Identification of Determinants of Inverse Agonism in a Constitutively Active Parathyroid Hormone/Parathyroid Hormone-related Peptide Receptor by Photoaffinity Cross-linking and Mutational Analysis*

We have investigated receptor structural components responsible for ligand-dependent inverse agonism in a constitutively active mutant of the human parathyroid hormone (PTH)/parathyroid hormone-related peptide (PTHrP) receptor type 1 (hP1R). This mutant receptor, hP1R-H223R (hP1RCAM-HR), was originally identified in Jansen's chondrodysplasia and is altered in transmembrane domain (TM) 2. We utilized the PTHrP analog, [Bpa2,Ile5,Trp23,Tyr36]PTHrP-(1–36)-amide (Bpa2-PTHrP-(1–36)), which has valine 2 replaced byp-benzoyl-l-phenylalanine (Bpa); this substitution renders the peptide a photoreactive inverse agonist at hP1RCAM-HR. This analog cross-linked to hP1RCAM-HR at two contiguous receptor regions as follows: the principal cross-link site (site A) was between receptor residues Pro415–Met441, spanning the TM6/extracellular loop three boundary; the second cross-link site (site B) was within the TM4/TM5 region. Within the site A interval, substitution of Met425 to Leu converted Bpa2-PTHrP-(1–36) from an inverse agonist to a weak partial agonist; this conversion was accompanied by a relative shift of cross-linking from site A to site B. The functional effect of the M425L mutation was specific for Bpa2-containing analogs, as inverse agonism of Bpa2-PTH-(1–34) was similarly eliminated, whereas inverse agonism of [Leu11,d-Trp12]PTHrP-(5–36) was not affected. Overall, our data indicate that interactions between residue 2 of the ligand and the extracellular end of TM6 of the hP1R play an important role in modulating the conversion between active and inactive receptor states.

Inverse agonists are ligands that reduce receptor signaling activity to below the basal signaling level seen with the unoccupied receptor (11,12). These compounds are thus functionally different from agonists, which activate the receptor, and neutral antagonists, which have no efficacy of their own but can prevent the actions of both agonists and inverse agonists by a simple competitive mechanism (13). At least some constitutive activity of a receptor is required in order for inverse agonist activity to be detected; in general, such constitutive activity may be induced by certain receptor mutations or by overexpressing the wild-type form of a receptor (14 -16). By using either method to increase basal receptor signaling, many ligands that were previously classified as neutral antagonists were subsequently shown to be inverse agonists (14 -16). Some of these ligands have gained clinical significance, for example as histamine-1 blockers (17), histamine-2 blockers (18), betablockers (19), or antidepressants (16,20). How the inverse agonist activities of these compounds relate to their clinical effectiveness has not yet been determined. Furthermore, the molecular mechanisms by which inverse agonists reduce receptor signaling activity remain largely unknown, although the pharmacological behavior of these ligands has been discussed in theoretical terms (12,13,21). One crucial question is whether or not the receptor contact points that mediate inverse agonism are distinct from those that induce receptor activation.
Jansen's chondrodysplasia is a rare human disease caused by mutations in the P1R that result in constitutive activity (22). Patients with this disorder have skeletal abnormalities, hypercalcemia, and low serum PTH and PTHrP levels; clinical manifestations that are consistent with the important role that the P1R plays in skeletal development and calcium homeostasis (23). Three different P1R mutations have been identified in these patients: His 223 3 Arg (hP1R CAM-HR ), Thr 410 3 Pro (hP1R CAM-TP ), and Ile 458 3 Arg (hP1R CAM-IR ) (22,24,25), which alter residues in TM2, TM6, and TM7, respectively. Each of these receptor mutants exhibits significantly elevated basal cAMP signaling activity when tested in vitro (5-10-fold above wild-type receptor levels). The P1R receptor containing the His 223 3 Arg mutation was recently expressed in transgenic mice and shown to result in a substantial increase in trabecular bone volume and a decrease in cortical bone thickness (26). Thus, systems are now emerging for the study of P1R constitutive activity and, potentially, ligand inverse agonist activity in model disease states.
