Starfish (A. rubens) were collected from the Thanet Coast (Kent, UK) at low tide or were acquired from a fisherman based in Whitstable (Kent, UK) and transported to the School of Biological and Chemical Sciences, Queen Mary University of London. The starfish were maintained in circulating artificial seawater at approximately 12 °C in an aquarium and were fed mussels (Mytilus edulis). Animals ranging in diameter from 5 to 15 cm were used for in vitro and in vivo pharmacological experiments, whereas smaller animals (< 5 cm in diameter) were used for anatomical studies. Additionally, juvenile specimens of A. rubens (diameter 0.5–1.5 cm) were collected from the University of Gothenburg Sven Lovén Centre for Marine Infrastructure (Kristineberg, Sweden) and used for anatomical studies.
Mass spectrometric identification of the VP/OT-type neuropeptide asterotocin in A. rubens radial nerve cord extracts
Radial nerve cords were dissected from five adult specimens of A. rubens using a method described previously . Neuropeptides were then extracted in 1 ml of 80% acetone on ice . Acetone was removed by evaporation using nitrogen, with the aqueous fraction centrifuged at 11,300g in a MiniSpin® microcentrifuge (Eppendorf) for 10 min, and the remaining supernatant stored at − 80 °C. Prior to mass spectrometry (MS), the extract was thawed (with an aliquot diluted tenfold with 0.1% aqueous formic acid) and filtered through a 0.22-μm Costar Spin-X centrifuge tube filter (Sigma-Aldrich) to remove particulates. In comparison with synthetic asterotocin (PPR Ltd., Fareham, UK), the radial nerve cord extract was analysed by nanoflow liquid chromatography (LC) with electrospray ionisation (ESI) quadrupole time-of-flight tandem MS (nano LC-ESI-MS/MS) using a nanoAcquity ultra performance LC (UPLC) system coupled to a Synapt® G2 High-Definition Mass Spectrometer™ (HDMS) (Waters Corporation) and MassLynx v4.1 SCN 908 software (Waters Corporation). The mobile phases used for the chromatographic separation were 0.1% aqueous formic acid (termed mobile phase A) and 0.1% formic acid in acetonitrile (termed mobile phase B). An aliquot containing 15 μl of the A. rubens radial nerve cord extract was applied to a Symmetry C18® (180 μm × 20 mm, 5 μm particle size, 100 Å pore size) trapping column (Waters Corporation) using 99.9% mobile phase A at a flow rate of 10 μl min−1 for 3 min, after which the fluidic flow path included the HSS T3 (75 μm × 150 mm, 1.8 μm particle size, 100 Å pore size) analytical capillary column (Waters Corporation). A linear gradient of 5–40% mobile phase B over 105 min was utilised with a total run time of 120 min. The nanoflow ESI source conditions utilised a 3.5-kV capillary voltage, 25 V sample cone voltage and an 80 °C source temperature. The instrument was operated in resolution mode (~ 20,000 measured at full width and half height). A solution containing 500 fmol μl−1 Glu1-Fibrinopeptide B peptide in 50% v/v aqueous acetonitrile containing 0.1% formic acid was infused via a NanoLockSpray interface at a constant rate of 500 nl min−1, sampled every 60 s and used for lockmass correction (m/z 785.8426) enabling accurate mass determination.
A data-dependent acquisition was performed that would trigger an MS/MS scan on any singly charged peptide having a mass to charge ratio (m/z) of 960.3919, or a doubly charged peptide of m/z 480.6999, with a tolerance of 100 mDa allowed on the precursor m/z. MS/MS spectra, obtained from data-dependent acquisition, were processed using MassLynx™ software (Waters Corporation). Spectra were combined and processed using the MaxEnt 3 algorithm to generate singly charged, monoisotopic spectra for interpretation and manual validation.
