1857

BIOLOGY OF REPRODUCTION 63, 1857–1866 (2000)

Cloning, Tissue Distribution, and Central Expression of the Gonadotropin-ReleasingHormone Receptor in the Rainbow Trout (Oncorhynchus mykiss)1

Thierry Madigou,3 Evaristo Mananos-Sanchez,3,4 Sandra Hulshof,3,5 Isabelle Anglade,3 Silvia Zanuy,4

and Olivier Kah2,3

Endocrinologie Moleculaire de la Reproduction,3 UMR CNRS 6026, Campus de Beaulieu,35042 Rennes cedex, FranceInstituto de Acuicultura de Torre de la Sal,4 CSIC, Castellon, SpainResearch Group Comparative Endocrinology University of Utrecht,5 NL-3584 CH Utrecht, The Netherlands

ABSTRACT

A full-length cDNA encoding a GnRH receptor (GnRH-R) hasbeen obtained from the brain of rainbow trout. This cDNA en-codes a protein of 386 amino acids (aa) exhibiting the typicalarrangement of the G-protein-coupled receptors in seven trans-membrane domains. However, a second ATG could give rise toa receptor with a 30-aa longer extracellular domain. As alreadyshown in other fish and Xenopus, this protein possesses an in-tracellular domain, in contrast with its mammalian counterparts.In the case of rainbow trout, this intracellular carboxy-terminaltail consists of 58 residues. Northern blotting experiments car-ried out in the brain, the pituitary, and the liver only resulted ina single band of 1.9–2 kilobases in the pituitary, although re-verse transcription-polymerase chain reaction amplificationproducts were found in the brain, the pituitary, the retina, andthe ovary. In situ hybridization using a probe corresponding tothe full-length coding region of the receptor was performed onvitellogenic or ovulating females and allowed to detect a weakbut specific signal in the proximal pars distalis of the pituitary,the preoptic region, the mediobasal hypothalamus, and the optictectum. However, the strongest signal was consistently detectedin a mesencephalic structure, the nucleus lateralis valvulae, thesignificance of which is presently open to speculation.

anterior pituitary, central nervous system, GnRH, GnRH receptor

INTRODUCTION

Gonadotropin-releasing hormone plays a central role inthe neuroendocrine control of the reproductive process invertebrates, notably by stimulating synthesis and release ofgonadotropins. However, GnRH may also deserve otherfunctions centrally. Indeed, a major outcome of the numer-ous comparative studies based on biochemical, morpholog-ical, and physiological approaches in all vertebrate classesis the fact that the brain of many species expresses at leasttwo GnRH variants. One GnRH form is mainly synthesizedin the forebrain and varies according to the species, where-as the second GnRH variant, chicken GnRH-II (cGnRH-II),is highly conserved in terms of both structure and site of

1Supported by the CNRS, INRA, Ministere de la Recherche et de la Tech-nologie, Fondation Langlois, and European Union (Fair CT97-3785). Thefirst two authors have contributed equally to this study.2Correspondence: O. Kah, Endocrinologie Moleculaire de la Reproduc-tion, UMR CNRS 6026, Bat 13, Campus de Beaulieu, 35042 Rennes ced-ex, France. FAX: 33 2 99 28 67 94; e-mail: olivier.kah@univ-rennes1.fr

Received: 4 April 2000.First decision: 5 May 2000.Accepted: 7 August 2000.Q 2000 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

expression [1]. Indeed, cGnRH-II-expressing neurons havebeen consistently detected in the synencephalic-mesence-phalic area of many vertebrate groups, including primates[2]. If it is now clear that, at least part of the GnRH-ex-pressing neurons of the anterior brain sustain mainly a hy-pophysiotropic function, the role of the cGnRH-II mesen-cephalic neurons remains totally unknown. However, thehigh conservation of this system in the vertebrate lineagesuggests important neuromodulatory functions for this pep-tide. Because of the diversity of GnRH variants identifiedin teleosts, the GnRH systems of fish have been extensivelystudied by either immunohistochemistry and in situ hybrid-ization [3]. There is now extensive information about thelocalization of the GnRH neurons and their respective pro-jections in different species, but surprisingly, little attempthas been made to correlate this information with the ex-pression of GnRH-R. Indeed, most studies on GnRH-Rhave focused on the pituitary and, more recently, on thegonads.

The first fish GnRH receptor (GnRH-R) has been char-acterized in the goldfish pituitary, and the presence of twoclasses of binding sites, one with high affinity and low ca-pacity and one with low affinity and high capacity, wasdemonstrated [4]. However, in other species, such as thecatfish [5], winter flounder [6, 7], sea bream, and stickle-back [8], a single class of GnRH binding sites has beendescribed in the pituitary, whereas in the lamprey, twohigh-affinity binding sites were shown [9]. In addition, sim-ilar to what has been shown in mammals [10, 11], extra-pituitary sites of GnRH actions have been detected in anumber of reproductive and nonreproductive organs suchas the ovary, testis, brain, liver, and kidney in the goldfish[12].

The first functional GnRH-R was cloned using RNA pu-rified from the mouse gonadotrope cell line aT3-1 [13, 14].A number of subsequent publications have described thecloning of GnRH receptors in several mammalian speciesincluding the rat [15], sheep [16], human [17], cow [18],and pig [19]. Analysis of the primary sequence shows thatthe GnRH-R has the predicted structure characteristic of amember of the large rhodopsin-like G protein-coupled re-ceptor (GPCR) superfamily, consisting of a single polypep-tide chain containing seven hydrophobic transmembranedomains connected by hydrophilic extra- and intracellularloops. Several features conserved among the GPCR familyare altered in the mammalian GnRH-R, the most strikingone being the complete absence of an intracellular C-ter-minal domain. These structural differences have been re-viewed recently [20]. In addition to a structure-functionanalysis, the cloning of GnRH-R cDNAs allowed in situhybridization to localize the GnRH-R mRNA-expressing

1858 MADIGOU ET AL.

cells in the rat brain [21] and ovary [22] where it wasshown that the GnRH-R had the same primary structure asits pituitary counterpart [23].

Until now, GnRH-R have been cloned and characterizedfrom only two teleost species: the African catfish, Clariasgariepinus [24], and the goldfish, Carassius auratus [25].This latter species is until now the only one in which twoGnRH-R cDNAs, gf A and gf B, sharing 71% identity havebeen found. These two cDNAs share 71% and 82% iden-tity, respectively, with the catfish GnRH-R (cfGnRH-R) and43% with the human receptor. The corresponding receptorsshow marked differences in their ligand selectivity and tis-sue distribution [25]. In contrast with the mammalianGnRH-R, but in accordance with other GPCRs, the recep-tors cloned in fish and also in Xenopus [26] contain anintracellular carboxy-terminal domain.

