Starfish (A. rubens) were collected at low tide from the Thanet coast (Kent, UK) or were obtained from a fisherman based at Whitstable (Kent, UK), and then transported to Queen Mary University of London. The starfish were maintained in an aquarium with circulating artificial seawater at ~12°C and were fed with mussels (Mytilus edulis). In addition, juvenile specimens of A. rubens (diameter 0.5–1.5 cm) were collected at the University of Gothenburg Sven Lovén Centre for Marine Infrastructure (Kristineberg, Sweden) and were fixed in Bouin’s solution for immunohistochemical analysis (see below).

Transcripts encoding the ArGnRH precursor (ArGnRHP) and the ArCRZ precursor (ArCRZP) were first identified by analysis of A. rubens radial nerve cord transcriptome sequence data (36) and subsequently cDNAs comprising the complete open reading frame of ArGnRHP (accession number KT601712) and ArCRZP (accession number KT601713) have been cloned and sequenced (25). These cDNAs were used here for production of digoxigenin-labeled antisense and sense probes for ArGnRHP and ArCRZP, employing methods reported previously (39, 40). The templates for generation of the ArGnRHP and ArCRZP probes were 490 and 316 bases in length, respectively, and the probes were used without being hydrolyzed. Specimens of A. rubens were fixed, decalcified, and sectioned and then slide-mounted sections were processed for mRNA in situ hybridization using methods reported previously (39). To facilitate interpretation of staining, sections of A. rubens stained using Masson’s Trichrome method were also prepared, employing methods reported previously (41). Images of stained sections were captured with a QIClick™ CCD Color Camera (Qimaging) linked to a DMRA1 light microscope (Leica), using Volocity^® v.6.3.1 image analysis software (PerkinElmer) installed on an iMac (27-in., Late 2013 model with OXS Yosemite, Version 10.10). Adobe Photoshop CC (version 14.0, ×64) was used for removing dust, contrast adjustment, cropping images, and assembling images into montages.

To generate rabbit antisera to ArGnRH (pQIHYKNPGWGPG-NH₂) a C-terminal fragment of ArGnRH (KNPGWGPG-NH₂) was used as a peptide antigen. To generate rabbit antisera to ArCRZ (HNTFTMGGQNRWKAG-NH₂) two different peptides were tested as antigens. First, a peptide comprising the C-terminal region of ArCRZ (KYGQNRWKAG-NH₂), with an N-terminal lysine residue incorporated to provide two reactive amine groups for coupling to a carrier protein (thyroglobulin). A tyrosine residue was also incorporated at the N-terminus so that the peptide could, if needed, be radiolabeled with iodine-125 (e.g., for use in a radioimmunoassay). Second, an analog of the C-terminal region of ArCRZ (KYGQNRWRAG-NH₂) in which the lysine corresponding to the thirteenth residue of ArCRZ was replaced with arginine. The rationale for testing this second antigen peptide was a concern that chemical reaction of the free amine group in the lysine side chain with glutaraldehyde used for production of antigen peptide–carrier protein conjugates (see below) may affect antigenicity. The antigen peptides were synthesized by Peptide Protein Research Ltd. (Hampshire, UK) and were conjugated to the carrier protein thyroglobulin (Sigma-Aldrich) using 5% glutaraldehyde (v/v) in phosphate buffer (PB, 0.1 M sodium phosphate dibasic and 0.1 M sodium phosphate monobasic, pH = 7.2) as a coupling reagent. Peptide–thyroglobulin conjugates were then dialyzed using phosphate-buffered saline (PBS), aliquoted, and stored at −20°C.

Rabbit immunization and serum collection was performed by Charles River Labs (Margate, England, UK) according to the following protocol. On day 0, preimmune serum was collected and the first immunization (~100 nmol of conjugated antigen peptide in Freund’s complete adjuvant) was administered. Booster immunizations (~50 nmol of conjugated antigen peptide in Freund’s incomplete adjuvant) were administered on days 28, 42, and 56. Antiserum samples were collected on days 37 and 51 and a final bleed was collected on day 70.

