Full Activation of Chimeric Receptors by Hybrids between Parathyroid Hormone and Calcitonin EVIDENCE FOR A COMMON PATTERN OF LIGAND-RECEPTOR INTERACTION*

Calcitonin (CT) and parathyroid hormone (PTH), whose receptors belong to the same family of G protein-coupled receptors, share no amino acid sequence homo- logy and selectively activate either CT or PTH receptors. We now show, however, that reciprocal hybrid ligands (CT/PTH and PTH/CT), which do not activate the “wild-type” receptors, activate PTH/CT and CT/PTH receptor chimeras, respectively. Our findings indicate that PTH and CT share a similar architecture with at least two functional, receptor-specific domains. These domains are sufficiently independent to permit synthetic hybrid ligands to efficiently activate appropriate receptor chimeras. Therefore, both ligands follow, despite their very different primary sequences, a common pattern of li-gand-receptor interaction. The isolation of cDNAs encoding the receptors for secretin (1), calcitonin (CT) 1 (2), and parathyroid hormone (PTH)/PTH-related peptide (PTHrP) (3, 4) established a new family of G protein-coupled receptors (GPRs) (5, 6). Members of this family, including some invertebrate receptors (7), 2,3 share, in ad-dition to their overall structure with seven membrane-span-ning helices, approximately 50 strictly conserved amino acids, including 8 important cysteines, an almost invariant amino acid sequence of the seventh membrane-spanning domain, and a similar intron/exon organization (10–16). These findings suggest that this family of GPRs may have evolved from a (800,000 cpm/well, except 200,000 cpm/well for the PR), 125 I-sCT-(1–32) (800,000 cpm/well, except 200,000 cpm/well for the CR), or 125 I-P (1– 13)/C (12–32) (800,000 cpm/well), or at 15 °C for 4 h with 125 I-C (1–11)/P (15–34) (800,000 cpm/well) and various concentrations of unlabeled competing peptides (31). Ligand association studies demonstrated that under these conditions binding reached equilibrium. For subsequent determination of cell-associated radioactivity, rinsed was was determined sub-tracting nonspecific binding was the presence of max- imal cold competing

The isolation of cDNAs encoding the receptors for secretin (1), calcitonin (CT) 1 (2), and parathyroid hormone (PTH)/PTHrelated peptide (PTHrP) (3,4) established a new family of G protein-coupled receptors (GPRs) (5,6). Members of this family, including some invertebrate receptors (7), 2,3 share, in ad-dition to their overall structure with seven membrane-spanning helices, approximately 50 strictly conserved amino acids, including 8 important cysteines, an almost invariant amino acid sequence of the seventh membrane-spanning domain, and a similar intron/exon organization (10 -16). These findings suggest that this family of GPRs may have evolved from a common ancestral precursor. The ligands which activate these GPRs are similar in length, but lack, with the exception of the first 13 residues in PTH and PTHrP, any amino acid sequence homology. For both PTH-  and CT- , the importance of the amino terminus for bioactivity has been recognized, whereas the carboxyl-terminal portion contributes predominantly to receptor binding (17)(18)(19)(20)(21)(22)(23). Based on the limited data available, we and others had previously proposed that the carboxyl-terminal portion of these ligands determines specificity for the aminoterminal, extracellular receptor domains, while the amino termini of most ligands functionally interact with the membraneembedded receptor region (23)(24)(25)(26)(27)(28)(29). We have now tested this hypothesis more directly by constructing reciprocal chimeric ligands composed of portions of PTH and CT, and reciprocal chimeric receptors composed of portions of the PTH/PTHrP and CT receptor. These chimeras were designed such that the carboxyl-terminal portion of the ligand would match the aminoterminal, extracellular domain of the receptor, while the amino-terminal portion of the hybrid ligand would correspond to the membrane-embedded domains of the receptor and connecting loops. Our functional analysis of these hybrid ligands and chimeric receptors strengthens the proposed model of ligand-receptor interaction. Moreover, our studies imply that the receptors and their ligands are composed of functionally independent domains.
Construction of Chimeric Receptors and Expression in COS-7 Cells-Expression cloning of the rat PTH/PTHrP receptor (PR) and the porcine calcitonin receptor (CR) using the mammalian expression vector pcDNAI (Invitrogen, San Diego, CA) was described previously (2)(3)(4). The human influenza virus hemagglutinin (HA) epitope tag (32) which had been introduced into the PR (replacing residues 93-101) (33) was also inserted into the CR (between residues 66 and 67, PYDVPDYA) using site-directed mutagenesis (34). To distinguish them from chimeric receptors, these receptors are referred to as "wild-type." Corresponding regions of the PR and CR were identified by aligning the complete amino acid sequences of the two receptors using the GCG Pile-up program (Genetics Computer Group, Madison, WI).
A PstI restriction site in the CR, engineered into the PR at the corresponding location (residue 177) by site-directed mutagenesis, was used to construct the hybrid receptor PCR1 that contains residues 1-177 of the PR, joined to residues 143-482 of the CR. The reciprocal receptor hybrid CPR1 (amino terminus of the CR (residues 1-142) annealed at position 178 to the carboxyl terminus of the PR) was constructed by introducing an XmaI site into the CR; the introduction of this site into the PR was reported previously (23). A second set of reciprocal receptor hybrids was constructed at a BspEI site within the HA-tag: PCR2 contains residues 1-97 of the PR, joined to residues 67-482 of the CR; CPR2 contains the corresponding amino terminus of the CR (residues 1-66), joined to residues 98 -591 of the PR. The described restriction sites and a novel BspH1 site in CR (residue 140) were also used to construct the receptors CR-(1-66)/PR-(98 -175)/CR-(141-482) (CPC) and PR-(1-97)/CR-(67-142)/PR-(178 -591) (PCP).
Plasmids encoding the wild-type and hybrid receptor constructs were transfected into COS-7 cells, grown in 10-cm dishes, using the diethylaminoethyl (DEAE)-dextran method (23,33). The cells were subcultured into 24-well plates 18 -24 h after transfection, and functionally evaluated after 72 h.
Evaluation of Receptor Cell Surface Expression-The use of an HAepitope to assess receptor surface expression in intact COS-7 cells was described previously (32,33). Briefly, transfected COS-7 cells were incubated at 15°C for 2 h with binding buffer (50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 5% heat-inactivated horse serum, 0.5% heat-inactivated fetal bovine serum) containing 5 g/ml anti-HA monoclonal antibody (12CA5, Boehringer Mannheim). After extensive washing, the anti-HA antibody binding was detected by incubation with a 125 I-labeled sheep anti-mouse (FabЈ) 2 fragment (NEX-162, DuPont NEN) (200,000 cpm/well) which was diluted in binding buffer for an additional 2 h at 15°C. The cells were then rinsed twice, lysed with 1 N NaOH, and the entire lysate was counted for ␥-irradiation. The cell surface expression levels of the different receptors, as determined by specific anti-HA antibody binding, were equalized by varying the amount of transfected plasmid DNA (Fig. 1A). The amounts transfected per 10-cm dish were: PCR1 ϭ 0.14 g; PCR2 ϭ 150 g; CPR1 ϭ 80 g; CPR2 ϭ 2 g; CPC ϭ 1.1 g; PR ϭ 0.8 g; CR ϭ 0.17 g.
Cyclic AMP Assay-24-well plates containing transfected COS-7 cells were chilled on ice, the cells were rinsed with phosphate-buffered saline, and stimulated in cAMP assay buffer (Dulbecco's modified Eagle's medium, 2 mM 3-isobutyl-1-methylxanthine, 0.1% bovine serum albumin, 20 mM HEPES (pH 7.4)) containing various concentrations of peptide for 15 min at 37°C. The reaction was then stopped by removing the buffer and chilling the plates on dry ice. Intracellular cAMP was extracted with 50 mM HCl and measured by radioimmunoassay as described (3).
Data Calculation-EC 50 values and IC 50 values were determined from plots of log(S/S max Ϫ S) versus log[peptide] and calculation of the x-intercept. S represents the cAMP accumulation or the specific binding at a given peptide concentration [peptide], respectively. The data shown represent the mean Ϯ S.E. of at least three independent experiments that were performed in duplicate.

