Eel gonadotropin receptor genes.

Genes encoding gonadotropin receptors were sought by Blast searches [34] using the Japanese eel (Anguilla japonica) fshr and lhcgr cDNA sequences (AB360713, EU635883, AY742794, AY742795, [27–29]), against the European and Japanese eel genomic sequence database (Eelgenome.org; NCBI) [24–25]. Exons and introns were annotated using the package CLC Main Workbench (CLC Bio, Qiagen, Denmark).

Gonadotropin receptors from other teleosts.

Using the FSHR and the two LHR protein sequences identified in the European eel in this study as queries (tBlastn), fshr and lhcgr genes were sought and annotated (if not yet predicted) in the increasing number of teleost genomes available in Ensembl or NCBI (see S1 Table).

Gonadotropin receptor genes from other vertebrates.

In addition to the already known vertebrate gonadotropin receptors, fshr and lhcgr genes were also sought in genomes of other vertebrate species of key-phylogenetic positions (S1 Table): in two chondrichthyans, an elasmobranch, the little skate, Leucoraja erinacea and a holocephalan, the elephant shark, Callorhinchus milii, in a basal sarcopterygian fish, the coelacanth, Latimeria chalumnae, in various sauropsids including avians, crocodilians, chelonians and squamates (e.g. chicken, Gallus gallus; white-throat sparrow, Zonotrichia albicollis; American alligator, Alligator mississipiensis; Chinese alligator, Alligator sinensis; Chinese soft-shelled turtle, Pelodiscus sinensis; painted turtle, Chrysemys picta; green anole, Anolis carolinensis; Burmese python, Python bivittatus; king cobra, Ophiophagus hannah) and in a non-teleost actinopterygian, the spotted gar, Lepisosteus oculatus.

Gene coding sequences (CDS) were annotated or corrected by comparison with well-known orthologous CDS from other species, in respect with the canonical splice site rule using CLC Main Workbench (S1 Table).

RNA samples and cDNA synthesis.

Tissue distribution analysis was performed on RNA samples from 8 female silver migrating eels caught in the Loire River in France and previously prepared [32–33]. Total RNA extracted from ovary, olfactory bulb, mesencephalon and diencephalon, telencephalon, cerebellum, medulla oblongata, pituitary, eyes, liver, intestine, muscle, adipose tissue, gills, and thyroid follicles were used. Previously prepared RNA samples from testis [26] from 8 male silver eels were also analyzed. The silver stage is the last continental phase of the eel life cycle. The silver eels are ready to perform the oceanic reproductive migration, but are still immature, the ovary being at the pre-vitellogenic stage, and the testis at the spermatogonial stage. Four hundred ng of total RNA were reverse transcribed using Superscript III (Invitrogen) and random primers (50 ng) (Invitrogen), at 50°C for 60 min after an initial step of 25°C for 10 min.

Quantitative PCR primer design and analyses.

Eel specific primer sets for each gonadotropin receptor were designed using Primer3 [35] on two successive exons, or on exon junctions, of the extracellular domain. Various primer set combinations were tested and compared. Efficiency and amplification specificity were checked for each primer set. The sets with the higher efficiency (100% for fshr and lhcgr1 and 87% for lhcgr2) were chosen for the following quantification experiment (S2 Table). Quantitative PCR (qPCR) product was sequenced to control the primer specificity.

Messenger RNAs were assayed using Light Cycler System (Roche) with the LightCycler FastStart Master plus SYBR Green I kit (Roche) as recommended by the manufacturer. The final primer concentration used was 500 nM. Each sample was run in duplicate. The PCR conditions were 95°C for 10 min followed by 50 cycles at 95°C for 5 sec, 60°C for 10 sec and 72°C for 10 sec. The specificity of amplified qPCR products, was checked by melting curve analysis. Relative transcript levels were quantified using standard curves prepared with cDNA from tissue samples in which they were abundant, i.e. ovarian, telencephalon, or testis, for lhcgr1, lhcgr2, and fshr, respectively. Results were expressed as arbitrary units of gonadotropin receptor transcript level / total RNA level, and normalized to the mean expression value in the ovary, considered as 1.

Prediction of a single fshr gene in both European and Japanese eel draft genomes.

Blast searches using the fshr cDNA sequence cloned in the Japanese eel (AB360713; [28]) allowed us to retrieve one scaffold covering the full-length fshr cDNA in the Japanese eel genome, as well as in the European eel genome (S1 Fig).

Prediction of two lhcgr genes in both European and Japanese eel draft genomes.

