Functional characterization of a second pedal peptide/orcokinin‐type neuropeptide signaling system in the starfish Asterias rubens

Abstract Molluscan pedal peptides (PPs) and arthropod orcokinins (OKs) are prototypes of a family of neuropeptides that have been identified in several phyla. Recently, starfish myorelaxant peptide (SMP) was identified as a PP/OK‐type neuropeptide in the starfish Patiria pectinifera (phylum Echinodermata). Furthermore, analysis of transcriptome sequence data from the starfish Asterias rubens revealed two PP/OK‐type precursors: an SMP‐type precursor (A. rubens PP‐like neuropeptide precursor 1; ArPPLNP1) and a second precursor (ArPPLNP2). We reported previously a detailed analysis of ArPPLNP1 expression in A. rubens and here we report the first functional characterization ArPPLNP2‐derived neuropeptides. Sequencing of a cDNA encoding ArPPLNP2 revealed that it comprises eleven related neuropeptides (ArPPLN2a‐k), the structures of several of which were confirmed using mass spectrometry. Analysis of the expression of ArPPLNP2 and neuropeptides derived from this precursor using mRNA in situ hybridization and immunohistochemistry revealed a widespread distribution, including expression in radial nerve cords, circumoral nerve ring, digestive system, tube feet and innervation of interossicular muscles. In vitro pharmacology revealed that the ArPPLNP2‐derived neuropeptide ArPPLN2h has no effect on the contractility of tube feet or the body wall‐associated apical muscle, contrasting with the relaxing effect of ArPPLN1b (ArSMP) on these preparations. ArPPLN2h does, however, cause dose‐dependent relaxation of cardiac stomach preparations, with greater potency/efficacy than ArPPLN1b and with similar potency/efficacy to the SALMFamide neuropeptide S2. In conclusion, there are similarities in the expression patterns of ArPPLNP1 and ArPPLNP2 but our data also indicate specialization in the roles of neuropeptides derived from these two PP/OK‐type precursors in starfish.

has revealed the occurrence of multiple precursor proteins that each comprise "cocktails" of structurally related peptides (Moroz et al., 2006;Veenstra, 2010Veenstra, , 2011. Investigation of the effects of PPs in mollusks has revealed that these neuropeptides regulate foot muscle activity and increase the ciliary beat frequency of cells in the foot epithelium, actions that are indicative of physiological roles in the control of locomotor activity (Araki, Liu, Zhang, Takeuchi, & Munekata, 1995;Hall & Lloyd, 1990;Malyshev, Norekian, & Willows, 1999;Willows, Pavlova, & Phillips, 1997).
OK was identified in the crayfish Orconectes limosus as a neuropeptide that enhances the frequency and amplitude of spontaneous contractions of the hind-gut (Stangier, Hilbich, Burdzik, & Keller, 1992). Subsequent studies have revealed that OKs also alter the rhythmic motor output of the stomatogastric nervous system in crustaceans (Li et al., 2002;Skiebe, Dreger, Meseke, Evers, & Hucho, 2002). Furthermore, experimental studies on insects have revealed that OKs cause a phase-dependent shift in circadian locomotor activity in cockroaches (Hofer & Homberg, 2006), regulate "awakening" behavior in the beetle Tribolium castaneum (Jiang, Kim, & Park, 2015) and regulate ecdysis in the kissing bug Rhodnius prolixus (Wulff et al., 2017). Thus, PP/OK-type neuropeptides have a variety of actions in mollusks and arthropods but with stimulatory actions appearing to be a common theme.
Discovery of two genes in the sea urchin Strongylocentrotus purpuratus (phylum Echinodermata) that encode precursors of PP-like neuropeptides (SpPPLNP1, SpPPLNP2) provided the first evidence that PP/OK-type peptides also occur in deuterostomes (Rowe & Elphick, 2012). However, nothing is known about the physiological roles of PP/OK-type neuropeptides in sea urchins. Recently, a neuropeptide isolated from the starfish Patiria pectinifera that acts as a muscle relaxant ("starfish myorelaxant peptide" or SMP) was identified as a PP/OK-type neuropeptide (Kim et al., 2016). Cloning and sequencing of a cDNA encoding the SMP precursor revealed that it comprises 12 copies of SMP and 7 copies of three other SMP-like peptides. Analysis of the distribution of SMP precursor transcripts in P. pectinifera using quantitative PCR revealed a widespread pattern of expression. Furthermore, in vitro pharmacological experiments revealed that SMP causes relaxation of three preparations from P. pectinifera-the apical muscle, tube feet, and cardiac stomach. Collectively, these data indicated that SMP may have a general role as a muscle relaxant in starfish (Kim et al., 2016).
A homolog of the P. pectinifera SMP precursor comprising 8 copies of SMP-like PP/OK-type peptides has been identified in the common European starfish Asterias rubens (Kim et al., 2016). Furthermore, a partial sequence of a second PP/OK-type precursor has also been identified in A. rubens (Semmens et al., 2016). Therefore, we have named the A. rubens SMP-type precursor A. rubens PP-like neuropeptide precursor 1 (ArPPLNP1) and we have named the other A. rubens PP/OK-type precursor A. rubens PP-like neuropeptide precursor 2 (ArPPLNP2).
Recently, we reported a detailed analysis of the anatomy of ArPPLNP1 expression in A. rubens using mRNA in situ hybridization and immunohistochemistry . Consistent with qPCR data from P. pectinifera (Kim et al., 2016), a widespread pattern of expression of ArPPLNP1 and the neuropeptides derived from this precursor was revealed in A. rubens. Thus, cells expressing ArPPLNP1 transcripts were detected, for example, in the radial nerve cords, circumoral nerve ring, esophagus, cardiac stomach, pyloric stomach, pyloric caeca, tube feet, and body wall . Immunohistochemical analysis using antibodies to one of the neuropeptides derived from ArPPLNP1 (ArPPLN1b or ArSMP) revealed a pattern of immunostained cells consistent with data obtained using mRNA in situ hybridization. Furthermore, immunohistochemistry revealed extensive networks of immunostained processes in the radial nerve cords, circumoral nerve ring, digestive system, tube feet, coelomic lining, apical muscle, and body wall . One of the most striking features of the immunostaining was the presence of labeled processes in close association with interossicular muscles that link the calcite ossicles of the endoskeleton. Consistent with this pattern of expression, intense immunostaining was also observed in the lateral motor nerves, components of the starfish nervous system that were first described based on analysis of histochemical staining (Smith, 1937). Investigation of the in vitro pharmacological effects of ArPPLN1b revealed that it causes dose-dependent relaxation of three preparations from A. rubens-apical muscle, tube feet and cardiac stomach , consistent with previous findings from P. pectinifera (Kim et al., 2016). Thus, collectively the data obtained from experimental studies on SMP/PPLN1-type neuropeptides in P. pectinifera and A. rubens indicate that a physiological role of these peptides is to act as inhibitory neuromuscular transmitters or modulators in starfish.
The objective of this study was to investigate the expression pattern and pharmacological actions of neuropeptides derived from the second PP/OK-type neuropeptide precursor in A. rubens, ArPPLNP2.

