Sequence similarity searches using the sequences of the principal GPCR signaling components of the human HPG axis were performed against the C. elegans genome. This analysis indicated the presence of two proteins, one of 401 amino acids [GenBank: NP_491453] and another of 379 amino acids [GenBank: NP_506566] sharing 46.9% and 44.7% nucleotide similarity to that of human GnRHR1 and GnRHR2, respectively (Table 1). In addition, a leucine-rich GPCR previously reported in C. elegans [GenBank: NM_073147] has 47.6% nucleotide similarity to human follicle stimulating hormone receptor (FSHR) 9. This paper is primarily focused on the 401 amino acid protein [GenBank: NP_491453] orthologous to human GnRHR1. Amino acid sequence alignment of human GnRHR1 with the C. elegans orthologue (Ce-GnRHR) and 3 other class A GPCRs demonstrated considerable sequence similarity (Fig. 1). Like its human orthologue, 3 separate prediction algorithms used to analyze the structure of Ce-GnRHR predicted a GPCR belonging to the rhodopsin family (Figs. 1 and 2A; See additional file 1: TM_predictions.eps; 10). These prediction programs indicate that Ce-GnRHR contains several structural motifs similar to that of human GnRHR1, including 7 transmembrane domains, 3 intracellular and extracellular loops, and amino acid residues representing PKC phosphorylation sites and G-protein coupling sites (Figs. 1 and 2A; See additional file 1: TM_predicitions.eps; 11). As has been reported for all non-human GnRHR1s 11, the Ce-GnRHR has a long C-terminal tail (Figs. 1 and 2A) characteristic of a gene ancestral to GnRH and AKH receptors. Although the overall amino acid sequence identity between C. elegans and human GnRHR1 was only ~21%, closer analysis revealed that 56% of the functionally important amino acid residues (FIRs) of human GnRHR1 were conserved in Ce-GnRHR (Figs. 1 and 2A; Table 2). The highest degree of conservation was observed in those amino acids involved in receptor activation (83%), and G_q11 coupling (62%; Table 2). Lesser conservation was observed in amino acids important to binding pocket formation (54%), G_s coupling (50%), and PKC phosphorylation (50%). Ligand binding sites were least conserved (36%), indicating the possibility that the ligand of Ce-GnRHR is significantly different from that of human GnRHR1. Indeed, in Drosophila melanogaster, a cloned GnRHR 12 was subsequently identified as AKHR from ligand binding studies indicating AKH was the binding partner of this receptor in insects 13. In this connection, Ce-GnRHR had the highest overall sequence similarity (45%) and the highest FIR similarity (66%) with Drosophila melanogaster AKHR (Dm-AKHR; Table 2; 13), a receptor involved in metabolism (Table 2). In comparison, there was less conservation of overall sequence and FIRs in two other human class A GPCRs, rhodopsin (12% and 41%; respectively) and vasopressin receptor type 1A (18% and 44%; respectively) when compared with C. elegans (21% and 56%; respectively; Table 2). This comparative analysis suggests Ce-GnRHR is orthologous to insect AKHR and human GnRHR. The relatively greater conservation of FIRs observed between Ce-GnRHR, Dm-AKHR and human GnRHR1 (56–66% similarity; Table 2), suggests these proteins are all derived from a single ancestral GPCR, and, are thus orthologous to each other.

Despite the orthology of these sequences, the genomic organization of Ce-GnRHR, human GnRHR1, and Dm-AKHR are radically different (Fig. 2B). The coding region of each gene contains a different number of exons: 6 in Ce-GnRHR, 5 in Dm-AKHR, and 3 in human GnRHR1. Intronic sequence length is also variable. While the 2 introns of human GnRHR1 total more than 10 kilobases, the 4 introns of Dm-AKHR only amount to ~1 kilobase. Additionally, sequence corresponding to the 3' end of Ce-GnRHR exon 5 and all of Ce-GnRHR exon 6 is absent from human GnRHR. In order to deduce the genomic organization of a gene ancestral to all three of these GnRHR(-like) genes, a considerably larger number of genes would have to be analyzed. Yet, at least one coincidental feature of the gene maps shown in Fig. 2B might be indicative of ancestral gene structure: exon 1 of human GnRHR and exon 3 of DM-AKHR terminate at the same point.