Blast of the lhcgr cDNA sequences previously reported in Japanese eels (AY742795; [27] Jeng et al. 2007 and EU635883; [29]) against the Japanese eel genome, allowed the prediction of two lhcgr genes located on two distinct scaffolds (S1 Table). One lhcgr gene corresponds to the cDNA sequence identified by Kazeto et al. [29]. We named it lhcgr1 according to phylogenic and synteny analyses (see sections below). The other lhcgr showed high sequence identity with the partial cDNA sequence identified by Jeng et al. [27] and was named lhcgr2. Blast of the two Japanese eel predicted lhcgr sequences against the European eel genome also allowed the identification of two lhcgr genes located on two distinct scaffolds. European eel lhcgr2 sequence could be only partially predicted, due to low assembly resolution or ambiguities of the European eel draft genome. To confirm the presence of three gonadotropin receptors in the eels and complete the predicted coding sequences, the cDNA for fshr, lhcgr1 and lhcgr2 were cloned in the European eel.

Cloning of the cDNAs of the three European eel gonadotropin receptors.

Fshr cDNA (2040 bp) was cloned using testis RNA (S2 Fig). The cloned sequence shared 99.7% identity with the predicted European eel fshr CDS and the recently published European eel fshr cDNA cloned from ovarian RNA by Minegishi et al. [31] (AB700600). Few mismatches observed between the predicted genes and the corresponding transcripts may reflect some polymorphism. These differencies concerned the third base position of the codons, with no impact on amino acid sequences. A partial cDNA of 1913 bp encoding lhcgr1 was cloned from ovarian RNA (S3 Fig). This sequence covered 90.4% of the predicted CDS (2115 bp) of European eel lhcgr1. It shared 99% identity with the European eel lhcgr1 CDS and 98.1% identity with the Japanese eel lhcgr1 cDNA (EU635883; [29]). The full-length coding sequence of the European eel lhcgr2 was cloned (2104 bp) using testis RNA (S4 Fig). It shared 98.5% identity with our predicted Japanese eel lhcgr2 CDS. It is the first time that duplicated lhcgr transcripts are isolated in a vertebrate species.

Prediction of the protein domains of the eel gonadotropin receptors.

European eel fshr CDS encoded a 660 amino acids (aa) FSHR including a 17 aa signal peptide (S2 Fig). Lhcgr1 and lhcgr 2 CDS encoded a 704 aa LHR1 including a 26 aa signal peptide, and a 700 aa LHR2 including a 17 aa signal peptide, respectively (S3 and S4 Figs). The three eel gonadotropin receptors showed the typical topology of the glycoprotein hormone receptors with a long N-terminal extracellular domain (ECD), corresponding to about the half of the whole amino acid sequence (322 aa, 349 aa and 359 aa, for FSHR, LHR1 and LHR2, respectively) (Fig 1). The N-terminal cysteine box (cb1) of ECD of the eel receptors included 4 cysteines for both eel LHR as for mammalian glycoprotein hormone receptors, but only 2 cysteines for FSHR as in some other teleosts. The typical leucine-rich repeat domain (LRRD) of the ECD was composed of 11 leucine-rich repeats (LRR) for the three eel receptors as in mammals (Figs 1 and 2). The succession of LRR confers to the ECD a concave shape that is involved in the hormone binding [40–41,47–50]. The C-terminal part of the ECD, which corresponds to the hinge region, was delimited in the three eel receptors by two conserved cysteine boxes (cb2 and cb3) containing 3 cysteines each. The transmembrane domain (TMD) of the three eel receptors (263 aa, 267 aa, and 264 aa, for FSHR, LHR1 and LHR2 respectively) was composed, as for all GPCR, of highly conserved seven transmembrane alpha-helices (TH), connected by 3 extracellular (el) and 3 intracellular (il) loops. The receptors terminated with an intracellular domain (ICD) corresponding to a C-terminal cytosolic tail (58 aa, 62 aa, and 60 aa, for FSHR, LHR1 and LHR2 respectively). The ICD is involved in GPCR signal transduction and receptor intracellular trafficking and desensitization (for review [51]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Gene organization of gonadotropin receptors in human, eel, spotted gar and elephant shark.

Numbers cover the 11 exons of the lhcgr and the 10 exons of the fshr. On the coding sequences are indicated: signal peptide, extracellular domain including β-strand of leucin-rich repeat (LRR), transmembrane helices (TH), intracellular and extracellular loops (il and el, respectively) and intracellular domain. Cysteine residues, potential N-glycosylation and sulfation sites are also illustrated.

https://doi.org/10.1371/journal.pone.0135184.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Phylogeny relationships of the gonadotropin receptors.

Phylogram of maximum likelihood relationships between LHR and FSHR amino acid sequences of representative gnathostome species. Bootstrap values over 500 replicates (%) are given next to each node in red and SH-like aLRT values (%) are given in black (when different from the boostrap value). Asterisks indicate partial sequences. The actinopterigyan FSHR and LHR clades are highlighted to facilitate the phylogram examination. The sequence of the glycoprotein hormone receptor (GpHR) of an early chordate, the lancelet, was designated to root the tree.

https://doi.org/10.1371/journal.pone.0135184.g002

Comparison of the gene structure of the three eel gonadotropin receptors.