To assess production of antibodies during the immunization protocol and following collection of a terminal bleed, antisera were tested for antibodies to the antigen peptides using enzyme-linked immunosorbent assay (ELISA). 100 µl of a 1 µM solution of antigen peptide dissolved in carbonate/bicarbonate buffer (25 mM anhydrous sodium carbonate, 25 mM sodium bicarbonate, pH 9.8) was pipetted into wells of a PVC microtiter plate (Starlab), which was then covered with Parafilm M^® (Starlab) and incubated overnight at 4°C. The following day the liquid contents of the plate were disposed of, the wells were rinsed with 200 µl PBS (3×, 10 min), and then the plate was drained on blotting paper. Then 200 µl of a blocking solution containing 5% goat antiserum/PBS was added to each well and left at room temperature for 2 h. The blocking buffer was discarded and each well was washed with PBS containing 0.1% Tween-20 (PBST; 200 µl; 3×, 10 min) and then the plate was drained, as above. Varying dilutions of preimmune serum or antiserum (10⁻³ to 10⁻⁸) diluted in 5% goat serum (Sigma-Aldrich)/PBS was added to each well and incubated overnight at 4°C. The antiserum was discarded and each well was washed with PBST (200 µl; 3×, 10 min) and then the plate was drained. Alkaline phosphatase-conjugated goat anti-rabbit IgG secondary antibodies (ThermoFisher Scientific; diluted 1:3,000 in 5% goat antiserum/PBST) were added to each well and incubated for 3 h at room temperature. After washing the plate with PBST (4×, 10 min), 100 µl p-nitrophenylphosphate alkaline phosphatase substrate (Vector Laboratories) made up in carbonate/bicarbonate buffer was added to each well and then after a 20 min incubation at room temperature absorbance at 415 nm was measured using FLUOstar Omega (BMG LABTECH). Mean absorbance values were calculated and plotted using Prism 6.0c.

Antibodies to the antigen peptide were affinity-purified using AminoLink^® Plus Immobilization Kit (Thermo Scientific). 1 mg of the ArGnRH/ArCRZ antigen peptides were dissolved in 2 ml of PBS and coupled to a column of beaded agarose, according to the manufacturer’s instruction. After washing with 5 ml of PBS (2×), 1 ml of ArGnRH antiserum or ArCRZ antiserum was added to the column and incubated at room temperature for 2 h on a rocking table. After washing with 5 ml of PBS (6×), bound antibodies were eluted using 7 ml of glycine elution buffer [6.3 ml of 100 mM glycine (VWR) and 0.7 ml of Tris (1 M, pH = 7.0)] and then 7 ml of triethylamine elution buffer [6.3 ml of 100 mM triethylamine (Sigma-Aldrich) and 0.7 ml of Tris (1 M, pH = 7.0)]. The glycine eluate and triethylamine eluate were then dialyzed separately in PBS at 4°C for at least 48 h. After addition of 0.1% sodium azide (VWR), the dialyzed antibodies were stored at 4°C.

