This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Millar, R. P. Articles by Maudsley, S. R. Search for Related Content PubMed PubMed Citation Articles by Millar, R. P. Articles by Maudsley, S. R.

Gonadotropin-Releasing Hormone Receptors

Medical Research Council Human Reproductive Sciences Unit (R.P.M., Z.-L.L., A.J.P., K.M., S.R.M.), Centre for Reproductive Biology, Edinburgh EH16 4SB, Scotland, United Kingdom; and Division of Medical Biochemistry and Department of Medicine (R.P.M., C.A.F.), University of Cape Town Faculty of Health Sciences, Cape Town 7925, South Africa

Correspondence: Address all correspondence and requests for reprints to: Professor Robert P. Millar, Medical Research Council Human Reproductive Sciences Unit, The Chancellor’s Building, 49 Little France Crescent, Edinburgh EH16 4SB, Scotland, United Kingdom. E-mail: r.millar{at}hrsu.mrc.ac.uk

	Abstract

Top Abstract I. Introduction II. Structure of GnRHs... III. Structure of GnRH... IV. Binding of GnRH... V. Binding Interactions of... VI. Receptor Activation VII. GnRH Receptor Mutations... VIII. Structural Correlates of... IX. Conclusions and Future... References

 
GnRH and its analogs are used extensively for the treatment of hormone-dependent diseases and assisted reproductive techniques. They also have potential as novel contraceptives in men and women. A thorough delineation of the molecular mechanisms involved in ligand binding, receptor activation, and intracellular signal transduction is kernel to understanding disease processes and the development of specific interventions. Twenty-three structural variants of GnRH have been identified in protochordates and vertebrates. In many vertebrates, three GnRHs and three cognate receptors have been identified with distinct distributions and functions. In man, the hypothalamic GnRH regulates gonadotropin secretion through the pituitary GnRH type I receptor via activation of G_q. In-depth studies have identified amino acid residues in both the ligand and receptor involved in binding, receptor activation, and translation into intracellular signal transduction. Although the predominant coupling of the type I GnRH receptor in the gonadotrope is through productive G_q stimulation, signal transduction can occur via other G proteins and potentially by G protein-independent means. The eventual selection of intracellular signaling may be specifically directed by variations in ligand structure. A second form of GnRH, GnRH II, conserved in all higher vertebrates, including man, is present in extrahypothalamic brain and many reproductive tissues. Its cognate receptor has been cloned from various vertebrate species, including New and Old World primates. The human gene homolog of this receptor, however, has a frame-shift and stop codon, and it appears that GnRH II signaling occurs through the type I GnRH receptor. There has been considerable plasticity in the use of different GnRHs, receptors, and signaling pathways for diverse functions. Delineation of the structural elements in GnRH and the receptor, which facilitate differential signaling, will contribute to the development of novel interventive GnRH analogs.

I. Introduction

II. Structure of GnRHs and Analogs

A. Structural variants of GnRHs

B. Structure of GnRH and peptide analogs

C. The evolutionarily conserved GnRH II

D. Nonpeptide GnRH antagonists

III. Structure of GnRH Receptors

A. Primary structures of GnRH receptors

B. Tertiary structure of the mammalian type I GnRH receptor

IV. Binding of GnRH to the Mammalian Type I GnRH Receptor

A. Aspartate^2.61(98) [D^2.61(98)]

B. Asparagine^2.65(102) [N^2.65(102)]

C. Lysine^3.32(121) [K^3.32(121)]

D. Asparagine^5.39(212) [N^5.39(212)]

E. Tyrosine^6.58(290) [Y^6.58(290)]

F. Aspartate^7.32(302) [D^7.32(302)]

G. Effects of mutations of other residues on the ligand binding pocket

H. Ligand docking to the receptor

V. Binding Interactions of Other GnRH Ligands and Other Receptors

A. GnRH II

B. Peptide agonists

C. Peptide antagonists

D. Nonpeptide antagonists

E. Binding sites in nonmammalian type I GnRH receptors

F. Binding sites in type II receptors

G. Utilization of binding sites common to the rhodopsin family of GPCRs

VI. Receptor Activation

A. Interaction of Asn^2.50(87)/Asp^7.49(319) in TM 2/7 in GnRH receptor activation

B. Disruption of TM3 Asp^3.49(138)/Arg^3.50(139) interaction in GnRH receptor activation

C. The triad of Glu^2.53(90)-Lys^3.32(121)-Asp^2.61(98)

D. Role of extracellular loop 2

E. Other residues possibly involved in receptor activation

F. Integrated model of GnRH receptor activation

VII. GnRH Receptor Mutations in Hypogonadotropic Hypogonadism

VIII. Structural Correlates of GnRH Receptor Coupling and Internalization

A. Coupling to multiple G proteins

B. Regulators of G protein signaling (RGS) proteins

C. GnRH receptor internalization

IX. Conclusions and Future Perspectives

	I. Introduction

Top Abstract I. Introduction II. Structure of GnRHs... III. Structure of GnRH... IV. Binding of GnRH... V. Binding Interactions of... VI. Receptor Activation VII. GnRH Receptor Mutations... VIII. Structural Correlates of... IX. Conclusions and Future... References

 
GnRH IS THE central regulator of the reproductive hormonal cascade and was first isolated from mammalian hypothalami as the decapeptide (pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly.NH₂) (1, 2, 3). GnRH is processed in hypothalamic neurons from a precursor polypeptide by enzymic processing and packaged in storage granules that are transported down axons to the external zone of the median eminence (4, 5). The peptide is released in synchronized pulses from the nerve endings of about 1000 neurons into the hypophyseal portal system every 30–120 min to stimulate the biosynthesis and secretion of LH and FSH from pituitary gonadotropes (4). Each GnRH pulse stimulates a pulse of LH release, but FSH pulses are less distinct. Although LH is stored and largely dependent on GnRH for secretion, FSH tends to be constitutively secreted and is more dependent on biosynthesis for its secretion. The frequency of pulses is highest at the ovulatory LH surge and lowest during the luteal phase of the ovarian cycle. The asynchronous patterns of LH and FSH release result from changes in GnRH pulse frequency, modulating effects of gonadal steroid and peptide hormones on FSH and LH responses to GnRH, and differences in the half-lives of the two hormones.