In cell-based settings, inverse agonist activity at both hP1R CAM-HR and hP1R CAM-TP has been demonstrated with the previously described antagonist peptides [Leu 11 ,D-Trp 12 ]hPTHrP-(7-34)-amide and [D-Trp 12 ]bPTH-(7-34)-amide (27). Recently, we reported (28) the characterization of several new PTH and PTHrP antagonist analogs that act as inverse agonists at each of the constitutively active P1R mutants. One of these analogs, [Bpa 2 ,Ile 5 ,Trp 23 ,Tyr 36 ]PTHrP-(1-36)-amide (Bpa 2 -PTHrP-(1-36)), contains a photolabile amino acid derivative (p-benzoyl-L-phenylalanine) at position 2 and functions as an inverse agonist at hP1R CAM-HR ; however, it is not an inverse agonist at hP1R CAM-TP or hP1R CAM-IR , and it is a weak partial agonist at the wild-type receptor (28). Importantly, the photolabile position 2 amino acid derivative itself is responsible for the inverse agonist activity of this analog at hP1R CAM-HR , as the otherwise isosteric parent peptide [Ile 5 ,Trp 23 ,Tyr 36 ]PTHrP-(1-36)-amide (valine at position 2) is an agonist at hP1R CAM-HR , as it is at the wild-type hP1R. The structural requirements for inverse agonist activity at position 2 in PTHrP-(1-36) are highly specific, because none of several other position 2 modifications that are structurally similar to Bpa, including tryptophan and the D-Bpa stereoisomer, confer inverse agonism to the peptide ligand (28).
Recent photoaffinity mapping studies performed with the wild-type hP1R have indicated that the benzophenone group of a similar antagonist analog, [Bpa 2 ,Ile 5 ,Arg 11,13 ,Tyr 36 ]PTHrP-(1-36)-amide, cross-links at or near Met 425 , in the extracellular end of TM6 (6). We hypothesized that mapping the cross-link site for Bpa 2 -PTHrP-  in hP1R CAM-HR would help identify receptor residues that play a role in mediating the inverse agonist activity of this analog, as well as residues that play a role in converting the receptor between active and inactive conformations. Thus, we physically mapped the cross-linking site of Bpa 2 -PTHrP-  in hP1R CAM-HR . We also showed by mutational methods that a residue within the mapped receptor interval is involved in mediating the inverse agonist effect of Bpa 2 -PTHrP-   ) were synthesized by the Protein and Peptide Core Facility at Massachusetts General Hospital (Boston) by the solid-phase method on PerkinElmer Life Sciences models 430A and 431A synthesizers. Peptides were purified by reverse-phase high pressure liquid chromatography, and their compositions were confirmed by amino acid analysis and mass spectroscopy.
Na 125 I (specific activity 2000 Ci/mmol) was purchased from PerkinElmer Life Sciences. Dulbecco's modified Eagle's medium, trypsin/EDTA, penicillin G/streptomycin, and horse serum were purchased from Life Technologies, Inc. Fetal bovine serum and Tricine were purchased from Sigma. Trifluoroacetic acid was purchased from Pierce. Cyanogen bromide (CNBr) was purchased from Serva Fine Chemicals/ Boehringer Ingelheim (Heidelberg, Germany). 14 C-Methylated protein molecular mass markers and DEAE-dextran were purchased from Am-ersham Pharmacia Biotech, and Biomax MS film was purchased from Eastman Kodak Co.