Specimens of A. rubens (diameter 0.5–6 cm) were fixed in Bouin’s solution (75% saturated picric acid in seawater, 25% formalin, 5% acetic acid) for at least 2 days, and decalcified with 2% ascorbic acid/0.15 M sodium chloride solution at 4°C for at least 1 week. The arms and central disk regions of decalcified larger specimens of starfish (>2 cm diameter) were separated and processed separately, whereas smaller specimens were processed intact. Starfish tissue samples were embedded in paraffin, sectioned at 8 µm (RM 2145 microtome, Leica) and mounted on chrome alum/gelatin coated microscope slides. Paraffin was removed by immersion of slides in xylene (3× 10 min) and then slides were immersed in 100% ethanol (3× 10 min). Endogenous peroxidase activity was quenched by placing slides in 0.3% hydrogen peroxide (VWR Chemicals) in methanol for 30 min. The tissue sections were then rehydrated through a series of 10 min ethanol washes (90, 70, and 50%) and finally rinsed in distilled water. Slides were blocked in 5% goat serum (Sigma-Aldrich) made up in PBST (2 h at room temperature) and then were incubated overnight at 4°C in affinity-purified primary antibodies diluted 1:5 or 1:10 in 5% goat serum/PBST. Following a series of washes in PBST, peroxidase-conjugated AffiniPure Goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) were used as secondary antibodies (1:3,000 dilution, in 2% goat serum/PBST) and were added to each slide (500 μl/slide) and incubated for 2 h at room temperature. Immunostaining was visualized using staining buffer [0.05% diaminobenzidine (VWR Chemicals), 0.05% nickel chloride, 0.015% hydrogen peroxide in PBS]. Slides were then washed in autoclaved distilled water when staining was observed. Following dehydration through washes in autoclaved distilled water, 50% ethanol (1× 10 min), 70% ethanol (1× 10 min), 90% ethanol (1× 10 min), and 100% ethanol (2× 10 min), slides were washed in xylene (2× 10 min) before being mounted with coverslips using DPX (Thermo Scientific) as a mounting medium. Control experiments were performed using antibodies preabsorbed with ArGnRH antigen peptide (Peptide Protein Research Ltd., Hampshire, UK). Preabsorption was done by incubating affinity-purified antibodies with 200 µM antigen peptide for 1–2 h at room temperature. Then the preabsorbed antibodies were diluted to 1:5 or 1:10 in PBS and tested as above. Images of immunostained sections were captured with a QIClick™ CCD Color Camera (Qimaging) linked to a DMRA2 light microscope (Leica), using Volocity^® v.6.3.1 image analysis software (PerkinElmer) installed on an iMac (27-in., Late 2013 model with OS X Yosemite, Version 10.10). Adobe Photoshop CC (version 14.0, ×64) was used for removing dust, contrast adjustment, cropping images, and assembling images into montages.

The predicted structures of ArGnRH (pQIHYKNPGWGPG-NH₂) and ArCRZ (HNTFTMGGQNRWKAG-NH₂) have been confirmed by mass spectrometric analysis of A. rubens radial nerve cord extracts (25). Accordingly, ArGnRH and ArCRZ were custom synthesized (Peptide Protein Research Ltd., Hampshire, UK) to enable testing of the pharmacological effects of these neuropeptides. Both peptides were tested on in vitro preparations of cardiac stomach, apical muscle, and tube foot dissected from at least five different specimens of A. rubens. The cardiac stomach was dissected as described previously (32, 38) and then cotton ligatures were tied around the esophagus and the aboral side of the cardiac stomach. The apical muscle was dissected from the aboral body wall of the arms of A. rubens, cut into ~1 cm long segments and tied at each end with cotton ligatures, as described previously (35). The tube foot preparations were dissected from starfish arms by cutting out a square-shaped piece of ambulacral body wall and removing all of the tube feet except one at the center, with its ampulla intact. The external epithelium of the tube foot was scraped off using a blunt scalpel blade to facilitate penetration of peptides when tested in vitro (see below) and then cotton ligatures were tied around the ambulacral body wall and the tube foot sucker, as described previously (35). The dissected preparations were tied at one end to a fixed metal hook in a 20 ml aerated organ bath containing circulating artificial sea water at ~11°C. The other ligature was attached to a high grade Isotonic Transducer (ADinstruments MLT0015), which was connected via a bridge amplifier (FE221 Bridge Amp, ADInstruments Pty Ltd.) to data acquisition hardware (Power Lab 2/36, AD Instruments Pty Ltd.). Data were collected and analyzed using LabChart (v8.0.7) software installed on a laptop computer (Lenovo E540, Windows 7 Professional).