As pointed out before, information on the central distri-bution of GnRH-R in vertebrates is very limited. Therefore,one of the objectives of this study was to clone one or moreGnRH-R in the rainbow trout (Oncorhynchus mykiss), toanalyze its (their) tissue distribution(s), and to perform adetailed in situ hybridization study of its (their) sites ofexpression within the brain.

MATERIALS AND METHODS

Animals

Mature female rainbow trout were obtained from theINRA experimental fish farm (Le Drennec, France) andkept under recycled water at 12–158C and natural photo-period (468 north) until use. The developmental stage of thefish was estimated by the time of the year and confirmedby the gonadosomatic index. Animals were treated in agree-ment with the European Union regulations concerning theprotection of experimental animals.

Cloning of a Full-Length cDNA Encodingthe Trout GnRH-R

Total RNA was prepared from brain using the TRIzolreagent (Gibco-BRL, Gaithersburg, MD) according to themanufacturer’s protocol. A first partial cDNA was obtainedafter two successive polymerase chain reaction (PCR) am-plifications performed with degenerate primers designedfrom conserved regions of known GnRH-R. The first am-plification was carried out using the following primers:For1, 59-GAYGGVATGTGGAAYATHAC-39; and Rev1,59-ACRTARTADGGNGTCCARCA-39. A 336-base pair(bp) product spanning from the first to the second extra-cellular loop was then obtained using primers For1 and asecond reverse primer Rev2 (59-GRAACATGTTRTA-KGCYGTTTCC-39).

The 39 extremity of the cDNA was cloned by 39 RACE(rapid amplification of cDNA extremity)-PCR, using a 59-39 RACE kit according to the manufacturer’s conditions(Boehringer Mannheim, Mannheim, Germany). Briefly, areverse transcription (RT) was performed on 2.5 mg of totalRNA using an oligo(dT)-anchor primer (Boehringer Mann-heim). A first amplification was then carried out usingprimer 1 (59-GGTGCAGTGGTATGGCGGAG-39) and theanchor-primer (Boehringer Mannheim). A nested PCR wasconducted using 1 ml of the first reaction as template withprimer 2 (59-GCCACGTGTAAGATGCTATGCTTCC-39)and the same anchor-primer. Primers 1 and 2 were designedfrom the 336-bp cDNA. The PCR products were purifiedand cloned in the EcoRV site of Bluescript plasmid forsequencing.

The 59 extremity was obtained from the brain by RT-PCR using specific primers designed from the 336-bpcDNA (primer 3, 59-TGCCATCGCTCTTTGAAGCT-39)and a genomic clone obtained by screening a genomic li-brary with a probe corresponding to the 336-bp sequence(primer 4, 59-GAGCAATGTAGAATAACTGACGGCA-39). The PCR products were then cloned in the EcoRV siteof Bluescript plasmid.

A cDNA containing the complete reading frame wasgenerated under the same conditions using the followingprimers (primer 5, 59-ACTTTGATGTGTATATGATAA-TAATGTATGTAATC-39 and primer 6, 59-CTCCAGT-CATCTACCGTCACCT-39) and was cloned into the Blues-cript plasmid (p-rtGnRH-R).

Northern Blot Analysis

Thirty micrograms of total RNA from brain, optic tec-tum, pituitary, and liver were heat-denatured (658C for 15min) in formaldehyde-formamide, electrophoresed on a 1%agarose gel containing formaldehyde, and transferred ontoa nylon membrane (Hybond-N; Amersham, Uppsala, Swe-den). The RNA was fixed by UV illumination (254 nm) for1 min and baking for 1 h at 808C. The membrane washybridized under the conditions previously described [27]with a 1372-bp probe corresponding to the coding regionof the cDNA, labeled with [a32P]dCTP by random priming.The membrane was then washed four times in 23 SSC (13SSC is 0.15 M NaCl plus 0.015 M sodium citrate), 0.1%SDS for 5 min at room temperature and three times in 0.23SSC, 0.1% SDS for 15 min at 508C, and exposed to aBiomax film (Eastman-Kodak, Rochester, NY) at 2808C.

Analysis of rtGnRH-R Tissue Distribution by RT-PCRand Southern Blotting

For the detection of rtGnRH-R mRNA, tissues were col-lected and total RNA was prepared and reverse transcribedas described above with random hexamers. A 300-bp rt-GnRH-R cDNA, corresponding to transmembrane domainsIII and IV, was amplified using the following primers: 300F,59-GGTGCAGTGGTATGGCGGAG-39 and 300R, 59-TGCCATCGCTCTTTGAAGCT-39. The PCR productswere transferred onto a nylon membrane (Hybond-N;Amersham) and hybridized as described [27] with the cor-responding probe labeled with [a32P]dCTP by randompriming.

In Situ Hybridization

Sense and antisense rtGnRH-R riboprobes were synthe-sized using the pBS-rtGnRH-R transcription vector con-taining the complete coding region, linearized with BamHIor HindIII as a template for T3 and T7 RNA polymerase,respectively.

Mature female rainbow trout, anesthetized with phen-oxyethanol (0.3 ml/L), were perfused through the heart with0.65% NaCl followed by a fixative solution (4% parafor-maldehyde, 0.1 M phosphate buffer, pH 7.4). Brains andpituitaries were collected, fixed overnight at room temper-ature, dehydrated, embedded in paraffin, and cut at 6 mm.Sections were mounted on Tespa-treated (2% Tespa; Sigma,St. Louis, MO) slides for subsequent processing using thein situ hybridization protocol described previously [28].

Following hybridization and washes, sections were de-hydrated, air-dried, and exposed to Biomax film (Amer-sham) for 5 days. Slides were then dipped in an autoradio-

1859GnRH RECEPTORS IN RAINBOW TROUT

FIG. 1. Schematic representation of the GnRH-R structure and position of the different cDNA clones. The transmembrane domains are indicated byblack boxes.

graphic emulsion (Ilford K5 Nuclear Track; Amersham)and stored at 48C for 4–5 wk before development. Sectionswere then stained with toluidine blue and photographedwith an Olympus Provis photomicroscope under darkfieldor brightfield illumination.

The nomenclature used for rainbow trout brain nuclei isfrom Meek and Nieuwenhuys [29].

RESULTS

Cloning of a Full-Length cDNA for rtGnRH-Receptor

The cloning strategy and the different clones obtainedare summarized in Figure 1. Using degenerate primers, afirst cDNA of 336 bp (p336), corresponding to a regionspanning from the first to the second extracellular loop, wasobtained. This cDNA showed about 80% identity with thecfGnRH-R or the two gfGnRH-Rs. Sequence identitieswere lower with the mammalian GnRH-Rs (about 45%).

