture, the growth cone motility and neurite elongation of neuron B19 are reversibly inhibited by 5-HT. This response is thought to be transduced by a cascade involving the activation of adenylate cyclase and elevation of cyclic AMP levels, activation of cyclic nucleotide-gated sodium channels, depolarization and action potential activity, influx of extracellular calcium, activation of calcium/calmodulin-dependent protein kinase (Polak et al., 1991), and cytoskeletal rearrangement (Welnhofer et al., 1999). This scheme leads to the prediction that 5-HT_7Hel receptors may mediate this response through their adenylate cyclase-stimulating activity. Unfortunately, both the anti-5-HT_1Hel and anti-5-HT_7Hel antibodies only stained neurites in the intact CNS, such that the absence of soma staining precluded a determination of whether the identifiable neuron B19 expressed these receptors. Staining cultured B19 neurons, or staining in situ preparations after neuron B19 has been injected with a fluorescent marker, could possibly reveal whether the 5-HT_7Hel receptors are expressed in these neurons. Interestingly, neither 5-HT₇ receptors nor members of the 5-HT₄ or 5-HT₆ receptor families, all of which induce elevations in cyclic AMP levels, have been cloned in Aplysia californica, in which serotonergic pathways involving elevations in cyclic AMP have been well characterized (Barbas et al., 2003).

Another well-studied 5-HT pathway in Helisoma underlies a behavioural response to hypoxia during embryonic development. Helisoma embryos contain a pair of unique sensorimotor neurons that detect hypoxia and stimulate ciliary beating in postsynaptic ciliary cells through the release of 5-HT (Kuang et al., 2002; Kuang and Goldberg, 2001). Cell culture studies have revealed that the cilioexcitatory response to 5-HT occurs through a highly complex signal transduction pathway that may include multiple 5-HT receptor subtypes (Doran and Goldberg, 2004), PLC and protein kinase C (PKC) activation (Christopher et al., 1999; Doran and Goldberg, 2006), constitutive nitric oxide activity (Doran et al., 2003) and release of calcium from intracellular stores (Christopher et al., 1996; Doran and Goldberg, 2004). Since inhibitors of PLC or PKC only partially block the cilioexcitatory response (Christopher et al., 1999; Doran and Goldberg, 2006), there may be more than one type of 5-HT receptor expressed on ciliary cells, with each one mediating a component of the entire response through different pathways. The current in situ hybridization experiments and previous immunofluorescence experiments (Doran et al., 2004) together suggest that both 5-HT_1Hel and 5-HT_7Hel receptors are expressed in embryonic ciliary cells, thus supporting the multi-receptor model. However, the expected coupling mechanisms of these types of 5-HT receptors do not easily fit into what is currently known about the signal transduction of the cilioexcitatory response. Previous experiments indicated that the response is not mediated by an elevation of cyclic AMP levels, which would be expected from a 5-HT₇ receptor (Christopher et al., 1996). Furthermore, the possibility that a 5-HT₁-mediated decrease in cyclic AMP levels is involved in the response has not yet been explored. One possibility is that while an elevation in cyclic AMP may not be involved in producing the primary response, it is involved in producing the long duration plateau in the response that persists beyond the removal of 5-HT (Gallin et al., 2006), as well as the response facilitation that occurs upon repeated exposure of the animal to anoxia (Kuang et al., 2002). Determining whether the 5-HT_1Hel and 5-HT_7Hel receptors couple typically to the cyclic AMP system or operate through atypical signal transduction pathways, and their specific roles in the cilioexcitatory responses to 5-HT, awaits a functional examination of expressed receptors and molecular knockout experiments on ciliary cells.

The in situ hybridization experiments in this study and previous immunolocalization experiments on embryos (Doran et al., 2004) revealed widespread expression of both receptor subtypes in all ciliated regions. Whether this expression indicates roles for these receptors in the regulation of ciliary activity, as proposed above, or more fundamental roles in regulating the development of these tissues remains to be determined. Expression was also observed in regions of the embryo associated with neural tissue, such as the regions containing ENC1 somata at early stages of embryonic development (data not shown), and the regions where ENC1 neurons innervate pedal ciliary cells at later stages (Fig. 3). In the adult CNS, both receptor proteins were expressed widely and selectively in neurons throughout all ganglia of the CNS, corroborating the data from embryos that both of these are neural receptors. To our surprise, CNS expression was entirely limited to neurites within neuropiles and connectives, with none seen in neuronal somata. Although our preliminary Western blot experiments did not confirm that each antibody recognizes only a single antigen, the striking restriction of CNS immunoreactivity to neurites argues that the antibodies were highly selective for receptor proteins. In an autoradiographic study of lysergic acid diethylamide binding in the CNS of Aplysia californica, Kadan and Hartig (Kadan and Hartig, 1988) reported that 5-HT receptors were most prominently localized to the neuropile, with relatively few neurons displaying somatal expression. In the present study on tissue sections, the intense immunoreactivity seen in numerous neurites suggests that the absence of cell body staining correctly reflected an absence of actual expression. Perhaps the small number of labeled cell bodies seen in Aplysia represents the expression of a 5-HT receptor subtype distinct from 5-HT₁ or 5-HT₇ receptors. On the other hand, the expression of 5-HT_1Hel and 5-HT_7Hel immunoreactivity in the cell body of cerebral ganglion neuron C1 only when it was regenerating in culture suggests that expression in the cell body may be more likely to occur under conditions of neuronal growth, such as during development or regeneration. This will be tested in future immunolocalization experiments on Helisoma embryos to confirm the expression in embryonic neuronal somata. In any event, the widespread neuritic expression of 5-HT_1Hel and 5-HT_7Hel receptors throughout the CNS supports the hypothesis that these molecules are critical to numerous physiological processes in the Helisoma CNS.