Mutagenesis of the hP1R-The mutations were introduced into single-stranded plasmid DNA encoding the wild-type human P1R by oligonucleotide-directed site-specific mutagenesis (29). The oligonucleotide primers were synthesized on an Applied Biosystems model 380A DNA synthesizer. Mutants were verified by automated nucleotide sequence analysis of single-stranded plasmid DNA. Radioligand binding and cAMP activation responses to rPTH-(1-34) for all of the mutant receptors generated for these studies were comparable with those of the two corresponding parent receptors, hP1R or hP1R CAM-HR (Table I and data not shown).
Cell Culture/DNA Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin G, and 100 g/ml streptomycin in a humidified atmosphere containing 95% air and 5% carbon dioxide. Cells were seeded in 24-well plates for radioreceptor and cAMP assays and 6-well plates for cross-linking experiments. Once the monolayer of COS-7 cells reached ϳ90% confluency, cells were transfected by the DEAE-dextran method as described (30) using 200 ng plasmid/well in 24-well plates or 800 ng of plasmid/well in 6-well plates. After 4 days, cells were used for experiments.
Radiolabeling of Peptides-Radiolabeled peptides were prepared by chloramine-T iodination, followed by high pressure liquid chromatography purification using a 30 -50% acetonitrile gradient in 0.1% trifluoroacetic acid over 30 min.
Radioligand-Receptor Binding and cAMP Accumulation Assays-Binding assays were performed as described previously (31). In brief, the 125 I-labeled radioligand bPTH-  was incubated with cells expressing wild-type or mutant P1R in the presence of varying concentrations (0 -10 Ϫ6 M) of unlabeled peptide. After a 4-h incubation at 15°C, the binding mixture was removed, and the cells were lysed, and the entire lysate was counted for ␥-irradiation. Intracellular cAMP accumulation after 30-min treatments at room temperature with or without ligand was measured by radioimmunoassay as described previously (9).
Photoaffinity Labeling of the P1R-Cells transiently expressing wildtype or mutant P1R were incubated with 125 I-labeled Bpa 2 -PTHrP-(1-36) (3 million cpm/well in 6-well plates) for 6 h at 4°C. Cells were rinsed twice and covered with cold buffer. The plates were placed on an ice tray under a UV light source (Blak Ray long-wave lamp, 366 nm, 7 milliwatts/cm 2 ; UV Products, San Gabriel, CA) at a distance of ϳ5 cm for 15 min. Cells were lysed using 1% Triton buffer and centrifuged at 1500 ϫ g for 10 min. The supernatant was then mixed 1:1 with 2ϫ SDS-PAGE sample buffer to give final concentrations of 4% SDS, 80 mM Tris-HCl (pH 6.8), 20% glycerol, 0.2% bromphenol blue, and 100 mM dithiothreitol.
SDS-PAGE Analysis/Purification-The samples in SDS-PAGE sample buffer were incubated at room temperature for 2 h and then subjected to SDS-PAGE analysis (10% acrylamide) performed according to the method of Laemmli (32). For visualization of the intact cross-linking products prior to digestion, the gels were dried and subjected to autoradiography at Ϫ80°C. For purification of cross-linked ligand-receptor complexes, wet gels were cut into strips and counted for ␥-irradiation, and the gel strips with peak counts were subjected to electroelution in a dialysis bag (molecular mass cut-off ϭ 12,000 Da) at 100 V for 2 h. Eluted samples were concentrated using Centricon-10 tubes (Millipore Co., Bedford, MA).
Chemical Cleavage and Size Analysis-For cleavage of the receptor at methionine residues ( Fig. 1), the gel-purified radiolabeled ligandreceptor complexes were incubated with CNBr (100 mM) in 70% formic acid at 20°C for 24 h. After digestion, CNBr and formic acid were removed by repetitive lyophilization. Samples were suspended in SDS-PAGE sample buffer, incubated at room temperature for 2 h, and then analyzed by Tricine/SDS-PAGE (12% acrylamide) performed according to the method of Schä gger and von Jagow (33). Dried gels were subjected to autoradiography at Ϫ80°C. For all studies shown, similar results were obtained in at least three separate experiments.