| Animals
Starfish (A. rubens) with a diameter >4 cm were collected at low tide from the Thanet coast (Kent, UK) or were obtained from a fisherman based at Whitstable (Kent, UK). These animals were maintained in a circulating seawater aquarium at 128C in the School of Biological and Chemical Sciences at Queen Mary University of London and were fed on mussels (Mytilus edulis). Smaller juvenile specimens of A. rubens (diameter 0.5-1.5 cm) were collected at the University of Gothenberg Sven Lov en Centre for Marine Infrastructure (Kristenberg, Sweden) and fixed in Bouin's solution.

| cDNA cloning and sequence analysis
Analysis of A. rubens radial nerve cord transcriptome sequence data has enabled identification of a contig encoding the partial sequence (325 residues) of PP-like neuropeptide precursor 2 (ArPPLNP2), which comprises a 31-residue N-terminal signal peptide and seven putative neuropeptides bounded by dibasic cleavage sites (Semmens et al., 2016;GenBank Accession number KT601720). To facilitate cloning and sequencing of the complete open reading frame (ORF) of ArPPLNP2, ovarian transcriptome sequence data obtained from other starfish species were analyzed (Reich, Dunn, Akasaka, & Wessel, 2015; http://www.echinobase.org/ Echinobase/Blasts) and a contig (GAUS01023767.1) comprising the 3 0 region of PPLNP2 was identified in Asterias forbesi. Combining the partial PPLNP2 transcript sequences from A. rubens and A. forbesi, forward and reverse primers (5 0 -GAGACATCGAGGGTGGTTTTG-3 0 , 5 0 -GGCCCGG CCTAAGAATCAT-3 0 ) were designed to enable PCR amplification of a full-length ArPPLNP2 cDNA from A. rubens.
The methods used for RNA extraction, cloning and sequencing were as described previously (Lin, Mita et al., 2017), but with minor modifications. Zero Blunt TOPO PCR cloning kit (Invitrogen, Carlsbad, CA) was used to ligate the PCR product into the pCR-Blunt II with TOPO vector for sequencing of a ArPPLNP2 cDNA comprising the entire ORF. T-easy vector (Promega) was employed for cloning a cDNA comprising a partial ORF of ArPPLNP2, using the following primers: 5 0 -CTATTCTGTCTGGCTCTT-3 0 (forward) and 5 0 -GACAGCTTCTTCTCT CTTA-3 0 (reverse).