Once set up in the organ bath, preparations were incubated in artificial seawater, with several washes until a stable baseline length was achieved. Then synthetic peptides (ArGnRH or ArCRZ) were added to the organ bath to achieve concentrations ranging from 10⁻⁶ to 10⁻¹⁰ M. Known contractants of the preparations tested, NGFFYamide (for cardiac stomach, 10⁻⁷ M) (38) or acetylcholine (ACh; for apical muscle and tube foot preparations, 10⁻⁶ M), were tested to check the viability of preparations and to enable normalization of responses to ArGnRH or ArCRZ in different experiments. Thus, the effects of ArGnRH or ArCRZ on cardiac stomach preparations were normalized to the maximal effect observed with NGFFYamide (10⁻⁷ M) and the effects of ArGnRH or ArCRZ on apical muscle and tube foot preparations were normalized to the maximal effect observed with ACh (10⁻⁶ M). Dose–response curves were generated using the non-linear regression (curve fit) model in Prism 6 (GraphPad 6.0c). The maximal response (E_max) was expressed as the percentage (%) of the maximal contraction induced by each peptide compared to the maximal contraction induced by 10⁻⁷ M of NGFFYamide (for cardiac stomach) or 10⁻⁶ M of ACh (for apical muscle and tube foot preparations). Statistical analysis was performed using t-tests when carrying out pairwise comparisons between peptides tested at the same concentration.

Twenty specimens of A. rubens, which had been withheld from a food supply for 1 week, were used to investigate if ArGnRH and ArCRZ can induce cardiac stomach retraction in vivo. As reported previously (38), cardiac stomach eversion was induced by placing starfish in a glass tank containing 2% magnesium chloride (MgCl₂) dissolved in seawater, which acts as a muscle relaxant in marine invertebrates (42). Then 10 µl of 1 mM ArGnRH or ArCRZ peptide was injected into the perivisceral celom at two sites in the aboral body wall of the arms proximal to their junctions with the central disk, as reported previously (38). Starfish injected with ArGnRH or ArCRZ were video recorded for 5 min. Then as a positive control, NGFFYamide (10 µl of 100 nM), which is a potent cardiac stomach contractant in vitro and in vivo (38), was injected into the perivisceral celom (at two sites in the aboral body wall different to those used for ArGnRH or ArCRZ) and the starfish were video recorded for a further 5 min.

Detailed descriptions of the anatomy of the starfish A. rubens (Figures 1A,B) and other starfish species can be found in textbooks (43), reviews (44) and research articles (41, 45–48). However, to aid interpretation of images shown in Figures 2–6 (mRNA in situ hybridization) and Figures 7–12 (immunohistochemistry), in Figure 1 we show images of trichrome-stained sections of the arm (Figure 1C) and central disk (Figure 1D) in A. rubens, which were prepared using methods reported previously (41).

The main components of the nervous system in A. rubens are the radial nerve cords that extend along the oral (lower) side of each arm, with two rows of tube feet on either side of the radial nerve cord (Figure 1C). The radial nerve cords are linked by the circumoral nerve ring in the central disk (Figure 1D). Running parallel with the radial nerve cords are smaller marginal nerves, which are located lateral to the outer row of tube feet on each side of the arm (Figure 1C).

The digestive system of A. rubens comprises a mouth located on the underside of the central disk and surrounded by a contractile peristomial membrane, which is continuous aborally with a short tubular esophagus (Figure 1D). The stomach comprises two compartments: (1) the large and highly folded cardiac stomach, which is everted through the mouth during feeding (Figure 1D). (2) the smaller pyloric stomach (Figure 1D), which is linked via a short rectum to the anus located on the aboral surface of the central disk. Paired digestive glands or pyloric caeca are located in each arm (Figure 1C) and these are connected to the pyloric stomach by pyloric ducts (Figure 1D).

The body wall skeleton of A. rubens comprises calcite ossicles that are interconnected by muscles and collagenous tissue. The external surface of the body wall has a variety of appendages that include protective pincer-like pedicellariae and spines as well as thin-walled gas exchange organs known as papulae. A specialized tube foot-like organ located at the tips of each arm is the terminal tentacle, a sensory organ with a photosensitive and pigmented optic cushion located at its base. The internal surface of the body wall is lined by a coelomic epithelium, which is underlain by longitudinally and circularly orientated muscle layers. Along the midline of each arm the longitudinal muscle layer of the coelomic lining is thickened to form the apical muscle (Figure 1A), which facilitates flexion of the arm.