Complementary DNA sequences allowed us to assess the intron-exon gene structure of the eel gonadotropin receptors. The European eel fshr gene consisted of 10 exons and showed the typical exon-intron structure described for the human fshr [52], (Fig 1). The coding part of first exon (156 bp) encoded the signal peptide and the N-terminal region of the ECD, including the first cysteine box (cb1) and one of the LRR. A succession of 7 small exons (69-81bp) and a larger 9th exon (188 bp) encoded the major part of the ECD, including all the other LRRs, and two cysteines of the second cysteine box (cb2). The large 3’-terminal “rhodopsin like” exon, covered the C-terminal region of the ECD, including the third cysteine of the cb2, the hinge region and the cb3, together with the TMD and ICD. Prediction of the fshr gene structure in other vertebrates such as chondrichthyans, non-teleost actinopterygian and various teleosts (Fig 1 and S1 Table) showed a similar typical 10 exon structure, with the exception of salmonid and percomorph fshr, which present several additional exons as previously reported [21,53].

The European eel duplicated lhcgr genes were both organized in 11 exons as the human lhcgr (Fig 1) [4,54]. As compared to eel fshr gene, they included an additional exon between the 9th exon and the large terminal exon. As for human lhcgr, this additional small exon (72 bp and 69 bp for eel lhcgr1 and lhcgr2, respectively) encoded the third cysteine of the cb2, and a part of the hinge region. This typical 11-exon structure was also predicted for lhcgr from various vertebrates, including chondrichthyans, a non-teleost actinopterygian and various teleosts [21] (Fig 1 and S1 Table) (see next section).

A single fshr and one or two lhcgr are present in teleosts.

The screening of the available representative teleost genomes revealed, as in eels, the presence of one fshr and of both lhcgr1 and lhcgr2 genes in seven other species, e.g. in a characiforme, the Mexican tetra, Astyanax mexicanus, in a salmonid, the Atantic salmon, Salmo salar, in two atherinomorphs, the platyfish, Xiphophorus maculatus and the guppy, Poecilia reticula, and two cichlids, the Nile tilapia, Oreochromis niloticus, and zebra mbuna, Maylandia zebra, and in a pomacentridae, the bicolor damselfish, Stegastes partitus (S1 Table). The coexistence of duplicated lhcgr in various teleost species clearly refutes one previous assumption of two mutually exclusive lhcgr in this lineage [21]. In Atlantic salmon, in addition to the lhcgr2 gene, we succeeded to annotate duplicated lhcgr1 genes (S1 Table and S5 Fig). The large number of lhcgr1 in this species may result from the recent salmonid tetraploidization (4R) [55].

In contrast, we found a single fshr gene and alternatively only one or the other of the two lhcgr genes, in seven other teleost annotated genomes. One lhcgr homologous to eel lhcgr2 (see section Phylogeny) was identified in a cypriniforme, the zebrafish, Danio rerio, while lhcgr homologous to eel lhcgr1 (see section Phylogeny) was identified in a gadiforme, the Atlantic cod, Gadus morua, in a beloniformes, the Japanese medaka, Oryzias latipes, in a gasterosteiforme, the stickleback Gasterosteus aculeatus, and in two tetraodontiformes, the pufferfish, Tetraodon nigroviridis and the fugu, Takifugu rubripes. In another cypriniforme, the common carp, Cyprinus carpio carpio, two putative lhcgr2 were identified (one full-length and one partial with potential frameshifts), that may result from the recent cyprinid tetraploidization (4R) [56–57] (S1 Table).

A single fshr and two lhcgr are present in a non-teleost actinopterygian.

We could predict one fshr and two lhcgr (S1 Table), homologous to eel lhcgr1 and lhcgr2, respectively (see section Phylogeny), in the genome of a holostean actinopterygian, the spotted gar. The fshr consisted of a succession of 10 exons, as human and eel fshr (Fig 1), and encoded a 719 aa receptor. Both lhcgr were organized in 11 exons as eel lhcgr1 and lhcgr2, and human lhcgr (Fig 1), and encoded a LHR1 (696 aa) showing 53% identity with eel LHR1, and a LHR2 showing 59% identity with eel LHR2, respectively (Table 1). Considering that the holostean lineage has diverged before the teleost-specific 3R, the duality of lhcgr in the gar indicates that a gene duplication event has already occurred before the teleost 3R.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Overall and subdomain amino acid sequence comparison between the European eel LHR1, LHR2 and FSHR and representative species (% of identity).

https://doi.org/10.1371/journal.pone.0135184.t001

A single fshr and a single lhcgr are present in sarcopterygians.