Identification and cloning of a cDNA encoding an A. rubens VP/OT-type receptor
To identify an A. rubens VP/OT-type receptor, the amino acid sequence of the sea urchin (Strongylocentrotus purpuratus) VP/OT-type receptor  was submitted as a query in a tBLASTn search of A. rubens radial nerve cord transcriptome sequence data using SequenceServer software . The top hit (contig 1122053) was a 2710-bp transcript encoding a 428-residue protein, which based on reciprocal BLAST analysis was identified as an ortholog of the S. purpuratus VP/OT-type receptor. A cDNA encoding the A. rubens VP/OT-type receptor was then cloned using A. rubens radial nerve cord cDNA as a template for PCR amplification, employing the use of primers (see Additional file 2) that were designed using Primer3 (http://primer3.ut.ee). The cDNA was then incorporated into the pBluescript SKII (+) vector and sequenced (Eurofins Genomics).
Phylogenetic analysis of the A. rubens VP/OT-type receptor
Phylogenetic analysis of the relationships between the A. rubens VP/OT-type receptor and VP/OT-type receptors from other species was performed using the maximum-likelihood method. Firstly, receptor protein sequences were aligned in MEGA7 (v.7170509) using MUltiple Sequence Comparison by Log-Expectation (MUSCLE) . Once aligned, poorly aligned regions were removed using the Gblocks server (using the least stringent settings) to optimise the alignment for phylogenetic analysis . PhyML (version 3.0) was then used to generate a maximum-likelihood tree . The LG model was automatically selected, and the bootstrap was manually set to 1000. FigTree version 1.4.3 (http://tree.bio.ed.ac.uk/software/figtree/) was used to visualise and re-root the tree generated by PhyML. NPS/CCAP/NG peptide-type, and GnRH/AKH/ACP/CRZ-type receptor sequences were used as outgroups in the tree.
Testing asterotocin as a ligand for the A. rubens VP/OT-type receptor
A pBluescript SKII (+) vector containing the A. rubens VP/OT-type receptor cDNA as an insert (see above) was used as a template to amplify by PCR the open reading frame of the VP/OT-type receptor. To accomplish this, the oligonucleotides 5′-ggatccCACCATGACGCCCTC-3′ (upstream) and 5′-cccgggCTACATGTGAGCGGAAGCA-3′ (downstream) were used as primers and the PCR product was subcloned into the eukaryotic expression vector pcDNA 3.1+, which had been cut with the restriction enzymes BamHI and ApaI. For the upstream primer, a partial Kozak translation initiation sequence (CACC) was introduced before the start codon to optimise the initiation of translation.
To determine if asterotocin acts as a ligand for the A. rubens VP/OT-type receptor, a cell-based assay was used where Ca2+-induced luminescence is measured. For this assay, Chinese hamster ovary (CHO-K1) cells stably expressing the Ca2+-sensitive bioluminescent GFP-aequorin fusion protein—G5A (CHO-K1/G5A cells ) are co-transfected with plasmids containing a G-protein coupled receptor cDNA insert and plasmids containing an insert encoding the promiscuous G-protein Gα16. If a candidate ligand activates the expressed receptor, the promiscuous Gα16 protein triggers the activation of the IP3 signalling pathway, causing an increase in intracellular Ca2+ and luminescence. We recently reported the use of this assay to demonstrate that the luqin-type neuropeptide ArLQ acts as a ligand for the A. rubens G-protein coupled receptors ArLQR1 and ArLQR2 , and therefore, here only a brief outline of methods employed is described. CHO-K1/G5A cells were transfected with 5 μg of the pcDNA 3.1+ plasmid containing the A. rubens VP/OT-type receptor cDNA and 1.5 μg of plasmid containing an insert encoding the promiscuous Gα16 subunit using the Lipofectamine 3000® Transfection Kit, (Invitrogen). The cells were then loaded with the aqueorin substrate coelenterazine-H (Thermo Fisher Scientific). Asterotocin at concentrations ranging from 10−4 to 10−12 M (n = 3) was pipetted into the wells of clear-bottomed 96-well plates (Sigma-Aldrich), then a FLUOstar Omega Plate Reader (BMG LABTECH) was used to inject a fixed amount of transfected CHO-K1 cells into each well sequentially, and luminescence was measured for a 35-s period after injection. Luminescence measurements were normalised to the maximum response obtained in each experiment (100%) and the response obtained with the vehicle media (0%). These data (three repeats per experiment and at least three independent transfections) were used to construct a dose-response curve utilising non-linear regression analysis in Prism 6.0c (GraphPad, La Jolla, USA) and displayed on a semi-logarithmic plot. The half maximal effective concentration (EC50) was calculated from the dose-response curve in Prism 6.0c. For negative control experiments, CHO-K1/G5A cells were transfected with the empty pcDNA 3.1+ vector. Human vasopressin and oxytocin and the A. rubens neuropeptide NGFFYamide were also tested to assess the specificity of asterotocin as the candidate ligand for the A. rubens VP/OT-type receptor. A Wilcoxon signed-rank test was performed to compare luminescence of cells exposed to asterotocin, vasopressin, oxytocin and NGFFYamide (at 10−4 M) with luminescence of cells in basal media (control).