Experiments testing antisense probes on sections of arms from A. rubens revealed stained cells in the radial nerve cords (Figure 2A). Importantly, the specificity of staining with antisense probes was demonstrated by control experiments with sense probes, where no staining was observed (Figure 2A inset). Analysis of staining in the radial nerve cords revealed that ArGnRHP is expressed by cells in both the ectoneural and hyponeural regions. In the ectoneural region, stained cells are sparsely distributed throughout the subcuticular epithelial layer of the radial nerve cords, as seen in both transverse (Figure 2B) and parasagittal longitudinal (Figures 2C,D) sections of arms. In the hyponeural region, there are bilaterally symmetrical clusters of strongly stained cells (Figure 2B) and analysis of parasagittal longitudinal sections of arms revealed that these stained cell clusters are distributed along the length of the radial nerve cord, interrupted by regions without stained cells. Consistent with the pattern of expression in the radial nerve cords, ArGnRHP-expressing cells were revealed in both the ectoneural and hyponeural regions of the circumoral nerve ring (Figure 2E). Anatomically, the circumoral nerve ring corresponds to one half of the V-shaped radial nerve cord and accordingly only single clusters of stained cells are observed in the hyponeural region in transverse sections of the circumoral nerve ring. In the ectoneural region of the circumoral nerve ring, stained cells located in the subcuticular epithelial layer are largely grouped in a distinct band that is located approximately mid-way between the junction of the circumoral nerve ring with the peristomial membrane and the junction with perioral appendages (Figure 2E). ArGnRHP-expressing cells are also present in the marginal nerves (Figure 2F).

In the tube feet, ArGnRHP-expressing cells are located in a subepithelial position along the length of the tube foot stem (Figure 3A) and associated with the basal nerve ring (Figure 3B). At the tips of each arm ArGnRHP-expressing cells are present in the optic cushion (Figure 3C), the terminal tentacle (Figure 3D), flaps of tissue known as lateral lappets located adjacent to the terminal tentacle (Figure 3D), and the body wall epithelium surrounding the terminal tentacle (Figure 3D).

A very sparse population of cells expressing ArGnRHP is present in both the cardiac stomach (Figures 3E,F) and pyloric stomach (Figure 3G), with elongate stained cells located in the mucosal layer (Figure 3E) and roundish stained cells located in close association with the basiepithelial nerve plexus (Figure 3F). No ArGnRHP-expressing cells were observed in the pyloric ducts and the pyloric caeca, but stained cells were present in mesenteries that attach the pyloric caeca to the aboral body wall (Figure 3H).

Experiments testing antisense probes on sections of arms from A. rubens revealed stained cells in the radial nerve cords (Figure 4A) and the specificity of staining with antisense probes was demonstrated by control experiments with sense probes, where no staining was observed (Figure 4A inset). Analysis of staining in the radial nerve cord revealed that ArCRZP is expressed by cells in the ectoneural region but not in the hyponeural region. The ArCRZP-expressing cells in the ectoneural region were sparsely distributed throughout the subcuticular epithelial layer both transversely (Figures 4A,B) and longitudinally (Figure 4C). Consistent with this pattern of expression in the radial nerve cords, ArCRZP-expressing cells were likewise observed only in the ectoneural region of the circumoral nerve ring. These cells were mostly distributed sparsely within the subcuticular epithelial layer (Figure 4D), but some cell clusters were also observed (Figure 4E).

Proximal to the outer rows of tube feet, ArCRZP-expressing cells were detected in the marginal nerves (Figure 5A). In the tube feet, ArCRZP-expressing cells were observed close to the junction between adjacent podia (Figure 5B). Stained cells were also sparsely distributed in a subepithelial location along the length of the tube foot stem and closely associated with the basal nerve ring (Figure 5C). Stained cells were observed in the external epithelium of the body wall and these cells were located at the junction between adjacent spines (Figure 5D). Stained cells were also associated with the apical muscle, with stained cells most prominent on the oral side of the muscle just beneath the coelomic epithelium. However, stained cells were also evident on the aboral side of the muscle close to the circular muscle layer (Figure 5E).