Low doses of synthetic GnRH delivered in a pulsatile fashion to simulate the endogenous GnRH levels in the portal vessels (picograms per milliliter) restore fertility in hypogonadal men and women and are also effective in the treatment of undescended testes and delayed puberty (6, 7, 8, 9, 10, 11, 12). However, high doses of GnRH or agonist analogs desensitize the gonadotrope with resultant decrease in LH and FSH and a decline in ovarian and testicular function (6, 7, 8, 9, 10, 11, 12, 13). This desensitization phenomenon is extensively applied in clinical medicine for the treatment of a wide range of diseases (6, 7, 8, 9, 10, 11, 12, 13) (Table 1). GnRH peptide antagonists also inhibit the reproductive system through competition with endogenous GnRH for receptor binding, but the doses required are higher than the desensitizing agonist doses, presenting challenges for administration in the treatment of chronic diseases (14). The development of novel delivery systems for peptide antagonists or the development of nonpeptide orally active GnRH antagonists (14) is therefore likely to replace agonist therapy and avoid the undesirable stimulation and disease flare that precedes desensitization. In addition to the therapeutic applications, GnRH analogs are predicted to be used as new generation male and female contraceptives in conjunction with steroid hormone replacement (15, 16, 17).

This article reviews our current knowledge on the structure, ligand interactions, and activation of the type I GnRH receptor. In addition, the array of novel GnRH receptors in mammals and nonmammals and their relationships and differences will be described. The review will only briefly cover current knowledge on GnRH structural variants, their possible functions, and structure-activity relations of GnRH analogs because these have been thoroughly reviewed previously and have not been subject to major new developments (18, 22, 23, 24). Nevertheless, information on the structures and subtype classification of naturally occurring GnRHs and GnRH analogs is provided because this is required for the discussion of ligand selectivity and ligand interactions of GnRH receptors. Considerable attention will be given to the molecular functioning of the GnRH receptor as new insights have recently emerged. The important area of GnRH-mediated intracellular signaling will not be covered because it has been the subject of recent comprehensive review (25, 26, 27, 28, 29, 30), but receptor structural elements involved in binding and activation of signaling proteins and internalization of receptors will be addressed in detail.

	II. Structure of GnRHs and Analogs

Top Abstract I. Introduction II. Structure of GnRHs... III. Structure of GnRH... IV. Binding of GnRH... V. Binding Interactions of... VI. Receptor Activation VII. GnRH Receptor Mutations... VIII. Structural Correlates of... IX. Conclusions and Future... References

 
A. Structural variants of GnRHs
Although mammalian GnRH isolated from the hypothalamus was thought to be a unique structure with a primary role in regulating LH and FSH, it became apparent that diverse forms exist in vertebrates (31, 32). This has led to the structural identification of 23 different forms (20, 21, 22, 23, 24, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42 42A ) (Fig. 1). These are distributed in a wide range of tissues in vertebrates in which they apparently have diverse functions, including neuroendocrine (e.g., GH release in certain fish species), paracrine (e.g., in placenta and gonads), autocrine (e.g., GnRH neurons, immune cells, breast and prostatic cancer cells), and neurotransmitter/neuromodulatory roles in the central and peripheral nervous systems (e.g., sympathetic ganglion, mid-brain) (6, 13, 18, 22, 23, 24, 33, 34, 35, 36, 37, 38, 43, 44, 45). Because none of this signaler/target cell communication is mediated through secretion of GnRH into the general circulation, a single form of GnRH is theoretically capable of serving all of these roles (see Type II GnRH receptor, Section III.A). However, it is evident that at least two, and usually three, forms of GnRH are present in the majority of the vertebrate species studied (18, 20, 21, 22, 23, 24, 33, 34, 35, 36, 37, 38). The most ubiquitous is chicken GnRH II, which was first isolated from chicken brain (46). Because the chicken GnRH II structure is totally conserved from bony fish to man, this is probably the earliest evolved form and has critical functions (see Section II.C). This form has been designated GnRH II, whereas the hypothalamic form is designated type I (18, 47). In many vertebrate species, a third conserved form of GnRH (salmon GnRH) is localized to the terminal nerve in the forebrain in teleost fish and is designated GnRH III (48). GnRH III exhibits complete sequence conservation but occurs only in teleosts, suggesting that the gene encoding this peptide originated after the divergence of teleosts from the vertebrate lineage. Interestingly, sockeye salmon possess two genes encoding GnRH III. Structural analysis of the genes encoding the GnRHs supports this general classification into three forms (48), and this conclusion is confirmed by a more extensive phylogenetic analysis (Fig. 2).

View larger version (57K): [in this window] [in a new window]   FIG. 1. Primary amino acid sequences of naturally occurring GnRH structural variants spanning approximately 600 million yr of evolution. The shaded regions show the conserved NH2- and COOH-terminal residues that play important functional roles. Nonconserved residues are either unimportant or convey ligand selectivity for a particular GnRH receptor. Note that the GnRHs are named according to the species in which they were first discovered, and they may be represented in more than one species. For example, mammalian GnRH is widely present in amphibians and primitive bony fish. Chicken GnRH II (Chicken II) is present in most vertebrate species, including man and salmon. GnRH (GnRH III) is probably present in all teleost fish. For reviews, refer to Refs. 20 21 22 23 24 and 33 34 35 36 37 38 . More recent discoveries of novel GnRHs from Rana (40 ), medaka (120 ), the sea squirt, Ciona (AV893326, AV974399, and Ref. 42 ), and octopus (49 ) are also shown. The octopus has two additional amino acids shown as an insert for alignment purposes.