The 39 extremity of the trout GnRH-R cDNA was ob-tained by 39 RACE-PCR. Using this method, we havecloned a PCR product of 1250 bp. This clone, pBS-1202,was shown to contain a part of the coding region spanningfrom transmembrane domain III to the intracellular COOH-terminus of the receptor and 429 bp of the 39 untranslatedregion. A polyadenylation signal AATAAA was found 35nucleotides upstream from the poly(A) tail.

The 59 extremity was obtained by RT-PCR using specificprimers designed from the cDNA cloned previously and agenomic clone containing the first exon of rtGnRH-R gene(T. Madigou, unpublished data). A 750-bp long cDNA(pATG-750) was obtained. This clone contained part of the59 untranslated region and an open reading frame encodingthe extracellular NH2 terminus of the rtGnRH-R and trans-membrane domains I to IV. The two clones (pATG-750 andpBS-rtGnRH-R1202) overlapped and were 100% identicalto a region of 300 bp corresponding to transmembrane do-mains III and IV. The complete coding sequence of the troutGnRH-R was deduced from these two clones, and in orderto check if these clones corresponded to the two extremitiesof the same RNA, a cDNA of about 1.3 kilobases (kb)containing the entire reading frame was generated usingtwo specific primers located in the 59- and the 39-untrans-lated regions, respectively. The sequence of this cDNA hasbeen deposited in the EMBL Nucleotide Sequence Data-base (accession number AJ272116). Different attempts toobtain a second potential GnRH-R cDNA have failed untilnow.

Sequence Analysis and Comparison of rtGnRH-Rwith Other GnRH-R

The nucleotide sequence and the deduced amino acid(aa) sequence of the rtGnRH-R are shown on Figure 2.The potential translation initiation site, as predicted usingNetstart 1.0 neural network prediction server (Center forBiological Sequence Analysis, Denmark), corresponds tothe ATG present in all other species. If we consider thiscodon as the initiation site, the extracellular domain con-sists of 39 aa. However, another ATG is located 90 bpupstream from this first site. This second ATG could giverise to a receptor with a 30-aa longer extracellular domain.However, until the functionality of this first ATG as atranslation initiation codon is established, we will considerthe second methionine as Met1, giving rise to a protein of386 aa (Fig. 2). Hydrophobicity analysis of the deducedaa sequence, as determined by the method of Kyte andDoolittle [30], shows that the transmembrane domain isconstituted of 289 aa arranged as seven transmembranesegments typical of a GPCR. In contrast to its mammaliancounterparts, but like all nonmammalian GnRH-R de-scribed at this time, it contains an intracellular domainconsisting of 58 aa. In this domain, as in catfish and gold-fish receptors, serine residues that are potential phosphor-ylation sites (SXXS*) are conserved (Ser380 and Ser383).Similarly, a single cysteine (Cys369 in the trout receptor)that may be a site of palmitoylation is conserved in allfish GnRH-Rs.

Alignment of the predicted rtGnRH-R amino acid se-quence with the corresponding sequences described in cat-fish [24], goldfish [25], mouse [13], rat [15], sheep [16],human [17], bovine [18], Drosophila [31], and Typhlo-nectes natans (Ebersole et al., Genbank AF174481) wasperformed using Clustal W (Fig. 3), and a phylogenetictree was drawn (Fig. 4). These analyses showed that thertGnRH-R is closely related to the other fish GnRH-R andparticularly to the goldfish GnRH-R A (.80% identity).The rtGnRH-R shares about 40% identity with the mam-malian or amphibian receptors.

Tissue Distribution of the rtGNRH-R mRNA

The size of the rtGnRH-R mRNA was determined byNorthern blot analysis on total RNA extracted from varioustissues of an ovulating female. This analysis revealed thepresence of a single mRNA of approximately 1.9–2 kb in

1860 MADIGOU ET AL.

FIG. 2. Nucleotide and deduced amino acid sequence of the rainbow trout GnRH-R. Numbers on the left indicate nucleotide positions. Amino acidsare indicated below their respective codons and numbered on the right. The start codon as predicted by Netstart 1.0 neural network prediction server(Center for Biological Sequence analysis, Denmark) is in a stippled box.

the pituitary, but no signal could be detected in the brain,even on poly(A)1 RNA, or the ovary (Fig. 5A). The ab-sence of signal in these latter tissues might indicate thatmRNA levels were below the detection limits of this meth-od. Thus, we used the more sensitive RT-PCR and Southernblotting technique. Products of amplification of the expect-ed size were found in extrapituitary tissues such as thebrain, particularly the optic tectum, the retina, and the ovary(Fig. 5B). Surprisingly, a weaker signal was obtained in thepituitary. However, the tissues used for this analysis werecollected from vitellogenic females.

Localization of rtGnRH-R mRNA in the Brainand Pituitary of the Rainbow Trout

The precise localization of rtGnRH-R mRNA in thebrain and pituitary of female rainbow trout was studied byin situ hybridization using the full-length coding region ofthe rtGnRH-R cDNA as a probe. In general, in situ hybrid-ization showed a very weak expression of rtGnRH-RmRNA throughout the brain and pituitary, confirming ourresults of Northern blotting and RT-PCR. Nevertheless, spe-cific hybridization signals could be detected in the preoptic

1861GnRH RECEPTORS IN RAINBOW TROUT

FIG. 3. Alignment of the aa sequence of the rtGnRH-R and the other species GnRH-Rs. The seven transmembrane domains are indicated by overlineswith roman numerals. Dashes indicate sequence gaps introduced to maximize sequence homology. ’’ Symbols indicate aa conserved between troutand other species, whereas aa differences are shown in their corresponding position.

FIG. 5. A) Northern blot analysis of rtGnRH-R. Total RNA from femalepituitary separated by agarose-formaldehyde gel electrophoresis, trans-ferred and hybridized with the rtGnRH-R probe in high stringency. Thepositions of the 28s and 18s rRNAs are indicated on the right. B) Expres-sion of rtGnRH-R mRNA in central and peripheral tissues as detected byRT-PCR and Southern blotting. Amplification of b-actin was used as aninternal control.

FIG. 4. Phylogenetic analysis of the GnRH-Rs. A sequence alignmentwas performed according to the Lipman-Pearson algorithm using the Day-hoff matrix for aa homology. From this alignment, distances for the dif-ferent receptors were calculated with the neighbor joining method. Onethousand bootstrap replicates were performed (expressed as percentages).Branch lengths are proportional to estimated divergence along eachbranch.

