Transmembrane residues of the parathyroid hormone (PTH)/PTH-related peptide receptor that specifically affect binding and signaling by agonist ligands.

Polar residues within the transmembrane domains (TMs) of G protein-coupled receptors have been implicated to be important determinants of receptor function. We have identified mutations at two polar sites in the TM regions of the rat parathyroid hormone (PTH)/PTH-related peptide receptor, Arg-233 in TM 2 and Gln-451 in TM 7, that caused 17-200-fold reductions in the binding affinity of the agonist peptide PTH-(1-34) without affecting the binding affinity of the antagonist/partial agonist PTH-(3-34). When mutations at the TM 2 and TM 7 sites were combined, binding affinity for PTH-(1-34) was restored to nearly that of the wild type receptor. The double mutant receptors, however, were completely defective in signaling cAMP or inositol phosphate production in response to PTH-(1-34) agonist ligand. The results demonstrate that Arg-233 and Gln-451 have important roles in determining agonist binding affinity and transmembrane signaling. Furthermore, the finding that residues in TM 2 and TM 7 are functionally linked suggests that the TM domain topology of the PTH/PTH-related peptide receptor may resemble that of receptors in the rhodopsin/beta-adrenergic receptor family, for which structural and mutagenesis data suggest interactions between TMs 2 and 7.

The PTH 1 /PTHrP receptor (PPR) is a member of a distinct G protein-coupled receptor family that includes the receptors for secretin, calcitonin, glucagon, and several other unmodified peptide hormones of intermediate size (25-60 amino acids) (1,2). The receptors in this group display no apparent sequence homology to receptors in the larger G protein-coupled receptor family represented by rhodopsin and the ␤-adrenergic receptor. Current work in our laboratory and several others is focussed on understanding how the receptors in the PTH/secretin/calcitonin family interact with their cognate peptide ligands and mediate transmembrane signaling.
In previous studies in which we used mutant and chimeric PTH/PTHrP receptors, we identified small regions and some specific residues widely scattered throughout the extracellular portion of the receptor that affect the binding of PTH analogs (3)(4)(5). Studies on chimeras formed between other members of this peptide hormone receptor family have reinforced the notion that ligand binding involves multiple receptor domains acting in concert (6 -8).
For members of the ␤-adrenergic receptor family, there is much evidence that indicates that polar residues within the TM domains contribute importantly to binding affinity and specificity toward agonist and antagonist ligands (9 -11). This includes receptors in this family that bind peptide ligands such as substance P (12,13), the endothelins (14,15), angiotensin II (16), and lutropin (17). In addition to ligand interaction, such functionally important polar residues within the TM domains of G protein-coupled receptors could have roles in maintaining receptor structure (9). Although the three-dimensional structure of any G protein-coupled receptor is still unknown, models of the transmembrane portions of bacteriorhodopsin (18) and vertebrate rhodopsin (19) suggest that in each of these proteins, TM domains 2 and 7 are close to each other and could, therefore, interact. Recent mutagenesis data on several members of the ␤-adrenergic receptor family support the notion that residues in transmembrane domains 2 and 7 of these G proteincoupled receptors interact (12, 20 -23).
The TM domains of the PTH/PTHrP receptor contain a number of polar residues that could likewise be important for receptor structure and/or function. In the course of evaluating some of these residues in the rat PPR, we observed that mutations at two sites, Arg-233 in TM 2 or Gln-451 in TM 7 had marked effects on ligand binding. In this paper we show that point mutations at either site markedly impair the binding of the agonist, PTH-(1-34), without affecting the binding of the antagonist PTH- . Furthermore, when the mutations at the two sites are combined, then agonist binding affinity is restored, although signal transduction capability is severely impaired. The data suggest that residues in TM 2 and TM 7 of the PTH/PTHrP receptor are functionally linked and that this putative interaction is important for transmission of the hormone's signal across the cell membrane. cDNA sequence carried on the pCDNA-1 plasmid vector. The rWT-HA DNA sequence encodes the rat PPR containing the HA (human influenza virus hemagglutinin) epitope tag in the extracellular E2 region (4). Point mutations were incorporated into rWT-HA single-stranded plasmid DNA by oligonucleotide-directed mutagenesis (25). The mutagenesis procedure was random with a bias that favored conservative polar substitutions; thus the codon corresponding to Arg-233 (CGC) was changed to (A/C)A(A/C), (His, Gln, Asn, and Lys), and the codon corresponding to Gln-451 (CAG) was changed to (A/G)A(G/C) (Asn, Lys, Asp, and Glu), where the slash denotes an equal mixture of two nucleotides. Mutations at the TM 2 and TM 7 sites were combined by exchanging the appropriate ApaLI-ApaLI restriction enzyme DNA fragments. Mutations were verified by nucleotide sequence analysis of single-stranded plasmid DNA.