| Mass spectrometric identification of peptides derived from ArPPLNP2 in extracts of A. rubens radial nerves
The structures of neuropeptides derived from ArPPLNP2 were determined by MS/MS analysis of A. rubens radial nerve cords extracted in 90% methanol/9% acetic acid, employing use of methods described previously Lin, Mita et al., 2017).

| Localization of ArPPLNP2 expression in A. rubens using mRNA in situ hybridization
A cDNA comprising a partial ORF of ArPPLNP2 was used as a template for synthesis of probes for mRNA in situ hybridization. The methods employed for (a) production of anti-sense and sense DIGlabeled RNA probes, (b) preparation of sections of fixed specimens (diameter 4-6 cm) of A. rubens and (c) visualization of ArPPLNP2 transcripts in sections of A. rubens were the same as reported previously in a study investigating expression of the relaxin-like gonad-stimulating peptide precursor in A. rubens (AruRGPP) (Lin, Mita et al., 2017).

| Production and characterization of rabbit antibodies to ArPPLNP2-derived neuropeptides
The peptide ArPPLN2h (GRTSLSGSSGLTHLSSGFH) was selected as a target antigen for production of rabbit antibodies because it is the most abundant of the ArPPLN2-derived neuropeptides, occurring in triplicate whilst other peptides occur singly in the precursor. Furthermore, ArPPLN2h shares between 58 and 95% sequence identity with the other ArPPLNP2-derived neuropeptides. A peptide comprising the C-terminal octapeptide of ArPPLN2h but with the addition of N-terminal lysine and tyrosine residues (KYTHLSSGFH) was designed as an antigen (ag) peptide (ArPPLN2h-ag) and synthesized by Peptide Protein Research Ltd. (Hampshire, UK). Rabbit immunization with the ArPPLN2h-ag-thyroglobulin conjugate and serum collection was performed by Charles River Labs (Margate, UK), as described previously for ArPPLN1b .
To assess production of antibodies during the immunization protocol and following collection of a terminal bleed, antisera were tested for antibodies to ArPPLN2h-ag using Enzyme-Linked ImmunoSorbent Assays (ELISA), employing the same methods as described previously for ArPPLN1b .

| Localization of ArPPLN2h in A. rubens using immunohistochemistry
The methods employed for (a) preparation of sections of fixed specimens (diameter 4-6 cm) of A. rubens, (b) visualization of ArPPLN2h expression in sections of A. rubens using immunohistochemistry (including pre-absorption tests) and (c) photography of immunostained sections and preparation of montages were the same as those reported previously for ArPPLN1b .

| In vitro and in vivo pharmacology
Consistent with the selection of ArPPLN2h as a representative neuropeptide for generation of antibodies (see above), ArPPLN2h was also selected as a representative ArPPLNP2-derived peptide for in vitro and in vivo pharmacological experiments. ArPPLN2h was synthesized by Peptide Protein Research Ltd. (Hampshire, UK) and was tested on in vitro preparations of the cardiac stomach, apical muscle and tube foot from A. rubens, using the same methods reported previously for ArPPLNP1b . The SALMFamide neuropeptide S2 (SGPYSFNSGLTF-NH 2 ), which is a known relaxant of the cardiac stomach, apical muscle, and tube feet in A. rubens (Melarange & Elphick, 2003;Melarange, Potton, Thorndyke, & Elphick, 1999), was tested in parallel with ArPPLN2h both as a positive control and to enable comparative analysis of potency/efficacy. A previous study demonstrated that S2 causes eversion of the cardiac stomach in A. rubens when injected in vivo (Melarange et al., 1999). Therefore, here the effect of in vivo injection of ArPPLN2h in A. rubens was investigated by intracoelomic injection of starfish (n 5 10) with 100 mL of 1 mM ArPPLN2h.