1862 MADIGOU ET AL.

FIG. 6. In situ hybridization of GnRH-Rs in the brain and the pituitaryof rainbow trout. A, B) Adjacent transverse sections at the level of theanterior preoptic region hybridized with the antisense (A) and sense (B)probes. A specific labeling is detected in the pars parvicellular of theanterior preoptic nucleus (Ppa) bordering the ventrolateral aspects of thepreoptic recess (rpo). C, D) Adjacent transverse sections at the level ofthe mediobasal hypothalamus hybridized with the antisense (C) and sense(D) probes. A hybridization signal is detected at the level of the nucleuslateralis tuberis (nlt) and in the proximal pars distalis of the pituitary (ppd),whereas the neurohypophysis (nh) is devoid of labeling. III, third ventricle.E) Transverse section through the optic tectum hybridized with the anti-sense probe. A specific labeling is detected over the periventricular layer(SPV), whereas the other layers are unstained. SAC, Stratum album cen-trale; SGC, stratum griseum centrale; SFGS, stratum fibrosum and griseumsuperficiale. For all micrographs, bar 5 300 mm.

region (parvicellular preoptic nucleus; Fig. 6A), the me-diobasal hypothalamus (nucleus lateralis tuberis; Fig. 6C),the pituitary (proximal pars distalis; Fig. 6C), and the peri-ventricular layer of the optic tectum (Fig. 6E). Surprisingly,the strongest signal was consistently detected in the nucleuslateralis valvulae, a mesencephalic structure bridging thedorsal tegmentum with the valvula of the cerebellum (Fig.7). This nucleus contains small cells strongly stained bytoluidine blue and larger elements. Figure 7F shows thatthe labeling preferentially concerns the large pale cells. Thespecificity of the signal was systematically checked on ad-jacent sections hybridized with the corresponding senseprobe showing only uniform background (Figs. 6, B and Dand 7, E and G).

DISCUSSION

In this study, we have cloned a cDNA encoding aGnRH-R from rainbow trout and used it as a probe to studythe localization of the corresponding mRNA in differentorgans by RT-PCR and in the brain and pituitary by in situhybridization. This cDNA, containing the full-length cod-ing sequence of the rainbow trout GnRH-R, was obtainedby RT-PCR. In this cDNA, a methionine codon is found atthe same position as in all the GnRH-Rs. However, anotherin-frame ATG, 90 bp upstream of this methionine, couldgive rise to a protein with a putative extracellular domain30 aa longer than those found in the other GnRH-Rs. Mul-tiple transcription start sites have already been reported inmany genes, including GnRH-R as for example in human[32] or sheep [33]. At the present stage, it is not knownwhether messengers lacking this ATG are generated in rain-

bow trout. It is interesting to note that such a potentialtranslation initiation site is also present upstream of thepredicted first ATG in the gonadotropin receptor of theamago salmon (Oncorhynchus rhodurus) [34]. The se-quence identity with the catfish and goldfish GnRH-R werefound to be around 80%, whereas it was only about 40%with the mammalian receptors. In the goldfish, recent datahave demonstrated the presence of two GnRH-R, Gf A andGf B, sharing 71% identity [25]. Phylogenetic analyses re-veal that the rtGnRH-R is more closely related to the gold-fish Gf A than to the Gf B. Until now, there is no indicationfor the presence of two different receptor subtypes in rain-bow trout.

The alignment of the aa sequence of the rtGnRH-R withthe other receptors shows that a number of key features areconserved. In most of the GPCR, conserved cysteines inthe first and the second extracellular loops form a disulfidebound that is necessary for the correct folding of the re-ceptor [35, 36]. The presence of such a bridge was dem-onstrated in the rat GnRH-R between Cys114 and Cys195

[37]. In the trout GnRH-R, two cysteines are also found inthe corresponding position (Cys119 and Cys186). However,another cysteine is present in the second extracellular loop.In the same way, the putative ligand contact sites Asp98,Asn102, Lys121, and Glu301 described for the mouse [20] areconserved in the rtGnRH-R as Asp103, Asn107, Lys120, andGlu304.

Structural elements such as an Asp-Arg-Tyr (DRY) mo-tif in the cytoplasmic end of TMIII or Asn-Pro-X-X-Tyr(NPXXY) in TMVII are highly conserved in the rhodopsin-like GPCR family. The DRY motif was shown to be im-plicated in receptor activation and the coupling with G-proteins [38]. In mammalian GnRH-Rs, the Tyr is replacedby a Ser except in the marmoset in which a Phe is found[39]. Several studies have shown that mutation of the thirdaa of this motif may affect receptor functionality. Indeed,substitution of Ser by Tyr or Phe induces an increase inagonist binding affinity and internalization rate [39, 40],whereas replacement by Ala affects neither internalizationnor signal transduction [41]. The trout GnRH-R, like allother fish GnRH-Rs, presents a substitution of the Tyr bya His in this DRY motif. Because Phe and Tyr have thesame effect on receptor internalization, it has been sug-gested that aromatic residues at this site could be importantin regulating internalization [39]. Although the presence ofan aliphatic residue, Ser or Ala, in this motif seems to re-duce receptor internalization, the effect of a substitution bya heterocyclic amino acid such as His has not been evalu-ated. Another conserved domain is the NPXXY motif lo-cated in TMVII that has also been shown to be involved inthe internalization of some GPCR including the GnRH-R[42]. Although the sequence is modified to DPXXY in allthe GnRH-R, the Tyr of this motif, critical for receptoractivation and signal transduction, is conserved [42].

The three-dimensional model of the GPCRs and recip-rocal mutation of conserved loci in GnRH-R predicts thatTMII and TMVII are juxtaposed [43]. Two residues, Aspin TMII and Asn in TMVII, highly conserved amongGPCRs but interchanged in all mammalian GnRH-Rs, arelikely to be implicated in interhelical interaction [44].Moreover, the mutation of Asn in TMII by Asp in mouse[43] or rat [45] GnRH-R eliminated detectable ligand bind-ing. However, this seems to be specific to the mammalianreceptor because in all other nonmammalian receptors in-cluding the rtGnRH-R, two Asp are found at these posi-tions.

1863GnRH RECEPTORS IN RAINBOW TROUT

FIG. 7. In situ hybridization of GnRH-Rs in the nucleus lateralis valvulae (nlv) of rainbow trout. A, B) Darkfield (A) and brightfield (B) micrographs ofthe same transverse section at the level of the anterior part of the nlv. Note that the hybridization signal is restricted to the nlv and that the denselypacked cells of the granular layer of the valvula of the cerebellum (VCg) are devoid of labeling; VCm, molecular layer of the valvula of the cerebellum;nIII, oculomotor nucleus. C, D) Darkfield (A) and brightfield (B) micrographs of the same transverse section at the level of the central part of the nlv.The signal clearly overlaps with the nlv, whereas the granular layer of the valvula of the cerebellum (VCg) is devoid of labeling. E) Darkfield micrographof a transverse section adjacent to that shown in C hybridized with the sense probe. F, G) High-power view of two adjacent sections hybridized withthe antisense (F) and sense (G) probes at the level of the nlv. Note that silver grains are preferentially associated with the large pale cells (arrowheads).VCm, Molecular layer of the valvula of the cerebellum; nIII, oculomotor nucleus. A–E, bar 5 150 mm; F, G, bar 5 50 mm.

