Abstract
Mammalian gonadotropin-releasing hormone (GnRH) receptors preferentially bind mammalian GnRH, which has Arg in position eight. The Glu7.32(301) residue, which determines selectivity of the mouse GnRH receptor for Arg8-containing GnRH, is Asp7.32(302) in the human GnRH receptor. We have confirmed that Asp7.32(302) confers selectivity of the human GnRH receptor for Arg8 of GnRH and investigated the mechanism of this specificity using site-directed mutagenesis and ligand modification. We find that although Arg8 and Asp7.32(302) are required for high-affinity binding of GnRH, conformationally constrained peptides, with d-amino acid substitutions in position six or with a 6,7 γ-lactam, bind the human GnRH receptor with high affinity, which is independent of the presence of Asp7.32(302) in the receptor or Arg8 in the ligand. The ability of the ligand constraints to compensate for the absence of both Arg8 and Asp7.32(302) indicates that these residues both have roles in stabilizing a high affinity ligand conformation and that their roles are complementary. This suggests that the Arg8 and Asp7.32(302) side chains interact to induce a high affinity conformation of native GnRH. Thus, Asp7.32(302) of the human GnRH receptor determines selectivity for mammalian GnRH by its ability to induce a high affinity conformation of its native ligand. However, this initial interaction seems not to contribute to the final ligand-receptor complex. We propose that Arg8 interacts transiently with Asp7.32(302) to induce a high-affinity ligand conformation of GnRH, which then interacts with a binding pocket that is common for both constrained and unconstrained analogs of GnRH.
Gonadotropin-releasing hormone [GnRH, also called luteinizing hormone releasing hormone or luliberin] is a decapeptide that is synthesized in the hypothalamus and interacts with GnRH receptors on gonadotrope cells in the anterior pituitary. GnRH stimulates the biosynthesis and release of luteinizing hormone and follicle-stimulating hormone, which in turn are required for steroidogenesis and gametogenesis, respectively. Because of this central role in reproduction, GnRH analogs have been used in a variety of therapeutic applications (Millar et al., 1987).
Although an X-ray diffraction analysis of rhodopsin has recently been published (Palczewski et al., 2000), most other G protein-coupled receptors (GPCR) are more difficult to purify. Consequently, understanding of the structure of these GPCRs is likely to depend on indirect methods, such as computational modeling and mutagenesis, for some time to come. Considerable advances have been made in understanding how GnRH interacts with its receptor. In the human GnRH receptor, residues Asp2.61(98), Trp2.64(101), Asn2.65(102), Lys3.32(121), and Asn5.61(212) (residue numbering is described under Materials and Methods) have been shown to have roles in ligand binding (Zhou et al., 1995; Davidson et al., 1996; Flanagan et al., 2000; Hoffmann et al., 2000). Some of these receptor residues have been proposed to form part of the ligand binding pocket, interacting with the amino and carboxyl termini of GnRH in a computational model of the receptor-ligand complex (Sealfon et al., 1997). Asp2.61(98) is proposed to interact with His2 of GnRH, whereas Asn2.65(102) interacts with Gly10-NH2. In the mouse GnRH receptor, Glu7.32(301) was shown to have a role in recognizing the Arg8 residue of GnRH (Flanagan et al., 1994). However, Glu7.32(301) is not completely conserved in mammalian GnRH receptors. In the human and other nonrodent GnRH receptors the equivalent residue is Asp7.32(302) (Kakar et al., 1992; Chi et al., 1993; Illing et al., 1993; Sealfon et al., 1997; Cui et al., 2000). Although this is a conservative substitution, it is surprising that such a functionally important residue is not absolutely conserved. In the monoamine receptors, the Asp3.32 residue, which is important for ligand binding, is conserved as Asp not only in different species but also in different receptor subtypes that recognize the same ligand and in different receptors that recognize distinct ligands ranging through acetylcholine, adrenaline, serotonin, and histamine (Probst et al., 1992).