| Cloning and sequencing of a cDNA encoding ArPPLNP2
A cDNA encoding ArPPLNP2 was cloned and sequenced (GenBank accession number KT601720). The ArPPLNP2 ORF encodes a 563 amino acid residue protein comprising a 31-residue signal peptide and 13 copies of predicted PP/OK-type peptides, which we have named located N-terminally and C-terminally, respectively. Furthermore, comparison of ArPPLN2h, as a representative peptide, with other PP/OK-type peptides revealed sequence conservation at several positions, but most notably three hydrophobic residues: Leu 5 located in the N-terminal region and Leu 14 and Phe 18 located in the C-terminal region (Figure 1c).
FIG URE 1 Sequence of ArPPLNP2 and alignments of predicted ArPPLNP2-derived peptides. (a) Sequence of ArPPLNP2 with the predicted signal peptide marked in blue, predicted cleavage sites marked in green, and peptides ArPPLN2 a to k marked in red. (b) Alignment of ArPPLN2h, which occurs in triplicate in ArPPLNP2, with other peptides derived from ArPPLNP2, which occur singly in ArPPLNP2, as shown in (a). All of the peptides comprise nineteen residues, and residues that are identical to the corresponding residue in ArPPLN2h are shown with a black dot. Eight of the residues in ArPPLN2h are conserved in the ten other peptides (highlighted in yellow). (c) Manual alignment of ArPPLN2h, as a representative ArPPLNP2-derived peptide, with ArPPLN1b and with PP/OK-type peptides from other species. Hydrophobic residues (Phe, Leu or Met) are conserved at three positions (highlighted in yellow), one located in the N-terminal region and two located in the C-terminal region. Other residues that are conserved between peptides from at least one deuterostome (echinoderm) species and at least one protostome species are highlighted in gray. Citations and/or accession numbers for the peptide sequences included here are as follows: ArPPLN2h (this article; KT601720); ArPPLN1b (Kim et al., 2016;KT870153); PpeSMPa (Kim et al., 2016;KT870152); SpPPLN1d (Rowe & Elphick, 2012; XP_785647); SpPPLN2h (Rowe & Elphick, 2012; XP_003727926); AjPPLN2a (Rowe et al., 2014;Isotig 17873); PP1-peptide (Lloyd & Connolly, 1989;Moroz et al., 2006;NP_001191585); PP3-peptide (Moroz et al., 2006;NP_001191625); NLP14-peptide (Nathoo, Moeller, Westlund, & Hart, 2001; NP_001257067); NLP15 peptide (Nathoo et al., 2001;T20275); OK (Stangier et al., 1992;P37086) 3.2 | Mass spectrometric identification of ArPPLNP2derived peptides in A. rubens radial nerve cords extracts Ten of the eleven predicted peptides derived from ArPPLNP2 ( Figure   1a,b) were detected by MS/MS in an acidified methanol extract of radial nerve cord without additional protease (trypsin) treatment. All precursor ions were detected with less than 2 ppm mass deviation from predicted sequences. ArPPLN2a, e, f, h, i, j were detected with high quality MS/ MS data ( Figure 2). ArPPLN2b, d, k were detected with poor fragmentation in the MS/MS data, only partial coverage of ArPPLN2g was observed and ArPPLN2c was not detected at all (data not shown). The cardiac stomach is attached to the ambulacrum in each of the five arms of A. rubens by extrinsic retractor strands. The paired extrinsic retractor strands coalesce to form a nodule that connects the extrinsic retractor strands with the intrinsic retractor strands located within the wall of the cardiac stomach (Anderson, 1954). Large numbers of ArPPLNP2-expressing cells were observed in the coelomic lining of the nodule and in the ceolomic lining of extrinsic retractor strands and mesenteries that are attached to the ambulacral ossicles (Figure 5a 3.4 | Characterization of a rabbit antiserum to ArPPLN2h using ELISA ELISA analysis revealed that ArPPLN2h-ag immunoreactivity was detected with antiserum at dilutions ranging from 10 23 to 10 26 and no cross-reactivity with ArPPLN1b. Furthermore, with an antiserum dilution of 10 24 the ArPPLN2h-ag peptide was detected in the range from 10 210 to 10 212 moles, indicative of a high titer of specific antibodies to the ArPPLN2h-ag peptide. The rabbit antiserum to ArPPLN2h has been assigned the RRID AB_2716765.

| Localization of ArPPLNP2 transcripts in
3.5 | Immunohistochemical localization of neuropeptides derived from ArPPLN2 in A. rubens

| Tube feet and terminal tentacle
Immunostaining is present throughout the sub-epithelial nerve plexus of tube foot stems (Figure 7a,b) and in the basal nerve ring located in inset of (a) shows absence of immunostaining in a radial nerve cord section incubated with ArPPLN1b antiserum pre-absorbed with the antigen peptide (ArPPLN2h-ag), demonstrating the specificity of immunostaining observed with the ArPPLN2 antiserum. (b) Immunostaining in a transverse section of the circumoral nerve ring; here immunostained processes can also be seen here in the peristomial membrane, an adjacent oral tube foot and in a Tiedemann's body. A high magnification image of the boxed region can be seen in (d). (c) High magnification image of the ectoneural region of the radial nerve cord showing immunostained bipolar cells in the sub-cuticular epithelium (arrowheads) and densely packed immunostained processes (asterisks) in the underlying neuropile region. (d). Immunostained monopolar shaped cells (black arrowheads) in the hyponeural region of the circumoral nerve ring, with stained processes (white arrowhead) projecting into a fiber layer that is adjacent to the unstained collagenous tissue layer (arrow). The intensely stained ectoneural neuropile is labeled here with an asterisk. (e) The continuity of immunostaining in the ectoneural region of the radial nerve and in the basiepithelial nerve plexus of an adjacent tube foot (asterisks) can be seen here. The stained process(es) (arrowhead) of a hyponeural neuron(s) can be seen projecting over the roof of the peri-hemal canal in close association with the transverse infra-ambulacral muscle (arrow). (f) Immunostaining in the marginal nerve (arrow) and in axonal processes of the lateral motor nerve (arrowheads). CONR, circumoral nerve ring; CuL, cuticle layer; Ec, ectoneural region; Hy, hyponeural region; PM, peristomial membrane; RHS, radial hemal strand; TB, Tiedemann's body; TF, tube foot. Scale bars: 50 lm in (a), (f); 200 lm in (a) inset; 10 lm in (b), (d); 100 lm in (c); 20 lm in (e)