View larger version (50K): [in this window] [in a new window]   FIG. 2. Unrooted phylogenetic tree constructed from primary amino acid sequences of cloned GnRH ligand precursors in the Genbank database. Clustal alignments were generated using GeneJockey II software (Biosoft UK, Cambridge, UK), and phylogenetic trees were generated using a topological algorithm with PHYLIP software available at the Russian EMBnet Node: http://www.genebee.msu.su/emb.html. Bootstrap values are not indicated, and branch lengths are approximated. Identification of the source of the genes is beyond the scope of this review and will appear elsewhere.

The protochordate (chordate ancestor) sea squirt (Ciona) gene is interesting in that three GnRH forms are encoded in tandem within single genes. The most ancient of the GnRHs identified is a homolog identified in the octopus (49). This molecule exhibits the characteristic pGlu and Pro⁹Gly¹⁰.NH₂ but has an additional two amino acids inserted in the middle region of the molecule. Nevertheless, it is capable of stimulating LH release from quail pituitary cells (49).

B. Structure of GnRH and peptide analogs
The conservation of the length of the peptide (10 amino acids) and the NH₂ terminus (pGlu-His-Trp-Ser) and COOH terminus (Pro-Gly.NH₂) (Fig. 1) indicates that these features are critically important for receptor binding and activation. This is borne out from structure-activity data from several thousand analogs that were developed largely on an empirical basis. Indeed, cognizance of the evolutionary constraints on GnRH structure identify the functionally important residues and would have obviated a considerable degree of the endeavor to produce agonists and antagonists. The considerable variation in position 8 of natural GnRHs (Arg, Gln, Trp, Ser, Thr, Asn, Leu, Tyr, Lys, Ala, Trp) suggests that virtually any residue is tolerated in this position. However, this is clearly not the case for the mammalian pituitary type I GnRH receptor (18, 50), which requires Arg in position 8 for high-affinity binding. Recent work on cloned nonmammalian receptors also indicates certain specificities for the amino acid in this position (35, 51, 52, 53). Thus, the residue in position 8 seems to play an important role in ligand-selectivity of the different GnRH receptors. The roles of the individual amino acids comprising GnRH and the structure-activity relations of agonist and antagonist analogs have been extensively reviewed (18, 50) and will not be repeated here.

Short peptides such as GnRH are highly flexible in solution and exist as an equilibrium between numerous conformations (54, 55, 56, 57). However, among these conformations are so-called bioactive conformations that represent preferred structures for interaction with the receptor. For GnRH, the bioactive conformation is the product of a number of influences, which include intramolecular interactions, local influences of solvents (water), lipids, and initial receptor interactions that conform the ligand. In addition, membrane and intracellular proteins associate with the receptor and alter its conformation and selectivity for ligand conformations. Studies on GnRH and its receptor are increasingly pointing to a multiplicity of bioactive conformations of both the ligand and receptor. These in turn result in differential activation of intracellular signaling pathways (our unpublished observations).

The first studies on GnRH structure by conformational energy analysis of the NH₂-terminal 1–6 and carboxyl-terminal 6–10 amino acids identified a low energy CC conformer that featured a ß-II’ type turn involving Tyr⁵-Gly⁶-Leu⁷-Arg⁸ such that the NH₂ and COOH termini are closely apposed (55) (Fig. 3). This conclusion has subsequently been supported by a variety of experimental data with synthetic GnRH analogs (58, 59, 60, 61), interactions with region-specific antibodies (62), and a range of physicochemical studies. These studies have been extensively reviewed (18, 50) and will only be briefly covered here. A recent study using electron capture dissociation mass spectrometry has confirmed the presence of the ß-II’ type turn and interaction of the NH₂ and COOH termini (N. C. Polfer, personal communication).

View larger version (18K): [in this window] [in a new window]   FIG. 3. Schematic representation of mammalian GnRH in the folded conformation in which it is bound to the GnRH pituitary receptor. The molecule is bent around the flexible glycine in position 6. Substitution with D-amino acids in this position stabilizes the folded conformation, increases binding affinity, and decreases metabolic clearance. This feature is incorporated in all agonist and antagonist analogs (Fig. 5). The NH2 (red) and COOH (green) termini are involved in receptor binding. The NH2 terminus alone is involved in receptor activation and substitutions in this region produce antagonists (see Fig. 5). [Adapted from R. P. Millar, Reproductive medicine: molecular cellular and genetic fundamentals (edited by B. C. J. M. Fauser), Parthenon Publishing, Lancaster, UK, 2002, pp 199–224 (21 ).]

View larger version (18K): [in this window] [in a new window]   FIG. 4. Schematic of NMR analysis of the structure of mammalian and chicken GnRH I (54 ). Mammalian GnRH exhibits three major families of conformers similar to the one shown in panel A. All three show a ß-II’ turn about Gly6 and the NH2 and COOH termini in close proximity. Although several hydrogen bonds were found in the three structures, only one between the carbonyl oxygen of Ser4 and the side chain amino group hydrogen of Arg8, and another between the Gly10.NH hydrogen and the pGlu1 carbonyl oxygen are present in all three conformers. The Arg8 side chain is involved in at least one other hydrogen bond, with either the His2 side chain or the Tyr5 side chain. The chicken GnRH I structures fall into four main families of conformers that differ from each other to a greater extent but are all extended forms as represented by panel B. None of the conformers have the hydrogen bonding of the mammalian GnRH, but a series of other hydrogen bonds of which a Ser4 bond to pGlu1 is the only one present in all four conformers. [Adapted from J. C. Maliekal et al.: S Afr J Chem 50:217–219, 1997 (54 ).]