Cell Culture and DNA Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (HyClone Laboratories, Logan, Utah), 20 units/ml penicillin G, 20 g/ml streptomycin sulfate, and 0.05 g/ml amphotericin B in a humidified atmosphere containing 95% air/5% CO 2 . Transfections were performed with plasmid DNA that had been purified on cesium chloride gradients. Cells were seeded in 24-well plates (50,000 cells/well) and transfected (100 ng DNA/well) by the DEAE/dextran method (26) when the monolayer reached 90 -95% of confluence. Assays were performed on the intact transfected COS-7 cells 72 h after transfection, at which point the cell density was ϳ550,000 cells/well. Antibody Binding to Epitope-tagged Receptors-The use of PPRs containing the HA epitope tag in a nonessential extracellular region to assess receptor surface expression has been described previously (4). Briefly, transfected COS-7 cells were rinsed with binding buffer (50 mM Tris-HCl, pH 7.7, 100 mM NaCl, 5 mM KCl, 2 mM CaCl, 5% heatinactivated horse serum, 0.5% fetal bovine serum), and 250 l of binding buffer containing the monoclonal antibody 12CA5 (1.25 g/well) was added. After 2 h at 15°C, the cells were rinsed with binding buffer, and 250 l of binding buffer containing 125 I-labeled goat anti-mouse IgG was added (ϳ350,000 cpm/well). After 2 h at 15°C, the cells were washed with binding buffer and lysed with 0.5 ml of 5 M NaOH, and the entire lysate was counted. Nonspecific binding was determined with mock-transfected cells and was 1,100 Ϯ 100 cpm. Maximum specific binding to cells expressing rWT-HA was 1.2 Ϯ 0.1% of total radioactivity added (n ϭ 17). The radioactivity specifically bound by each mutant receptor was divided by the radioactivity specifically bound to the rWT-HA control (defined as 100%) included in each experiment and multiplied by 100.
Radioligand-Receptor Binding-PTH binding assays were performed as described previously (27). Radiolabeled bPTH-(3-34) analog was prepared by chloramine-T iodination and HPLC purification. Binding reactions were performed in 24-well plates (final volume, 300 l/well) and contained 26 fmol of 125 I-bPTH-(3-34) (100,000 cpm) and successive doses (0.4 -300 pmol) of an unlabeled competitor ligand and were incubated for 4 h at 15°C. After washing with binding buffer, the cells were lysed with NaOH and counted for radioactivity. Nonspecific binding, determined in wells containing mock transfected COS-7 cells, was ϳ1% Cyclic AMP Production-24-well plates containing transfected cells were chilled on ice, rinsed with 500 l of binding buffer, and treated with 200 l of ice-cold cAMP assay buffer (Dulbecco's modified Eagle's medium containing 2 mM 3-isobutyl-1-methylxanthine, 1 mg/ml bovine serum albumin, 35 mM Hepes-NaOH, pH 7.4) and 100 l of binding buffer containing various concentrations of PTH ligand. After 30 min at 37°C, the buffer was removed, and the cells were lysed by placing the plates on powdered dry ice and adding 0.5 ml of 50 mM HCl. The cAMP content of a 5-50-l aliquot of the diluted lysate (1:6 with H 2 0) was determined by radioimmunoassay.
Inositol Phosphate Production-48 h after transfection, cell medium was replaced by 0.5 ml of inositol-free Dulbecco's modified Eagle's medium containing [ 3 H]myo-inositol (2 Ci/ml), and incubation continued overnight at 37°C. Cells were washed with 0.5 ml of binding buffer containing LiCl 2 (30 mM), and then 300 l of binding buffer-LiCl 2 with or without PTH ligand was added, and the cells were incubated at 37°C for 40 min. The buffer was then removed and replaced by 0.5 ml of ice-cold trichloroacetic acid (5%). After 1 h on ice, the acid lysate was extracted twice with ether and then applied to a column packed with 0.5 ml of Dowex AG 1-X8 resin (Bio-Rad) in 10 mM myo-inositol. The resin was washed with 10 ml of 10 mM myo-inositol followed by 10 ml of 5 mM Borax/60 mM ammonium formate; the total inositol phosphates were eluted in 3 ml of 1 M ammonium formate/100 mM formic acid (28) and quantified by scintillation counting. Fig. 1 shows the location of Arg-233 and Gln-451 in TMs 2 and 7 of the rat PPR and the conservation of these residues in the PTH/secretin/calcitonin receptor family. These two residues were targeted for analysis by random mutagenesis. To avoid possible impairment of receptor expression that might occur with nonconservative changes, the substitutions introduced at these positions were limited to other polar residues.