| Digestive system
The external epithelial layer of the peristomial membrane that surrounds the mouth contains immunostained cells and immunostained processes are present in the underlying basiepithelial nerve plexus (Figure 9a,b). A similar pattern of immunostaining is observed in the short esophagus that is located between the peristomial membrane and the cardiac stomach (Figure 9a).

| Body wall, body wall-associated appendages and interossicular muscles
ArPPLNP2h-immunoreactive fibers were observed in the coelomic lining of the body wall ( Figure 10a). Thus, stained fibers are present in the coelomic basiepithelial nerve plexus that is located beneath the coelomic epithelium and which is closely associated with a longitudinally orientated muscle layer (Figure 10a,b). Furthermore, immunostained fibers were also observed within the circularly orientated muscle layer of the coelomic lining (Figure 10a,b). Immunostaining of nerve fibers was also observed in the sub-epidermal plexus of the body wall ( Figure   10b,c). This staining extends into papule, hollow finger-shaped organs located above voids in the calcite skeleton that enable gas exchange between the external seawater and the coelomic fluid. Another type of body wall appendage in starfish is pedicellariae, pincer-like organs that keep the body surface clear of encrusting material. In pedicellariae, immunostaining was observed in processes associated with the adductor muscles (Figure 10d). Immunostained fibers were also revealed in association with interossicular muscles that link adjacent ossicles of the body wall endoskeleton (Figure 10e,f).
3.6 | ArPPLN2h has no effect on the contractility of in vitro preparations of apical muscle and tube feet from A. rubens but is a potent relaxant of cardiac stomach preparations ArPPLN2h was tested at a high concentration (10 26 M) on three in vitro preparations from A. rubens: apical muscle, tube feet and cardiac stomach. No effects on the contractile state of apical muscle and tube foot preparations were observed, whereas ArPPLN2h caused relaxation of cardiac stomach preparations. ArPPLN2h caused dose-dependent relaxation of cardiac stomach preparations and comparison of the effects of ArPPLN2h with the SALMFamide neuropeptide S2 revealed that the potency and efficacy of ArPPLN2h was very similar to that of S2 ( Figure   11). Thus, relaxing effects were observed with ArPPLN2h and S2 in the concentration range 10 28 and 10 26 M and the maximum effect (E max ) observed at 10 26 M was 40.05 6 5.07% and 48.03 6 8.17% reversal of KCl-induced contraction for ArPPLN2h and S2, respectively.

| ArPPLN2h does not induce cardiac stomach eversion
Previous studies revealed that the SALMFamide S2 causes cardiac stomach relaxation in vitro and induces cardiac stomach eversion when injected in vivo (Melarange et al., 1999). Here we observed that ArPPLN2h and S2 cause cardiac stomach relaxation in vitro with similar potency and efficacy. Therefore, we compared the effects of ArPPLN2h and S2 in vivo, with 100 mL of peptide at a concentration of 1 mM injected at two or three sites in the aboral body wall of the arms proximal to the junctions with the central disk region. Cardiac stomach eversion was observed within 5 min of injection in seven out of ten starfish injected with S2 but stomach eversion was not observed in any of ten animals injected with ArPPLN2h.

| D I SCUSSION
Here we report the structural, anatomical and functional characterization of neuropeptides derived from A. rubens PP-like neuropeptide precursor 2 (ArPPLNP2). ArPPLNP2 comprises 13 copies of 11 PP/OK-type peptides (ArPPLN2a-k), with three copies of ArPPLN2h and one copy each of the other ten peptides. Alignment of ArPPLNP2-derived peptides with other PP/OK-type neuropeptides reveals several conserved structural characteristics (see Figure 1c). The functional significance of the conserved residues remains to be determined for starfish (e, f) Immunostaining in the body wall at the junction between two arms in a horizontal section of a juvenile starfish. Immunostained fibers can be seen associated with muscles that link adambulacral ossicles (arrowheads). Stained fibers are also evident in thickenings of the sub-epithelial nerve plexus of the body wall (arrows) and in the tube feet. A high magnification image of the boxed region is shown in (f), which shows immunostained processes (arrowheads) in muscles that link adambulacral ossicles. BW, body wall; CBNP, coelomic basiepithelial nerve plexus; CMLNP, circular muscle layer nerve plexus; CT, collagenous tissue; Os, ossicle; Pa, papulae; SNP, sub-epidermal nerve plexus; TF, tube foot. Scale bars: 20 lm in (a), (c), (d), (f); 100 lm in (b), (e) PP/OK-type peptides. However, it is noteworthy that structure-activity analysis has revealed that N-and C-terminally located phenylalanine residues appear to be critically important for the bioactivity of OK (Bungart, Kegel, Burdzik, & Keller, 1995), consistent with the evolutionary conservation of hydrophobic residues at these positions in PP/OKtype peptides from other protostomian invertebrates and from echinoderms.