View larger version (53K): [in this window] [in a new window]   FIG. 5. GnRH agonist and antagonist analogs in clinical practice or in clinical development.

The amino-terminal residues of GnRH are involved in receptor activation, and modification of these residues in GnRH produces analogs with antagonistic properties (18, 50) (Figs. 3 and 5). As in agonists, substitution of Gly⁶ with a D-amino acid enhances the activity of the antagonists. Because the antagonists have high binding affinity, the loss of amino-terminal contacts in agonists that activate the receptor is presumably compensated for by new contacts made by the substituted amino acids in antagonists.

The first generation of potent GnRH antagonists were characterized by high histamine-releasing properties as a result of the presence of basic residues (basic-X-basic sequence). Elimination of basicity produced analogs with lower histaminic properties but poorer solubilities and the tendency to form gels. This has resulted in difficulties in formulation that continue to be a problem in GnRH antagonists that are currently under clinical investigation. The primary structures of GnRH analogs that are extensively employed therapeutically, or are in clinical development, are shown in Fig. 5.

GnRH II is an intriguing exception to the general conclusion that a D-amino acid substitution for Gly⁶ enhances the binding affinity of GnRHs (72) (Table 2). An explanation may be that GnRH II is already stabilized in the ß-II’ turn conformation and that incorporation of a D-amino acid in position 6 does not further stabilize this conformation. Residues His⁵, Trp⁷, and Tyr⁸ are proposed to contribute to the stabilization of GnRH II (72). Gly⁶ is essential to allow assumption of the folded conformation, and the NH₂- and COOH-terminal sequences are essential for receptor binding and activation. Thus, all the amino acids appear to have crucial roles, and this offers an explanation for the total conservation of GnRH II structure over 500 million yr of evolution.

View this table: [in this window] [in a new window]   TABLE 2. Enhancement1 of binding affinity by constraint with D-amino acid substitution of Gly6 or 6, 7 -lactam bridge

C. The evolutionarily conserved GnRH II
As mentioned earlier, a second form of GnRH identified from chicken brain (chicken GnRH II, GnRH II) (Fig. 1) is ubiquitous in vertebrates from primitive bony fish to man (22, 23, 24, 33, 34, 35, 36, 37, 38, 73, 74). This complete conservation of structure over 500 million yr suggests that GnRH II has an important function and a discriminating receptor (or receptors) that has selected against any structural change in the ligand. This points to essential functions that have yet to be definitively identified. The wide distribution of GnRH II in the central and peripheral nervous systems suggests a neurotransmitter/neuromodulatory role. This has been thoroughly demonstrated in the inhibition of M currents in the bullfrog sympathetic ganglion, which sensitizes neurons to depolarization (75, 76). GnRH II was identified in amphibian sympathetic ganglia, and the receptors present are highly selective for the peptide (77).

Because GnRH had been shown to have direct effects on sexual arousal in rodents (78, 79, 80) and type II GnRH is localized in brain areas associated with reproductive behavior, it was suggested that this may be a role for the peptide (22, 23, 24, 33, 79, 80). GnRH II and a GnRH II analog were both found to be potent stimulators of reproductive behavior in ring doves (23, 24), song sparrows (81), and the musk shrew (82). Recently, the cognate receptor for GnRH II was cloned from the marmoset and found to be distributed in those areas of primate brain associated with reproductive behaviors (45, 83). Infusion of GnRH II into the third ventricle of female marmosets increased sexual behavior (D. Abbott, personal communication).

In addition to its apparent role as a neuromodulator in the nervous system, GnRH II and its receptor are present in reproductive tissues (45). GnRH binding sites and antiproliferative effects of GnRH analogs have also been described in reproductive tissue tumors and their cell lines (see reviews in Refs. 6 , 8 , 11 , 13 , and 45). Interestingly, GnRH and analog binding, signaling, and pharmacological effects were not characteristic of classical hypophysial type I GnRH receptors but are more similar to type II GnRH receptors (13, 83). Although this suggests that the antitumor effects in human tissues are mediated via the type II GnRH receptors (84), the human type II receptor gene is disrupted by a frame-shift and a stop codon, and a transcript that could encode a full-length receptor or the expressed receptor protein has not been identified (85, 86 86A ). It has now become apparent that the different pharmacology of GnRH analogs in affecting pituitary function and in inhibiting proliferation of tumor cell lines can both be mediated by the human type I GnRH receptor by coupling through different signaling pathways (viz. G_q for pituitary and G_i for tumor cells). This differential coupling can be accomplished through both ligand selectivity and intracellular milieu and will be discussed later in this review (see Section VIII.A).

D. Nonpeptide GnRH antagonists
The development of nonpeptide GnRH antagonists has seen intense endeavors from the pharmaceutical industry. Representative compounds are shown in Fig. 6. The first described nonpeptide GnRH antagonist (compound 1) is a fused tetracyclic benzodiazepine that blocks ovulation in rats when given at a dose of 0.5 mg/kg (87). The antifungal drug ketoconazole (Nizoral, Janssen Pharmaceutica, Beerse, Belgium) (compound 2) was found to bind and inhibit the rat pituitary GnRH receptor with an apparent IC₅₀ of 2 µM. Addition of a number of groups to this core structure, such as dipeptides and tripeptides related to GnRH, improved affinity to approximately 500 nM (88).