Data Calculation-Calculations were performed using Microsoft Excel. Nonlinear regression analyses of binding and cAMP dose-response data were performed using the four-parameter equation: y p ϭ Min ϩ ((Max Ϫ Min)/(1 ϩ (IC 50 /x) slope )). The Excel Solver function was utilized for parameter optimization, as described previously (9,34). Surface receptor density (B max ) was calculated by the method of Scatchard assuming a single class of binding site and equivalent binding of radiolabeled and unlabeled peptides. The statistical significance between two data sets was determined using a two-tailed Student's t test, assuming unequal variances for the two sets.

Mapping of the Bpa 2 -PTHrP-(1-36) Cross-link Site-For
both hP1R CAM-HR and wild-type hP1R, cross-linking of 125 I-Bpa 2 -PTHrP-(1-36) followed by SDS-PAGE autoradiography of the resulting COS-7 cell lysates yielded a single large (ϳ79 kDa) protein band that was consistent with specific crosslinking to intact glycosylated PTH-1 receptors ( Fig. 2A). No cross-linking was observed in COS-7 cells transfected with vector alone (Fig. 2A, lane 3). To map the cross-linked sites in the two receptors, we gel-purified the ligand-receptor complexes, treated the resulting complexes with cyanogen bromide (CNBr) in formic acid to cleave at the carboxyl-end of methionine residues (Fig. 1), size-separated the digestion products on tricine gels, and visualized the results by autoradiography. Digestion of the complexes formed between Bpa 2 -PTHrP-(1-36) and either hP1R CAM-HR or hP1R with CNBr yielded similar but not identical banding patterns. With each receptor, a prominent radiolabeled protein band of ϳ3.5 kDa and a slightly weaker band of ϳ5 kDa was observed (Fig. 2B, lanes 2 and 4). In hP1R CAM-HR , but not in hP1R, a third band of ϳ17 kDa was observed (Fig. 2B, lane 2).
The origin of the ϳ3.5-kDa band was assessed but proved to be refractory to identification. Treatment of the purified com-plexes with formic acid alone confirmed that the bulk of the radioactivity in this band arose from CNBr-dependent mechanisms (Fig. 2B, lanes 1 and 3). The mobility of the ϳ3.5-kDa band was indistinguishable from that of free radioligand (see Fig. 4). The release of free ligand by a CNBr-dependent cleavage process is known to occur if the ligand has cross-linked to the side chain methyl group of a methionine residue (35,36). In fact, such CNBr-dependent release of free ligand was observed previously by others (6) for a related Bpa 2 -containing PTHrP analog ([Bpa 2 ,Ile 5 ,Arg 11,13 ,Tyr 36 ]PTHrP-(1-36)-amide) crosslinked to the wild-type hP1R, and it was concluded in this study that cross-linking occurred to Met 425 because mutation of this residue to leucine markedly reduced cross-linking efficiency. In our current study, introduction of the Met 425 3 Leu mutation in hP1R resulted in a moderate reduction in the yield of cross-linking product (Fig. 3A, lanes 3 and 4); however, when introduced into hP1R CAM-HR , this same mutation resulted in a comparable if not greater yield of cross-linking product, relative to that obtained with the control hP1R CAM-HR (Fig. 3A,  lanes 1 and 2). Upon CNBr digestion, both hP1R-M425L and hP1R CAM-HR -M425L complexes yielded a prominent ϳ3.5-kDa band (Fig. 3B). These results imply that within both hP1R and hP1R CAM-HR at least some cross-linking occurred to a residue other than Met 425 that can also yield free ligand upon CNBr digestion. We examined the possibility that Bpa 2 -PTHrP-(1-36) cross-linked to one of the other methionine residues in the juxtamembrane region (e.g. at positions 224, 231, 414, 441, and 445; cf. Fig. 1) by constructing hP1R mutants with these residues mutated individually or in combination. Each of the resulting mutant receptor complexes continued to show a prominent ϳ3.5-kDa band after CNBr digestion (Fig. 4); thus, this band could not be attributed to cross-linking of the ligand to the methyl group of any of the methionine residues tested. Whereas the origins of the ϳ3.5-kDa band remain uncertain, one possibility is that Bpa 2 -PTHrP-(1-36) cross-links to another oxidation-sensitive amino acid (e.g. Trp, Asn, Gln, and Tyr) that could be cleaved by CNBr, although such a mechanism has not previously been appreciated.
