Evidence for Gonadotropin-Releasing Hormone-Like Peptides in a Cnidarian Nervous System

Home , Renilla, Renilla koellikeri, Sea pansy

General and Comparative Endocrinology 119, 317–328 (2000) doi:10.1006/gcen.2000.7524, available online at http://www.idealibrary.com on

Evidence for Gonadotropin-Releasing Hormone-like Peptides in a Cnidarian Nervous System

Michel Anctil De´partement de Sciences Biologiques and Centre de Recherche en Sciences Neurologiques, Universite´ de Montre´al, Montre´al, Que´bec, Canada H3C 3J7; and Whitney Laboratory, University of Florida, St. Augustine, Florida 32086

Accepted May 25, 2000

There is increasing evidence that peptides of the gonado- resented in invertebrates by native forms (De Loof et tropin-releasing hormone (GnRH) family, long consid- al., 1989), suggesting that peptide families are con- ered a vertebrate preserve, are also present in inverte- served through evolutionary lineages from invertebrate (molluscan) nervous systems. The possibility was brates to vertebrates. This conservation may also exist examined that GnRHs are present and bioactive in cni- in the lowest phylum (Cnidaria) in which neurons are darians, considered to be representatives of the most identifiable. For example, neurons of several cnidarian primitive animals possessing a nervous system. Immuno- species contain peptides belonging to the FMRFamide reactive GnRH was detected in endodermal neurons of family (Grimmelikhuijzen et al., 1992; Grimmelikhui- two anthozoans, the sea pansy Renilla koellikeri and the jzen and Westfall, 1995), all but one (Yasuda et al., sea anemone Nematostella vectensis. In the sea pansy, 1993) unique to this phylum. immunoreactivity was detected throughout the auto- The gonadotropin-releasing hormones (GnRHs) zooid polyps, including gamete-producing endoderm. constitute a peptide family whose members are re- High-performance liquid chromatography and radioim- markably conserved in length and composition of munoassays of extracts from whole sea pansy colonies amino acid sequences and appear to have a distribu- yielded two elution peaks exhibiting GnRH immunoreac- tion restricted to vertebrates (Sherwood, 1987). The tivity with antisera raised against shark or mammalian possibility that GnRHs are also represented in inver- GnRH. Vertebrate GnRHs as well as the two sea pansy tebrates is supported by evidence showing that GnRH-like factors inhibited the amplitude and frequency GnRH-like immunoreactive material is present in neu- of peristaltic contractions in the sea pansy, and these rons of the gastropod mollusc Helisoma trivolvis and 1 actions were blocked by the LHRH analog [D-pGlu ,D- that extracts from its nervous system induced electro- 2 3,6 Phe ,D-Trp ]-LHRH. These results suggest that the physiological responses in Helisoma neurons and go- GnRH family of neuropeptides is more widespread in nadotropin release from dispersed goldfish pituitary metazoans than previously thought. Although our physi- cells (Goldberg et al., 1993). In addition, GnRH-like ological data are preliminary, they point to a role for immunoreactive material was detected in several GnRHs as inhibitory modulators of neuromuscular trans- protochordates (Georges and Dubois, 1980; Dufour mission in the sea pansy. © 2000 Academic Press et al., 1988; Kelsall et al., 1990; Mackie, 1995; Cameron Key Words: Cnidaria; GnRH; neuropeptide; sea pansy; et al., 1999) and two new peptides of the GnRH molecular evolution. family were identified in the tunicate Chelyosoma productum (Powell et al., 1996), thus demonstrating that Several of the families of structurally related neu- vertebrate GnRHs had ancestors in the protochordate ropeptides first identified in vertebrates are also rep- lineages.