In the digestive system, ArCRZP expression was detected in the cardiac stomach, pyloric stomach and pyloric ducts (Figure 6). An extensive population of strongly stained cells was revealed in the pyloric stomach and the aboral part of the cardiac stomach and this was the most prominent expression of ArCRZP in the entire animal (Figure 6A). Stained cells in the cardiac stomach (Figure 6B) and the pyloric stomach (Figure 6C) included cells with an elongate shape located in the mucosal layer and roundish-shaped cells located close to the position of the basiepithelial nerve plexus. A similar pattern of expression was observed in the pyloric duct, but stained cells were only observed on the oral (lower) side of the pyloric duct (Figures 6D–F).

Analysis of the in vitro pharmacological effects of ArGnRH and ArCRZ revealed that both peptides cause contraction of cardiac stomach, tube foot, and apical muscle preparations, but with differing potency/efficacy.

ArGnRH was more potent and effective than ArCRZ as a contractant of cardiac stomach preparations (Figures 13A,B). Thus, ArGnRH caused dose-dependent contraction of cardiac stomach preparations at concentrations between 10⁻⁶ and 10⁻⁸ M (Figure 13B). The E_max for ArGnRH at a concentration of 10⁻⁶ M was 71 ± 6.7% in comparison with the effect of NGFFYamide at 10⁻⁷ M (defined as 100% contraction). ArCRZ caused contraction of cardiac stomach preparations at 10⁻⁶ and 10⁻⁷ M and the E_max for ArCRZ at 10⁻⁶ M was only 43.7 ± 5.9% when compared with NGFFYamide (10⁻⁷ M). Accordingly, comparison of the effects of ArGnRH and ArCRZ when tested at 10⁻⁶ and 10⁻⁷ M revealed that the effect of ArGnRH was significantly larger than the effect of ArCRZ, supported by p values (Student’s t-test) of 0.012 and 0.002, respectively.

ArCRZ was more potent and effective than ArGnRH as a contractant of apical muscle preparations (Figures 13C,D). Thus, ArCRZ induced dose-dependent contraction of apical muscle preparations at concentrations between 10⁻⁶ and 10⁻⁹ M (Figure 13D). The E_max for ArCRZ at a concentration of 10⁻⁶ M was 112.4 ± 10% in comparison with the effect of ACh at 10⁻⁶ M (defined as 100% contraction) (Figure 13D). ArGnRH caused contraction of apical muscle preparations at 10⁻⁶ and 10⁻⁷ M but the E_max for ArGnRH at 10⁻⁶ M was only 26.3 ± 4% when compared with ACh (10⁻⁶ M). Accordingly, comparison of the effects of ArCRZ and ArGnRH when tested at 10⁻⁶ and 10⁻⁷ M revealed that the effect of ArCRZ was significantly larger than the effect of ArGnRH, supported by p values (Student’s t-test) of 3.943 × 10⁻⁵ and 6.000 × 10⁻⁷, respectively.

Both ArGnRH and ArCRZ caused contraction of tube foot preparations at 10⁻⁶ M (Figures 13E,F), but effects were not observed when tested at lower concentrations (10⁻⁷ to 10⁻¹⁰ M). In comparison with the effect of ACh at 10⁻⁶ M, the E_max for ArGnRH was 106.5 ± 11.21% whereas the E_max for ArCRZ was 83.7 ± 19.5% (Figure 13F). However, statistical analysis revealed that there was no significant difference in the effects of ArGnRH and ArCRZ at 10⁻⁶ M (p = 0.328; Student’s t-test).

In 10 specimens of A. rubens that were injected with ArGnRH, no cardiac stomach retraction was observed in the 5 min observation period, whereas NGFFYamide induced cardiac stomach retraction in all 10 of these animals (but with variability in the rate and extent of the retraction; data not shown). Likewise, cardiac stomach retraction was not observed in any of the 10 animals injected with ArCRZ, whereas NGFFYamide induced cardiac stomach retraction in all 10 of these animals (but with variability in the rate and extent of the retraction; data not show).