To determine whether Ce-GnRHR was transcribed in the worm, we isolated RNA, and, using two gene-specific primers, amplified the region depicted in Fig. 2B. The expected 946 bp cDNA fragment encompassing exons 2 through 6 was detected (Fig. 2C). The sequence of the amplified cDNA [GenBank; NM_059052], matched the genomic sequence (chromosome 1; [GenBank: NC_003279]), minus intronic sequence, demonstrating that the amplified cDNA was amplified from Ce-GnRHR mRNA template. This confirms that Ce-GnRHR is actively transcribed in adult C. elegans.

To determine whether Ce-GnRHR protein was expressed in C. elegans, we generated a polyclonal antibody against the C-terminus (amino acids 386 to 401) of Ce-GnRHR and the nematode homogenate was subjected to immunoblot analyses. The affinity purified anti-Ce-GnRHR antibody recognized a 46-kDa band (Fig. 3A), the calculated molecular weight of Ce-GnRHR, as well as higher molecular weight (116-kDa, 237-kDa) variants that might represent mature or otherwise post-translationally modified species and lower molecular weight (27-kDa and 23-kDa) variants that might represent truncated protein products, as we have reported for human GnRHR1 14. Despite limited similarity between the N-terminus of the human and C. elegans receptor, a monoclonal antibody raised against the N-terminus of human GnRHR1 identifies a single band of 46-kDa in the C. elegans homogenate (See : Immunoblot 1.eps) and produces similar immunolocalization (see below); however, it remains unclear whether this antibody is detecting Ce-GnRHR. The specificity of the polyclonal antibody against Ce-GnRHR for Ce-GnRHR compared to the putative orthologue to human GnRHR2 [GenBank; NP_506566] (orthologue 2) is indicated by the fact that the antibody does not recognize a band at 43.5-kDa, the molecular weight of the immature form of orthologue 2 (Fig. 3A). Additionally, comparison of the amino acid sequences and genomic organization of Ce-GnRHR and orthologue 2 revealed limited similarity between the two genes, especially at the C-terminus of Ce-GnRHR – against which the anti-Ce-GnRHR antibody was raised (See additional file 3: Ortho.eps).

To examine the localization of Ce-GnRHR, eggs were immunostained with the monoclonal antibody against human GnRHR1, or the anti-Ce-GnRHR antibody (Fig. 3B). Both the N-terminus (anti-human-GnRHR1 antibody) and C-terminus (anti-Ce-GnRHR antibody) antibodies immunolabeled the same structures (see also later, Fig. 4), indicating that Ce-GnRHR is expressed in eggs. With both antibodies, intense staining was observed within the egg, but not the cuticle (Fig. 3B). Staining was not apparent in eggs treated with secondary antibody alone. The authenticity of the anti-human GnRHR1 antibody was verified by positive staining of gonadotropes in sections of human pituitary (data not shown).

To better localize cellular staining in worms, we permeabilized worms and performed whole-mount fluorescent immunohistochemistry. These experiments indicated that Ce-GnRHR was localized to the nucleus of maturing oocytes and intestinal cells (Fig. 4i–iii), to sperm, pharyngeal muscles (see later), but not other cells such as hypodermal cells. Similar staining of these structures was evident for both the anti-human-GnRHR1 and anti-Ce-GnRHR antibodies. Ce-GnRHR staining clearly illustrates an increase in the size of the oocyte nuclei as they mature along the gonadal arm (Fig. 4i, 4iii). Upon fertilization, Ce-GnRHR staining becomes diffuse throughout the developing egg, although staining appears to increase during egg development (see Figs. 3B for staining of laid eggs). Only uniform background autofluorescence was apparent in worms treated with secondary antibody alone (Figs. 4iv, 4v).