This paper addresses whether GnRH-like peptides diluted (30%) local sea water at the Whitney Laboratory. occur in invertebrates other than molluscs and tuni- They were fed three times weekly with brine shrimps, at cates and, focusing on cnidarians, whether the evolu- which times their culture medium was changed. tionary history of the GnRH family is traceable to the Immunohistochemistry. Ten sea pansy colonies earliest invertebrates known to possess a nervous sys- were prepared for whole-mount immunohistochemis- tem. The primary cnidarian used to investigate these try. Sea pansies were anesthetized in equal parts of questions was the sea pansy, Renilla koellikeri, an octo- 0.37 M MgCl2 and ASW. Autozooid polyps and their corallian belonging to the sea pen group (Pennatula- recessed reproductive tissue inside the colonial mass cea). The organization of the pennatulacean nervous (rachis) were excised and immersed in Zamboni’s fix- system is reasonably well known (Buisson, 1970; Sat- ative (4% paraformaldehyde and 7.5% saturated picric terlie et al., 1980), and the sea pansy was the focus of acid) freshly prepared in 0.1 M phosphate buffer with extensive mapping of neurotransmitter-specific neu- 2.4% NaCl (PBS) at pH 7.4. Fixation proceeded at rons (Umbriaco et al., 1990; Pani et al., 1995; Anctil and room temperature for 1 h and at 4° for the following Ngo Minh, 1997; Mechawar and Anctil, 1997). In ad- 24 h. After washing in PBS, tissues were incubated for dition to classical neurotransmitters, the sea pansy 48 h at 4° in either of three different primary antisera: contains large amounts of the peptide Antho-RFamide an antiserum raised in rabbit against chicken GnRH-II which is widely distributed in its nervous system and graciously provided by Dr. J. P. Chang, University (Grimmelikhuijzen and Groeger, 1987). of Alberta (Goldberg et al., 1993) and two other anti- A combination of immunohistochemical and chro- sera raised in rabbit against LHRH, one from a com- matographic techniques was used to gain evidence for mercial source (Sigma Chemicals) and the other ob- the presence of GnRH-like peptides in the sea pansy. tained as a courtesy from Dr. Robert Benoıˆt (LR-1: In addition, the physiological function of GnRHs was Benoıˆt et al., 1987). All three antisera were diluted investigated by comparing the activity of synthetic 1:1000 in PBS with 1% normal goat serum or 1% GnRHs with chromatographically purified sea pansy bovine serum albumin and 0.3% Triton X-100. After GnRH-like fractions on sea pansy muscle activity. Fi- further washing in PBS, tissues were treated for 1 h nally, confirmatory immunohistochemical data were with Cy3-conjugated goat anti-rabbit IgG (Jackson Im- obtained in another anthozoan species, the starlet sea munochemicals) diluted 1:100 in PBS containing 0.3% anemone Nematostella vectensis. Together, the results Triton X-100. After a final wash, tissues were mounted from these approaches provide strong evidence for the in a 1:3 mixture of glycerol and PBS and examined presence and physiological relevance of GnRHs in with a Wild-Leitz Dialux 20 fluorescence microscope. Cnidaria. Preliminary results of this study were pre- Controls were performed by replacing the primary sented to the XIIIth International Congress of Com- antisera with nonimmunized serum and by preab- parative Endocrinology (Anctil, 1997). sorption of the primary antisera with 0.1 mM LHRH or chicken GnRH-II. The double-labeling immunohistochemistry proce- MATERIALS AND METHODS dure of Wu¨ rden and Homberg (1993) was used to determine whether FMRFamide-like immunoreactivity (Fair), reflecting the presence of native Antho-RF- Colonies of sea pansy (Renilla koellikeri Pfeffer) were amide, colocalizes with GnRH-like immunoreactivity purchased from Marinus Inc. (Long Beach, CA) and (GnRHir) in the same cells. Tissues were treated for maintained in aquaria with recirculating, filtered arti- 48 h in the LR-1 LHRH antiserum diluted 1:1000 fol- ficial sea water (ASW, Instant Ocean) at 12–16° and pH lowed by the Cy3-conjugated secondary antiserum as 8.0. The unfed colonies were exposed to a light–dark previously described. After washing in three 1-h cycle according to the photoperiod length experienced changes of rabbit IgG (Jackson Immunochemicals) di- in their environment before shipment and were pro- luted 1:25 in PBS, the tissues were treated for 24 h at 4° cessed within 2 weeks of shipment. in a biotinylated FMRFamide antiserum (also raised in Starlet sea anemones (Nematostella vectensis) were rabbit, Incstar), diluted 1:800 in PBS. The FMRFamide maintained in culture dishes at room temperature in antiserum was biotinylated by the succinimide ester

Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. GnRH-like Hormones in a Cnidarian 319 method of Harlow and Lane (1986). Tissues were then noassay (RIA) in some of the fractions, more extract incubated for3hinstreptavidin-conjugated fluores- was subsequently processed by HPLC to select those cein isothiocyanate (FITC, Jackson Immunochemicals) specific fractions and pool them for further purifica- diluted 1:20. tion steps. Starlet sea anemones were processed similarly to sea The selected fractions were run as above but on a pansies except that anesthesia was performed in equal smaller column (Brownlee RP-300, 2.1 ϫ 220 mm) at mixtures of 30% sea water and 0.12 M MgCl2, and PBS 0.5 ml/min, and a stronger ion-pairing agent, hepta- was prepared with 0.8% NaCl to reflect the salinity con- fluoro-butyric acid (HFBA) was substituted for TFA at ditions under which the animals were cultured. A dilu- the same dilution. Three successive purification steps tion of 1:500 of the primary LHRH antiserum (Sigma) were performed with this procedure. was used, and either Cy3 (1:100) or FITC (1:80) was used Radioimmunoassay. Iodination and RIA protocols as fluorophore conjugated to secondary antibodies. followed Price et al. (1987). Two iodinated GnRH pep- Peptide extraction. Whole sea pansies were ex- tides, mammalian GnRH and salmon GnRH, were used tracted during the reproductive season (May–August) in the RIA methods tested. Iodination was performed by by the method of Goldberg et al. (1993) after modifi- mixing 2–3 ␮lof125I (1–1.5 mCi) with 1 ␮l of peptide cations. Sea pansies were deflated to remove residual solution (1 mM in 0.5 M phosphate buffer, pH 7.0), sea water from their gastrovascular channels, followed by addition of 5 ␮l of 0.2% chloramine T and weighed, and immediately frozen in liquid nitrogen. 100 ␮l of 0.5% sodium metabisulfite prepared in the The frozen colonies were rapidly crushed in a coffee same buffer. After mixing, the iodination mixture was grinder and the resulting slush was added to a mix- retained in C18 SepPak cartridges (Waters) and eluted ture of acetone and1NHCl(100:3) on wet ice. Extrac- with a solution of 80% ACN and 0.1% TFA. For assays, tion proceeded for 3–4 h with continuous shaking at a dilution of this trace on the order of 1:1000 was made 4°. After separating the soluble extract by filtration, into RIA buffer (0.01 M disodium phosphate, 0.9% NaCl, the residue was further extracted in a 4:1 mixture of 0.73% EDTA, 0.01% Thimerosal, and 1% bovine serum acetone and 0.01 N HCl. The combined filtrates were albumin at pH 7.5) to yield 100,000–130,000 cpm/ml. washed five times by adding petroleum ether to the This diluted trace was stable for at least 1 week at 4°. filtrate in the ratio 1:4 to remove hydrophobic materi- Several antisera were used in attempts to optimize als. The final aqueous phase was stored frozen at Ϫ80° RIA sensitivity. Four were graciously provided by Dr. until used. Just before use, the extracts were concen- Nancy Sherwood, University of Victoria (British Colum- trated in a rotary vacuum evaporator and centrifuged bia): FP-3, a salmon GnRH antiserum; 7CR-10, a dogfish at 10,000 rpm and 10° for 10 min to remove particu- GnRH antiserum; BLA-5, a lamprey GnRH-I antiserum lates. A total of 120 g of tissue from 30 colonies were (Kelsall et al., 1990); and Jasmine-6, a tunicate GnRH-I thus extracted to a final volume of 40 ml. antiserum (Powell et al. 1996); all raised in rabbit. The Extract purification. Extracts were first loaded mammalian GnRH (LHRH) antiserum was obtained onto a reverse-phase Brownlee RP-300 column (4.6 ϫ from Peninsula Laboratories. The most sensitive assay 220 mm, 5-␮m particles) at 2 ml/min using a Gilson method was obtained with mammalian GnRH trace, 714 HPLC system. The column was loaded with an antiserum (1:75000), and standard (mammalian RIA), extract from the worm Urechis to ascertain that no yielding a detection limit of 12.6 pg (B/Bo ϭ 80%). The immunoreactive GnRH was retained from previous only other workable method used salmon GnRH trace loadings of GnRH standards (LHRH, chicken GnRH- and standard with dogfish GnRH antiserum (1:10000) II, and salmon GnRH; all from Peninsula Laboratories) (fish RIA), allowing a detection limit of 51.6 pg. or Renilla extracts. Fractionation of extracts was by a The assay reaction was started by adding 100 ␮lof linear gradient of 10 to 50% acetonitrile (ACN) in trace and 100 ␮l of antiserum to 50 ␮l of buffer (Bo), 0.05% trifluoroacetic acid (TFA) over 40 min. Fractions GnRH standard, or HPLC fraction and allowed to were collected at 30-s intervals, dried in a SpeedVac equilibrate at 4° overnight. One milliliter of charcoal concentrator (Savant), and reconstituted with 50 ␮lof slurry (0.25% activated charcoal and 0.025% Dextran radioimmunoassay buffer (see below). After immuno- with 0.01% merthiolate in phosphate buffer, pH 7.4– reactive GnRH (irGnRH) was detected by radioimmu- 7.6) was then added to assay tubes and let stand for 10

Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. 320 Michel Anctil min. Assay tubes were centrifuged at 3000g and 4° for 15 min and decanted in radiocounting vials. Bioassay. Triangular pieces of rachis bearing several autozooids were excised from unanesthetized sea pansies, cutting from the margin of the colony to the point of emergence of the pedoncle into the rachis. The pieces were pinned on the Sylgard-coated bottom of a recording chamber and attached to a Grass FT-03C force displacement transducer at a wedge end. Isomet- ric tension signals were transmitted to a Grass 7P122 amplifier and digitally stored for analysis with either a BIOPAC MP100WS (Goleta, CA) or a Grass Instru- ments PolyView (Astro-Med Inc., Longueuil, Canada) data acquisition and analysis system. The preparations were allowed to relax tension for 0.5–2 h in a fresh sea water bath before beginning experiments. Synthetic peptides or purified sea pansy extracts were delivered to the chamber with a syringe at a concentration 10 times higher than the final concentration in the bath. The same volume of ASW containing reconstituted HPLC fractions without irGnRH was similarly delivered as additional control. Baths were evacuated by gravity through an outlet at the bottom of the chamber, and fresh sea water was almost simulta- neously added from the top of the chamber to main- tain fluid level. All experiments were conducted at room temperature (20–24°). Experiments began by recording tension activity at least 30 min after addition of normal sea water (con- FIG. 1. GnRH-immunofluorescent neurons shown in whole- trols). Peptides or extracts were then introduced and mount preparations of polyps from the sea pansy. (A) Reactive cell tension was recorded for another 30 min or more. The bodies (short arrows) and varicose neurites (long arrow) are present bath was evacuated and entirely replaced with fresh both in the trunk and in the pinnules (pin) of tentacles. (B) Close-up of tentacle oriented to show reactive neurites extending in the sea water, and the preparation left to stand for1hto endoderm (thin arrows), below the ectoderm (Ect). Note also neurite let tension return to the initial baseline before measur- extending through the ectoderm (thick arrow) and ending near the ing tension again for a minimum of 30 min (recovery). surface with a small terminal swelling. (C) Reactive multipolar Each contraction assay preparation was used once. neuron in column of polyp showing cell body (arrow) and several Both amplitude and rate of several consecutive peri- neurites oriented mostly circumferentially along the column. staltic waves were measured for each experimental condition in time bins of 30 min. Data were averaged and normalized as percentage of controls. proved to be the most sensitive and provided the most consistent staining. The immunofluorescent cells were exclusively located in the endodermal layer and were RESULTS identified as neurons according to conventional mor- phological criteria as applied to cnidarians (Saripalli and Westfall, 1996). No GnRH-like reactivity was Localization of GnRH-immunoreactive neurons. present in control preparations. Although all three GnRH antisera used in sea pansy Specific GnRH-immunoreactive (GnRHir) neurons tissues gave positive results, the antiserum LR-1 were detected throughout the autozooid polyps, with