View larger version (23K): [in this window] [in a new window]   FIG. 6. Examples of nonpeptide GnRH antagonists.

Merck has described both indole (93) (compound 6) and quinolone-based (94, 95, 96) (compound 7) small molecule antagonists, and Abbott’s description of the ketoconazole was followed by the discovery of an erythromycin A derivative (97) (compound 8). Takeda has reported a new series (92) (compound 9) based on compound 5, and Alanex Corp. developed compound 10 (98). The most recent reports are a further series of derivatives of compound 6 by Merck (compound 11) with IC₅₀ in subnanomolar concentrations and excellent oral bioavailability (99) and a series of imidazol-pyrimid-5-ones from Neurocrine (compound 12) that bind the human receptor in the low nanomolar range (100, 101). Recently, a new series of small molecule antagonists have been developed by Pfizer (102).

Although there are exceptions, the majority of the small molecule GnRH antagonists conform to a simple pharmaco-phore model. This comprises a requirement for a basic protonatable nitrogen group (optionally substituted by lipophilic groups), one or two aromatic groups, and an aliphatic lipophilic group arranged in a putative ß-turn mimetic configuration (92). Some of the small molecule antagonists have progressed to clinical trial. Takeda’s thienopyrimidinedione (TAK-013) is in phase two for endometriosis and uterine fibroids, whereas their thienopyridine-one (TAK-810) is in phase one. Neurocrine’s pyrolopyrimidone (NBI-42902) is in phase one trials for a range of reproductive indications.

	III. Structure of GnRH Receptors

Top Abstract I. Introduction II. Structure of GnRHs... III. Structure of GnRH... IV. Binding of GnRH... V. Binding Interactions of... VI. Receptor Activation VII. GnRH Receptor Mutations... VIII. Structural Correlates of... IX. Conclusions and Future... References

 
A. Primary structures of GnRH receptors
The amino acid sequence of the GnRH receptor was first deduced for the mouse receptor cloned from the pituitary T3 gonadotrope cell line (103). This sequence was confirmed (104), and it provided the basis for the cloning of GnRH pituitary receptors from the rat (105, 106, 107), human (108, 109) (Fig. 7), sheep (110, 111), cow (112), and pig (113) that share over 80% amino acid identity. Homologs of the mammalian GnRH receptors have also been cloned from a marsupial (possum) (114 114A ), catfish (65), two forms from the goldfish (51), bullfrog (67), brown frog (115), clawed toad (66), chicken (71), medaka (116), striped bass (117), trout (118), salmon (118A ), cichlid (119), Japanese eel (120), amberjack (CAB 65407), rubber eel (AD 49750), and seasquirt (120A ). The nonmammalian receptors with greatest homology to the mammalian pituitary receptors have 42–47% amino acid identity with the mammalian receptors but 58–67% identity among each other. These are all designated as type I GnRH receptors (Figs. 8 and 9). It is not altogether clear from homology comparisons that the classification of the mammalian and nonmammalian type I together is correct, but similarities in microdomains (e.g., EC3) support this. Because the evolutionary time separating amphibians and mammals is similar to that separating amphibians and bony fish, the poor conservation of sequence of the mammalian type I GnRH receptor with the nonmammalian receptors implies a sudden acceleration in evolutionary change in the mammals. This may have been driven by the loss of the carboxyl-terminal tail in the mammalian type I receptor, which is unique among G protein-coupled receptors (GPCRs). In the goldfish (51), zebra fish (47), catfish (121, 122), and salmon (P. Swanson, unpublished observations), there are two isoforms (type Ia and type Ib) that have 70% amino acid identity. In the goldfish, they differ in type Ia having a putative SH3 binding domain (poly proline sequence) in the carboxyl-terminal tail, which potentially conveys the possibility of coupling to MAPKs (S. R. Maudsley, unpublished observations).

View larger version (36K): [in this window] [in a new window]   FIG. 7. Two-dimensional representation of the human GnRH receptor showing TM domains (boxed) connected by ECs and ICs. Putative ligand binding sites (red) and residues thought to be important in receptor structure or binding pocket formation are shown in green letters. These include disulfide bond formation and glycosylation sites. Residues involved in receptor activation are shown in blue. Residues in squares are the ones highly conserved throughout the rhodopsin family of GPCRs. Residues involved in coupling to G proteins are shown in orange. Protein kinase C (PKC) and protein kinase A (PKA) phosphorylation sites are indicated.

View larger version (200K): [in this window] [in a new window]   FIG. 8. Clustal alignment of primary amino acid sequences of cloned vertebrate type I, II, and III GnRH receptors. The TM domains are boxed, and the ICs and ECs are indicated. The consensus for the most characteristic domain (EC3 going into TM7) of the three receptor types is shown. This domain was used to clone the three receptor types. The amino acids are colored green for charged/polar, blue for nonpolar, red for nonpolar aromatic, and black for nonpolar sulfhydryl. Note that TM domains are predominantly hydrophobic, and loop domains are hydrophilic.

View larger version (37K): [in this window] [in a new window]   FIG. 9. Unrooted phylogenetic tree of GnRH receptor sequence relationships generated using a topological algorithm with PHYLIP software (EMBnet). Bootstrap values are not indicated, and branch lengths are approximated. Identification of the source of the genes is beyond the scope of this review and will appear elsewhere.

First, the frame-shift and stop codon are accommodated posttranscriptionally and during translation. Numerous mechanisms of mRNA editing have been described and could potentially repair the frame-shift and stop. In this regard we have noted transcripts that splice out the stop. Alternatively, the stop codon can be translated as a selenocysteine by means of a specific tRNA and a selenocysteine insertion sequence motif (127), which is present in the 3' untranslated region of the gene encoding human type II receptor. However, we were unable to demonstrate selenocysteine incorporation (86).