Similar to other nonmammalian GnRH-R, the trout re-ceptor contains an intracellular tail that in this case consistsof 58 residues. This intracellular tail is important for re-ceptor functioning because the truncation of this domainproduces the loss of both the ability for agonist binding andGnRH-stimulated cAMP production [46]. Likewise, addi-tion of cfGnRH-R C-terminal tail to the rat GnRH-R affectsits expression, regulation, and the differential coupling withG proteins [47], producing a rapid desensitization of ino-sitol phosphate production as well as increased internali-zation rate [48].

Northern blot analysis of total RNA revealed the pres-ence of a single GnRH-R mRNA species only in the pitu-itary of ovulating female trout. This mRNA is shorter thanthose described in mammals in which a major band of 4–5 kb is detected, together with other shorter bands [13, 16–18]. The structure of the GnRH-R gene has been elucidatedin several mammalian species. Unlike many GPCRs, theGnRH-R gene contains introns [32, 33, 49, 50], allowingthe existence of alternative splicing, which may explain thepresence of multiple bands on Northern blot analysis. In allthese species, the gene is constituted of three exons andtwo introns. In the mouse, several variant transcripts lack-ing exon 2 or exon 3 have been described [49]. Although

only one band is detected by Northern blotting in the trout,we have also found in the brain by RT-PCR an mRNAvariant that contains only exons 1 and 2 and may encodea truncated protein corresponding to the NH2 terminus andTM I to V (unpublished data). Unfortunately, brain expres-sion levels of GnRH-R messengers appear too low for al-lowing detection by Northern blot, even on poly(A)1 RNA.Truncated receptors have already been reported in humansand were shown to be unable to bind GnRH and to activatesignal transduction [51]. However, when coexpressed withthe wild-type receptor, this truncated protein caused im-paired insertion of the wild type into the plasma membrane[51].

Because Northern blotting failed to detect GnRH-RmRNA in other tissues than the pituitary, we used the moresensitive method of RT-PCR followed by Southern blottingand hybridization. The PCR products were found in extra-pituitary tissues, namely the brain, the retina and the ovaryof vitellogenic females. These results confirm those ob-tained in the gonads of other teleosts. Indeed, Scatchardanalysis revealed the presence of one class of high affinitybinding sites in the ovary of common carp [52], Africancatfish [53], and immature goldfish, whereas two classes ofGnRH binding sites with characteristics similar to that of

1864 MADIGOU ET AL.

the pituitary were identified in the ovary of mature goldfish[54]. In this latter species, in situ hybridization demonstrat-ed expression of Gf A receptor mRNA in the interstitialtissue and the theca-granulosa cell layers surrounding theoocytes [25]. The precise role of GnRH in the gonads offish remains unclear, although in vitro studies indicate thatit may act as a meiosis-stimulating factor in the oocyte [55].Similarly, the presence of gonadal GnRH-R has been dem-onstrated in mammals, and, in the rat, the primary structureof the ovarian receptor was shown to be identical to thatof the hypohyseal receptor [23].

The presence of GnRH-Rs in the retina correlates withprevious studies demonstrating an innervation of the retinaby GnRH-immunoreactive fibers [56]. In cichlid, poecilid,and centrarchid fishes, GnRH-immunoreactive neurons arefound in the nucleus olfactoretinalis and processes of theseneurons project to the contralateral retina [57]. A more re-cent study in Atlantic salmon showed that some GnRH-positive cells clustered within the medial component of theolfactory nerve project to the retina where the GnRH fiberssynapse on the dopaminergic interplexiform cells [58]. Ithas been shown that the effects of GnRH on horizontal cellactivity mimicked the effects of dopamine on these cells,suggesting that GnRH acts by stimulating the release ofdopamine from interplexiform cells [59, 60].

In this study, we have also analyzed the distribution ofGnRH-R mRNA in the pituitary by in situ hybridizationusing a probe corresponding to the full-length coding re-gion. The signal was restricted to the proximal pars distalisbut remained weak in mature females or ovulating females.This is in agreement with the weak signal detected by RT-PCR (this study) and with the fact that binding studies us-ing different GnRH analogs failed to evidence a strong spe-cific binding in the trout pituitary (C. Weil and L.W. Crim,personal communication) in contrast with other species [5–8]. This is surprising because GnRH is well known to beactive in vivo and in vitro in stimulating LH secretion inthe trout [61]. The weak but significant signal observed inthe pituitary by Northern blotting tends to indicate that thenumber of GnRH-R mRNA copies per cell is low but stillsufficient to be detected by Northern blot. In addition, par-allel studies performed in sea bass using the same techniqueallowed detection of a very strong GnRH-R hybridizationsignal in the pituitary of mature, and even immature, males[62].

To date, most of the information concerning the locali-zation of GnRH-R expressing cells in the brain has beenobtained by autoradiography of binding of 125I-labeledGnRH analogs and in situ hybridization in mammals. In therat brain, the mRNA encoding the GnRH-R is present inselected regions associated with the regulation of GnRHrelease from the median eminence and with the generationof reproductive behaviors [21].

In fish, apart from a possible neuromodulatory function,GnRH has been shown to stimulate sexual behavior [63];however, the precise central sites of GnRH actions are un-known. This paper reports for the first time the precise dis-tribution of GnRH-R mRNA in the brain of a teleost fish.In the brain of trout, a specific signal could be consistentlydetected in two regions clearly implicated in the neuroen-docrine control of the pituitary, the preoptic area, and themediobasal hypothalamus [64]. In teleosts, the main GnRH-positive fiber tracts cross the preoptic region and the ventralhypothalamus on their way to the pituitary [56, 65] and onemay speculate that these varicose fibers establish contacts

of passage with some cell populations and influence theiractivities.

A significant GnRH-R hybridization signal has also con-sistently been detected in the optic tectum in agreementwith the RT-PCR results. The optic tectum of teleosts is ahighly laminar retino-recipient structure in which GnRH-RmRNAs were restricted to the piriform neurons of the peri-ventricular layer shown recently to strongly express mela-tonin receptor in trout [28], and to be the site of a circadianclock in zebrafish [66]. The potential functions of GnRHin the tectum remain to be established. However, the ex-pression of GnRH-R in the tectum is consistent with thesignificant GnRH innervation of this structure [56, 65].