FIG. 2. Whole-mounts of GnRH-immunofluorescent neurons in the zooecium of polyps, deep inside the rachis of the sea pansy. (A) Strongly fluorescent cell bodies in the zooecial wall, with their neurites running parallel to each other. (B) One neuron as in A is seen to span the endoderm layer of the zooecial wall, from the cell body bordering a gastrovascular cavity (GVC) across the wall with an axon-like neurite (arrow) that bifurcates in the circular muscle layer. (C) In septal filaments arising from the zooecial wall, neurons are polymorphous, with rugose neurites coursing jaggedly in various orientations. Their cell bodies are ciliated (small arrows). (D) Group of intertwined neurons at the distal end of a septal filament to which is attached an ovocyte (arrow).

greater density in the tentacles (Fig. 1A). In the latter appeared to lie over the circular muscle layer of the they were bipolar with small round somata (7–8 ␮m) body wall. and coarse varicose neurites ϳ150 ␮m in length. Var- The eight gastrovascular cavities of the autozooid icosities were prominent but varied greatly in size column extend beyond the pharynx within the rachis within neurites at the endoderm/mesoglea boundary where they form the zooecium. Numerous bipolar and (Fig. 1B). In the trunk of the tentacles, many neurons multipolar GnRHir neurons were detected in the walls appeared to be incompletely stained and to run in a of these cavities (Fig. 2A). Their oblong somata, circular orientation, whereas neurites in the pinnules ϳ14–18 ␮m, extend from the external surface of the branching from the trunk were better stained and myoepithelium inward where they taper to a neurite could be seen to follow a course perpendicular to the bifurcating near the myoepithelium/mesoglea border long axis of the trunk (Fig. 1A). Some neurites branch (Fig. 2B) at distances up to 150 ␮m. At their apex they off into the ectodermal layer but do not appear to appear to bear a cilium (Fig. 2C). Neurites of the reach the surface (Fig. 1B). bipolar neurons tend to run parallel to each other (Fig. GnRHir neurons in the wall of the autozooid col- 2A), whereas those of multipolar neurons appear in- umn appeared to be fewer than in tentacles. The shape tertwined (Fig. 2C). The neurites are slightly varicose and size of their somata were similar to those in ten- and exhibit terminal knobs (Fig. 2C, large arrow). tacles. They were bi- or multipolar, with branching GnRHir neurons were also present in mesenteric ﬁla- neurites oriented primarily along the circular axis of ments bearing gametes (Fig. 2D). Their somata resem- the column (Fig. 1C). These neurites had smaller di- ble those observed in the wall of the zooecium, but ameters and varicosities than those in tentacles and their neurites appear to be shorter.

In the starlet sea anemone, GnRHir neurons were also confined to the endoderm. They were sparsely distributed along the mesenteries of the upper region of the column, near the pharynx (Fig. 3C). Their somata had an oblong shape, were 10–12 ␮m in length, and appeared to be either bipolar or tripolar (Fig. 3D). Their neurites were very thin and relatively smooth, exhibiting little or no varicose swellings, and they appeared to lie over a muscle layer as in the sea pansy. GnRHir neurons, usually in clusters, were also present in mesenteric filaments bearing gametes (Fig. 3E). These differed from the stained neurons in the upper column in that their neurites were conspicously varicose, some of which reached to the cortex of ovocytes (Fig. 3E). Characterization of sea pansy GnRH-like factors. Preliminary HPLC separations of sea pansy extracts (representing 2.6 g of tissue) using the 4.6 ϫ 220 mm column and the mammalian RIA (see Materials and Methods) resulted in the detection of irGnRH in two fraction clusters, 38–40 (19–20 min elution) and 41–43 (20.5–22.5 min elution). None of these fractions coeluted with mammalian GnRH or chicken GnRH-II, but salmon GnRH coeluted with fraction 41. Subsequent separations with larger amounts of extracts (8 g of tissue) allowed the detection of irGnRH also in fractions 38–40 using the fish RIA. The last separation was done with an extract representing 60 g of tissue. From this extract fractions 38–39 (early GnRH-like peak; 19–19.5 min elution) and 42–44 (late GnRH-like peak; 21–22 min elution), con- FIG. 3. Whole-mounts of GnRH-immunoreactive (GnRHir) and FMRF- taining the largest amount of irGnRH determined by amide-immunoreactive (Fair) neurons in the endoderm of the sea pansy (A,B) and starlet sea anemone (C–E). (A,B) GnRHir neurons in the zooecial both mammalian and fish RIAs, were selected for ϫ wall (A) display no Fair (B) where instead other neurons are labeled. further purification using a 2.1 220 mm column, a Arrows point to corresponding area in both panels. (C) Low-power view mobile phase containing HFBA instead of TFA, and a of upper column of a starlet sea anemone in which a few sparsely distrib- flow rate of 0.5 ml/min (Fig. 4). The fractions contain- uted neurons are seen. (D) Higher magnification of neurons in the upper ing the highest irGnRH from each peak were fractions column of a starlet sea anemone. Note the bifurcated neurite in one such neuron (arrow). (E) Small dense groups of neurons (large arrows) are 65–67 of early peak (32.5–33.5 min elution) and frac- found in this mesenteric filament of the starlet sea anemone that impinges tions 68–69 of late peak (34–34.5 min elution); these on an ovocyte. Note the varicose neurites that extend from these group- were passed again through the same HPLC column ings toward the follicle epithelium of the ovocyte (small arrows). and resulted in the elution profiles shown in Fig. 5. The major peaks for the two GnRH-like factors eluted Although Fair neurons were present throughout the at 23.4 (factor 1) and 25.3 min (factor 2). Contrary to endoderm, double-immunofluorescence experiments earlier purification steps, none coincided with the elu- failed to reveal colocalization of Fair and GnRHir in tion peak of salmon GnRH (Fig. 5). This immunoreac- the same cells. Figures 3A and 3B provide such an tive material was used in the following bioactivity example in the wall of the zooecium. assays.