Second, a partial type II receptor is elaborated and is functional. A Kozac consensus start site follows the stop, and when this partial receptor sequence is expressed in COS cells it down-regulates the expression of the type I receptor (A. J. Pawson, unpublished observations). Interestingly, the full-length cDNA (including frame-shift and stop) increases type I expression. Could it be that the function of type II transcripts is to regulate type I expression and coupling? Their coexpression in gonadotropes suggests that this is feasible. The stop codon has arisen independently in evolution in the same vicinity and no other place in unrelated species. This might suggest that there is an advantage in having the stop codon and producing partial receptor sequences. The presence of type II receptors with or without the stop codon in closely related species suggests that any advantage of the stop is only marginal.

GnRH receptor orthologs have been identified in Drosophila melanogaster (128) and Caenorhabditis elegans (P. Swanson, unpublished observations), indicating a very early evolutionary origin. However, the cognate ligand for the Drosophila GnRH receptor homolog is not a GnRH, but is an adipokinetic hormone that has a similar structure in its length and the presence of the NH₂-terminal pGlu and a COOH-terminal amide (129).

We have examined the suggested classifications of GnRH receptors by constructing phylogenetic trees. The primary amino acid structures of cloned GnRH receptors were aligned and used to generate an unrooted tree using a topological algorithm that optimizes tree structure before determining branch lengths. This revealed that the receptors can be grouped into distinct classes: types I, II, and III (Fig. 9). Type I and type II GnRH receptors form elongated clusters. Type III GnRH receptors are more closely related to type II receptors than to type I receptors, suggesting that type II and type III receptors may have arisen from duplication of an ancestral gene in lower vertebrates. The numerical naming of individual cloned receptors in the Genbank database frequently does not comply with the phylogenetic relatedness because researchers have named them by pharmacological characteristics, order of discovery, or tissue expression (67, 115, 116, 122). For example, Bogerd et al. (122) recently named a novel catfish receptor R2, although it has greatest homology with goldfish 1a receptor (51). Similarly, Wang et al. (67) named the receptor cloned from bullfrog pituitary as bullfrog I, although its greatest homology is with type III receptors. We have retained the original authors’ designation in Fig. 9 to highlight these discrepancies. Clearly, a more systematic and consistent approach is required.

GnRH receptors have the characteristic features of GPCRs (Figs. 7 and 8). The NH₂-terminal domain is followed by seven -helical TM domains connected by three EC domains and three intracellular loop (IC) domains. The extracellular domains and superficial regions of the TMs are usually involved in binding of peptide hormones such as GnRH, and the TMs are believed to be involved in receptor configuration and conformational change associated with signal propagation (receptor activation). These changes are thought to propagate into conformational changes in the intracellular domains involved in interacting with G proteins and other proteins for intracellular signal transduction.

A unique feature of the mammalian type I GnRH receptor is the absence of a carboxyl-terminal tail present in all other GPCRs and in all of the nonmammalian and mammalian type II GnRH receptors. This is, therefore, a recently evolved feature that presumably serves an important role in the functioning of the mammalian GnRH receptor (see Section VI.F).

The conservation of amino acids during evolution from bony fish to mammals is likely to identify those residues that are crucial for GnRH receptor function (18, 35). These include residues thought to be involved in GnRH receptor binding Asp^2.61(98), Asn^2.65(102), Lys^3.32(121), Asn^5.39(212), Tyr^6.58(290), and Asp^7.32(302) (Figs. 7 and 8). The conserved residues include those conserved or conservatively substituted throughout the rhodopsin family of GPCRs. These are shown in Fig. 7 in squares and are the residues used as reference points for the consensus numbering of the rhodopsin family of GPCRs [i.e., Asn^1.50(53), Asn^2.50(87), Arg^3.50(138), Trp^4.50(164), Pro^5.50(223), Pro^6.50(282), and Pro^7.50(320) (Ref. 18)] (see Section III.B).

B. Tertiary structure of the mammalian type I GnRH receptor
A knowledge of the three-dimensional structure of the mammalian GnRH receptor is essential for an understanding of its molecular functioning. The only direct structural information at atomic resolution of a GPCR is derived from x-ray analysis of the ground state of rhodopsin (130). Previous structural information on GPCRs was predicted from low resolution electron microscopy of bacteriorhodopsin and rhodopsin (131, 132, 133). Structural information for all other GPCRs has relied on molecular models (18, 19, 134, 135, 136, 137, 138) based on the rhodopsin structure. A model (139) incorporating structural information derived from the analyses of approximately 500 sequences in the rhodopsin-like family of GPCRs ultimately turned out to be very similar to the structure of rhodopsin, suggesting a similar structure of the seven TM domains of all GPCRs in the rhodopsin family. In contrast, the EC and IC domains are highly variable in amino acid sequence and probably in their tertiary structure. In addition to this limitation, no direct structural information of the activated state of any GPCR is available. Consequently, an understanding of the conformational changes associated with receptor activation has relied on biophysical and biochemical studies.