In a basal sarcopterygian, the coelacanth, one partial lhcgr could be predicted on scaffold JH127072 (S1 Table). This partial sequence consisted of 10 exons giving a 632 aa protein, the first exon encoding for the signal peptide being missing. In addition, we could identify one fshr gene (S1 Table) on scaffold JH127072, positioned in tandem with the lhcgr. The fshr sequence consisted of 10 exons and encoded for a 690 aa receptor. As in the coelacanth, a single fshr and a single lhcgr are present in the other sarcopterygians, including tetrapods. No additional lhcgr could be found in any sarcopterygians. In the course of our investigation of genome databases, we observed that some sauropsid species possess fshr but may have lost lhcgr. In birds, fshr and lhcgr are still present and positioned in tandem, as in coelacanth and mammals. In contrast, we could not find any lhcgr in the genome of three squamates, the anole lizard, the Burmese python and the king cobra. Furthermore, we found several premature stop codons and frameshifts in the putative lhcgr sequences of a chelonian, the painted turtle and of crocodilians, the Chinese and American alligators (S1 Table). This suggests that a loss of a functional LHR would have occurred several independent times during the sauropsid radiation, in squamate, chelonian and crocodilian lineages [58–59]. Similarly, our recent studies on the kisspeptin system revealed multiple and independent losses of kiss and kiss receptor genes during the sauropsid radiation [60].

A single fshr and a single lhcgr are present in chondrichthyans.

In the elephant shark genome, a single fshr and a single lhcgr could be predicted on scaffold 134 and scaffold 49, respectively (S1 Table). They showed the typical gene organization consisting of 10 exons and 11 exons (Fig 1), respectively. Fshr encoded a 741 aa receptor, and lhcgr a 709 aa receptor. From the little skate contig data, we succeeded to assemble a single partial fshr encoding 570 aa and a single lhcgr encoding 691 aa (S1 Table).

In conclusion, genome searches indicated that among all gnathostomes investigated, the actinopterygians, including teleost and non-teleost species, were the only group with two lhcgr. This suggests that duplicated lhcgr may represent a specific feature of the actinopterygian lineage.

Divergence in the extracellular domain between the duplicate LHR.

Amino acid sequences of ECD, a region mainly involved in ligand binding, revealed some divergence between eel duplicated LHR (43.6% identity). As expected, further divergence was observed between each eel LHR and eel FSHR (34.1 and 32.1%, for LHR1 and LHR2, respectively). Among actinopterygians, ECD sequence identity was slightly higher between LHR orthologs (49.7 to 57.8% for LHR1, 46.5 to 52% for LHR2) than between LHR paralogs (39.4 to 44%). The conservation of the ECD between all actinopterygian LHR paralogs was still higher than between LHR and FSHR (30.1 to 34.5% identity) (Table 1).

N-glycosylation sites could be predicted in the ECD of duplicate eel LHR, 3 for LHR1 and 5 for LHR2 (S7 Fig). These sites included for both LHR, a site on LRR7 common to all glycoprotein hormone receptors [19,63], and a site on cb2 conserved among all LHR, except spotted gar LHR2. A N-glycosylation site on LRR3 appeared specific to actinopterygian LHR2. In human, directed-mutagenesis of potential N-glycosylation sites of gonadotropin receptors, or disruption of glycosylation, demonstrated that glycosylation plays a role in the proper folding of the nascent protein into a mature receptor able of binding hormone and signaling [64–65].

An O-sulfation tyrosine site could be predicted in the hinge region of eel LHR2 but not eel LHR1 (S7 Fig). It corresponds to the sulfated tyrosine in the hinge region of human LHCGR (Tyr³³¹), FSHR (Tyr³³⁵) and TSHR (Tyr³⁸⁵), that has been postulated to be a prominent feature, conserved among all the GPHR, and involved in receptor activation [40–41,66–67]. This sulfation tyrosine site was predicted in some but not all teleost LHR2, and in none of teleost LHR1 nor teleost FSHR [68]. The lack of potential sulfation tyrosine site in most teleost gonadotropin receptors suggests alternative receptor activation mechanisms than the one requiring the sulfation of the hinge region, as described in mammals [41].