FIG. 4. GnRH immunoreactivity in HPLC eluates from two fraction clusters obtained in a previous purification step of sea pansy extracts (see Materials and Methods). The top panel represents the eluate profile of fractions 38–39 (early GnRH-like activity) and the bottom panel fractions 42–44 (late GnRH-like activity) from the earlier purification step. The straight line represents the acetonitrile gradient of 10 to 50% in the mobile phase. Absorbance is shown at 210 nm (solid trace) and at 280 nm (dotted trace).

Effects of GnRHs and sea pansy GnRH-like extracts LHRH was 10 times more potent than chicken on sea pansy peristalsis. Both LHRH and chicken GnRH-II in eliciting these responses at a threshold Ϫ GnRH-II caused a reduction of the amplitude and concentration of 10 7 M and maximal responses at Ϫ frequency of rachidial peristaltic waves in a dose- 10 4 M (Fig. 7). The responses appeared 2–3 min after dependent and reversible manner (Figs. 6 and 7). peptide addition to the bath and vanished 10–15 min

FIG. 5. HPLC chromatograms of the two GnRH-like factors semipuriﬁed from immunoreactive fractions shown in Fig. 4. Note elution times for mammalian (M), chicken (C), and salmon (S) GnRHs. See Fig. 4 for other explanations.

after peptide removal. The LHRH peptide antagonist and antagonist-sensitive manners (Figs. 6 and 8). In [D-pGlu1,D-Phe2,D-Trp3,6]-LHRH often enhanced the contrast to the responses to authentic GnRHs, those of amplitude of peristaltic contractions while blocking the sea pansy factors were of shorter duration, waning the response to LHRH (Fig. 8). before the extracts were removed from the bath The lyophilized sea pansy factors 1 and 2 corre- (Fig. 6). sponding to the HPLC peaks shown in Fig. 5 were reconstituted each in 300 ␮l sea water (representing 0.9 and 0.6 nmol of equivalent LHRH activity, respec- DISCUSSION tively) and tested with the sea pansy peristalsis bioassay. Both factors produced effects qualitatively and quantitatively similar to those of authentic GnRHs, In this study evidence for the presence of GnRH-like reducing peristaltic contractions in dose-dependent peptides in cnidarians was obtained by visualizing

sues. Apart from the innervation of gonad tissues, the density of stained neurons was greater in the upper than in the lower column of the individuals: in the tentacles of the sea pansy or at the pharynx level in the starlet anemone. This is in contrast to other cnidarian peptidergic neurons such as Fair neurons, which are more widely distributed over the whole animal and in all tissue layers (Grimmelikhuijzen et al., 1996). Al- though Fair neurons are present in areas where the sea pansy GnRHir neurons are located, no evidence of colocalization of GnRHir and Fair in the same neurons was revealed by double-labeling. Thus, the GnRHir neurons constitute a discrete subset of neurons which appear to serve specific functions based on the association of their neurites with body wall muscles and with gamete-bearing tissues (see below). These neurons exhibit a few intriguing morpholog- ical features. In the sea pansy, GnRHir neurons in the tentacles were observed to project a neurite in the ectoderm. As the neurite does not appear to reach the FIG. 6. Effects of GnRH peptides on peristaltic contractions of a outer surface, it is doubtful that it serves a direct sea pansy rachis. Arrows indicate time of addition of peptides. Factor 1 refers to reconstituted sea pansy factor 1 (see Fig. 5). Ant., sensory role. However, it is possible that these neu- LHRH antagonist [D-pGlu1,D-Phe2,D-Trp3,6]-LHRH. Note that con- rites connect with sensory cells known to be abundant tractions are multiphasic. in the tentacle ectoderm (Lyke, 1965; Umbriaco et al., 1990). Many GnRHir neurons possess a cilium. The apex of the somata of these neurons reaches the gas- neurons in two anthozoan species with three different trovascular cavity so that the cilium may assist in GnRH antisera, by separating GnRH-like factors in sea detecting chemical or mechanical signals in the cavity pansy extracts using HPLC–RIA methods, and by as- fluids. Furthermore, some of the neurons in which saying the extracted GnRH-like factors in a sea pansy neurites lack conspicuous varicosities possess large bioassay in which peristaltic contractions are inhibited by GnRHs. Along with reports of GnRH-like neuropeptides and GnRH-like physiological activity in mollusks (Goldberg et al., 1993; Young et al., 1999; Pazos and Mathieu, 1999) and tunicates (Powell et al., 1996; Craig et al., 1997), this evidence points not only to the existence and functional involvement of peptides of the GnRH family in invertebrates, but also to their ancestral association with representatives of the earliest nervous systems to have arisen in metazoans. In both the sea pansy and the starlet anemone GnRH-like immunoreactivity was associated exclusively with a population of neurons that share simi- larities of morphology, distribution, and innervation targets. All stained neurons were either bi- or multipolar with more or less varicose neurites extending in FIG. 7. Dose-response curves for the reducing effect of LHRH, the endoderm either over parietal muscle sheets or chicken GnRH-II, and LHRH ϩ antagonist (10 ␮M) on sea pansy over mesenteric filaments associated with gonad tis- peristalsis (n ϭ 6). Data points are mean Ϯ SE.