The development of the first published GnRH receptor molecular model was based on initial alignment and positioning of the TM helices as indicated in the projection map of the electron density of rhodopsin followed by refinement of the angles, kinking, and side chain orientation of the TMs based on the specific amino acids comprising the GnRH receptor TMs (18). The validity of the model and proposed interactions of the TM side chains was tested by site-directed mutagenesis. An example is the observation that two residues that are highly conserved in GPCRs, Asp^2.50 in TM2 and Asn^7.50 in TM7, appear to have undergone reciprocal mutation to Asn^2.50(87) and Asp^7.49(318) in the mouse GnRH receptor (Asp^7.49(319) in human) (Figs. 7 and 8). This suggested that the two residues interact with each other. Mutation of Asn^2.50(87) in TM2 to Asp abolished receptor function, but a second mutation in TM7, recreating the arrangement found in other GPCRs [Asp^2.50(87) and Asn^7.49(318)], restored ligand binding (140). Cook et al. (141) reported that the reciprocal mutant was totally inactive, but the original observation of good binding activity (140) has been confirmed (142, 143, 144). This restoration of ligand binding by reciprocal mutation demonstrates that the side chains of two residues in TMs 2 and 7 have complementary roles in maintaining the structure of the receptor and occupy the same microenvironment within the receptor helical bundle. This experimentally derived conclusion was subsequently supported in the structural analysis of inactive rhodopsin, which shows that these residues are capable of interacting through a water molecule (145).

Recently, the rhodopsin x-ray structure has been used as a template for homology modeling of the TM domains of the GnRH receptor (146) (Z.-L. Lu, unpublished observations). Evidence gleaned from the mutagenesis of the reciprocal TM2/TM7 mutations (mentioned above) and the disulfide bridges [Cys⁽¹⁴⁾/Cys^5.27(200), Cys^3.25(114)/Cys^5.23(196), see below] was also used. A 2.5-nsec molecular dynamic simulation of the GnRH receptor in a water-vacuum-water box with no conformational restraints during the last 2 nsec was also undertaken. This revealed a hydrogen bond net of Glu^2.53(90)-Lys^3.32(121)-Asp^2.61(98) between TM2 and TM3 that may represent a component of intramolecular interactions that stabilize a receptor conformation similar to that of rhodopsin in the inactive state (146).

The seven TM helical domains are known from physical structural studies in the rhodopsins to be arranged in a tight bundle enclosing a hydrophilic pocket and surrounded by the hydrophobic membrane environment (18, 19, 25, 130, 133, 134, 135, 136, 137) (see schematic in Fig. 10). The evolutionary conservation of residues along a distinct face of the TM domains in the various GnRH receptors is evident (compare Figs. 7 and 8). This suggests that the conserved, more hydrophilic faces are orientated toward the hydrophilic pocket or the boundaries formed by the seven TM domains and potentially assists in the refinement of the molecular model. This proposal is supported by the studies on the TM2/TM7 interaction of Asn^2.50(87) and Asp^7.49(318) (140, 147) because Asn^2.50(87) is clearly part of the conserved hydrophilic face of TM2.

View larger version (33K): [in this window] [in a new window]   FIG. 10. A schematic representation of the human GnRH receptor. The receptor is viewed from above and shows the TM helices as a cluster of cylinders (yellow, going into the page) that encompass the hydrophilic pocket and are surrounded by the light hydrophobic membrane environment. The TM helices are connected by the ECs (red). The dark bands represent the disulfide bridges stabilizing extracellular domains. The binding pocket is defined by some putative binding sites, D2.61(98), N2.65(102), K3.32(121), N5.39(212), Y6.58(290), and D7.32(302) in the receptor. [Adapted from R. P. Millar, Reproductive medicine: molecular cellular and genetic fundamentals (edited by B. C. J. M. Fauser), Parthenon Publishing, Lancaster, UK, 2002, pp 199–224 (21 ).]

Although progress has been made in establishing molecular models of the TM helix bundle of the GnRH receptor, the proposed structure of the EC and IC is conjectural. Considerable effort has been directed at establishing programs to define loop structures (e.g., based on sequences for loop structures established from x-ray crystallography). However, these are not applicable to large loop sequences. Moreover, the known structures of the rhodopsin loops may be quite different from those of other GPCRs. Thus, a future challenge is the determination of the structure of the loops in the GnRH receptors. Some progress has been made in circular dichroism, NMR, and Raman spectral analysis of the structure of a synthetic peptide of EC3 anchored by cross-links similar to the distance between the anchoring TM6 and TM7 domains (69). Contrary to the suggestion that EC3 had an -helical structure (154), these techniques revealed the predominantly random structure of the loop. The NMR analysis showed a low incidence of a ß-hairpin structure (Fig. 11). When this structure is incorporated in the receptor model, Asp^7.32(302) is able to interact with Arg⁸ of GnRH when docked to the other interacting sites [Asp^2.61(98), Asn^2.65(102), and Lys^3.32(121)] (Fig. 12). Due to the low sequence homology of the EC and IC of the GnRH receptor with rhodopsin, Söderhall et al. (146) did not model the loops by using the rhodopsin structure alone but also by homology modeling of structural motifs found in the Protein Databank. These models of loop structures provide a point of departure for experimental testing. Despite the uncertainties of elements of the current GnRH molecular models, they are nevertheless providing key insight into putative mechanisms of ligand binding and receptor activation as a substrate for experimentation.

View larger version (41K): [in this window] [in a new window]   FIG. 11. Interaction of Arg8 in GnRH with Asp7.32(302) of EC3 of the human GnRH receptor. The GnRH receptor model was based on the rhodopsin structure and refined to accommodate known experimental data of interactions of TM domains. A ß-hairpin conformation of EC3 determined from the NMR structures of a cyclized EC3 peptide (69 ) was attached to TM6 and TM7 of the molecular model. Only TM6, EC3, and TM7 of the molecular model are shown for clarity. The GnRH molecule in its active ß-II’ turned conformation has been docked to Asp2.61(98), Asn2.65(102), and Lys3.32(121) cognate binding sites in the receptor, which are not shown for clarity. With these contacts in place, Arg8 of GnRH is able to interact with Asp7.32(302) of the receptor as shown. [Adapted from R. Petry: J Med Chem 45:1026–1034, 2002 (69 ).]