Activation of PTH-CT Receptor Chimeras by Hybrid Ligands 26470
activated cAMP accumulation was 420-fold higher with cells expressing CPR1 than with those expressing the wild-type PR (EC 50 : 460 Ϯ 83 and 1.1 Ϯ 0.28 nM, respectively) (Fig. 2, C and  D). Similarly, sCT-(1-32) stimulated cAMP accumulation not only in COS-7 cells expressing the CR, but also in cells expressing the chimeric receptor PCR1, while cells expressing CPR1 or PR showed no response (Fig. 1E). Comparable to the findings with PTH-(1-34), the EC 50 of sCT-(1-32)-activated cAMP accumulation was 40-fold higher with cells expressing the PCR1 than that observed for the native receptor CR (EC 50 : 19 Ϯ 1.3 and 0.5 Ϯ 0.13 nM, respectively) (Fig. 2, A and B). These results indicate that the membrane-embedded portion and associated loops of each receptor are sufficient to mediate, albeit with reduced potency, the biological activity of sCT-(1-32) and PTH- . The relatively high potency of sCT-(1-32) with the chimeric receptor PCR1 may be due to its exceptionally low dissociation constant (27,35). Indeed, human CT-(1-32)-NH 2 , which does not share this low dissociation constant, was only weakly active with the PCR1 chimera (EC 50 : 58 Ϯ 17 nM for CR, EC 50 : Ͼ10,000 nM for PCR1, Fig. 2, A and B).
Interestingly, in comparison to CR, sCT-(1-32) had reduced, but comparable potency on PCR1 and PCR2. Similarly, PTH-(1-34) was equally active with CPR1 and CPR2, but less active than with the PR. Determinants for ligand binding are, therefore, likely to be located at least in two distinct portions of the amino-terminal, extracellular domain of each receptor (25,33).
The members of the secretin/PTH/CT receptor family share significant amino acid sequence homology, and the precise spacing of conserved residues predicts that these GPRs all have a similar three-dimensional structure (5,6). Peptide ligands, which activate members of this receptor family, such as PTH and CT, are also predicted to have a similar conformation (36 -42) and likely display a two-site mode of interaction with their receptors (8, 9, 23, 26 -28, 33, 43).
In the present report, we show that the carboxyl-terminal portion of the ligands PTH or CT binds to the amino-terminal, extracellular domain of their receptor, while the amino terminus of the ligand interacts with the membrane-embedded domain of the receptor and its associated loops and thereby triggers receptor activation. Therefore, both ligands follow, despite their very different primary sequences, a common pattern of ligand-receptor interaction, which may apply also to other members of this receptor family and their ligands. Domains within ligands and receptors that interact with each other appear to be sufficiently independent, such that their modular exchange is possible; the resulting pairs of chimeras may be functional, if the interacting domains are appropriately matched. This approach may prove to be useful for the further analysis of ligand-receptor interactions.