FIG. 8. Histogram summarizing data on the effect of LHRH and sea pansy GnRH-like factors on peristaltic contractions of the sea pansy. Bar values are mean Ϯ SE. Asterisks indicate signiﬁcant differences compared with controls at P Ͻ 0.05 (*) or P Ͻ 0.01 (**) using paired t test. terminal swellings. Such an arrangement suggests an Although the available material was insufﬁcient for endocrine mode of neurosecretion in which neurons such purposes, it allowed testing their effect in a pre- release large amounts of peptide at one site, from viously developed sea pansy bioassay for synthetic which it diffuses widely in surrounding tissues. GnRHs (Anctil, 1997). Sea pansy colonies are traversed The characterization of two GnRH-like factors from by peristaltic waves of contraction every few minutes sea pansy extracts by HPLC–RIA provides supporting (Parker, 1920). These waves are modulated by indola- evidence that the GnRHir neurons described above mines, serotonin enhancing and melatonin reducing contain peptides of the GnRH family. The character- the amplitude of the contractions (Anctil, 1989; Anctil ization was validated by developing two different et al., 1991). As shown here, both sea pansy GnRH-like RIAs which consistently detected immunoreactive factors reduced the amplitude of peristaltic waves in GnRHs in the same elution fractions and by consistent the same manner as synthetic GnRHs (LHRH and detection of GnRHir materials upon repeated runs of chicken GnRH-II), including sensitivity to the block- elutions in different HPLC columns. None of the two ing action of the analog [D-pGlu1,D-Phe2,D-Trp3,6]- characterized factors coeluted with either LHRH, LHRH. The only departure from the response to syn- chicken GnRH-II, or salmon GnRH, contrary to ex- thetic GnRHs is the tendency of the response to the sea tracts from the gastropod Helisoma which coeluted pansy factors to wane faster. Although the reason for with mammalian GnRH (Goldberg et al., 1993). The this discrepancy is unclear, one possibility is that the detection of two forms of GnRH-like factors in the sea sea pansy factors may lack protective groups like the pansy follows a pattern that is common in vertebrate pyroglutamate or amidated C-terminus possessed by species. In several vertebrates the highly conserved the vertebrate GnRHs, thereby becoming more sus- chicken GnRH-II is often the second GnRH form and ceptible to degradation by peptidases. where present the two forms appear to serve different These results suggest that native sea pansy GnRHs functions (Sherwood et al., 1993; King and Millar, are physiologically active and play a role in the mod- 1995). ulation of peristalsis. GnRHs are primarily known as It will be desirable to determine the molecular struc- regulators of reproduction in vertebrates, although ture of the two sea pansy factors to assess their phy- other roles were suggested (Sherwood et al., 1993; logenetic divergence from known GnRHs and gain King and Millar, 1995). In prevertebrate chordates a insights into the early history of this peptide family. role in reproduction is suggested by the stimulatory

Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. GnRH-like Hormones in a Cnidarian 327 effect of native and other GnRH forms on estradiol REFERENCES levels in gonads of the tunicate Chelyosoma productum (Craig et al., 1997). There is some evidence to suggest that GnRH-like peptides are also involved in control Anctil, M. (1989). Modulation of a rhythmic activity by serotonin via cyclic AMP in the coelenterate Renilla ko¨llikeri. J. Comp. Physiol. B of reproduction in invertebrates. In cephalopod and 159, 491–500. gastropod molluscs GnRH immunoreactivity was re- Anctil, M. (1997). Gonadotropin releasing hormone-like factor in a ported to be associated with the innervation of either primitive metazoan: Immunohistochemical evidence and bioac- reproductive organs or their brain center (Di Cosmo tivity. In “Advances in Comparative Endocrinology” (S. Ka- and Di Cristo, 1998; Young et al., 1999) and synthetic washima and S. Kikuyama, Eds.), pp. 703–707. Monduzzi Editore, Bologna. GnRHs promoted gametogenesis in a bivalve mollusc Anctil, M., and Ngo Minh, C. (1997). Neuronal and nonneuronal (Pazos and Mathieu, 1999). In contrast, no GnRH immu- taurine-like immunoreactivity in the sea pansy, Renilla koellikeri noreactivity was found in brain extracts of an arthropod (Cnidaria, Anthozoa). Cell Tissue Res. 288, 127–134. (fleshfly) by radioimmunoassay (De Loof et al., 1995). Anctil, M., Pani, A. K., and Ali, M. A. (1991). Modulation of rhyth- In cnidarians the evidence for a role of GnRH-like mic contractions by melatonin via cyclic GMP in the coelenterate peptides in reproduction is less compelling. Experi- Renilla koellikeri. J. Comp. Physiol. B 161, 569–575. Benoit, R., Ling, N., Brazeau, P., Lavielle, S., and Guillemin, R. ments with ovary extracts from the jellyfish Spirocodon (1987). Strategies for antibody production and radioimmunoas- saltatrix led Ikegami et al. (1978) to suggest that an says. In “Neuromethods” Vol. 6 “Peptides” (A. A. Boulton, G. B. unidentified peptide of low molecular weight was Baker, and Q. J. Pittman, Eds.), pp. 43–72. Humana Press, Clifton, involved in oocyte maturation and spawning. Now NJ. the finding that GnRHir neurons innervate the mes- Buisson, B. (1970). Les supports morphologiques de l’inte´gration dans la colonie de Veretillum cynomorium Pall. (Cnidaria, Pennatu- enteric filaments bearing gametes in both the sea laria). Z. Morph. Tiere 68, 1–36. pansy and the starlet sea anemone provides support Cameron, C. B., Mackie, G. O., Powell, J. F. F., Lescheid, D. W., and for such an involvement by native GnRHs and a Sherwood, N. M. (1999). Gonadotropin-releasing hormone in framework for investigating their roles in reproduc- mulberry cells of Saccoglossus and Ptychodera (Hemichordata: En- tive functions. The present study shows that peristal- teropneusta). Gen. Comp. Endocrinol. 114, 2–10. sis, which is strongly enhanced during spawning in Craig, A. G., Fischer, W. H., Park, M., Rivier, J., Musselman, B. D., Powell, J. F. F., Reska-Skinner, S. M., Prakash, M. O., Mackie, the sea pansy (Satterlie and Case, 1979), is inhibited by G. O., and Sherwood, N. M. (1997). Sequence of two gonadotropin native and other GnRHs. It may be that sea pansy releasing hormones from tunicates suggest an important role of GnRHs are not involved, or are involved negatively, conformation in receptor activation. FEBS Lett. 413, 215–225. in spawning. Repression of peristalsis by GnRHs may De Loof, A., and Schoofs, L. (1989). Review: Homologies between serve another, as yet unknown, function, but an alter- the amino acid sequences of some vertebrate type peptide hormones and peptides isolated from invertebrate source. Comp. native reproductive role, such as in gametogenesis, Biochem. Physiol. B 95, 459–468. cannot be ruled out at present. De Loof, A., Bylemans, D., Schoofs, L., Janssen, I., Spittaels, K., Broeck, J. V., Huybrechts, R., Borovsky, D., Hua, U.-J., Koolman, J., and Sower, S. (1995). Folliculostatins, gonadotropins and a model for control of growth in the grey fleshfly, Neobellieria (Sar- cophaga) bullata. Insect Biochem. Mol. Biol. 25, 661–667. ACKNOWLEDGMENTS Di Cosmo, A., and Di Cristo, C. (1998). Neuropeptidergic control of the optic gland of Octopus vulgaris: FMRF-amide and GnRH immunoreactivity. J. Comp. Neurol. 398, 1–12. I thank Dr. Nancy Sherwood (University of Victoria, B.C., Can- Dufour, S., Monniot, F., Monniot, C., Baloche, S., Kerdelhue, B., and ada) for her generous gift of the FP-3, 7CR-10, BLA-5, and Jasmine-6 Fontaine, Y. A. (1988). Dosage radioimmunologique, chez Asci- antisera. Dr. Peter A. V. Anderson, Director of the Whitney Labo- diella aspersa, d’un facteur de type gonadolibe´rine (GnRH) de ratory and Dr. Michael J. Greenberg kindly provided laboratory poids molećulaire apparent supe´rieur a` celui du dećapeptide spaces and facilities. I am particularly indebted to Karen Doble for mammalien. C. R. Acad. Sci. Paris 306, 253–256. her technical assistance and advice at the Whitney Laboratory and Georges, D., and Dubois, M. P. (1980). Mise en e´vidence par les for suggestions to improve the manuscript. This research was sup- techniques d’immunofluorescence d’un antige`ne de type LH- RH ported by a grant from the Natural Sciences and Engineering Re- dans le syste`me nerveux de Ciona intestinalis (Tunicier ascidiace´). search Council of Canada to M.A. C. R. Acad. Sci. Paris 290, 29–31.