Interestingly, Ce-GnRHR also was detected along the myofilament lattice of the pharyngeal muscles as evidenced by 1). staining along the three parallel muscles that comprise the pharyngeal musculature beginning at the tip of the head and extending to the pharyngeal bulb (Fig. 5i; taken near the center of the body axis), 2). staining of all 8 pharyngeal muscle (pm1-8) domains, but not in the gaps between contractile zones of adjacent pharyngeal muscle domains (Fig. 5i), and 3). staining of myofilaments that run radially towards the lumen and which are most obvious in the bulbs (Fig. 5i; see online Handbook of Worm Anatomy 15 and 16 for details of pharyngeal anatomy). In Fig. 5ii taken at a higher focal plain, only one muscle lattice is observed (radial filaments are seen "end-on" pointing downward into the plane of section towards the lumen). These results indicate Ce-GnRHR is present on the pharyngeal musculature.

The specificity of binding of anti-Ce-GnRHR antibody to GnRHR was demonstrated by the lack of staining throughout the worm (including the germline, pharyngeal muscle and intestinal cells) when the antibody was preincubated with its antigen (the C-terminus amino acids 386 to 401 as described in the Methods; Fig. 5iii).

To confirm the nuclear localization of Ce-GnRHR, we stained C. elegans with Ce-GnRHR antibody and a nuclear stain (Hoechst dye) and superimposed the images. There was a clear overlap between Ce-GnRHR immunoreactivity and the nuclear stain in both oocytes and intestinal cell nuclei (Fig. 6i–vi), demonstrating that Ce-GnRHR was localized to the nuclear membrane of oocytes and intestinal cells. At higher magnification, separate chromosomes are evident with the Hoechst stain (Fig. 6ix) while Ce-GnRHR staining is uniform (Fig. 6viii), suggesting that Ce-GnRHR stains the membrane rather than intranuclear structures. This was again demonstrated by the scattered color of the superimposed image (Fig. 6x). In summary, these results indicate that Ce-GnRHR is present on the nuclear membrane of oocytes and intestinal cells. Further, these results corroborate earlier reports of GnRHR1 localization to the nucleus of proliferating cells 1718.

Maximum parsimony analysis of the Class A and B GPCR dataset (422 amino acids; See additional file 4: GPCRs_alignment.pdf) yielded 3 most parsimonious trees, the strict consensus of which resulted in the collapse of only 3 nodes (Fig. 7). Relatively robust support (68% boostrap proportion) was found for a group comprising all 3 vertebrate GnRHR types, and, though not achieving strong bootstrap support, all most parsimonious trees contained a group comprising tunicate GnRHR (Ciona intestinalis), insect corazonin receptors, and all vertebrate GnRHRs. Recovered in all of the most parsimonious trees, a monophyletic group of AKHRs was hypothesized as the sister group to the group Ce-GnRHR + C. briggsae-GnRHR. Together with the molluscan GnRHR (Crassotrea gigas), these receptors were hypothesized to be the sister group to the group of corazonin receptors + chordate GnRHRs discussed above. The group of all GnRHRs and GnRHR-like receptors was recovered more often than not (61% bootstrap proportion), to the exclusion of other Class A GPCRs included in the analysis.

The wild type N2 Bristol (Caenorhabditis Genetics Center, National Institutes of Health, National Center for Research Resources, MN) strain was cultured at 22–24°C under standard conditions on E. coli 51.