We next investigated the ϳ5.0-kDa band, which was observed prominently in the CNBr digestions of both the hP1R⅐Bpa 2 -PTHrP-(1-36) and hP1R CAM-HR ⅐Bpa 2 -PTHrP-(1-36) complexes. The size of this band was consistent with crosslinking of Bpa 2 -PTHrP-(1-36) to either of two predicted CNBrgenerated fragments that are contiguous and within the juxtamembrane region of the receptor, Pro 415 -Met 425 and Ala 426 -Met 441 . Mutation of Met 425 3 Leu in either hP1R CAM-HR or hP1R, however, did not alter the mobility of the ϳ5.0-kDa band (Fig. 3B). A possible explanation for this observation was that CNBr did not cleave at Met 425 , potentially because of steric interference from the cross-linked Bpa 2 -PTHrP- . In this case, the cross-linked receptor fragment would consist of Pro 415 -Met 441 . The ϳ5.0-kDa band was no longer observed after CNBr digest of the complex formed with hP1R-M414V (Fig. 4B, lane 1), suggesting that M414V is indeed the amino-terminus of the cross-linked interval.
The CNBr digestion of the hP1R-M441L mutant receptor did not yield a detectable shift in the ϳ5.0-kDa band (Fig. 4A, lane  4), but this result could be explained by the proximity of the next methionine residue at position 445. Consistent with this explanation, the ϳ5.0-kDa band was not detected in the digest of the complex formed with a triple mutant receptor altered at Met 425 , Met 441 , and Met 445 (Fig. 4A, lane 2). Furthermore, in the CNBr digestions of the complexes formed with single mutants hP1R-M441L and hP1R-M445I (Fig. 4A, lanes 4 and 5) (as well as the triple mutant; Fig. 4A, lane 2), a faint ϳ15-kDa band could be detected; the size of this band correlated with the  (Fig. 5A, lanes 1 and 2). These results thus established the identity of the ϳ5.0 kDa obtained with the wild-type hP1R as Bpa 2 -PTHrP-(1-36) cross-linked to the receptor fragment Pro4 15 -Met 441 .
To determine the origin of the ϳ5.0-kDa band in hP1R CAM-HR, we examined the effect of the M414V single mutation as well as that of the double mutation M441L/M445I in this receptor on the CNBr digestion pattern. As shown in Fig. 5B  (lane 2), the ϳ5.0-kDa band was not observed in the CNBr digestion of the complex formed between Bpa 2 -PTHrP-(1-36) and hP1R CAM-HR -M414V, and the expected new band of ϳ19 kDa was detected just above the ϳ17-kDa band (as discussed below, in the digest of this mutant receptor, the ϳ17-kDa band also showed a slight shift to ϳ18 kDa). The new ϳ19-kDa band correlated with the size predicted for Bpa 2 -PTHrP-(1-36) crosslinked to the receptor fragment Ala 313 -Met 441 (Fig. 1). Consistent with this interpretation, the ϳ5.0-kDa band was no longer observed in the CNBr digestion of the complex formed with hP1R CAM-HR -M441L/M445I (Fig. 5C, lanes 1 and 2). In these digests, the expected new larger band of ϳ15 kDa was not detected; nevertheless, the combined results confirmed that, as with hP1R, the ϳ5-kDa band obtained with hP1R CAM-HR corresponds to Bpa 2 -PTHrP-(1-36) cross-linked to the receptor fragment Pro 415 -Met 441 .