Localisation of the expression asterotocin precursor and asterotocin receptor transcripts in A. rubens using in situ hybridisation
The method employed for the production of the asterotocin precursor antisense and sense digoxygenin (DIG)-labelled RNA probes is as described in . Production of the asterotocin receptor antisense and sense DIG-labelled RNA probes was performed as follows. The plasmid containing the cloned asterotocin receptor cDNA was linearised using the restriction enzymes HindIII and EcoRI (NEB, Hitchin, Hertfordshire, UK). Once linearised, the plasmids were purified using phenol-chloroform/chloroform isomylalcohol (Sigma-Aldrich Ltd., Gillingham, UK) extraction. RNA probes were synthesised from the purified, linearised plasmid using a DIG-labelled nucleotide triphosphate mix (Roche, Nutley, NJ) supplemented with dithiothreitol (Promega), a placental RNase inhibitor (Promega), and RNA polymerases (New England Biolabs), according to the manufacturer’s instructions. To synthesise the antisense and sense probes, T3 and T7 RNA polymerases were used, respectively. Reaction products were digested with RNase-free DNase (New England Biolabs) to remove template DNA and then stored in 25% formamide made up in 2× saline-sodium citrate (SSC) buffer at − 20 °C.
The methods employed for preparation of sections of 4% paraformaldehyde-fixed specimens of A. rubens and visualisation of asterotocin precursor and asterotocin receptor transcripts in these sections were the same as those used previously for the analysis of AruRGPP expression, as reported in . For the visualisation of asterotocin precursor transcripts, the sections were incubated with antisense and sense RNA probes at a concentration of 800 ng/ml. Then, following incubation with anti-DIG antibodies, slides were incubated with the staining solution at room temperature for a few hours until strong staining was observed. For the visualisation of asterotocin receptor transcripts, the sections were incubated with antisense and sense RNA probes at a concentration of 1500 ng/ml. Then, following incubation with anti-DIG antibodies, the slides were incubated with the staining solution at room temperature for a few hours, left overnight at 4 °C and then were incubated at room temperature for several more hours until strong staining was observed.
Production, characterisation and purification of rabbit antibodies to asterotocin and guinea pig antibodies to the asterotocin receptor
To generate antibodies to asterotocin, a rabbit was immunised with a conjugate of thyroglobulin (carrier protein) and the peptide Lys-asterotocin (KCLVQDCPEG-NH2; disulphide bridge between the cysteine residues), with the N-terminal lysine providing a free amine group for glutaraldehyde-mediated coupling to thyroglobulin. To generate antibodies to the asterotocin receptor, a guinea pig was immunised with a conjugate of thyroglobulin and a peptide corresponding to the C-terminal region of the asterotocin receptor sequence (KFVSTTGTASAHM) but with an additional N-terminal lysine providing a free amine group for glutaraldehyde-mediated coupling to thyroglobulin.
The peptide antigens were synthesised by Peptide Protein Research Ltd. (Fareham, Hampshire, UK). Conjugation of the antigen peptides to thyroglobulin was performed as described in . Rabbit immunisation and serum collection were performed by Charles River Biologics (Romans, France) according to the following protocol. On day 0, pre-immune serum was collected and the first immunisation (~ 100 nmol of conjugated antigen peptide emulsified in Freund’s complete adjuvant) was administered. Booster immunisations (~ 50 nmol of conjugated antigen peptide emulsified in Freund’s incomplete adjuvant) were administered on days 28, 42 and 56. Samples of blood serum were collected on days 38 and 52, and the final bleed serum was collected on day 70. Guinea pig immunisation and serum collection were performed by Charles River Biologics (Romans, France) according to the following protocol. On day 0, pre-immune serum was collected and the first immunisation (~ 100 nmol of conjugated antigen peptide emulsified in Freund’s complete adjuvant) was administered. Booster immunisations (~ 50 nmol of conjugated antigen peptide emulsified in Freund’s incomplete adjuvant) were administered on days 14, 28 and 42. A serum sample was collected on day 38, and the final bleed serum was collected on day 56.