RESULTS
The effects of the resulting mutations at Arg-233 (Lys, Asn, Gln, and His) and Gln-451 (Lys) on receptor surface expression in transiently transfected COS-7 cells are reported in Table I. The WT PPR (rWT-HA) and each mutant receptor contained the phenotypically silent HA epitope tag in its amino-terminal extracellular domain. Mutant receptors with changes at Arg-233 bound anti-HA antibody to levels that were at least 80% of those bound by the rWT-HA, whereas the Q451K mutant displayed a mild expression defect (56% of rWT-HA). Scatchard analysis of homologous competition binding studies performed with labeled and unlabeled [Nle 8,18 ,Tyr 34 ]-bPTH(3-34)NH 2 confirmed that each of the mutant receptors was adequately expressed on the cell surface (Table I).
The maximum specific binding of the radiolabeled agonist peptide, 125 I-bPTH-(1-34) analog to the R223H or Q451K mutant receptor (6.6 and 2.3% of total added, respectively) was markedly reduced from the corresponding value seen for rWT-HA (17.6%). The Asn substitution at Arg-233 (R233N) caused a more modest decrease in 125 I-bPTH-(1-34) binding (13.6), whereas the Lys and Gln substitutions at this site  caused little or no change in the binding of the radioligand (Table I). In contrast to the severe effect of the mutations on the maximal binding of 125 I-bPTH-(1-34) analog, these same mutations had little or no effect on the maximal binding of the radiolabeled antagonist peptide, 125 I-bPTH-(3-34) ( Table I).
The small reduction in 125 I-bPTH-(3-34) binding observed for the Q451K receptor was consistent with the proportional reduction in surface expression (Table I) (Table II).
We then investigated the effects of combining the mutations at Arg-233 and Gln-451. As shown in Fig. 2D, the R233H/ Q451K double mutant receptor bound hPTH-(1-34) with higher affinity (IC 50 ϭ 50 nM) than the corresponding single mutant receptors (Table II). Similar increases in binding affinity occurred when the Q451K change was combined with the R233Q or R233N mutations (Table II). Thus, combining mutations in TM 2 and TM 7 did not lead to additive reductions in PTH-(1-34) binding affinity, but instead a substantial rescue of the binding defects caused by the respective individual point mutations.
The transmembrane domains of G protein-coupled receptors and the intracellular connecting loops are likely to play important roles in signal transduction and G protein coupling. The effects of the point mutations in TM 2 and TM 7 on the receptor's ability to stimulate cAMP production in response to PTH ligands are shown in Fig. 3. COS-7 cells expressing rWT-HA responded to hPTH-(1-34) with a 20-fold increase in intracellular cAMP, as compared with the basal level of cAMP measured in the untreated cells (EC 50 ϭ ϳ0.4 nM; Fig. 3). With the exception of the R233K mutant, which exhibited a cAMP response equivalent to that of rWT-HA, the responses of the mutant receptors were considerably weaker than those of rWT-HA. The low responses precluded the definition of precise response maxima and EC 50 values, although qualitative differences in the responses were consistently observed (Fig. 3). The R223H response curve was parallel and reached the maximum response of the WT PPR but was displaced 100-fold to the right. The Q451K, R233Q, and R233N mutants exhibited shallower responses reaching less than 50% of the WT maximum, and  these were shifted 10 -1000-fold to the right of the WT response (Fig. 3).
Remarkably, cells expressing the double mutant receptors, which bound hPTH-(1-34) with high affinities (Table II), were defective in mediating agonist-induced cAMP production; we detected less than a 2-fold increase in cAMP production in response to a maximum dose (1,000 nM) of hPTH-(1-34) (Fig. 3).