The size of the ϳ17-kDa band observed in the CNBr digests of hP1R CAM-HR was most consistent with Bpa 2 -PTHrP-(1-36) cross-linked to the receptor segment Ala 313 -Met 414 (Fig. 1). The apparent shift of this ϳ17 to ϳ18 kDa seen in the digest of the hP1R CAM-HR -M414V mutant receptor described above (Fig.  5B, lanes 1 and 2) verified this assignment, as it confirmed that Met 414 was a boundary residue for the cross-linked receptor interval. The overall mapping results obtained with hP1R CAM-HR therefore suggest that Bpa 2 -PTHrP-(1-36) can cross-link to either of two sites in two contiguous intervals of the juxtamembrane region of the receptor; one site (site A) is delimited by residues Pro 415 -Met 441 and gives rise to the ϳ5.0-kDa band, and the other (site B) is delimited by residues  1 and 2) or hP1R  (lanes 3 and 4) were treated with 70% formic acid in the absence (lanes 1 and 3) or presence (lanes 2 and 4)  Moreover, mutation of Met 425 to Leu in hP1R CAM-HR resulted in a reduction in the relative amount of cross-linking to site A and an increase in the relative amount of cross-linking to site B (Fig. 3B, lanes 1-2), a result which indicates that the residue at  Table I.  Table I and Fig. 6C). In the hP1R, the M425L mutation increased the agonist efficacy of Bpa 2 -PTHrP-(1-36) (Fig. 6B), again without affecting basal signaling or binding affinity (Fig. 6D and Table I). In contrast, the M414V mutation, analyzed as a control, had little or no effect on the signaling or binding properties of Bpa 2 -PTHrP-  in either hP1R CAM-HR or hP1R (Fig. 6, A-D). The binding and signaling properties of rPTH-(1-34) were not affected by either the M425L or the M414V mutation (Table I).
We also examined whether or not the M425L mutation would affect the signaling activity of a Bpa 2 -containing PTH analog, [Bpa 2 ,Nle 8,18 ,Nal 23 ,Arg 13,26,27 ,Tyr 34 ]bovine PTH-(1-34)-amide (Bpa 2 -PTH- (1-34)), which has been reported to function as a fully efficacious but reduced potency agonist at the wild-type hP1R (36). As with Bpa 2 -PTHrP-(1-36), this analog functioned as an inverse agonist at hP1R CAM-HR , and this inverse agonist activity was abolished by the M425L mutation (Fig. 7, A and B). In contrast to the effects on inverse agonist activity of the Bpa 2containing PTH and PTHrP analogs, the M425L mutation had no effect on the inverse agonist activity of [Leu 11 ,D-Trp 12 ]PTHrP (5-36), which lacks position 2 altogether (Fig. 7, A and B). Thus, the Leu mutation at Met 425 in hP1R CAM-HR specifically alters the inverse agonist activity of amino-terminally intact PTH or PTHrP ligands modified with Bpa at position 2.

DISCUSSION
This study was aimed at mapping the cross-linking site for a ligand determinant of inverse agonism in a constitutively active hPTH-1 receptor. Our data indicate that Bpa 2 -PTHrP-(1-36) can cross-link to more than one site in hP1R CAM-HR ; one of these sites (site A) occurs within an interval that spans the TM6/extracellular loop 3 (ECL3) boundary and is delimited by receptor residues Pro 415 and Met 441 , and the other (site B) occurs within an adjacent interval that contains TM4 and TM5 and is delimited by residues Ala 313 and Met 414 (Fig. 8). Crosslinking of Bpa 2 -PTHrP-  to site A in hP1R CAM-HR resulted in an ϳ5.0-kDa CNBr-generated band, whereas cross-linking to site B resulted in an ϳ17-kDa CNBr-generated band (Fig.  2B, lane 2). Cross-linking to site A was also detected in the wild-type hP1R, but there was little or no evidence for crosslinking to site B in this receptor (Figs. 2B and 3B). With respect    to the wild-type hP1R, cross-linking of our Bpa 2 -containing analog to site A is in agreement with the previously reported study of Behar et al. (6), in which it was concluded that a similar Bpa 2 -containing PTHrP-(1-36) analog cross-linked to an overlapping interval in TM6 and that the side chain methyl group of Met 425 was the principal contact site. Although in our study we could not confirm cross-linking to Met 425 , we did find that CNBr did not cleave at this residue, in either hP1R CAM-HR or hP1R, when Bpa 2 -PTHrP-(1-36) was cross-linked to site A. This result suggests that cross-linking in the site A interval occurred near enough to Met 425 to interfere with the CNBr cleavage reaction.