Goldberg, J. I., Garofalo, R., Price, C. J., and Chang, J. P. (1993). Parker, G. H. (1920). Activities of colonial animals. II. Neuromus- Presence and biological activity of a GnRH-like factor in the cular movements and phosphorescence of Renilla. J. Exp. Zool. 31, nervous system of Helisoma trivolvis. J. Comp. Neurol. 336, 571–582. 475–515. Grimmelikhuijzen, C. J. P., and Groeger, A. (1987). Isolation of the Pazos, A. J., and Mathieu, M. (1999). Effects of five natural gonado- neuropeptide pGlu-Gly-Arg-Phe-amide from the pennatulid Re- tropin-releasing hormones on cell suspensions of marine bivalve nilla ko¨llikeri. FEBS Lett. 211, 105–108. gonad: Stimulation of gonial DNA synthesis. Gen. Comp. Endocri- Grimmelikhuijzen, C. J. P., and Westfall, J. A. (1995). The nervous nol. 113, 112–120. systems of cnidarians. In “The Nervous Systems of Invertebrates: Powell, J. F. F., Reska-Skinner, S. M., Prakash, M. O., Fischer, W. H., An Evolutionary and Comparative Approach” (O. Breidbach and Park, M., Rivier, J. E., Craig, A. G., Mackie, G. O., and Sherwood, W. Kutsch, Eds.), pp. 7–24. Birkhoüser, Basel. N. M. (1996). Two new forms of gonadotropin- releasing hormone Grimmelikhuijzen, C. J. P., Leviev, I., and Carstensen, K. (1996). in a protochordate and the evolutionary implications. Proc. Natl. Peptides in the nervous systems of cnidarians: Structure, function, Acad. Sci. USA 93, 10461–10464. and biosynthesis. Int. Rev. Cytol. 167, 37–89. Price, D. A., Cobbs, C. G., Doble, K. E., Kline, J. K., and Greenberg, Grimmelikhuijzen, C. J. P., Carstensen, K., Darmer, D., Moosler, M. J. (1987). Evidence of a novel FMRFamide-related heptapep- A., Nothacker, H. P., Reinscheid, R. K., Schmutzler, C., Vollert, tide in the pulmonate snail Siphonaria pectinata. Peptides 8, 533– H., McFarlane, I. D., and Rinehart, K. L. (1992). Coelenterate 538. neuropeptides: Structure, action and biosynthesis. Am. Zool. 32, Saripalli, L. D., and Westfall, J. A. (1996). Classification of nerve cells 1–12. dissociated from tentacles of the sea anemone Calliactis parasitica. Harlow, E., and Lane, D. (1986). “Antibodies: A Laboratory Manu- Biol. Bull. (Woods Hole) 190, 111–124. al.” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, Satterlie, R. A., and Case, J. F. (1979). Development of biolumines- NY. cence and other effector responses in the pennatulid coelenterate Ikegami, S., Honji, N., and Yoshida, M. (1978). Light-controlled Renilla koˆllikeri. Biol. Bull. (Woods Hole) 157, 506–523. production of spawning-inducing substance in jellyfish ovary. Satterlie, R. A., Anderson, P. A. V., and Case, J. F. (1980). Colonial Nature 272, 611–612. coordination in anthozoans: Pennatulacea. Mar. Behav. Physiol. 7, Kelsall, R., Coe, I. R., and Sherwood, N. M. (1990). Phylogeny and 25–46. ontogeny of gonadotropin-releasing hormone: Comparison of Sherwood, N. M. (1987). The GnRH family of peptides. Trends guinea pig, rat and a protochordate. Gen. Comp. Endocrinol. 78, 479–494. Neurosci. 10, 129–132. King, J. A., and Millar, R. P. (1995). Evolutionary aspects of gona- Sherwood, N. M., Lovejoy, D. A., and Coe, I. R. (1993). Origin of mam- dotropin-releasing hormone and its receptor. Cell. Mol. Neurobiol. malian gonadotropin-releasing hormones. Endocr. Rev. 14, 241–253. 15, 5–23. Umbriaco, D., Anctil, M., and Descarries, L. (1990). Serotonin-im- Lyke, E. B. (1965). “The Histology of the Sea Pansies, Renilla reni- munoreactive neurons in the cnidarian Renilla koellikeri. J. Comp. formis (Pallas) and Renilla ko¨llikeri (Pfeffer) with a Note on the Fine Neurol. 291, 167–178. Structure of the Latter Species.” Ph.D. thesis, University of Wis- Wu¨ rden, S., and Homberg, U. (1993). A simple method for immu- consin, Madison. nofluorescent double staining with primary antisera from the Mackie, G. O. (1995). On the ‘visceral nervous system’ of Ciona. J. same species. J. Histochem. Cytochem. 41, 627–630. Mar. Biol. Ass. U.K. 75, 141–151. Yasuda, A., Naya, Y., and Nakanishi, K. (1993). Isolation of antho- Mechawar, N., and Anctil, M. (1997). Melatonin in a primitive RFamide related peptides from the eyestalks of blue crab. Comp. metazoan: Seasonal changes of levels and immunohistochemical Biochem. Physiol. B 104, 235–240. visualization in neurons. J. Comp. Neurol. 387, 243–254. Young, K. G., Chang, J. P., and Goldberg, J. I. (1999). Gonadotropin- Pani, A. K., Anctil, M., and Umbriaco, D. (1995). Neuronal localiza- releasing hormone neuronal system of the freshwater snails He- tion and evoked release of norepinephrine in the cnidarian Renilla lisoma trivolvis and Lymnaea stagnalis: Possible involvement in koellikeri. J. Exp. Zool. 272, 1–12. reproduction. J. Comp. Neurol. 404, 427–437.