In addition to the adenylyl cyclase system, WT PTH/PTHrP receptors also activate phospholipase-C-mediated signal transduction systems (2,29). Accordingly, PTH-(1-34) (1 M) stimulated a 4.5-fold increase in inositol phosphate (IP) levels in COS-7 cells expressing rWT-HA, as compared with the basal level of IPs in untreated cells (Fig. 4). No IP response was detected in cells expressing any of the mutant PTH/PTHrP receptors having either individual or combined substitutions in TM 2 and TM 7 (Fig. 4). At high doses (1 M), the truncated peptide hPTH-  was inactive in IP assays performed on COS-7 cells expressing either the WT PTH/PTHrP receptor or any of the mutant receptors (data not shown). A small, 7-9-fold increase in cAMP formation was detected for this ligand with rWT-HA and the R233K mutant receptor, but no response was observed with any of the other mutant receptors (data not shown). DISCUSSION These studies show that Arg-233 and Gln-451 of the rat PTH/PTHrP receptor play important roles in receptor function. These residues are likely to contribute to interactions involving the amino terminus of the hormone, based on the finding that mutations at these sites markedly impaired the binding of PTH-(1-34) analogs without affecting the binding of PTH-  analogs. Residues 1-2 of PTH are critical for activating the adenylyl cyclase reponse pathway, as shown by the ability of PTH-  analogs to antagonize the agonist activity of PTH-(1-34) (30). Although most mutations in the PTH/PTHrP receptor that we have examined so far have had comparable effects on the binding of PTH-(1-34) and PTH-(3-34) ligands, we recently showed that mutations at two sites in the third extracellular loop of the rat PPR (Trp-437 and Gln-440) resulted in similar dissociation of effects on the binding of the respective peptides (27). In other studies, we identified residues in TMs 5 and 6 of the PPR that contribute to altered binding and signaling responses to another amino-terminally modified ligand [Arg 2 ]PTH-(1-34)(5). Although direct biochemical evidence for any specific interaction between amino acid residues of the ligand and the PTH/PTHrP receptor has yet to be obtained, the cumulative mutagenesis data strongly point to a role for the receptor's membrane-spanning region and third extracellular loop in determining efficient interactions with the amino-terminal residues of PTH.
Perhaps the most revealing finding of our current work was that the PTH-(1-34) binding defects caused by individual point mutations at Arg-233 and Gln-451 could be "cured" by combining the mutations at the two sites. These results strongly suggest that residues in TMs 2 and 7 are functionally linked. Although the double mutant receptors did not fully regain the WT phenotype, it is difficult to understand the effects on PTH binding without considering potential interdomain interactions. This interpretation, if correct, would suggest that the TM domain structure of the PTH/PTHrP receptor bears similarity to the structures proposed for bacteriorhodopsin and rhodopsin (18,19), because in these models TMs 2 and 7 are proximal to each other. Furthermore, mutagenesis data for members of the adrenergic receptor family provide functional evidence in support of an interaction between TM 2 and TM 7 of these receptors (12, 20 -23). Although at the amino acid sequence level the members of the adrenergic/rhodopsin family of G protein-coupled receptors and the PTH/secretin/calcitonin family appear unrelated, the membrane-spanning segments of the proteins in each of these groups may be arranged similarly.
The strong dissociation between agonist binding affinity and signaling capability observed with the double mutant PTH/ PTHrP receptors imply that Arg-233 and Gln-451 have important roles in receptor activation in addition to their influence on agonist binding affinity. Current predictions based on multiple sequence alignment methods (31) suggest that Arg-233 and Gln-451 are near the middle of their respective membrane spanning domains (Fig. 1). These residues could, therefore, be in a suitable position to relay biochemical information from the extracellular surface to the intracellular G proteins. Current receptor theory suggests that the coupling of a receptor to a G protein induces or stabilizes a certain receptor conformation called R* that has enhanced affinity for agonist ligands, as opposed to a low affinity uncoupled form called R (32,33). By these models, neutral antagonist ligands do not distinguish between the R and R* conformations. These considerations raise the possibility that the mutations at Arg-233 and Gln-451 affect PTH binding by changing the receptor's ability to isomerize to the high affinity R* form. Such effects on receptor conformations have been recently discussed in relation to mutations in TM 5 of the angiotensin II receptor (16) and TM 2 of the substance P receptor (13) that impair binding of agonist ligands but not antagonist ligands. Additional studies will be necessary to directly assess how G protein coupling influences the binding of agonists and antagonists to WT and mutant PTH/PTHrP receptors.
That the two receptor residues that we have identified are highly conserved among the receptors in the PTH/secretin/ calcitonin receptor family (Fig. 1) suggests that they perform a function that is preserved in all members of the family. For example, these residues might be involved in maintaining the proper three-dimensional configuration of the membrane-spanning portions of the receptor. Given that the PTH/PTHrP receptor mutants that we have studied here were expressed sufficiently on the cell surface and retained high binding affinity for PTH-(3-34) ligands, it seems unlikely that the mutations severely altered the overall structure of the receptor. Nevertheless, it remains possible that mutations at these sites induce subtle conformational changes that alter other sites more directly involved in ligand binding and signal transduction.
In summary, we have identified two polar residues, Arg-233 in TM 2 and Gln-451 in TM 7 of the rat PTH/PTHrP receptor, that play important roles in determining agonist binding affinity and agonist-induced receptor activation. These residues are functionally linked and may therefore participate in interdomain interactions. Although additional work is needed to fully resolve how these residues contribute to receptor function, our current results demonstrate that receptor mutagenesis strategies combined with the use of agonist and antagonist peptide ligands can help identify and characterize critical residues in these peptide hormone receptors.