In GnRH, Arg8 is required for high-affinity binding to mammalian GnRH receptors. Substitution of this residue decreases GnRH potency and affinity for the receptor (Millar et al., 1989). Mutation of the Glu7.32(301) residue of the mouse GnRH receptor to Gln decreased the receptor affinity for GnRH, but not for analogs with substitutions for Arg8 (Flanagan et al., 1994). Subsequent models of GnRH receptor-ligand complexes have incorporated an interaction of the acidic residue of the receptor with Arg8 of the ligand (Chauvin et al., 2000; Flanagan et al., 2000; Hoffmann et al., 2000). However, a GnRH analog with d-Trp substituted in position six showed only a small decrease in affinity for the Glu7.32(301)Gln mouse receptor. This suggested that although Glu7.32(301) determines selectivity for native GnRH, the mechanism by which it does so may be more complex than a simple electrostatic interaction of Glu301with Arg8. It also indicates a need for caution in extrapolating experimental results to molecular models of GPCRs.
The lack of conservation indicates a need to determine whether Asp7.32(302) has the same function in the human GnRH receptor as Glu7.32(301) has in the mouse receptor. Furthermore, the incorporation of a direct interaction of Glu7.32(301)/Asp7.32(302)with Arg8 in models of receptor-ligand complexes, despite evidence that a direct interaction may not always occur, shows that better definition of the mechanism by which the Asp7.32(302) determines binding specificity for GnRH is needed. We now show that mutating Asp7.32(302) in the human GnRH receptor decreases affinity for GnRH, but not for analogs with substitutions for Arg8. In contrast, a series of peptides with different structural constraints that stabilize a high-affinity conformation of GnRH retain high affinity for the mutant receptor. This indicates that an electrostatic interaction is not involved in the binding of these constrained analogs. We also show that Arg8 is not required for high-affinity binding of constrained analogs. We interpret these results in terms of a sequential binding mechanism in which the Arg8side chain of native GnRH interacts transiently with Asp7.32(302), before interacting with a final ligand binding pocket that also binds constrained analogs.
Materials and Methods
Consensus Residue Numbering Scheme.
A consensus numbering scheme is used to facilitate the comparison of equivalent amino acid residues in the different rhodopsin-like GPCRs (Ballesteros and Weinstein, 1995). Amino acids were numbered relative to the most conserved residue in each transmembrane domain, which is assigned the number 50 (Fig. 1). Individual amino acid residues are identified by a generic identifier consisting of the transmembrane helix number, followed by the number representing its position relative to the most conserved residue in the helix. This is followed by its sequential number in the particular GPCR. For example, the most conserved residue in helix seven of the GnRH receptor is Pro, which is designated Pro7.50. In the GnRH receptor, Pro7.50 is residue number 320 and is designated Pro7.50(320). Asp302 is 18 amino acids closer to the amino-terminal than Pro7.50 and is therefore designated Asp7.32(302).
Site Directed Mutagenesis.
A polymerase chain reaction-based mutagenesis method was used to replace Asp7.32(302) with Asn in the human GnRH receptor. Primers contained the desired mutation and a silent restriction endonuclease site flanked by 12 bases of the wild-type receptor sequence on either side. Polymerase chain reaction products were digested with appropriate restriction enzymes, ligated using T4 DNA Ligase (Amersham Pharmacia Biotech, Piscataway, NJ), subcloned into theEcoRI and XhoI sites of the mammalian expression vector pcDNAI/AMP (Invitrogen, Carlsbad, CA), and transformed into competent XL-1 blue Escherichia coli. Plasmid DNA was extracted (Nucleobond Kit; Macherey-Nagel, Duren, Germany) from ampicillin-resistant clones and the mutation was confirmed by DNA sequencing (Epicentre Technologies, Madison, WI).
Transfection and Cell Culture.
COS-1 cells were transiently transfected using the DEAE-Dextran method (Keown et al., 1990), as described previously (Millar et al., 1995). After transfection, COS-1 cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal calf serum (Delta Bioproducts, Kempton Park, South Africa) and antibiotics (2 mg/ml streptomycin sulfate, 4000 U/ml sodium benzylpenicillin) in a 10% CO2 incubator at 37°C.
GnRH Analogs.