Enzyme-linked immunosorbent assays (ELISA) were performed to test the sera for the presence of antibodies to the antigen peptides, employing the same methods as described previously for ArGnRH antisera . Terminal bleed antisera were characterised by ELISA by testing antisera at a starting dilution of 1:500 and subsequent twofold serial dilutions down to 1:16,000 (Additional file 4). Then, antibodies to the antigen peptides were affinity-purified from terminal bleed antisera using AminoLink® Plus Immobilisation Kit (Thermo Scientific), employing the same methods as described previously for antibodies to ArGnRH .
Localisation of the expression of asterotocin and the asterotocin receptor in A. rubens using immunohistochemistry
The methods employed for preparation of sections of Bouin’s fixed specimens A. rubens and immunohistochemical localisation of asterotocin expression and asterotocin receptor expression were the same as those described previously for ArPPLN1 and ArGnRH [21, 48]. The sections were incubated overnight or for 3 days, respectively, with affinity-purified rabbit antibodies to asterotocin (1:4 dilution) and guinea pig antibodies to the asterotocin receptor (1:4 dilution). Then, bound antibodies were visualised using diaminobenzidine as a substrate for peroxidase-conjugated AffiniPure Goat anti-rabbit immunoglobulins (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) or peroxidase-conjugated AffiniPure Donkey anti-guinea pig immunoglobulins (Jackson ImmunoResearch Laboratories, West Grove, PA, USA).
To visualise asterotocin and the asterotocin receptor in the sections using double-labelling fluorescence immunohistochemistry, the sections were first incubated with affinity-purified asterotocin antibodies (1:4 dilution) overnight at 4 °C. Then, following washes (4 × 5 min) in phosphate-buffered saline (PBS) containing 0.1% Tween-20 (PBST) and washes in PBS (4 × 10 min), the sections were incubated with affinity-purified antibodies to the asterotocin receptor (1:4 dilution) for 2 weeks at 4 °C. Then, following washes with PBST and PBS (as described above), the sections were incubated for 3 to 4 h with Cy2-labelled goat anti-rabbit immunoglobulins and Cy3-labelled goat anti-guinea pig immunoglobulins (Jackson ImmunoResearch Europe Ltd., UK), which were both diluted 1:200 in 2% normal goat serum in PBS. Following washes in PBST and PBS (as described above), the slides were mounted with coverslips using Fluoroshield Mounting Medium with DAPI (Abcam, Cambridge, UK).
To capture images of sections labelled using fluorescence immunohistochemistry, a Leica SP5 confocal microscope was used in combination with the Leica Application Suite Advanced Fluorescence (LAS AF; version 18.104.22.16866) programme. Argon and DPSS 561 lasers were used for the detection of green fluorescence (asterotocin) and red fluorescence (asterotocin receptor), respectively. For Beam Path Settings, FITC and TRITC were selected, and in the case of TRITC, the PMT was set to Cy3. While visualising immunofluorescence on slides, the smart gain, smart offset and z position were adjusted to generate optimal immunofluorescent images. The settings for capturing images were as follows: image format, 1024 × 1024; scan speed, 200 Hz; frame average, 6; and line average accumulation, 3. The FITC and TRITC images were taken separately, and the colour channels were merged using ImageJ to produce double-labelled images. Contrast and levels were adjusted in ImageJ, and montages were created in Adobe Photoshop CC 2017.1.1.
Analysis of the in vitro pharmacological effects of asterotocin on cardiac stomach, apical muscle and tube foot preparations from A. rubens
To investigate if asterotocin affects muscle contractility in A. rubens, synthetic asterotocin (custom synthesised by PPR Ltd., Fareham, Hampshire, UK) was tested on three in vitro preparations: the cardiac stomach, apical muscle and tube feet. These three preparations have been used to examine the effects of other neuropeptides, and the methods employed have been described in detail previously [30, 33, 34, 73]. Preliminary tests revealed that asterotocin caused relaxation of the cardiac stomach and apical muscle preparations but had no effect on the contractile state of tube foot preparations. Therefore, the effects of asterotocin on cardiac stomach and apical muscle preparations were examined in more detail.