Amino acid comparison of the ECD revealed residues specific of one or the other duplicate LHR, which may constitute fingerprints of each LHR type (S7 Fig). These fingerprints were more marked in teleosts than in spotted gar, indicating that the duplicate LHR may have further diverged in teleosts. Some of these specific residues could be observed in the LRRD domain (S7 Fig). We also found specific features in the hinge region, the LHR1 showing a conserved motif negatively charged (AFHTWRR), that was deleted in the eel LHR2 and substituted in the other teleost LHR2 (S7 Fig). At the level of the Cb3, eel and other teleost LHR1 possess a proline residue (Pro³⁵⁹ for eel LHR1), as sarcopterygian LHR as well as spotted gar LHR1 and LHR2, while proline is substituted by alanine (Ala³⁶⁰) in eel LHR2, or by alanine or glutamic acid in other teleost LHR2 (S7 Fig). Minor residue replacements were also observed between eel duplicated LHR, as for instance in the LRR6, the substitution of Glu¹⁶⁰ in LHR1 by aspartic acid (Asp¹⁶¹) in LHR2 (S7 Fig). In human LHCGR, replacement by site directed mutagenesis of the corresponding Glu¹⁵⁴ by Asp was shown to provoke an alteration of the response intensity [69].

All these variations in ECD sequences suggest that duplicated LHR may have diverged in their binding and activation properties in teleosts. So far functional studies on teleost recombinant gonadotropin receptors, using transiently transfected mammalian cell lines (Cos-7 or HEK), were performed either on FSHR, or on one or the other of the duplicate LHR, depending of the LHR type isolated [29–31,68,70–71]. Data on ligand selectivity for eel LHR receptor are only available for eel LHR1 and not for LHR2. Japanese eel LHR1 is activated by eel recombinant LH (recLHR) and not to by eel recFSH, but can also be cross-activated by heterologous gonadotropins such as human chorionic gonadotropin (hCG), human FSH and trout LH [29–31,68]. Binding specificity toward homologous LH, with no cross-reaction of homologous FSH, has also been reported for LHR in other teleosts, whatever the LHR type: for LHR1 in channel catfish [15], Nile tilapia [68] and medaka [72]; for LHR2 in African catfish [73–74], zebrafish [75], salmon [13–14], European sea bass, Dicentrarchus labrax [76], and Senegalese sole, Solea senegalensis [21]. This indicates that both duplicate LH receptor types would be specific of LH in teleosts. Further functional studies, using recombinant receptors in species having conserved duplicated LHR, are still required to determine whether the dissimilarities between duplicated LHR ECD may affect ligand binding and receptor activation.

Conservation of the transmembrane domain between the duplicate LHR.

The amino acid comparison of the TMD showed a high identity between the duplicated eel LHR (68.4%) as well as with FSHR (69.9 and 63.9%) (Table 1). A strong amino acid identity was particularly observed among gnathostome gonadotropin receptors for the transmembrane helices 2, 3, 6 and 7, which include key residues participating to the network involved both in the transition of the inactive toward active receptor conformation and in binding the protein G [77] (S7 Fig). Nevertheless, some specific residue divergences between the duplicate LHR could be also noticed within the TMD (S7 Fig). As discussed above for ECD, these fingerprints in the TMD were more marked in teleosts than in spotted gar, indicating that the TMD of the duplicate LHR have further diverged in teleosts (S7 Fig). The phosphorylation site (TVR) in the il3 of the mammalian FSHR was highly conserved in the other gnathostome FSHR including the eel FSHR (S7 Fig). Noticeably, this phosphorylation site was retrieved in spotted gar LHR1, eel LHR1 and some other teleost LHR1, but not in any LHR2 nor in any sarcopterygian LHR (S7 Fig). In the rat, phosphorylation of FSHR il3 was identified as an important determinant in arrestin association [78–79]. Some differences between duplicate LHR TMD sequences might thus confer differential properties in the regulation of signal transduction.

High divergence in the cytosolic tail between the duplicate LHR.

The sequence comparison of the ICD showed a very low identity between the duplicated eel LHR (16.9%). A higher conservation was noticed among teleost LHR1 (43.6–49.3% identity) as compared to teleost LHR2 (14.7–32%) (Table 1). Specific signatures for teleost LHR1 included amino acids shared with other gnathostome LHR but not LHR2 (such as R⁶⁵² of eel LHR1), as well amino acids conserved only among actinopterygian LHR1 (Cys⁶⁹⁷) or among teleost LHR1 (AYHIK) (S7 Fig). Nevertheless, some sites were highly conserved among all gonadotropin receptors, such as the characteristic cysteine palmitoylation sites of the ICD of the GPCR (e.g. Cys⁶⁴³-Cys⁶⁴⁴ for human LHCGR), which are common to all LGR among GPCR, and involved in receptor internalization and recycling [80–81]. Gonadotropin receptor signaling in mammals is mainly mediated by adenylyl cyclase /cAMP pathway, but alternative signaling, such as phospholipase C/inositol phosphate (IP) and beta-arrestin pathways, are also requested for global activation of the response [82]. To our knowledge, investigations on teleost recombinant gonadotropin receptors mainly focused on the cAMP pathway, while a single study also addressed the IP pathway [73]. The large dissimilarity in cytosolic tail sequences between the duplicated LHR may reflect divergence in downstream intracellular signaling cascade and/or receptor internalization and desensitization process. Such differences may represent selective forces that have contributed to the conservation of the duplicated LHR. There are still limited data on the characterization of the signaling pathway activation and receptor trafficking in teleost fishes. Future investigations on the duplicated LH receptors are needed to characterize their respective activation specificities, and further infer the functional evolution of the LHR.