GnRH (pGlu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-GlyNH2), [His5,d-Tyr6]-GnRH, [d-Ala6,N-Me-Leu7,Pro9-NHEt]-GnRH, [d-Trp6,Pro9-NHEt]-GnRH, [d-Trp6,Gln8,Pro9-NHEt]-GnRH, [Gln8]-GnRH, GnRH II ([His5,Trp7,Tyr8]-GnRH), and antagonist 27 ([Ac-d-3-(2-naphthyl)alanine1,d-Me-4-Cl-Phe2,d-Trp3,Ipr-Lys5,d-Tyr6,d-Ala10-NH2]-GnRH) were prepared by conventional solid phase methodology and purified by preparative C-18 reverse phase high-performance liquid chromatography in our Cape Town laboratory. Antagonist 129–62([Ac-d-3-(2-naphthyl)alanine1,d-4-Cl-Phe2,d-Trp3,3-(3-pyridyl)alanine5,6,7 γ-lactam,Ipr-Lys8,d-Ala10-NH2]-GnRH) and [Glu8]-GnRH were prepared by solid phase synthesis on a 4-methylbenzhydrylamine HCl resin using Boc/Benzyl chemistry. The Boc-γ-lactam (Freidinger et al., 1980) was added as one amino acid unit. After removal from the resin by hydrogen fluoride, the peptides were purified to homogeneity by reverse-phase high-performance liquid chromatography on a C-18 preparative column. Antagonist 26 ([Ac-d-4-Cl-Phe1,2,d-Trp3,d-Lys6,d-Ala10-NH2]-GnRH) was a gift from David Coy (Tulane University School of Medicine, New Orleans, LA). [6,7 γ-lactam]-GnRH was a gift from Roger Freidinger (Merck & Co., West Point, PA).
Phosphatidyl Inositol Hydrolysis.
Transfected COS-1 cells (2 × 105 cells/well) in 12-well plates were incubated overnight with myo-[2- 3H]inositol (1 μCi/well; Amersham) in 0.5 ml Medium 199 (Invitrogen) with antibiotics. Labeled cells were incubated with various concentrations of ligand for 1 h at 37°C in the presence of LiCl as described previously (Millar et al., 1995). Aspirating the medium and addition of 10 mM formic acid (1 ml/well) terminated the incubation. Inositol phosphates were extracted from the formic acid extract on DOWEX-1 ion exchange columns and eluted into scintillation liquid (Quicksafe; Zinsser Analytical, Frankfurt, Germany) and the radioactivity was counted.
Radioligand Binding.
Membrane binding assays were performed because this method makes it possible to optimize receptor concentration by varying the amount of transfected membranes used in the assay (Millar et al., 1995). The agonist peptide, [His5,d-Tyr6]-GnRH, was radioiodinated by the Chloramine-T method as described previously (Flanagan et al., 1998). Specific activity ranged between 900 and 1800 μCi/μg and 69% of the radioactivity could be bound by GnRH receptors. Using this high-affinity label allowed accurate determination of IC50 values for the mutant receptor, which had low total binding (Flanagan et al., 1998). Transfected COS-1 cells were homogenized in binding buffer (1 mM EDTA, 10 mM HEPES, pH 7.4, 0.1% bovine serum albumin) and centrifuged at 15,000g for 30 min at 4°C. The resultant crude membrane pellet was resuspended in binding buffer. The membrane suspension was incubated overnight at 4°C with125I-[His5,d-Tyr6]-GnRH (50,000 cpm, 50 pM) and varying concentrations of unlabeled GnRH analogs. We have found previously that equilibrium binding is achieved after 21 h and stable for up to 30 h (Flanagan et al., 1998). The incubation was terminated by the addition of cold polyethylenimine (0.01%; PEI) and immediate filtration through glass-fiber filters (GF/C; Whatman, Maidstone, UK) which were presoaked in 1% PEI. The filters were washed twice with 0.01% PEI and the retained radioactivity was counted. Nonspecific binding was determined in the presence of 1 μM antagonist 27.
Data Reduction.