Very surprisingly, the highest GnRH-R hybridizationsignal was consistently found in a mesencephalic structure,the nucleus lateralis valvulae, that has been well studied inthe carp [67]. In this species, this nucleus consists of smalladendritic cells, projecting to both the corpus and the val-vula of the cerebellum, and larger dendritic elements [29,67]. Our study tends to indicate that expression of theGnRH-R is restricted to these larger cells. In the cyprinidPhoxinus phoxinus, part of the cell bodies in the nucleuslateralis valvulae have been shown to be cholinergic on thebasis of acetylcholinesterase histochemistry and cholineacetyltransferase immunohistochemistry [68]. There is noprecise information on the functions of the nucleus lateralisvalvulae, but it could be implicated in the processing ofsensory and motor information [29]. The nucleus lateralisvalvulae receives afferents from the telencephalon and pos-sibly from the nucleus of the medial longitudinal fascicle[67], whose large neurons are well known for expressingcGnRH-II in salmonids [69] and other species. Althoughno particular cGnRH-II innervation has ever been reportedin the nucleus lateralis valvulae, a potential cGnRH-II pro-jection from the nucleus of the medial longitudinal fascicleto the nucleus lateralis valvulae deserves considerable in-terest, as there is no precise information on the potentialrole of cGnRH-II nor on its sites of action in fish and othervertebrates. It is important to point out that although ini-tially discovered in fish [56, 70], cGnRH-II neurons in themesencephalon have now been identified in all vertebrategroups, including mammals [2]. Because the contributionof these cells to the pituitary in fish or median eminencein tetrapods seems to be low or nonexistant, cGnRH-II cer-tainly plays some still highly enigmatic but crucial rolewithin the brain in view of the high conservation of thecGnRH-II system. Interestingly, a recent study in Xenopusshowed that the central expression of a GnRH-R is mainlyassociated with the midbrain in this species [26].

In conclusion, a GnRH-R cDNA has been cloned fromthe rainbow trout and may encode a protein with a 30-aalonger extracellular domain. As already reported in catfishand goldfish, this receptor contains an intracellular domain,in contrast with its mammalian counterparts. The corre-sponding mRNAs are expressed in the retina, the brain, thepituitary, and the ovaries. In the brain the highest expres-sion is clearly associated with a mesencephalic structure,the nucleus lateralis valvulae, whose precise function is un-known but may concern the processing of sensory or motorinformation.

REFERENCES

1. Carolsfeld J, Powell JF, Park M, Fischer WH, Craig AG, Chang JP,Rivier JE, Sherwood NM. Primary structure and function of threegonadotropin-releasing hormones including a novel form, from an an-cient teleost, herring. Endocrinology 2000; 141:505–512.

1865GnRH RECEPTORS IN RAINBOW TROUT

2. Lescheid DW, Terasawa E, Abler LA, Urbansky HF, Warby CM, Mil-lar RP, Sherwood NM. A second form of gonadotropin-releasing hor-mone (GnRH) with characteristics of chicken GnRH-II is present inthe primate brain. Endocrinology 1997; 136:5618–5629.

3. Gonzalez-Martınez D, Madigou T, Zmora N, Anglade I, Zanuy S,Zohar Y, Elizur A, Munoz-Cueto JA, Kah O. Differential expressionof three different prepro-GnRH (gonadotrophin-releasing hormone)messengers in the brain of the European sea bass (Dicentrarchus la-brax). J Comp Neurol 2000; (in press).

4. Habibi HR, Peter RE, Sokolowska M, Rivier JE, Vale WW. Charac-terization of gonadotropin-releasing hormone (GnRH) binding to pi-tuitary receptors in goldfish (Carassius auratus). Biol Reprod 1987;36:844–853.

5. De Leeuw R, Van‘t Veer C, Goos HJ, Van Oordt PG. The dopami-nergic regulation of gonadotropin-releasing hormone receptor bindingin the pituitary of the African catfish, Clarias gariepinus. Gen CompEndocrinol 1988; 72:408–415.

6. Crim LW, St Arnaud R, Lavoie M, Labrie F. A study of LH-RHreceptors in the pituitary gland of the winter flounder (Pseudopleu-ronectes americanus Walbaum). Gen Comp Endocrinol 1988; 69:372–377.

7. Weil C, Crim LW, Wilson CE, Cauty C. Evidence of GnRH receptorsin cultured pituitary cells of the winter flounder (Pseudopleuronectesamericanus W.). Gen Comp Endocrinol 1992; 85:156–164.

8. Habibi HR, Peter RE. Gonadotropin-releasing hormone (GnRH) re-ceptors in teleost. In: Scott AP, Sumpter JP, Kime DE, Rolfe MS(eds.), Reproductive Physiology of Fish. Fish Symposium. Sheffield,UK: University East Anglia Printing Unit; 1991: 109–113.

9. Knox CJ, Boyd SK, Sower SA. Characterization and localization ofgonadotropin-releasing hormone receptors in the adult female sea lam-prey, Petromyzon marinus. Endocrinology 1994; 134:492–498.

10. Hsueh AJW, Schaeffer JM. Gonadotropin releasing hormone as aparacrine hormone and neurotransmitter in extra-pituitary sites. J Ste-roid Biochem 1985; 23:757–764.

11. Chieffi G, Pierantoni R, Fasano S. Immunoreactive GnRH in hypo-thalamic and extrahypothalamic areas. Int Rev Cytol 1991; 127:1–55.

12. Habibi HR, Pati D. Extrapituitary gonadotropin releasing hormone(GnRH) binding sites in goldfish. Fish Physiol Biochem 1993; 11:43–49.

13. Tsutsumi M, Zhou W, Millar RP, Mellon PM, Roberts JL, FlanaganCL, Dong K, Gillo B, Sealfon SC. Cloning and functional expressionof a mouse gonadotropin releasing hormone receptor. Mol Endocrinol1992; 6:1163–1169.

14. Reinhart J, Mertz LM, Catt KJ. Molecular cloning and expression ofcDNA encoding the murine gonadotropin-releasing hormone receptor.J Biol Chem 1992; 267:21281–21284.

15. Eidne KA, Sellar RE, Couper G, Anderson L, Taylor PL. Molecularcloning and characterisation of the rat pituitary gonadotropin-releasinghormone (GnRH) receptor. Mol Cell Endocrinol 1992; 90:R5–R9.

16. Brooks J, Taylor PL, Saunders PTK, Eidne KA, Struthers WJ,McNeilly AS. Cloning and sequencing of the sheep pituitary gonad-otropin-releasing hormone receptor and changes in expression of itsmRNA during estrous cycle. Mol Cell Endocrinol 1993; 94:R23–R27.

17. Chi L, Zhou W, Prikhozhan A, Flanagan C, Davidson JS, GolemboM, Illing N, Millar RP, Sealfon SC. Cloning and characterization ofthe human GnRH receptor. Mol Cell Endocrinol 1993; 91:R1–R6.