View larger version (35K): [in this window] [in a new window]   FIG. 12. Molecular model of GnRH interactions with the human GnRH receptor (Z.-L. Lu, unpublished observations). GnRH was docked in the ß-II’ folded conformation (18 54 57 ) to the human GnRH receptor model built by homology modeling using the rhodopsin x-ray structure as a template. The model accommodates the experimentally determined or putative interactions of GnRH (black) and receptor (yellow/blue) residues [pGlu1 with Asn5.39(212); His2 with Asp2.61(98)/Lys3.32(121); Trp3 with Trp6.48(280); Tyr5 with Tyr6.58(290); Arg8 with Asp7.32(302); Pro9 with Trp2.64(101); and Gly10NH2 with Asn2.64(102)]. The hydrogen bonds are indicated by dashed lines. A, View of GnRH docked to its receptor. GnRH is shown in gray, the interacting residues of the receptor in yellow, and the seven TM helices in blue. B, Stereo view of the above model. The GnRH and receptor interacting residues are shown in white, and the seven TM helices in orange.

	IV. Binding of GnRH to the Mammalian Type I GnRH Receptor

Top Abstract I. Introduction II. Structure of GnRHs... III. Structure of GnRH... IV. Binding of GnRH... V. Binding Interactions of... VI. Receptor Activation VII. GnRH Receptor Mutations... VIII. Structural Correlates of... IX. Conclusions and Future... References

 
Binding of ligand is a major component of receptor function, and this interaction is the primary determinant of whether a receptor initiates signaling within the cell. Binding of agonist ligands may be considered the first step in receptor activation or receptor-mediated transduction of a hormone signal across the cell membrane. Thus, understanding ligand binding interactions is an important component of an understanding of the fundamental mechanism of receptor function. The binding of GnRH and GnRH agonist analogs to mammalian type I GnRH receptors has attracted considerable experimental study and is reviewed in detail in this section. However, the interactions of type I receptors with peptide and nonpeptide antagonists and the ligand binding interactions of nonmammalian and type II GnRH receptors have not been as well studied. Consequently, these are considered in the next section in the context of what is known about GnRH binding to type I receptors.

Although site-directed mutagenesis has been a useful experimental tool in defining ligand binding interactions of GnRH receptors and most GPCRs, it is important to distinguish direct receptor-ligand interactions from changes in receptor structure or conformation that indirectly affect ligand binding. One approach to this is to modify the ligand in parallel with receptor mutation (157). For example, if a mutation is thought to decrease binding affinity by disrupting a specific interaction with the ligand, then a ligand that lacks the interacting group should have similar affinity for wild-type and mutant receptors. In the course of the targeted mutation of almost one third of all the amino acid residues of the GnRH receptor, considerable advances have been made in identifying putative ligand contact sites in the mammalian GnRH receptor (Figs. 7, 12, and 13A, and Table 3). As is the case for other GPCRs that bind small peptides (158), amino acid residues in the ECs and exofacial parts of the TM helices of GnRH receptors are thought to participate in ligand binding interactions. Specific interactions of three residues, Asp^2.61(98), Asn^2.65(102), and Asp^7.32(302), have been defined in detail, whereas residues Trp^2.64(101), Lys^3.32(121), Asn^5.39(212), and Tyr^6.58(290) have been shown to be important for binding of agonist ligands but not antagonists, and specific interactions have been proposed for these residues. These residues are discussed in numerical order, whereas other residues for which less experimental evidence is available are discussed at the end of this section.

View larger version (37K): [in this window] [in a new window]   FIG. 13. Molecular models of a GnRH agonist and antagonist interactions with the human GnRH receptor (146 ). A, Starting from a distance of 30–40 Å outside of the defined binding pocket, the mass centers of interacting amino acid residues were restrained to gradually approach each other within a distance of 3–5 Å (146 ). After the simulated annealing phase (5 psec heating up to 1500°K, 20 psec hot phase, and 25 psec slow cooling phase followed by complex minimization), the lowest energy docked structure was selected. For D-Trp6 GnRH, the ligand was satisfactorily docked in the ß-II’ turned conformation to the identified contact sites in a putative active conformation of the receptor in which all of the experimentally identified binding interactions are accommodated (146 ). D-Trp6 GnRH agonist interactions with the activated receptor include a hydrogen bond between pGlu1 and Asn212; a hydrogen bond between His2 and Lys3.32(121)/Asp2.61(98) that disrupts the hydrogen bond network of Glu2.53(90)-Lys3.32(121)-Asp2.61(98); -stacking between Trp3, Tyr5, and Trp6.48(280), Phe5.43(216); -stacking between D-Trp6 and Trp6.57(289), which is close to the Cys14/Cys5.27(200) disulfide bridge; hydrogen bonds between Arg8 and Asp7.32(302) and Gly10NH2 and Asn2.61(102). A ß-II’ turn is formed by an intramolecular hydrogen bond between Tyr5 and Arg8 of the agonist. B, Cetorelix GnRH antagonist interactions with the inactive receptor include a hydrogen bond between the NH2-terminal acetyl group and Asn5.39(212), -stacking between D-Cpa and Trp6.48(280), a hydrogen bond between D-Pal3 and Lys3.32(121) without effect on the Glu2.53(90)/Lys3.32(121)/Asp2.61(98) hydrogen bond network, D-Cit6 near the Cys14/Cys5.27(200) disulfide bridge, hydrogen bonds between Arg8 and Asp7.32(302) and D-Ala10NH2 and Asn2.65(102). A ß-II’ turn is formed between Tyr5 and Arg8 of the antagonist. [Adapted with permission from Ref. 146 .]