IP assays were performed at least three times in duplicate and competition binding assays in triplicate. Four-parameter nonlinear curve fitting (Prism; GraphPad Software Inc., San Diego, CA) was used to estimate the peptide concentrations required to stimulate half-maximal IP production (EC50) and to half-maximally inhibit the binding of the radioligand (IC50). K i values were calculated using the Cheng-Prusoff equation (Cheng and Prusoff, 1973). K d andB max values were determined using nonlinear curve-fitting (Prism) of homologous competition binding assays (Munson and Rodbard, 1980; Klotz, 1982; Motulsky, 1999). The high nonspecific binding of the125I-[His5,d-Tyr6]-GnRH tracer at high concentrations makes saturation binding assays unreliable. Data in figures are from single experiments that are representative of at least three independent experiments. Data in tables are the mean ± S.E. of at least three experiments.P values were calculated using unpaired two-tailedt tests performed on the negative log ofK i and EC50 values, and paired two-tailed t tests on theE max counts.
Results
Mutation of Asp7.32(302) Decreases Affinity for GnRH.
The wild-type GnRH receptor bound [His5,d-Tyr6]-GnRH, which was used as a radiolabeled ligand, with high affinity (K d = 0.35 ± 0.06 nM). The Asp7.32(302)Asn mutant receptor showed 2.8-fold lower affinity for [His5,d-Tyr6]-GnRH (K d 0.99 nM ± 0.01 nM). Receptor number was unaffected by the Asp7.32(302)mutation (wild-type, 1.31 ± 0.23 × 105 sites/cell; Asp7.32(302)Asn mutant, 1.31 ± 0.06 × 105 sites/cell). The similar expression suggests that the lower total binding of the mutant receptor reported in initial experiments (Flanagan et al., 1998) results from the slightly decreased affinity for125I-[His5,d-Tyr6]-GnRH.
In competitive ligand binding assays, the affinities of the wild-type human GnRH receptor for uncharged, [Gln8]-GnRH (K i = 923 ± 222 nM) and negatively charged, [Glu8]-GnRH (>10,000 nM) were lower than that for Arg8-containing GnRH (K i = 6.79 ± 1.08 nM) (Table1, Fig. 2). This shows that the human GnRH receptor preferentially binds Arg8-containing GnRH. Mutating Asp7.32(302) to Asn decreased affinity for native GnRH (31.2-fold), but the affinities for uncharged [Gln8]-GnRH or negatively charged [Glu8]-GnRH were unchanged (Table 1, Fig. 2). This indicates that Asp7.32(302) determines the specificity of the wild-type GnRH receptor for native Arg8-containing GnRH. This is consistent with the proposal that the carboxyl side chain of Asp7.32(302) may be involved in an electrostatic interaction with the positively charged Arg8 side chain in GnRH. However, the mutant receptor retained higher affinity for native GnRH than for [Gln8]-GnRH (6-fold) and did not show enhanced affinity for [Glu8]-GnRH. This result suggests that the mechanism by which the receptor selects for Arg8-containing GnRH may be more complex than a simple electrostatic interaction of Arg8 with Asp7.32(302).
GnRH Analogs with d-Amino Acid Substitutions in Position Six of GnRH Enhance Affinity and Overcome the Absence of Arg8 and/or Asp7.32(302).
We found previously that binding of the conformationally constrained analog, [d-Trp6,Pro9-NHEt]-GnRH, was not affected by the Glu7.32(301) Gln mutation in the mouse GnRH receptor (Flanagan et al., 1994). To test whether this is a general phenomenon in mammalian GnRH receptors, binding affinities of a series of GnRH analogs with differentd-amino acid substitutions in position six were characterized in the wild-type human receptor and in the Asp7.32(302)Asn mutant (Fig.3). As expected for the wild-type receptor (Karten and Rivier, 1986; Sealfon et al., 1997), three GnRH analogs, [His5,d-Tyr6]-GnRH, [d-Ala6,N-Me-Leu7,Pro9-NHEt]-GnRH, and [d-Trp6,Pro9-NHEt]-GnRH, showed higher affinity (7.0- to 45-fold) than native GnRH (Tables 1 and2, Fig. 3). The affinity of the wild-type receptor for the uncharged but conformationally constrained analog [d-Trp6,Gln8,Pro9-NHEt]-GnRH (K i = 1.9 ± 0.19 nM) was 486-fold higher than the affinity for [Gln8]-GnRH (K i = 923 ± 222 nM) (Table 1). This shows that, in the wild-type receptor, incorporation of a d-amino acid in position six enhances the affinity of [Gln8]-GnRH more than the affinity of Arg8 -containing GnRH and thus compensates for the absence of Arg8 (Table 2). However, the affinity of [d-Trp6,Gln8,Pro9-NHEt]-GnRH (K i = 1.9 ± 0.19 nM) remained 13-fold lower than the affinity of [d-Trp6,Pro9-NHEt]-GnRH (K i = 0.15 ± 0.04 nM).