18. Kakar SS, Rahe CH, Neill JD. Molecular cloning, sequencing andcharacterizing the bovine receptor for gonadotropin releasing hormone(GnRH). Domest Anim Endocrinol 1993; 10:335–342.

19. Weesner GD, Matteri RL. Nucleotide sequence of luteinizing hor-mone-releasing hormone (LHRH) receptor cDNA in the pig pituitary.J Anim Sci 1994; 72:1911–1914.

20. Sealfon SC, Weinstein H, Millar RP. Molecular mechanisms of ligandinteraction with the gonadotropin-releasing hormone receptor. EndocrRev 1997; 18:180–205.

21. Jennes L, Woolums S. Localization of gonadotropin-releasing hor-mone receptor mRNA in the rat brain. Endocrine 1994; 2:521–528.

22. Kogo H, Kudo A, Park MK, Mori T, Seiichiro K. In situ detection ofgonadotropin releasing hormone (GnRH) receptor mRNA expressionin the rat ovarian follicles. J Exp Zool 1995; 272:62–68.

23. Moumni M, Kottler ML, Counis R. Nucleotide sequence analysis ofmRNAs predicts that rat pituitary and gonadal gonadotropin releasinghormone receptor proteins have identical primary structure. BiochemBiophys Res Commun 1994; 200:1359–1366.

24. Tensen C, Okuzawa K, Blomenrohr M, Leurs R, Rebers F. Bogerd J,Schulz R, Goos H. Distinct efficacies for two endogenous ligands on

a single cognate gonadoliberine receptor. Eur J Biochem 1997; 243:134–140.

25. Illing N, Troskie BE, Nahorniak CS, Hapgood JP, Peter RE, MillarRP. Two gonadotropin-releasing hormone receptor subtypes with dis-tinct ligand selectivity and differential distribution in brain and pitu-itary in the goldfish (Carassius auratus). Proc Natl Acad Sci U S A1999; 96:2526–2531.

26. Troskie BE, Hapgood JP, Millar RP, Illing N. Complementary deox-yribonucleic acid cloning, gene expression, and ligand selectivity ofa novel gonadotropin-releasing hormone receptor expressed in the pi-tuitary and midbrain of Xenopus laevis. Endocrinology 2000; 141:1764–1771.

27. Thomas P. Hybridization of denatured RNA and small DNA fragmentstransferred to nitrocellulose. Proc Natl Acad Sci U S A 1980; 77:5201–5205.

28. Mazurais D, Brierley I, Anglade I, Drew J, Randall C, Bromage N,Michel D, Kah O, Williams LM. Central melatonin receptors in therainbow trout: comparative distribution of ligand binding and geneexpression. J Comp Neurol 1999; 409:313–324.

29. Meek J, Nieuwenhuys R. Holosteans and teleosts. In: NieuwenhuysR, ten Donkelaar HJ, Nicholson C (eds.), The Central Nervous Systemof Vertebrates, vol. II. New York: Springer; 1997: 761–937.

30. Kyte J, Doolittle RF. A simple method for displaying the hydropathiccharacter of a protein. J Mol Biol 1982; 157:105–132.

31. Hauser F, Sondergaard L, Grimmelikhuijzen CJP. Molecular cloning,genomic organization and developmental regulation of a novel recep-tor from Drosophila melanogaster structurally related to gonadotro-pin-releasing hormone receptors from vertebates. Biochem BiophysRes Commun 1998; 249:822–828.

32. Fan NC, Peng C, Krisinger J, Leung PCK. The human gonadotropinreleasing hormone receptor gene: complete structure including mul-tiple promoters, transcription initiation sites, and polyadenylation sig-nals. Mol Cell Endocrinol 1995; 107:R1–R8.

33. Campion CE, Turzillo AM, Clay CM. The gene encoding the ovinegonadotropin releasing hormone (GnRH) receptor: cloning and initialcharacterization. Gene 1996; 170:277–280.

34. Oba Y, Hirai T, Yoshiura Y, Yoshikuni M, Kawauchi H, NagahamaY. Cloning, functional characterization, and expression of a gonado-tropin receptor cDNA in the ovary and testis of amago salmon (On-corhynchus rhodurus). Biochem Biophys Res Commun 1999; 263:584–590.

35. Karnik SS, Sakmann JP, Chen HB, Khorana HG. Cysteine residues110 and 187 are essential for the formation of correct structure inbovine rhodopsin. Proc Natl Acad Sci U S A 1988; 85:8459–8463.

36. Perlman JH, Wang W, Nussenzveig DR, Gershengorn MC. A disulfidebound between conserved extracellular cysteines in the thyrotropin-releasing hormone receptor is critical for binding. J Biol Chem 1995;270:24682–24685.

37. Cook JVF, Eidne KA. An intracellular disulfide bound between con-served extracellular cysteines in the gonadotropin releasing hormonereceptor is essential for binding and activation. Endocrinology 1997;138:2800–2806.

38. Schoneberg T, Schultz G, Gudermann T. Structural basis of G-protein-coupled receptor function. Mol Cell Endocrinol 1999; 151:181–193.

39. Byrne B, McGregor A, Taylor PL, Sellar R, Rodger FE, Fraser HM,Eidne KA. Isolation and characterisation of the marmoset gonadotro-phin releasing hormone receptor: Ser140 of the DRS motif is replacedby Phe. J Endocrinol 1999; 163:447–456.

40. Arora KK, Sakai A, Catt KJ. Effects of second intracellular loop mu-tations on signal transduction and internalization of the gonadotropinreleasing hormone receptor. J Biol Chem 1995; 270:22820–22826.

41. Arora KK, Cheng Z, Catt KJ. Mutations of the conserved DRS motifin the second intracellular loop of the gonadotropin releasing hormonereceptor affect expression, activation and internalization. Mol Endo-crinol 1997; 11:1203–1212.

42. Arora KK, Cheng Z, Catt KJ. Dependence of agonist activation on anaromatic moiety in the DPLIY motif of the gonadotropin releasinghormone receptor. Mol Endocrinol 1996; 10:979–986.

43. Zhou W, Flanagan CA, Ballesteros JA, Konvicka K, Davidson JS,Weinstein H, Millar RP, Sealfon SC. A reciprocical mutation supportshelix 2 and helix 7 proximity in the gonadotropin-releasing hormonereceptor. Mol Pharmacol 1994; 45:165–170.

44. Ballesteros J, Kitanovic S, Guarnieri F, Davies P, Fromme BJ, Kon-vicka K, Chi L, Millar RP, Davidson JS, Weinstein H, Sealfon SC.Functional microdomains in G-protein-coupled receptors: the con-served arginine-cage motif in the gonadotropin-releasing hormone re-ceptor. J Biol Chem 1998; 273:10445–10453.