| Functional interpretation of the expression patterns and actions ArPPLN2-type peptides in A. rubens
ArPPLNP2 mRNA expressing cells and ArPPLN2h-immunoreactive cells and processes were detected in both the ectoneural and hyponeural regions of the radial nerve cords and circumoral nerve ring and in the marginal nerves of A. rubens. Collectively, these findings indicate that neuropeptides derived from ArPPLNP2 have a widespread role in mediating neuronal signaling in the starfish nervous system. Little is known about the functional organization of the ectoneural nervous system in starfish (or other echinoderms), although it is thought to comprise sensory neurons, interneurons and motor neurons (Cobb, 1970;Cobb, 1987;Mashanov, Zueva, Rubilar, Epherra, & Garcia-Arraras, 2016). On the other hand, the hyponeural division of the nervous system is recognized as having a purely motor function, based on both anatomical and functional studies (Cobb, 1970;Cobb, 1987;Mashanov et al., 2016). Therefore, the presence ArPPLNP2-derived neuropeptides in hyponeural cell bodies and processes can be interpreted as evidence of a role in motoneuronal function in starfish. Consistent with this notion, ArPPLN2h-immunoreactive processes are present in the lateral motor nerves, which innervate interossicular FIG URE 11 ArPPLN2h causes dose-dependent relaxation of in vitro preparations of the cardiac stomach from A. rubens. (a) Representative recording showing the dose-dependent relaxing effect of ArPPLN2h (10 210 to 10 26 M) on a cardiac stomach preparation pre-contracted with artificial seawater containing 30 mM added KCl (leftward arrow). Following tests with ArPPLN2h, the preparation was washed with artificial seawater (downward arrow). (b) Graph showing concentration-dependent relaxing effect of ArPPLN2h (green) on cardiac stomach preparations. Each point represents the mean 6 SEM from at least four separate experiments, with the effect calculated as the percentage reversal of contraction induced by KCl. The effectiveness of ArPPLN2h as a cardiac stomach relaxant is much greater than that of the ArPPLN1b (red) but similar to that of the SALMFamide neuropeptide S2 (blue) muscles in the starfish body wall (Smith, 1937(Smith, , 1950. Accordingly, the widespread occurrence of ArPPLN2h-immunoreactive processes in interossicular muscles and other muscles in A. rubens is suggestive of a general role for ArPPLNP2-derived neuropeptides in neuromuscular signaling.
The presence of ArPPLNP2-expressing cells and ArPPLN2himmunoreactive processes in the basiepithelial nerve plexus of tube feet in A. rubens suggested a potential role for ArPPLN2-derived neuropeptides in the regulation of the contractility of these motile and muscular organs and therefore ArPPLN2h was tested for myoactivity on in vitro preparations of tube feet. ArPPLN2h had no effect on tube foot preparations, contrasting with the ArPPLNP1-derived neuropeptide ArPPLN1b, which has a pattern of expression in tube feet that is very similar to that of ArPPLN2h but which causes dose-dependent relaxation of tube foot preparations . These findings are indicative of differences in the physiological roles of ArPPLNP1-derived and ArPPLNP2-derived neuropeptides in A. rubens.
Both ArPPLN1b and ArPPLN2h are also present in the tube foot basal nerve ring and in processes that innervate the adhesive tube foot sucker, which could be indicative of roles in regulation of the secretion of proteins that mediate tube adhesion or detachment (Hennebert, Jangoux, & Flammang, 2013;Hennebert, Wattiez, Waite, & Flammang, 2012;Santos, Haesaerts, Jangoux, & Flammang, 2005).
Each tube foot stem is linked via a tubular canal that extends through the body wall to a bulb-shaped muscular ampulla, which is located in the perivisceral coelom. Contraction of the ampulla facilitates tube foot extension, whereas relaxation of the ampulla enables accommodation of fluid from the lumen of the podium when a tube foot contracts (Hennebert et al., 2013;Hennebert et al., 2012;Santos et al., 2005). Both ArPPLN1b-immunoreactive  and ArPPLN2h-immunoreactive (this study) fibers are present in ampullae and are located in close proximity to the muscle layer. Previous studies have reported that the ampullae are innervated by hyponeural neurons (Smith, 1950) and therefore it is likely that ArPPLN1b-and ArPPLN2himmunoreactive fibers in the ampulla originate from ArPPLNP1-and ArPPLNP2-expressing hyponeural motoneurons.
The terminal tentacle is a tube foot-like sensory organ located at the tip of each arm in A. rubens and in other starfish. ArPPLN2himmunoreactive cells and fibers are present in the wall of the terminal tentacle and in the body wall epithelium and lateral lappets that surround the terminal tentacle. The physiological roles of the lateral lappets (Smith, 1937) are not known, although a possible function would be to act as chemosensory organs. Turning to a different sensory modality, located at the base of the terminal tentacle is a pigmented photosensory organ known as the optic cushion and here ArPPLNP2expressing cells are present in the photoreceptor cell layer. It is noteworthy that OKs have a physiological role in mediating light entrainment of circadian activity in insects (Hofer & Homberg, 2006) and therefore it is possible that PP/OK-type peptides have a similar role in starfish.
Cells expressing ArPPLNP2 transcripts were detected in the coelomic epithelial layer in many regions of the digestive system. In accordance with this pattern of precursor expression, antibodies to ArPPLN2h labeled processes located in the visceral nerve plexus underlying the coelomic epithelium. Extensive immunoreactive fibers in the extrinsic retractor strand nodule of the cardiac stomach were particularly striking, consistent with the abundance of ArPPLNP2expressing coelomic epithelial cells lining the nodule. Consistent with the extensive expression of ArPPLN2-type neuropeptides in the cardiac stomach and other regions of the digestive system, we found that ArPPLN2h caused dose-dependent relaxation of in vitro cardiac stomach preparations. We previously reported the expression of ArPPLNP1 and neuropeptides derived from this precursor in the cardiac stomach of A. rubens and found that ArPPLN1b, like ArPPLN2h, causes relaxation of in vitro preparations of the cardiac stomach . Here we compared the bioactivity of ArPPLN1b and ArPPLN2h as cardiac stomach relaxants and found that ArPPLN2h is much more effective than ArPPLN1b. This difference in the bioactivity of the two peptides may be associated with differences in the patterns of expression of ArPPLNP1-derived and ArPPLNP2-derived peptides in the cardiac stomach. Both types of peptides are expressed by mucosal cells and their processes in the basiepithelial plexus, but only ArPPLNP2derived peptide (ArPPLN2h) expression was detected in the visceral nerve plexus that underlies the coelomic epithelium of the cardiac stomach. This is significant because the visceral nerve plexus is closely associated with the visceral muscle layer of the cardiac stomach, whereas the basiepithelial nerve plexus is separated from the visceral muscle layer by a layer of collagenous tissue.
Previous studies have revealed that the SALMFamide neuropeptides S1 and S2 cause relaxation of in vitro cardiac stomach preparations from A. rubens, but with S2 ten times more potent than S1 (Melarange et al., 1999). Therefore, here we compared the bioactivity of ArPPLN2h with S2 and found that the potency/efficacy of ArPPLN2h is very similar to that of S2 in vitro. In vivo pharmacological tests with S2 have revealed that injection of this peptide into the perivisceral coelom of A.
rubens triggers eversion of the cardiac stomach, indicating that S2 may be involved in physiological mechanisms that mediate control of stomach eversion when starfish feed extraorally on prey (e.g., mussels) (Melarange et al., 1999), whereas here we found that ArPPLN2h does not trigger cardiac stomach eversion in A. rubens. Thus, whilst S2 and ArPPLN2h exhibit similar cardiac stomach relaxing activity in vitro, this appears not to be complemented by the same bioactivity in vivo. This difference in in vivo bioactivity may reflect differences in regions of the cardiac stomach musculature that are regulated by S2 and ArPPLN2h as muscle relaxants and/or differences in the stability of these peptides in vivo. With regard to the latter, it is noteworthy that S2 is a Cterminally amidated peptide, whereas PP/OK-type peptides in starfish are not amidated.