As highlighted in the Section “Introduction,” the discovery of ArGnRH and ArCRZ as respective ligands for GnRH-type and CRZ-type receptors in the starfish A. rubens provided an important new insight into the evolution of GnRH/CRZ-type neuropeptide signaling (25). It demonstrated that the evolutionary origin of these paralogous neuropeptide signaling pathways can be traced back to the common ancestor of protostomes and deuterostomes, but with subsequent loss in some lineages (e.g., loss of CRZ signaling in vertebrates, nematodes, and urochordates) and duplication of the GnRH-signaling system in arthropods to give rise to the paralogous AKH-type and ACP-type signaling pathways. Informed by these findings, we have reviewed published research on GnRH/CRZ-type neuropeptide signaling and have proposed a standardized nomenclature wherein peptides are either named “GnRH” or “corazonin,” with the exception of the paralogous “AKH/RPCH”-type and “ACP”-type neuropeptides that arose by gene duplication in the arthropod lineage (26). Reviewing published literature in this context reveals that some neuropeptides hitherto referred to as GnRH or invertebrate-GnRH are in fact ligands for CRZ-type receptors and therefore we have proposed that these should instead be referred to as CRZs. Furthermore, GnRH-type peptides in non-arthropodan protostomian invertebrates have often been named “AKH,” whereas we have proposed that a more suitable name for these peptides would be GnRH because AKH is a unique product of gene duplication in the arthropod lineage (26). With these nomenclature issues in mind, here we present a selective review of what is known about the physiological roles of GnRH/CRZ-type neuropeptides in bilaterian animals so that the findings reported here from A. rubens can be considered in a broader comparative and evolutionary context.

As the prototype of the GnRH/CRZ neuropeptide family, the physiological role of GnRH in mammals and other vertebrates as regulator of pituitary release of the gonadotropic hormones luteinizing hormone and follicle stimulating hormone may have influenced expectations of the physiological roles of homologous peptides in other taxa. Interestingly, evidence of roles of GnRH/CRZ-type neuropeptides in regulation of reproductive physiology have been obtained in some invertebrates, including the urochordate C. intestinalis (6) and several molluscan species (68). In this context it is interesting that an ortholog of ArGnRH was found not to induce release of the gonadotropin relaxin-like gonad-stimulating peptide (RGP) in the starfish Patiria pectinifera when tested in vitro (69). Furthermore, we have observed that ArGnRH does not induce spawning of gonadal fragments from A. rubens in vitro (M. Mita, S. Tian, M.R. Elphick, unpublished data) when tested in parallel with RGP, which does induce spawning (39). Therefore, based on the data currently available it appears that GnRH-type neuropeptides may not be involved in regulation of reproductive physiology in starfish.

Much of what we know about the physiological roles of GnRH/CRZ-type neuropeptides in invertebrates is based on findings from insects and to a lesser extent other arthropods. The first GnRH-type peptide to be identified in an invertebrate was the crustacean hormone RPCH (13), which causes alterations in body coloration due to pigment migration in chromatophores (13, 70). Subsequently, an RPCH-like peptide that triggers lipid mobilization was identified in the locust Schistocerca gregaria, and hence it was named AKH (11, 12, 14). However, more recent studies have revealed that AKH also has a variety of other physiological roles in insects (26). Less is known about the physiological roles of ACP, the paralog of AKH in arthropods, and this in part due to the fact that it was discovered more recently and because ACP signaling has been lost in the principal arthropod model system—Drosophila. Nevertheless, the data that are available are indicative of physiological roles distinct from those of AKH (71, 72).

Corazonin was first identified in the cockroach Periplaneta americana on account of its cardioacceleratory effect (15) and then from locusts using an immunoassay (73). Interestingly, CRZ was also identified independently as a peptide that triggers dark pigmentation in locusts (74). Furthermore, as with AKH, more recent studies have revealed that CRZ also has a variety of other physiological roles in insects (26).