All four GnRH analogs with d-amino acid substitutions in position six had similar high affinity for the Asp7.32(302)Asn mutant receptor compared with the wild-type receptor (Table 1). In the mutant receptor, the affinities of the three analogs with d-amino acids in position six and Arg in position eight were 137- to 784-fold higher than the affinity for (native) GnRH (Table 2). The affinity of the uncharged [d-Trp6,Gln8,Pro9-NHEt]-GnRH was 744-fold higher than for [Gln8 ]-GnRH (Fig.3, Table 2).
Incorporation of a 6,7 γ-Lactam Enhances Binding to the Receptor.
Because d-amino acid substitutions in position six are thought to stabilize a high-affinity conformation of GnRH (Monahan et al., 1973), the high affinity of the mutant receptor for peptides with d-amino acids in position six suggests that conformationally constrained peptides may be less sensitive to the Asp7.32(302) mutation. However, part of the enhanced affinity may be contributed by an interaction of the amino acid side chain (e.g., d-Trp) with a receptor residue. To test whether the high affinity is caused predominantly by the conformational constraint of the d-amino acid, peptides with a conformational constraint, in which there is no side chain, were examined. Introduction of a γ-lactam moiety in place of residues six and seven is reported to impose a peptide conformation, which is similar to that stabilized by d-amino acid modifications (Freidinger et al., 1980).
Consistent with the previous report (Freidinger et al., 1980), [6,7 γ-lactam]-GnRH exhibited higher affinity than GnRH for the wild-type GnRH receptor (5.0-fold; Table 2). This result is similar to the increase found with d-amino acid substitutions in position six of GnRH (Table 2). [6,7 γ-Lactam]-GnRH also had high affinity for the Asp7.32(302)Asn mutant GnRH receptor (K i = 4.66 ± 0.36 nM; Table1), which was 45.5-fold higher than the affinity for native GnRH (Table 2). This shows that the γ-lactam constraint enhances the affinity of GnRH for both the wild-type and mutant receptors (Fig. 4, Table 1). [6,7 γ-Lactam]-GnRH had similar affinity for both the wild-type receptor (K i = 1.37 ± 0.07 nM) and the mutant receptor (K i = 4.66 ± 0.36 nM, Table 1). Similar to the d-amino acid substitution in position six, incorporation of 6,7 γ-lactam enhanced binding affinity more in the mutant receptor than in the wild-type receptor. These results show that when the conformation of GnRH is constrained (6,7 γ-lactam or d-amino acid in position six), Asp7.32(302) of the receptor is not required for high-affinity binding. Furthermore, the high affinity of [d-Trp6,Gln8,Pro9-NHEt]-GnRH, which has a constraint but no Arg8, shows that Arg8 also is not required for high-affinity binding of conformationally constrained agonists.
Another class of conformationally constrained GnRH analogs includes the GnRH antagonists. The novel antagonist, antagonist 129–62, which has a 6,7 γ-lactam, and antagonist 26, which hasd-Lys6, had similar high affinities for the wild-type and mutant receptors (Table 1, Fig.5).
GnRH II, which occurs naturally in the human brain (White et al., 1998) does not contain Arg8, but binds the GnRH receptor with higher affinity (K i =193.7 ± 28.5 nM) than [Gln8]-GnRH (K i = 923 ± 222 nM; Table 1, Fig. 4) or [Tyr8]-GnRH (Millar et al., 1989). This suggests that the combination of substitutions in positions five, seven, and eight allows the peptide to bind with relatively high affinity, independently of interactions that involve Arg8, possibly by stabilizing a high affinity conformation (de L Milton et al., 1983). GnRH II had similar affinity for the wild-type (K i = 193.7 ± 28.5 nM) and mutant receptors (K i =213.3 ± 72.4 nM; Table 1). Consequently, the Asp7.32(302)Asn mutant receptor retained higher affinity for GnRH II (K i = 213.3 ± 72.4 nM) than for [Gln8]-GnRH (K i = 1240 ± 227 nM).