| Comparison of the expression and bioactivity of ArPPLN1-type and ArPPLN2-type peptides in A. rubens
Comparison of the expression patterns of ArPPLN1-type  and ArPPLN2-type precursors/neuropeptides reveals some general similarities, which may reflect a common evolutionary origin. However, there are also distinct differences in the expression patterns and bioactivity of ArPPLN1-type and ArPPLN2-type precursors/neuropeptides. Some of these differences have already been discussed above and here we highlight others. The prototype of PPLN1type peptides in starfish is SMP, which was discovered on account of its relaxing effect on the apical muscle of P. pectinifera (Kim et al., 2016) and accordingly ArPPLN1b causes dose-dependent relaxation of apical muscle preparations from A. rubens .
Furthermore, consistent with the bioactivity of ArPPLN1b, antibodies to ArPPLN1b reveal that this peptide is present in axonal processes that ramify amongst the muscle fibers of the apical muscle . It is interesting, therefore, that in this study no ArPPLN2h-immunoreactivity was detected in the apical muscle of A.
rubens and in accordance with this finding ArPPLN2h does not affect the contractile state of the apical muscle in vitro. Thus, whilst both PPLN1-type and PPLN2-type peptides act as muscle relaxants in starfish, it appears that they have specialized with respect to the muscle systems that they regulate. Thus, PPLN1-type peptides act as relaxants of the apical muscle and tube feet, whereas PPLN2-type peptides do not cause relaxation of these muscle preparations in vitro. Conversely, PPLN2-type peptides cause relaxation of cardiac stomach preparations with an efficacy and potency comparable to that of the SALMFamide neuropeptide S2, whilst PPLN1-type peptides are much less efficacious as cardiac stomach relaxants.