GnRHR1 monoclonal antibody (F₁G₄) raised against the N-terminus 1–29 amino acids of human GnRHR1 was a kind gift from Dr. Anjali Karande of the Indian Institute of Science, Bangalore, India 52. Rabbit polyclonal antibodies were raised against C-terminus amino acids 386 to 401 (Ac-GIDKRNHNVQLEIIDFC-OH) of Ce-GnRHR (The New England Peptides Inc., Gardner, MA), a region not found in human GnRHR1. This sequence was chosen based on antigenicity and to limit cross-reactivity with other proteins. The C-terminus included a cysteine for coupling purposes and the N-terminus was blocked by acetylation. Rabbits were immunized with this HPLC-purified antigen (structural confirmation was determined by mass spectrometry) using standard procedures (The New England Peptides Inc., Gardner, MA). Titer levels were monitored periodically in the animals and the animals bled after 60 days. The serum was then affinity purified (The New England Peptides Inc., Gardner, MA) for use in immunodetection assays. Secondary antibodies including goat anti-mouse IgG-HRP (sc-2055; for the GnRHR1 monoclonal antibody), goat anti-rabbit IgG-HRP (sc-2054; for the Ce-GnRHR antibody) as well as the Western blotting luminal reagent were from Santa Cruz Biotechnology, Inc., Santa Cruz, CA.

A BLASTp 53 search was performed against the C. elegans genome using human GnRHR1 and GnRHR2 as query sequences. For each query, the hit with the lowest E-value was aligned with human GnRHR1 using ALIGN 54. Transmembrane regions of Ce-GnRHR were predicted using the following programs: TMHMM V.2.0 55, DAS-TMfilter 56 and PSIPRED 57. In order to determine the level of conservation of functionally important amino acid residues, Ce-GnRHR was aligned with human GnRHR1 [GenBank: NP_000397], Dm-AKHR [GenBank: AAC61523], human vasopressin receptor type 1A [GenBank: NP_000697], and human rhodopsin [GenBank: NP_000530]. Homologous FIRs were identified through comparison to a previously reported structural prediction of human GnRHR1 11. In pairwise comparisons, FIRs were scored as 'similar' based on the BLOSUM62 substitution matrix – i.e., if the two residues belonged to any one of the following groups of similar amino acids, which have positive scores in the BLOSUM62 matrix: [WYF], [ST], [AS], [RKQ], [NSHD], [DE], [QKE], [YH], or [VIML]. Percent similarity for each type of FIR was calculated using the formula: (identical comparisons + (0.5 * similar comparisons))/total comparisons) * 100. Comparative genomic organization of Ce-GnRHR, human GnRHR1, and Dm-AKHR was deduced from comparison of our sequence alignment with the exon and intron lengths retrieved from GenBank sequence annotations.

Total RNA was isolated from 10 adult worms using TRIzol reagent (Invitrogen, Carlsbad, CA) according to manufacturer's instructions. Ce-GnRHR cDNA was synthesized and amplified using the SuperScript III One-Step RT-PCR system (Invitrogen, Carlsbad, CA). Both cDNA synthesis and PCR amplification were carried out using gene specific primers: 5' – GGT AAA AGT TCG ACG GGT GCA - 3' and 5' - GTT ATT TGT TTT GCC GCC GTC A - 3'. PCR product was run on a 4% Metaphor agarose gel (Cambrex Bio Science, Rockland, ME), and DNA was extracted from bands using a QIAquick gel extraction kit (Qiagen, Valencia, CA). Extracted PCR product was cycle sequenced using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Madison, WI) and automated sequencing was performed at the University of Wisconsin Biotechnology Center (Madison, WI). To ensure that sequenced PCR products were indeed Ce-GnRHR cDNA and not the result of amplification of residual genomic DNA in the RNA sample, sequenced PCR products were aligned with C. elegans cosmid F54D7 [GenBank: AF039712] and checked for the absence of intronic sequence.