Comparison of the physiological roles of AKH/ACP-type and CRZ-type signaling in arthropods reveals some similarities, although this varies between different lineages. For example, both AKH and CRZ affect responses to stress in Drosophila (75–78), while in the crayfish Procambarus clarkii both CRZ and RPCH affect pigment migration (79, 80), Presumably these and other similarities in the physiological roles of AKH/ACP and CRZ signaling in arthropods (26) are a reflection of their common evolutionary origin, while differences may reflect neofunctionalization or subfunctionalization.

A similar mix of similarities and differences in the physiological roles of GnRH/CRZ-type signaling have been revealed by an elegant series of experimental studies on the marine mollusk Aplysia californica. However, interpretation of the published data is complicated by nomenclature issues. Two GnRH/CRZ-type neuropeptide signaling systems have been identified in Aplysia, one referred to as GnRH and the other referred to as AKH (81–83). However, Aplysia “GnRH” is actually the ligand for a CRZ-type receptor (83) and hence we have proposed that this peptide should be named Aplysia californica corazonin (AcCRZ), and Aplysia “AKH” should be renamed Aplysia californica GnRH (AcGnRH) (26). Employing use of the proposed revised nomenclature, below we summarize what is known about the physiological roles of GnRH/CRZ-type neuropeptides in Aplysia.

Neurons expressing the AcGnRH precursor are located in the abdominal, cerebral, and pleural ganglia, but the processes of these neurons are present in all ganglia, indicating potential roles in regulation of a variety of physiological processes. Injection of AcGnRH inhibits feeding, reduces body mass, increases defecation, and reduces gonadal mass and oocyte diameter (82). Analysis of the expression of the AcCRZP revealed expression in the pedal, cerebral and abdominal ganglia (84). Furthermore, in vivo pharmacological experiments revealed that AcCRZ triggers parapodial opening, inhibition of feeding, and promotion of substrate attachment but has no effect on oocyte growth, ovotestis mass, reproductive tract mass, egg-laying, egg-laying hormone accumulation and secretion, and penile eversion (81). Thus, there are distinct differences in the physiological roles of GnRH-type and CRZ-type neuropeptides in Aplysia but there are also some similarities (e.g., inhibitory effect on feeding).

In conclusion, the discovery of similarities and differences in the functions of the GnRH-type and CRZ-type signaling systems in the starfish A. rubens, as reported here, is consistent with findings from other animals. However, clearly there is still much to be learnt about the physiological roles of the GnRH/CRZ-type neuropeptide signaling systems in starfish. The functional characterization of neuropeptides is dependent on the physiological and/or behavioral assays employed to investigate their roles. Here, we employed bioassays for muscle contractility, revealing that ArGnRH and ArCRZ are both myostimulatory peptides. This finding, combined with anatomical analysis of their expression patterns, represents a significant first advance in characterizing GnRH/CRZ-type neuropeptide signaling in echinoderms. In the future use of other bioassays may reveal additional functions of GnRH/CRZ-type neuropeptides in A. rubens. Furthermore, the findings reported here provide a basis for analysis of GnRH/CRZ-type neuropeptide signaling in other starfish species; for example, in the starfish species P. pectinifera and Acanthaster planci (the crown-of-thorns starfish), which are established or emerging model systems for neuropeptide research (69, 85–87). Looking beyond the class Asteroidea, orthologs of ArGnRH and ArCRZ have been identified in other echinoderms, including the sea urchin Strongylocentrotus purpuratus (88) and the brittle star Ophionotus victoriae (89) and therefore it will be of interest to investigate the expression and actions of GnRH/CRZ-type neuropeptides in these and/or other echinoderm species. The five extant echinoderm classes, the Asteroidea, Ophiuroidea, Echinoidea, Holothuroidea, and Crinoidea diverged during the Cambrian and Ordovician eras in a period estimated to be between 475 and 509 million years ago (90). Over the long period of evolutionary time that has elapsed since, there may have been clade-specific specialization in GnRH/CRZ-type neuropeptide function and therefore, a broad physiological/behavioral perspective may be necessary in selecting assays that reveal the actions of GnRH/CRZ-type neuropeptides in echinoderms.