In summary, these results show that mutating Asp7.32(302) of the human GnRH receptor to Asn decreases the affinity for GnRH in a manner that is specific to an interaction with Arg8, and that a specific conformation of GnRH is important for high-affinity binding of GnRH to its receptor.
Decreased IP Production in the Asp7.32(302)Asn Mutant GnRH receptor.
The Asp7.32(302)Asn mutant GnRH receptor coupled to the IP signaling pathway (Fig.6). GnRH displayed a 44-fold decrease in potency at the mutant receptor (EC50 = 12.6 ± 1.78 nM) relative to the wild-type receptor (EC50 = 0.29 ± 0.07 nM; Fig. 6). This decrease in potency is consistent with the decreased affinity of the mutant receptor for native GnRH. However, the mutant receptor also exhibited a decreased E max value for GnRH (Fig. 6, Table 3), suggesting that the mutation may induce partial uncoupling of the receptor from intracellular signaling. Surprisingly, ligands that showed no decrease in affinity for the mutant receptor also exhibited decreased IP production in the mutant receptor (Table 3, Fig. 6). This shows that the effect of the mutation on cytosolic signaling is distinct from its effect on binding affinity for GnRH.
Discussion
The basic Arg residue in position eight of GnRH is required for high-affinity binding to mammalian GnRH receptors (Millar et al., 1989). The proposal that Arg8 may be involved in an electrostatic interaction with an acidic residue in the GnRH receptor (Hazum, 1987) was examined in the mouse GnRH receptor, where it was found that the Glu7.32(301) residue confers specificity for GnRH with Arg in position eight (Flanagan et al., 1994). Despite the demonstrated functional importance of Glu7.32(301), this residue is not conserved in the human GnRH receptor, which has Asp7.32(302)instead (Chi et al., 1993). The carboxyl side chain of both residues suggests that Asp7.32(302) can potentially perform the same functions in the human GnRH receptor as Glu7.32(301) does in the mouse receptor and computational models of both rodent and human receptors have incorporated interactions of Glu or Asp with Arg8. However, the study in the mouse receptor suggested that the mechanism of receptor selectivity for Arg8-containing GnRH may be more complex than a simple electrostatic interaction in the receptor ligand complex. This demonstrates a need for detailed investigation of the mechanism by which Asp7.32(302) determines specificity for Arg8. We show that an electrostatic interaction is not required for high-affinity binding of certain GnRH analogs and we propose that Asp7.32(302) may stabilize a high-affinity conformation of GnRH.
Asp7.32(302) of the Human GnRH Receptor Determines Selectivity for GnRH.
The decreased affinity of the Asp7.32(302)Asn mutant for native GnRH, but not for [Gln8]-GnRH shows a loss of specificity for Arg8 of GnRH. This shows that, like the Glu7.32(301) residue of the mouse GnRH receptor, the Asp7.32(302) side chain determines receptor preference for the Arg8 side chain of GnRH. However, the 31-fold decrease in the affinity of the Asp7.32(302)Asn mutant for GnRH is smaller than would be expected for a loss of an electrostatic interaction (Wells et al., 1987). The low affinity of the negatively charged ligand, [Glu8]-GnRH could potentially result from a repulsive interaction with the negatively charged Asp7.32(302) of the wild-type receptor, but the Asp7.32(302)Asn human GnRH receptor mutant did not show increased affinity for [Glu8]-GnRH. This suggests that the mutation does not remove an unfavourable interaction with [Glu8]-GnRH; consequently, high-affinity binding of native GnRH may not arise from an electrostatic interaction of the Arg8 and Asp7.32(302) side chains.
Conformational Constraints Enhance GnRH Binding to the Wild-Type GnRH Receptor.