1866 MADIGOU ET AL.

45. Cook JVF, Faccenda E, Anderson L, Couper GG, Eidne KA, TaylorPL. Effects of Asn87 and Asp318 mutations on ligand binding andsignal transduction in the rat GnRH receptor. J Mol Endocrinol 1993;139:R1–R4.

46. Blomenrorh M, Bogerd J, Leurs R, Schulz RW, Tensen CP, Zandber-gen MA, Goos HJTh. Differences in structure-function relations be-tween nonmammalian and mammalian gonadotropin-releasing hor-mone receptors. Biochem Biophys Res Commun 1997; 238:517–522.

47. Lin X, Janovick A, Brothers S. Blomenrorh M, Bogerd J, Conn M.Addition of catfish gonadotropin-releasing hormone (GnRH) receptorintracellular carboxyl-terminal tail to GnRH receptor alters receptorexpression and regulation. Mol Endocrinol 1998; 12:161–171.

48. Heding A, Vrecl M, Bogerd J, McGregor A, Sellar R, Taylor PL,Eidne KA. Gonadotropin-releasing hormone receptors with intracel-lular carboxyl-terminal tails undergo acute desensitization of total ino-sitol phosphate production and exhibit accelerated internalization ki-netics. J Biol Chem 1998; 273:11472–11477.

49. Zhou W, Sealfon SC. Structure of the mouse gonadotropin-releasinghormone receptor gene: variant transcripts generated by alternativeprocessing. DNA Cell Biol 1994; 13:605–614.

50. Reinhart J, Xiao S, Arora KK, Catt KJ. Structural organization andcharacterization of the promoter region of the rat gonadotropin re-leasing hormone receptor gene. Mol Cell Endocrinol 1997; 130:1–12.

51. Grosse R, Schoneberg T, Schultz G, Gudermann T. Inhibition of go-nadotropin-releasing hormone receptor signaling by expression of asplice variant of the human receptor. Mol Endocrinol 1997; 11:1305–1318.

52. Pati D, Habibi HR. Characterization of gonadotropin releasing hor-mone (GnRH) receptors in the ovary of common carp (Cyprinus car-pio). Can J Physiol Pharmacol 1992; 70:268–274.

53. Habibi HR, Pati D, Ouwens M, Goos HJTh. Presence of gonadotropinreleasing hormone (GnRH) binding sites and compounds with GnRH-like activity in the ovary of african catfish, Clarias gariepinus. BiolReprod 1994; 50:643–652.

54. Pati D, Habibi HR. Characterization of gonadotropin releasing hor-mone receptors in goldfish ovary: variation during follicular devel-opment. Am J Physiol 1993; 264:R227–R234.

55. Nabissi M, Pati D, Polzonetti-Magni AM, Habibi HR. Presence andactivity of compounds with GnRH-like activity in the ovary of sea-bream Sparus aurata. Am J Physiol 1997; 272:R111–R117.

56. Kah O, Breton B, Dulka JG, Nunez-Rodriguez J, Peter RE, CoriganA, Rivier JJ, Vale WW. A reinvestigation of the Gn-RH (gonadotro-pin-releasing hormone) systems in the golfish brain using antibodiesto salmon Gn-RH. Cell Tissue Res 1986; 244:327–337.

57. Munz H, Claas B, Stumpf WE, Jennes L. Centrifugal innervation ofthe retina by luteinizing hormone releasing hormone (LHRH)-immu-

noreactive telencephalic neurons in teleostean fishes. Cell Tissue Res1982; 222:313–323.

58. Nevitt GA, Gorber MS, Marchaterre MA, Bass AH. GnRH-like im-munoreactivity in the peripheral olfactory sytem and forebrain of At-lantic salmon (Salmo salar): a reassessment at multiple life historystages. Brain Behav Evol 1995; 45:350–358.

59. Umino O, Dowling JE. Dopamine release from interplexiform cells inthe retina: effects of GnRH, FMRFamide, bicuculline, and enkephalinon horizontal cell activity. J Neurosci 1991; 11:3034–3046.

60. Behrens UD, Douglas RH, Wagner HJ. Gonadotropin releasing hor-mone, a neuropeptide of efferent projections to the teleost retina in-duces light-adaptative spinule formation on horizontal cell dendritesin dark-adapted preparations kept in vitro. Neurosci Lett 1993; 24:59–62.

61. Mananos E, Anglade I, Chyb J, Saligaut C, Breton B, Kah O. In-volvement of g-aminobutyric acid (GABA) in the control of GTH-1and GTH-2 secretion in male and female rainbow trout. Neuroendo-crinology 1999; 69:262–280.

62. Madigou T, Mananos-Sanchez E, Gonzalez-Martinez D, Hulshof S,Ferriere F, Le Drean G, Yu KL, Kah O. Molecular cloning and ex-pression of gonadotropin releasing hormone receptor (GnRH-R) in therainbow trout: comparison with the sea bass (D. labrax). In: Proceed-ings of the 6th international symposium on the reproductive physiol-ogy of fish; 1999; Bergen, Norway.

63. Volkoff H, Peter RE. Actions of two forms of gonadotropin releasinghormone and a GnRH antagonist on spawning behavior of the goldfishCarassius auratus. Gen Comp Endocrinol 1999; 116:347–355.

64. Anglade I, Zandbergen AM, Kah O. Origin of the pituitary innerva-tion in the goldfish. Cell Tissue Res 1993; 273:345–355.

65. Navas JM, Anglade I, Bailhache T, Pakdel F, Breton B, Jego P, KahO. Do gonadotropin releasing hormone neurons express estrogen re-ceptors in the rainbow trout? A double immunocytochemical study. JComp Neurol 1995; 363:461–474.

66. Whitmore D, Foulkes NS, Strahle U, Sassone-Corsi P. Zebrafish clockrhythmic expression reveals independent circadian oscillators. NatureNeurosci 1998; 1:701–707.

67. Ito H, Yoshimoto M. Cytoarchitecture and fiber connections of thenucleus lateralis valvulae in the carp (Cyprinus carpio). J CompNeurol 1990; 298:385–399.

68. Ekstrom P. Distribution of choline acetyltransferase-immunoreactiveneurons in the brain of a cyprinid teleost (Phoxinus phoxinus L.). JComp Neurol 1987; 256:494–515.

69. Amano M, Oka Y, Aida K, Okumoto N, Kawashima S, Hasegawa Y.Immunocytochemical demonstration of salmon GnRH and chickenGnRH-II in the brain of the masu salmon, Oncorhynchus masou. JComp Neurol 1991; 314:587–597.

70. Munz HP, Stumpf WE, Jennes L. LHRH systems in the brain of plat-yfish. Brain Res 1981; 221:1–13.