Cardiac stomach preparations from 16 starfish (8–13.5 cm in diameter) were incubated at 11 °C in an aerated organ bath containing artificial seawater with 3 × 10−2 M added KCl. This induces sustained contraction of the cardiac stomach, which facilitates recording of the effects of neuropeptides that act as muscle relaxants . The contractile state of preparations was monitored using a high-grade isotonic transducer (model 60-3001; Harvard Apparatus, Cambridge, UK) connected to a Goerz SE 120 chart recorder (Recorderlab, Sutton, Surrey, UK) or using a high-grade isotonic transducer (MLT0015; ADInstruments Ltd., Oxford, UK) connected to data acquisition hardware (PowerLab 2/26, ADInstruments Ltd.) via a bridge amplifier (FE221 Bridge Amp, ADInstruments Ltd.). The output from the PowerLab was recorded using LabChart (v8.0.7) software installed on a laptop computer (Lenovo E540, Windows 7 Professional). To obtain representative images of the effects of asterotocin, traces were exported from LabChart into Adobe Photoshop and the function ‘Select and Mask’ was utilised to remove the background grid pattern.
To investigate the dose dependence and potency of asterotocin as a cardiac stomach relaxant, it was tested at concentrations ranging from 3 × 10−11 M to 10−6 M (n = 16). The percentage relaxation at each concentration was calculated relative to the maximal relaxing effect produced by asterotocin in each preparation. To enable the assessment of the magnitude of the relaxing effect of asterotocin on cardiac stomach, experiments were performed to compare the effect asterotocin with the effect of the neuropeptide SALMFamide-2 (S2; SGPYSFNSGLTF-NH2), which has been shown previously to cause relaxation of A. rubens cardiac stomach preparations in vitro [33, 34]. Having established that the maximal relaxing effect of asterotocin is observed at concentrations ranging from 3 × 10−9 M to 10−6 M (with a mean of ~ 10−7 M), experiments were performed where the effects of 10−7 M asterotocin and 10−7 M S2 on cardiac stomach preparations (n = 5) were compared. For these experiments, the effect of 10−7 M S2 was defined as 100%, and the effect of 10−7 M asterotocin was calculated as a percentage of the effect of 10−7 M S2. The relaxing effects of asterotocin and S2 on cardiac stomach preparations were analysed statistically using the Mann-Whitney U test.
To examine the effects of asterotocin on apical muscle preparations, 10−6 M acetylcholine (ACh) was used to induce contraction prior to application of asterotocin. A relaxing effect of asterotocin was observed only at a high concentration (10−6 M) and therefore it was not possible to examine the dose dependency of this effect. However, the experiments were performed in which the time course of the relaxing effect of asterotocin was analysed by measuring the percentage reversal of 10−6 M ACh-induced contraction over a 50-s period after application of 10−6 M asterotocin (n = 4).
Analysis of the in vivo effects of asterotocin on A. rubens
During feeding in A. rubens, the cardiac stomach is everted out of the mouth over the digestible soft tissue of prey (e.g. mussels), and to accomplish this, the cardiac stomach must be relaxed. Previous studies have revealed that S2, a neuropeptide that induces relaxation of A. rubens cardiac stomach preparations in vitro, induces cardiac stomach eversion when injected in vivo . Having established that asterotocin causes relaxation of cardiac stomach preparations in vitro, experiments were performed to investigate if asterotocin also triggers cardiac stomach eversion when injected in vivo. First, a pilot experiment was performed using 30 starfish (diameter 6.4–7.5 cm) that had been starved for 1 week to normalise their physiological status. Then animals were injected with different doses of asterotocin (10 μl of 10−6–10−3 M; n = 5 for each dose) or with 10 μl of water (negative control; n = 5) or with 10 μl of 10−3 M S2 (positive control; n = 5). The animals were injected at a site located in the aboral body wall of an arm proximal to the junction with the central disc and adjacent to the madreporite. Care was taken to ensure that the tip of a Hamilton syringe used for injections was pushed through the body wall into the perivisceral coelom but not too deep so as to avoid injecting into the digestive organs. The doses of asterotocin injected were informed by our analysis of the volume of the perivisceral coelomic fluid in A. rubens. Thus, injection of 10 μl of 10−6–10−3 M asterotocin was estimated to achieve concentrations in the perivisceral coelom of ~ 10−9, 10−8, 10−7 and 10−6 M, respectively, which are the concentrations at which asterotocin was found to be effective as a cardiac stomach relaxant when tested in vitro. Following injection, the starfish were placed in a glass vessel containing seawater so that the mouth of the animal could be observed from below.