Comparative tissue distribution of the three eel gonadotropin receptors.

Development of specific qPCR for eel fshr and two lhcgr allowed analyzing their tissue distribution in European silver eels.

Eel fshr, lhcgr1 and lhcgr2 transcripts are all expressed in the gonads. The three gonadotropin receptors were easily detectable in the immature ovary of female silver eels. The mean ovarian transcript level was used as a reference for comparison with the other tissues (Fig 5). The three receptors were also well detectable in immature testis of male silver eels, with a much higher expression of fshr in testis as compared to ovary. Previous qPCR data in Japanese eel also indicated the expression of fshr and lhcgr in both ovary and testis [27–28,30]. Early radioreceptor and autoradiographic studies in the European eel ovary, revealed binding sites for carp LH and hCG in both granulosa and external theca layers [83]. In Japanese eel, FSHR was immunolocalized in the testis, in both Sertoli cells surrounding spermatogonia cysts and interstitial Leydig cells [28].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Tissue distribution of fshr, lhcgr1 and lhcgr2 transcripts in European eel.

Messenger RNA levels for fshr (B), lhcgr1 (C) and lhcgr2 (D) were assayed by qPCR in various tissues of silver female European eels: Ovary (Ov); brain dissected in five parts: olfactory bulbs (Ob) telencephalon (Tel), di- and mes-encephalon (Di/Mes), corpus cerebellum (Cb), and medulla oblongata (Mo); pituitary (Pit); eyes; liver (Liv); intestine (Int); muscle (Mus); adipose tissue (AT); gills and thyroid follicles (TF). Receptor mRNA levels in testis of silver male European eels were also assayed. Data are normalized to total RNA and the expression level in the ovary was set as 1. Means are given ± SEM (n = 8 eels).

https://doi.org/10.1371/journal.pone.0135184.g005

In other female teleosts, in situ hybridization (ISH) revealed that fshr transcript is strongly expressed in granulosa cells, and weakly in theca cells, in chub mackerel, Scomber japonicus [84], Atlantic halibut, Hippoglossus hippoglossus [85] and Atlantic salmon [86]. Lhcgr1 transcript is also expressed in granulosa and weakly in theca cells in chub mackerel [84] and [85] Atlantic halibut. Similarly lhcgr2 transcript is highly expressed in granulosa cells and weakly in theca cells in Atlantic salmon [86]. In medaka, immunocytochemistry (ICC) revealed LHR1 protein labeling in granulosa and theca cells [72].

In male teleosts, fshr transcript is localized by ISH in Sertoli cells, and a weak expression is also detected in Leydig cells in African catfish [87]. ISH and ICC studies also showed FSH expression in both Leydig and Sertoli cells in Senegalese sole [21,88]. Lhcgr2 is expressed in Leydig cells in African catfish [87], and in both Sertoli cells and Leydig cells in zebrafish [89]. In Senegalese sole, ISH and ICC revealed lhcgr2 transcript and LHR2 in Leydig cells and free spermatids, supporting a direct role of LH on spermiogenesis [21,88,90].

In tetrapods, as in the other vertebrates, fshr and lhcgr are mainly expressed in the gonads [52]. The present study supports that fshr as well as duplicated lhcgr1 and lhcgr2 have all conserved a role in the control of gonadal function in the eel. Future studies, including ISH, should aim at further analyzing the cellular localization of the expression of the duplicated lhcgr in the eel, as well as in the other fish species, which have also conserved duplicate lhcgr.

Eel fshr, lhcgr1 and lhcgr2 are differentially expressed in the brain (Fig 5). The three gonadotropin receptors are also expressed in non-gonadal tissues, as analyzed by qPCR in the female European silver eel. Lhcgr2 showed a remarkable expression in the whole brain, higher than the expression measured in the ovary, while lhcgr1 transcript was at the limit of detection. Fshr was highly expressed in some brain regions, such as olfactory bulb and telencephalon (Fig 5). High brain expression of fshr was also reported in female Japanese eel at previtellogenic and late vitellogenic stages [30].

Fshr transcript is also found in the brain of males or females from other teleost species, such as Atlantic salmon [86], and Atlantic halibut [85]. Lhcgr2 transcript is expressed in the brain in various teleosts, including African catfish, Atlantic salmon and European sea bass [20,73,86,91]. Lhcgr1 is also expressed in the brain of the Korean rockfish, Sebastes schlegelii [92] and Atlantic halibut [85], while its expression is not detectable in the brain of the channel catfish [15].