With both hP1R CAM-HR and hP1R we found abundant CNBrdependent release of free ligand, which has been observed in other studies (35,36) when cross-linking occurred to the side chain methyl group of a methionine residue. In our study, we were unable to verify that cross-linking occurred to the side chain methyl group of any of the methionine residues in the juxtamembrane region of the receptor, as we continued to obtain adequate cross-linking to mutant receptors altered at one or several of these methionines, and we observed abundant CNBr-dependent release of free ligand from each of the mutant receptor complexes (Figs. [3][4][5]. Whether this free ligand originated from a cross-link to a site within the TM6 interval which gives rise to the ϳ5.0-kDa band (6), or to a site within another interval, could not be determined from our current data. In any case, it seems likely that this release of free ligand involves a cleavage mechanism that differs from those that have been described previously for CNBr action at methionine residues.
Within the site A cross-linking interval of hP1R CAM-HR , we identified Met 425 as a functional determinant of the inverse agonist activity of Bpa 2 -PTHrP-(1-36); mutation of this residue to leucine eliminated the inverse agonist activity that Bpa 2 -PTHrP-(1-36) exhibited on the constitutively active receptor and resulted instead in a weak partial agonist response, with-out affecting basal signaling activity. The M425L mutation in hP1R CAM-HR also eliminated the inverse agonist activity of Bpa 2 -PTH-(1-34), but it had no effect on the inverse agonist activity of the amino-terminally truncated analog [Leu 11 ,D-Trp 12 ]PTHrP (5-36) (Fig. 7). These observations are consistent with the hypothesis that Bpa 2 -PTHrP-(1-36) and [Leu 11 ,D-Trp 12 ]PTHrP (5-36) mediate inverse agonism through distinct mechanisms that involve critical contacts to different sites in the receptor (28). In the wild-type hP1R, the M425L mutation increased the agonist efficacy of Bpa 2 -PTHrP-(1-36), suggesting that the methionine residue at this position in the wild-type receptor plays a role in mediating the antagonist action of the ligand, as suggested previously (6). Consistent with this hypothesis, the human PTH-2 receptor has valine at the corresponding TM6 position (37) and elicits a full agonist response to Bpa 2 -PTHrP-(1-36) (9).
The M425L mutation in hP1R CAM-HR resulted in a reduction of Bpa 2 -PTHrP-(1-36) cross-linking to site A and an increase in cross-linking to site B. In the wild-type receptor the M425L mutation also resulted in a small but reproducible increase in the amount of Bpa 2 -PTHrP-(1-36) cross-linking to site B (Figs. 3B and 4A). The rank order of cross-linking to site B (relative to site A) seen in these receptors, hP1R CAM-HR -M425L Ͼ hP1R CAM-HR Ͼ hP1R-M425L Ͼ hP1R (Fig. 3B), correlated with the amount of receptor signaling activity that would be expected for these receptors in the presence of the low concentration (ϳ0.5 nM) of Bpa 2 -PTHrP-(1-36) used in the cross-linking experiments (Fig. 6, A and B, and Table I). This raises the intriguing possibility that the benzophenone group of Bpa 2 -PTHrP-(1-36) contacts site A when the receptor is in an inactive conformation and contacts site B when the receptor is in an active conformation. Further investigations are needed to assess this possibility more directly.