Worms raised in liquid culture were pelleted by centrifugation at 800 g for 5 min., washed 3 times with S-basal and then filtered through a Whatman filter paper No. 1 (70 mm) under mild suction in order to remove adherent bacteria. Worms were collected, washed in S-basal and pelleted by centrifugation at 800 g for 3 min. Worms and bacteria were collected separately, re-suspended in a small volume of S-basal containing protease inhibitors (1 mM phenyl methane sulfonyl fluoride, 10 μg/ml each of aprotinin and leupeptin, 1 μg/ml of Pepstatin A; Roche Diagnostics, Indianapolis, IN), and the samples then subjected to 4 cycles of sonication at 30 Hz with intermittent cooling. Following protein assay, 40 μg of total nematode and bacterial protein (control) were resuspended in sample buffer (50 mM Tris-HCl, pH 6.8, containing 2% (w/v) SDS, 10% glycerol, 1.25% β-mercaptoethanol and 0.1% bromophenol blue) and separated using polyacrylamide gel electrophoresis (10–20% Tricine gels, Invitrogen, Carlsbad, CA). Following electrophoresis and electrophoretic transfer (Immobilon-P transfer membrane pore size 0.45um, Millipore) membranes were probed with antibodies using standard techniques as previously described 58.

Isolated eggs and hermaphrodite worms were washed in M9 buffer prior to mounting on poly-L lysine (100%) coated slides and cover slipped. Following the wicking of residual liquid from the slide with Whatman paper, worms were freeze fractured according to Miller and Shakes 59. Briefly, slides were frozen on dry ice for 30 min., the cover slip was quickly removed using a razor blade and the slide then immersed in cold methanol followed by cold acetone (5 min. each). Immunostaining was performed as per standard procedures. Briefly, worms were washed in 1% NGS (in TBS) for 10 min., 10% NGS for 30 min. and 1% NGS for 1 min. Slides were incubated with the anti-human GnRHR1 antibody F₁G₄ (dilution: 1:250) or with the affinity purified anti-Ce-GnRHR antibody (1:125) overnight at 4°C. Slides were then rinsed in 1%, 10% and 1% NGS for 10 min. each prior to incubation with secondary antibody for 30 min. at room temperature. Slides were washed with 1% NGS three times prior to DAB staining and mounting with Vectashield (Vector Laboratories, Inc. Burlingame, CA). Controls treated with secondary antibody alone were similarly processed. Images were acquired using a Zeiss Axiovert 200 inverted microscope connected to a Fluo Arc light source and an Axio Cam MRC-5 camera. Images were visualized and scaled using Axio Vision 4.0.

For whole-mounted fluorescent immunohistochemistry, hermaphrodite worm cuticles were permeabilized using Tris-Triton buffer with 1% mercaptoethanol and cellular contents fixed according to Finney and Ruvkun 60 as modified by Miller and Shakes 59. Worms were immunoprobed with GnRHR antibodies and fluorescently tagged secondary antibodies and images captured as described above. To localize nuclei, Hoechst dye (20 μl of 7 μg/μl) was added to the mounting medium and fluorescent staining detected using a DAPI filter 59.

The amino acid sequences of 52 Class A and B GPCRs were retrieved from the GPCR data base 61 including representatives of vertebrate GnRHRs, insect cardioregulatory peptide receptors, corazonin receptors, and AKHRs, and a variety of rhodopsin, vasopressin, and oxytocin receptors from a wide variety of animal species. Sequences were aligned in ClustalX 62 using default gap penalties, and any positions of the initial alignment represented by less than half of the sequences were excised in BioEdit 63. Excision resulted in a 422 amino acid dataset centered on the seven transmembrane domains and the intervening intra- and extracellular domains. Maximum parsimony analysis (1000 random sequence addition replicates with TBR perturbation) of the dataset was performed in PAUP*4.0b10 64. An unrooted, strict consensus tree was constructed from the most parsimonious trees. Bootstrapping, also performed in PAUP, was used to assess the support for relationships represented in the consensus tree: 500 boostrap replicates, 10 random sequence addition each.