Having established that asterotocin induced cardiac stomach eversion in all of the animals injected with 10 μl 10−4 M or 10 μl 10−3 M asterotocin, an experiment was performed to examine the time course of asterotocin-induced cardiac stomach eversion in A. rubens. For this experiment, 20 adult specimens of A. rubens (13.1–17.1 cm in diameter) were selected and starved for 1 week before the experiment. Then, the animals were injected with 10 μl of distilled water or 10 μl of 10−3 M synthetic asterotocin. To enable observation of cardiac stomach eversion, following injection, starfishes were placed individually in a petri dish containing 90 ml of artificial seawater. A petri dish was used as it has a shallow depth, which prevented the starfish from climbing vertically and therefore enabling recording of the whole oral surface of the starfish over time using a video camera (Canon EOS 700D) positioned beneath the petri dish. The oral surface of starfish was video recorded for 15 min, and static images from the video recordings were captured at 30-s intervals from the time of injection to 10 min post-injection, during which time maximal stomach eversion was observed. The two-dimensional area of the everted cardiac stomach was measured from the images using the ImageJ software and normalised as a percentage of the area of the central disc of the animal, which was calculated as the area of a circle linking the junctions between the five arms.
While examining the effect of asterotocin in inducing cardiac stomach eversion in A. rubens, it was also observed that asterotocin induced arm flexion and/or a “humped” posture resembling the posture that starfish have when feeding on prey. To investigate the influence of this effect of asterotocin on whole-animal behaviour in starfish, experiments were performed to compare the righting behaviour of asterotocin-injected animals with non-injected, water-injected and S2-injected animals. Starfish righting behaviour occurs if they are upturned so that their underside (oral surface) is uppermost. One or more of the arms or rays then twist (active rays) until the tube feet can make contact with and adhere to the floor surface. Then, the other inactive rays and the central disc are flipped over in a somersault-like manoeuvre to bring the entire oral surface back into contact with the floor surface [37, 74]. For an experiment investigating if asterotocin affects the righting behaviour of A. rubens, 20 adult animals (10.2–14.9 cm in diameter) were first starved for 1 week prior to the experiment to normalise their physiological status. First, the righting behaviour of starfish without injection was observed, and if righting behaviour did not occur or took more than 5 min, the animal was categorised as unhealthy and unsuitable for the experiment. Following a 30-min recovery period, starfish were either injected with 10 μl of 10−3 M asterotocin (n = 10) or 10 μl of distilled water (n = 10). Then, after 5 min, each starfish was placed with their oral side lowermost in a large glass container fully immersed in artificial seawater for 10 min, after which they were turned upside down with their oral side uppermost and the time taken to right was measured. For consistency, the time taken for righting was determined by noting when the central disc and all five arms touched either the bottom or the side of the glass tank.
To examine the specificity of the effects of asterotocin on starfish righting behaviour, an experiment was performed where the righting time of starfish (5.5–9.6 cm in diameter; starved for 4 weeks prior to testing) was measured before injection (n = 60; pooled from the three treatment groups) and after injection with 5 μl 10−3 M asterotocin (n = 20) or 5 μl 10−3 M S2 (n = 20) or 5 μl distilled water (n = 20). The time taken to right in seconds as well as the percentage difference between righting with and without injection was calculated. The percentage time difference between righting with and without injection in water-injected, asterotocin-injected and S2-injected animals was determined and plotted as a separated box and whiskers graph in Prism 6 (GraphPad 6). The effect of neuropeptides on righting behaviour was analysed statistically using the Mann-Whitney U test or a one-way ANOVA test with a post hoc Dunnett’s multiple comparisons test.