Lhcgr expression has been reported in the brain of amphibians, birds and mammals including human [93–95]. In mammals, various functions for brain LHR have been suggested such as GnRH regulation [96], sensory modulation, or fetal neurogenesis [97]. In African clawed frog, Xenopus laevis, LH controls courtship songs via its cognate receptor expressed in forebrain vocal nuclei [95]. To our knowledge there is no report on FSHR in the brain of tetrapods. The fact that fshr and lhcgr are highly expressed in brain in the eel as in some other teleosts, opens new research avenues on the potential roles of gonadotropin signaling in the vertebrate brain. In the eel, the remarkable differential expression of the duplicated lhcgr in brain specifically confers to lhcgr2 major roles in brain functions.

Eel fshr, lhcgr1 and lhcgr2 are differentially expressed in other non-gonadal tissues. Both lhcgr1 and lhcgr2, but not fshr, transcripts were abundant in eye in the eel (Fig 5). Lhcgr2 transcript has also been reported in eye of the European sea bass and mummichog, Fundulus heteroclitus [91,98]. In human, lhcgr transcript and LHCGR protein have been identified in retinal photoreceptors, where LH receptor could be involved in local modulation of vision [99]. The expression of lhcgr1 and lhcgr2 in eye of the eel may reflect an ancestral role in this organ, which may have been conserved by both duplicated receptors.

Fshr and lhcgr1 were well expressed in the eel gills, while lower levels of lhcgr2 were found (Fig 5). Fshr transcript is also expressed in the gills of Senegalese sole [21], lhcgr1 in the gills of channel catfish [16] and Atlantic halibut [84], and lhcgr2 in the gills of Atlantic salmon, European sea bass and mummichog [20,91,98]. This suggests that gonadotropin receptors may participate in the complex multi-endocrine regulation of gill function in teleosts.

Fshr as well as both lhcgr1 and lhcgr2 were expressed at a low level in the eel adipose tissue. Lhcgr2 transcript was also reported in the fat tissue of European sea bass [91]. FSHR transcript and protein have been detected in the abdominal fat tissue of female chicken, where FSH stimulates lipid accumulation [100].

Eel fshr, lhcgr1 and lhcgr2 showed a very low expression, at the limit of detection level, in the other tissues investigated, such as pituitary, liver, intestine, muscle and thyroid. Lhcgr1 transcript is also undetectable by PCR in various non-gonadal tissues, including liver, intestine, heart and spleen, in Korean rockfish [92] and medaka [72]. In contrast, lhcgr1 transcript has been demonstrated in muscle, liver, stomach and heart in channel catfish [16] and in spleen in goby [101]. Lhcgr2 transcript has been observed also in liver, muscle and heart in African catfish [73], and in intestine and pituitary in European sea bass and mummichog [91,98]. This suggests variations in the non-gonadal target tissues and roles of LH, according to teleost species.

Evolutionary scenario of gonadotropin receptors.

Genome search, phylogeny and synteny analyses allowed us to revise the number and evolutionary history of gonadotropin receptors in gnathostomes (Fig 6). A single copy of fshr gene is present in gnathostomes. A single lhcgr gene is also present in sarcopterygians, including in the basal representative species, the coelacanth, and it is positioned in tandem with fshr. This tandem, also observed in a holostean actinopterygian, the gar (for fshr-lhcgr1), is likely the result of a local duplication of an ancestral gonadotropin receptor gene in early gnathostomes (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Current status and evolutionary scenario of gonadotropin receptors in gnathostomes.

The presence of gonadotropin receptors (fshr, lhcgr, lhcgr1, lhcgr2) in representative extant species of gnathostome lineages is shown on the right. In some species, the presence or absence of some gonadotropin receptors could not be assessed, due to the lack of genomic data, and these cases are indicated by a question mark. Two gonadotropin receptors, fshr and lhcgr are present in gnathostomes including chondrichthyans and osteichthyans. They were positioned in tandem in the early osteichthyes. The fshr-lhcgr tandem was maintained after the split between sarcopterygians and actinopterygians. Gene loss of lhcgr in various sauropsid groups led to the presence of only fshr in some extant sauropsids. A duplicated lhcgr was generated by a local gene duplication event in the actinopterygian lineage before the emergence of teleosts. Teleost whole-genome duplication (3R) generated duplicated fshr, lhcgr1 and lhcgr2. Multiple and selective gene losses after the 3R led to the maintenance of only a single copy of fshr, lhcgr1 and lhcgr2 and to the physical separation of the tandem fshr-lhcgr. Additional lhr losses occurred independently through the teleost radiation, i.e. lhcgr1 was lost in cyprinidae and lhcgr2 in some percomorphs. Recent tetraploidization (4R) event has generated additional copies of lhcgr1 in salmonids.