The precise molecular mechanisms by which Bpa 2 -PTHrP-(1-36) mediates inverse agonism with H223R are unknown, but contact of the benzophenone adduct of the ligand to a site in or near the extracellular end of TM6 seems to be involved. In the wild-type P1R, a shift in the position of TM6 away from that of TM3 has been postulated to be a key step in the activation process (38), as suggested for other G-protein-coupled receptors as well (38,39). In our other studies of the P1R, we have identified several point mutations in the TM6/ECL3/TM7 region (e.g. at Thr 427 , Trp 437 , Gln 440 , and Gln 451 , Fig. 1) that specifically affect the functional properties of PTH ligands modified at positions 1 and/or 2 (30,40). Interactions between the Bpa group of Bpa 2 -PTHrP-(1-36) and the extracellular end of TM6 may facilitate or enable the binding of the ligand to an inactive state of hP1R CAM-HR and thus shift the equilibrium of G-protein-coupled and -uncoupled receptors in the direction of the latter (12,13). The lack of inverse agonism exhibited by Bpa 2 -PTHrP-(1-36) with hP1R CAM-TP , which contains a Thr 3 Pro mutation at position 410 in TM6 (Fig. 1), as well as the reduced overall apparent cross-linking efficiency observed with this ligand/receptor pair, as compared with that seen with hP1R CAM-HR (28), provides evidence to suggest that the topological configuration of the heptahelical bundle of hP1R CAM-TP is different from that of hP1R CAM-HR, particularly in regard to TM6. A recent computer simulation analysis of hP1R CAM-HR and hP1R CAM-TP suggests that whereas the two mutations in these receptors produce nearly the same increased solvent accessibility of intracellular loop 3, they do so by inducing different conformational changes and motions in the TM helices (41). Recent experimental data (28,42,43) provide support for the notion that activate-state PTH-1 receptors can assume different conformations, and this tertiary variability may give rise to FIG. 8. Two sites of Bpa 2 -PTHrP-(1-36) cross-linking to hP1R CAM-HR . Shown is a schematic of the amino-terminal portion of Bpa 2 -PTHrP-(1-36) (dark coil) in association with the heptahelical portion of hP1R CAM-HR , as viewed from the extracellular surface of the complex. As described in the text, the photoreactive benzoyl-phenylalanine side chain (dark-/light-shaded circle) at position 2 of the ligand was found to contact two different sites in the receptor: site A (as contacted by the dark-shaded circle) occurred within the interval Pro 415 to Met 441 , which included the TM6/ECL-3 boundary; site B (as contacted by the light-shaded circle) occurred within the interval Ala 313 to Met 414 , which included TM4 and TM5. Contact to site A correlated with the inverse agonist activity of the ligand, as the Met 425 3 Leu mutation at the extracellular end of TM6 abolished inverse agonist activity and caused a relative shift of Bpa 2 -PTHrP-(1-36) cross-linking from site A to site B. The site B contact is shown to occur to TM5 because of previous mutational studies that suggest functional interactions between residues in this helical domain (Ser 370 and Ile 371 ) and residue 2 in the ligand (40). altered ligand selectivity profiles (12), such as that observed with Bpa 2 -PTHrP-(1-36) (28).
In summary, we have identified two receptor sites of contact, one in TM6 and another in the TM4/TM5 region, between a ligand determinant of inverse agonism and a constitutively active PTH-1 receptor using a photoaffinity cross-linking approach. We identified a single residue (Met 425 ) in the TM6 contact region that, when mutated, changes the response induced by the ligand analog from that of inverse agonism to that of partial agonism, and the mutation results in a relative shift in the site of cross-linking from TM6 to the TM4/TM5 region. The results thus provide insights into the processes by which the PTH-1 receptor binds peptide ligands and isomerizes between active and inactive states.