Arg8 is proposed to stabilize an active conformation of GnRH (Shinitzky et al., 1976), which consists of a β-II-bend involving the Tyr5-Gly6-Leu7-Arg8residues (Monahan et al., 1973). Incorporation of a d-amino acid in position six (Momany, 1976) or a γ-lactam in positions six and seven (Freidinger et al., 1980) is proposed to stabilize this conformation and results in an increased GnRH potency (Monahan et al., 1973; Freidinger et al., 1980). The current study has shown that GnRH analogs with d-amino acid substitutions in position six, or with a 6,7 γ-lactam, have higher affinities than native GnRH for the wild-type human GnRH receptor. This shows that the human GnRH receptor has enhanced affinity for the conformational state of GnRH that is stabilized by incorporating either constraint.
To test whether preferential binding of Arg8-containing GnRH is retained in the presence of a conformational constraint in the wild-type receptor, affinities of conformationally constrained peptides with Arg8or Gln8([d-Trp6,Pro9-NHEt]-GnRH and [d-Trp6,Gln8,Pro9-NHEt]-GnRH) were compared. The modification enhanced the affinity of [Gln8]-GnRH more than that of GnRH (Table 2). Thus, a conformational constraint not only enhances the affinity of [Gln8]-GnRH but also compensates for the absence of Arg8. This shows that Arg8 is not required for high-affinity binding of conformationally constrained analogs to the wild-type human GnRH receptor.
Constrained Ligands Retain High Affinity for the Asp7.32(302)Asn Mutant.
In contrast to native GnRH, GnRH analogs with conformational constraints retained high affinity for the Asp7.32(302)Asn mutant receptor. Because of the lower affinity of the mutant receptor for GnRH, the enhancement of the affinity of the conformationally constrained analogs, compared with GnRH, was greater in the mutant receptor than in the wild-type receptor. The preservation of high affinity for constrained peptides in the mutant receptor shows that the Asp7.32(302)side chain is not required for high-affinity binding of conformationally constrained GnRH analogs.
Ligand Constraint Compensates for the Absence of Both Arg8 and Asp7.32(302).
The high-affinity binding of the Gln8-containing analog, [d-Trp6,Gln8,Pro9-NHEt]-GnRH, to the mutant receptor shows that the conformational constraint can compensate for the simultaneous absence of both Arg8 in the ligand and Asp7.32(302) in the receptor. This result suggests that native GnRH interacts with the receptor differently than conformationally constrained GnRH analogs. Thus, native GnRH and conformationally constrained GnRH analogs may occupy different (although overlapping) binding pockets on the receptor. Two other GnRH receptor residues, Asp2.61(98) and Asn2.65(102), determine recognition of His2 and Gly-NH2 of GnRH, respectively (Davidson et al., 1996a; Flanagan et al., 2000). Comprehensive analysis shows that the interaction of these receptor residues with native GnRH is similar to their interaction with constrained analogs. Mutation of these residues decreases receptor recognition of native and constrained analogs of GnRH to the same extent, suggesting that both native GnRH and constrained analogs interact with these residues (Davidson et al., 1996; Flanagan et al., 2000). Thus, the ability of the ligand conformational constraint to overcome a receptor mutation is specific for the Asp7.32(302)Asn mutant.
Asp7.32(302) and Arg8 Induce a High-Affinity Conformation of GnRH.