https://doi.org/10.1371/journal.pone.0135184.g006

In the elephant shark, fshr and lhcgr were not found positioned in tandem but each gene was on a distinct scaffold (S3 Table). This questions the origin of the fshr-lhcgr tandem. This suggests either that the tandem arose only in osteichthyan lineage after the emergence of the chondrichthyans, or that the tandem position was not maintained in the extant chondrichthyans. Further analysis using other chondrichthyan genomes would allow a better resolution of the glycoprotein hormone receptor evolution history.

Among tetrapods, genome search in sauropsids revealed multiple and independent lhcgr gene mutation or loss events in the squamate, crocodilian, and chelonian lineages. This suggests that gonadotropin action would be mediated only by fshr in these species. In contrast, birds have conserved lhcgr and fshr, positioned in tandem, as in the other tetrapods (Fig 3).

Our investigation revealed that duplicated lhcgr (named lhcgr1 and lhcgr2) are present in some actinopterygians, including the eel and some other teleosts, as well as in a holostean species, the gar. The coexistence of duplicated lhcgr in various actinopterygians, including extant teleosts, rules out a former hypothesis, that the duality of teleost receptors would result interallelic gene conversion and not from gene duplication [21]. We raised the hypothesis that these duplicated lhcgr genes may be the result of a second local gene duplication event, that would have occurred early in the actinopterygian lineage, before the split between holosteans and teleosts (Fig 6).

These duplicated lhcgr, which are present in a non-teleost actinopterygian, the gar, are clearly not the result of the teleost-specific whole-genome duplication (3R). Furthermore, phylogeny and synteny analyses show that there was no impact of 3R on the number of gonadotropin receptor genes in extant teleosts, due to multiple gene losses after the 3R. Our investigation in the eel, a representative species of a basal group of teleosts (elopomorph) suggests that these losses of 3R-duplicated gonadotropin receptor genes would have occurred soon after the 3R. Thus, only one copy of the 3R-duplicated fshr, lhcgr1 and lhcgr2 was retained, while the other one was lost early during teleost radiation (Fig 6), likely via the rediploidization process that follows polyploidization event. In agreement with previous studies [21,53], one consequence of these 3R-duplicate gonadotropin gene losses in teleosts, is the separation of the fshr-lhcgr tandem in all extant teleosts, with potential implications in the regulatory mechanisms of their expression.

Additional loss events of lhcgr1 or lhcgr2 have occurred independently throughout teleost radiation, so that some extant teleosts have conserved only a single lhcgr (lhcgr1 or lhcgr2), according to species (Fig 6). The loss of lhcgr1 in cyprinids and the preferential loss of lhcgr2 that occurred repeatedly through the percomorph lineage, suggest that the presence of two lhcgr may not be fully stabilized in teleosts. This diversity in lhcgr duplicate gene conservation or losses among teleosts, also raises the question of the possible presence of duplicate receptors in teleost species for which only one lhcgr has been characterized, while genomic data are still limited.

The recent tetraploidization events (4R) that occurred independently in cyprinids and salmonids, are reflected by some additional numbers of lhcgr genes, such as in Atlantic salmon (two lhcgr1 genes) (Fig 6). In contrast, loss of the other putative 4R-duplicated gonadotropin receptors in these species, illustrates the progressive gene fractionation occurring in polyploid genome [55].

Comparisons between gonadotropin receptor amino acid sequences revealed some divergences between teleost LHR1 and LHR2 paralogs, mainly concerning the ECD and ICD domains. These differences, which may affect receptor-binding properties, signaling pathways and recycling, might represent selective forces that have contributed to the conservation of duplicate LHR. We also compared the tissue distribution of the expression of fshr and duplicated lhcgr transcripts, using the eel as a model. All three gonadotropin receptors are expressed in the ovary and testis, and are thus involved in the mediation of the control of male and female reproductive function. In non-gonadal tissues, differential expression patterns were revealed. For instance, high expression of fshr and lhcgr2, but very low lhcgr1 expression was observed in the brain. Different gene regulatory environments could have driven the duplicated lhcgr to distinct extra-gonadal functions. This may act as evolutionary forces in the conservation of the duplicated lhcgr. The present discovery of duplicated lhcgr in actinopterygians, including various teleost species of biological, ecological and aquaculture relevance, opens new research avenues in basic and applied reproductive endocrinology, as well as in evolutionary biology. It will promote future investigation on the comparative function and regulation of gonadotropin receptors, in order to decipher the physiological significance of the conservation of duplicated LHR receptors in various species.