| Evolution and comparative physiology of PP/ OK-type neuropeptide signaling in the Bilateria
This study of PPLN2-type peptides in A. rubens together with our previous report on PPLN1-type peptides in A. rubens  are the first detailed analyzes of the anatomical expression patterns of PP/OK-type neuropeptides in an echinoderm. In the future, it would be interesting to perform similar investigations in other starfish.
Analysis of the genome/transcriptome of the crown-of-thorns starfish Acanthaster planci has revealed the presence of genes/transcripts encoding proteins that are orthologs of ArPPLNP1 and ArPPLNP2 (Smith et al., 2017). Furthermore, based on recent extensive molecular analysis of the phylogenetic relationships of extant Asteroidea (Linchangco et al., 2017), the occurrence of PPLNP1-type (SMP-type) and PPLNP2-type proteins in A. rubens (order Forcipulatida) and in A. planci (order Valvatida) suggests that both of these PP/OK-type precursors would have been present in the common ancestor of all extant starfish.
Previous studies have shown that PPLN1-type (SMP-type) peptides act as muscle relaxants in P. pectinifera (order Valvatida and in A. rubens (order Forcipulatida) and therefore it seems likely, based on asteroid phylogeny (Linchangco et al., 2017), that PPLN1-type peptides act as muscle relaxants in all extant starfish. This study is the first to reveal the myorelaxant activity of PPLN2-type peptides in a starfish species-A. rubens (order Forcipulatida). It will be interesting, therefore, to investigate if PPLN2-type peptides also act as muscle relaxants in starfish that belong to other starfish orders-for example, in P. pectinifera and/ or A. planci, both of which are species belonging to the order Valvatida.
Looking beyond starfish, it would be of interest to determine if PP/OK-type neuropeptides act as muscle relaxants in other echinoderms Addressing this issue is feasible because precursors of PP/OK-type neuropeptides have been identified in other echinoderms, including two precursors in the sea urchin S. purpuratus (SpPPLNP1 and SpPPLNP2; Rowe & Elphick, 2012), one precursor in the sea cucumber A. japonicus (Rowe, Achhala, & Elphick, 2014) and one or more precursors in several brittle star species (Zandawala et al., 2017).
As discussed in detail previously , comparative analysis of the physiological roles of PP/OK-type neuropeptides in the Bilateria is limited in as much as the echinoderms are the only deuterostomian phylum in which PP/OK-type neuropeptides have been identified. Thus, the functional characterization of PP/OK-type neuropeptides in starfish reported here and in two previous articles (Kim et al., 2016; represents the totality of our knowledge of physiological roles of PP/OK-type neuropeptides in deuterostomes. This contrasts with a much more extensive series of studies on PP-type neuropeptides in mollusks and OK-type neuropeptides in arthropods. It is clear from these studies that PP/OK-type neuropeptides have a wide variety of physiological roles in protostomes, including regulation of muscle and ciliary activity associated with locomotor activity in mollusks (Hall & Lloyd, 1990) and regulation of both the frequency and amplitude of hind-gut contraction in arthropods (Stangier et al., 1992). A unifying theme among the pharmacological actions of PP/OK-type neuropeptides in protostomes is stimulatory effects, with the effector tissues often being muscular in nature. Thus, this contrasts with the relaxing effects of PP/OK-type neuropeptides on starfish muscle reported here and previously (Kim et al., 2016;. We have speculated that this may reflect an excitatory-inhibitory transition in the roles of PP/OK-type neuropeptides as regulators of muscle activity that accompanied the divergence of protostomes and deuterostomes . To address this issue, investigation of the occurrence and physiological roles of PP/OK-type neuropeptides in a variety of deuterostomian and protostomian taxa is now needed. Furthermore, discovery of the receptors that mediate the effects of PP/OK-type neuropeptides will be necessary to gain a deeper understanding of the evolution PP/OK-type neuropeptide signaling in the Bilateria.

ACKNOWLEDGMENTS
This study was supported by a China Scholarship Council studentship awarded to ML and grants from the BBSRC (BB/M001644/1, awarded to MRE; BB/M001032/1 awarded to AMJ). We are grateful to Phil Edwards for his help with obtaining starfish and to Paul Fletcher for maintaining our seawater aquarium.