We have shown that substituting Arg8 of native GnRH or Asp7.32(302) of the receptor decreases binding affinity. The lack of an additive effect when both substitutions are combined suggests that these side chains interact with each other. However, conformational constraint of the ligand reverses the loss of affinity due to substitution of Arg8 and/or Asp7.32(302). This suggests that both residues have roles in stabilizing a high affinity conformation of unconstrained GnRH. Arg8 has been proposed to have two distinct roles in high-affinity binding: an intramolecular interaction that stabilizes a high-affinity peptide conformation (Shinitzky et al., 1976) and an intermolecular electrostatic interaction with an acidic group in the receptor (Hazum, 1987; Flanagan et al., 1994). The current results suggest that Arg8 both stabilizes peptide conformation and interacts with Asp7.32(302), and that Asp7.32(302) also affects peptide conformation. This, in turn, suggests that an interaction of Arg8 with Asp7.32(302)affects peptide conformation. The similar affinities of constrained peptides for the wild-type and mutant receptors (with and without Asp7.32(302)) suggest that once the ligand is in a high-affinity conformation, the putative Arg8-Asp7.32(302)interaction does not contribute to the binding energy of the final ligand-receptor complex. Although our results suggest that an interaction between Arg8 and Asp7.32(302) may be required to induce a high-affinity conformation in unconstrained, native GnRH, the absence of this interaction with constrained analogs suggests that the interaction that induces the high-affinity conformation is transient. It has been suggested that residues on the extracellular surface of the TRH receptor form a initial ligand recognition site (Perlman et al., 1997) and that TRH binds sequentially with the surface binding site, and then with the transmembrane binding pocket (Colson et al., 1998). GnRH may interact initially with Asp7.32(302) and then move to a final binding pocket, which involves Asn2.65(102) and Asp2.61(98) (Davidson et al., 1996; Flanagan et al., 2000), after assuming a high affinity conformation. Thus, contrary to the initial hypothesis of an electrostatic interaction in the ligand-receptor complex, we show that the basis of receptor selectivity for mammalian GnRH seems to be the ability of Asp7.32(302) to induce a high affinity conformation in native GnRH.
Cytosolic signaling assays showed that GnRH had decreased potency at the Asp7.32(302)Asn mutant receptor, consistent with its decreased affinity. However, theE max value for GnRH was lower in the mutant receptor, suggesting that the mutation destabilizes the activated receptor conformation (Samama et al., 1993). Several agonists, which showed no decrease in affinity for the mutant receptor, nevertheless showed decreased stimulation of IP production. This suggests that the Asp7.32(302) side chain has a role stabilizing the activated receptor conformation. The apparent decreased efficacy of ligands that had unchanged affinity suggests that this function is distinct from its role in ligand selectivity. A previous study reported that mutagenesis of extracellular loop three of the β2-adrenergic receptor also affected receptor activation (Zhao et al., 1998).
In conclusion, the wild-type human GnRH receptor recognizes and binds ligands in a specific conformation that can be stabilized by ligand modifications. We show that the Asp7.32(302) side chain determines selectivity for Arg8-containing GnRH. However, certain ligand conformational constraints overcome the decrease in affinity that results from substitution of Arg8 and Asp7.32(302). This suggests that Asp7.32(302) determines selectivity for Arg8-containing GnRH by its ability to induce a high-affinity conformation in the ligand. We propose that unconstrained, Arg8-containing, native GnRH interacts transiently with Asp7.32(302), which induces a high affinity conformation in the ligand, before it interacts with a final ligand binding pocket, which excludes Asp7.32(302). This further definition of the mechanism of ligand recognition improves our conceptual models of GnRH receptor-ligand interaction.
Acknowledgments
We thank David Coy for antagonist 26 and Roger Freidinger for [6,7 γ-lactam]-GnRH and Boc-γ lactam.
Footnotes
- Received February 2, 2001.
- Accepted August 15, 2001.
-
This research was supported by the Medical Research Council of South Africa, National Research Foundation of South Africa, University of Cape Town and the Wellcome Trust.
Abbreviations
- GnRH
- gonadotropin-releasing hormone
- GPCR
- G protein-coupled receptor
- HPLC
- high-performance liquid chromatography
- PEI
- polyethylenimine
- IP
- inositol phosphate
- antagonist 26
- [Ac-d-4-Cl-Phe1,2,d-Trp3,d-Lys6,d-Ala10-NH2]-GnRH
- antagonist 129–62
- [Ac-d-3-(2-naphthyl) alanine1,d-4-Cl-Phe2,d-Trp3,3-(3-pyridyl)alanine5,6,7 γ-lactam, Ipr-Lys8,d-Ala10-NH2]-GnRH
- Emax
- maximal agonist-stimulated inositol phosphate production
- Ipr-Lys
- Nε-isopropyllysine
- GnRH II
- [His5,Trp7,Tyr8]-GnRH
- The American Society for Pharmacology and Experimental Therapeutics