Altered Selectivity of Parathyroid Hormone (PTH) and PTH-Related Protein (PTHrP) for Distinct Conformations of the PTH/PTHrP Receptor

PTH and PTHrP use the same G protein-coupled receptor, the PTH/PTHrP receptor (PTHR), to mediate their distinct biological actions. The extent to which the mechanisms by which the two ligands bind to the PTHR differ is unclear. We examined this question using several pharmacological and biophysical approaches. Kinetic dissociation and equilibrium binding assays revealed that the binding of [ 125 I]PTHrP(1–36) to the PTHR was more sensitive to GTP (cid:1) S (added to functionally uncouple PTHR-G protein complexes) than was the binding of [ 125 I]PTH(1–34) ( (cid:1) 75% maximal inhibition vs. (cid:1) 20%). Fluorescence resonance energy transfer-based kinetic analyses revealed that PTHrP(1–36) bound to the PTHR more slowly and dissociated from it more rapidly than did PTH(1–34). The cAMP signaling response capacity of PTHrP(1–36) in cells decayed more rapidly than did that of PTH(1–34) (t 1/2 (cid:2) (cid:1) 1 vs. (cid:1) 2 h). Divergent residue 5 in the ligand, Ile in PTH and His in PTHrP, was identified as a key determinant of the altered receptor-inter-action responses exhibited by the two peptides. We conclude that whereas PTH and PTHrP bind similarly to the G protein-coupled PTHR conformation (RG), PTH has a greater capacity to bind

(t 1/2 ‫؍‬ ϳ1 vs. ϳ2 h). Divergent residue 5 in the ligand, Ile in PTH and His in PTHrP, was identified as a key determinant of the altered receptor-interaction responses exhibited by the two peptides. We conclude that whereas PTH and PTHrP bind similarly to the G protein-coupled PTHR conformation (RG), PTH has a greater capacity to bind to the G protein-uncoupled conformation (R 0 ) and, hence, can produce cumulatively greater signaling responses (via R 0 3RG isomerization) than can PTHrP. Such conformational selectivity may relate to the distinct modes by which PTH and PTHrP act biologically, endocrine vs. paracrine, and may help explain reported differences in the effects that the ligands have on calcium and bone metabolism when administered to humans. (Molecular Endocrinology 22: 156-166, 2008) P TH AND PTHrP PLAY distinct biological roles yet act through the same G protein-coupled receptor (GPCR), the PTH/PTHrP receptor (PTHR). PTH is a gland-secreted endocrine hormone that regulates calcium and phosphate homeostasis by acting primarily on target cells in bone and kidney. Biosynthetic PTH  increases bone mineral density and bone strength in humans and indeed is now considered to be one of the most effective treatments for osteoporosis (1). PTHrP acts in a paracrine/autocrine fashion to regulate cell proliferation and differentiation programs in developing tissues (2). In addition, PTHrP appears to play a role in regulating bone remodeling in adults (3,4).
PTH and PTHrP are most homologous in their amino-terminal (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14) signaling domains (eight amino acid identities), and show moderate homology in their 14-34 binding domains (three identities). It has generally been inferred that the fully active  portions of PTH and PTHrP interact with the PTHR via largely identical mechanisms (5,6). This mechanism is thought to consist of two principal components: an interaction between the C-terminal, binding domain of the ligand and the amino-terminal extracellular (N) domain of the receptor and an interaction between the amino-terminal, signaling domain of the ligand and the juxtamembrane (J) region of the receptor, which contains the extracellular loops and seven transmembrane helices (7)(8)(9)(10)(11). However, the extent, if any, to which the binding mechanisms used by the two ligands differ remains to be determined.
Interestingly, a recent series of clinical studies has revealed potential differences in the mechanisms by which PTH and PTHrP peptides function in vivo. Thus, PTHrP  was shown to increase bone mineral density to approximately the same extent as did PTH(1-34) but did not stimulate adverse bone-resorptive and hypercalcemic responses to the same extent as did PTH(1-34) (12)(13)(14). That such a difference is not due merely to a difference in pharmacokinetics is suggested by a direct comparison of the two peptides by steady-state infusions methods, which showed that PTHrP  is markedly less efficacious than PTH  for stimulating the renal synthesis of 1,25-(OH) 2 vitamin D 3 (13). Such results in humans could involve different mechanisms of action of the two ligands at the PTHR.
Here we explored more closely the mechanisms by which PTH and PTHrP interact with the PTHR. We focused particularly on the capacity of the two ligands to bind to different PTHR conformations. Of interest were two high-affinity conformations of the PTHR that can be distinguished based on their differing sensitivities to the nonhydrolyzable GTP analog GTP␥S that can be observed in radioligand binding assays (7,15,16). One conformation, termed RG, is sensitive to GTP␥S and is thus presumably coupled to a heterotrimeric G protein; the other, termed R 0 (17), is insensitive to GTP␥S and is thus functionally uncoupled from G protein (16). The following schema describes the working hypothesis regarding the PTH-PTHR binding mechanism and associated changes in receptor conformation that underlies our work: LϩR 7 LR N 7 LR 0 NJ 7 LR NJ G, where L is ligand, R is receptor, LR N is a low-affinity complex involving ligand interactions only to the PTHR N domain, LR 0 NJ is a highaffinity complex involving ligand interactions to both the N and J domains of the receptor, and LR NJ G is a high-affinity complex that is coupled to a heterotrimeric G protein (7,16). One prediction of this schema is that different PTHR ligands could have different capacities to stabilize the R 0 conformation, due to differences in their mode of interaction with the N and/or J domains of the receptor. We further hypothesize that R 0 is an intermediary between the classical R and RG states of the two-state or ternary complex models of GPCR action (18,19) and, as such, is a preactive receptor state that is primed to interact efficiently with a cognate heterotrimeric G protein. A prediction of this hypothesis is that the affinity with which a ligand binds to R 0 will be a determinant, in part, of the overall signaling response capacity of that ligand.
We explored these hypotheses specifically for PTH  and PTHrP  ligands. The results suggest the two ligands indeed differ in their capacities to stabilize R 0 , in that PTHrP  binds more weakly to this conformation than does PTH . We show that this difference in conformational selectivity can result in different biological outcomes in cells, in terms of the cumulative signaling response produced, and is determined strongly by the amino acid divergence at position 5 in the ligands.

RESULTS
We first performed kinetic dissociation experiments to examine the stability of complexes formed between PTH and PTHrP radioligand analogs and the human PTHR stably expressed in membranes prepared from HKRK-B7 cells. For each radioligand, dissociation was examined in the presence and absence of GTP␥S, so as to assess the effects of functionally uncoupling the receptor from heterotrimeric G proteins (Fig. 1). For [ 125 I]PTH(1-34) (Fig. 1A), the dissociation data, both in the absence and presence of GTP␥S, were better fit by a two-phase decay equation than by a single-phase equation. In the absence of GTP␥S (solid symbols), 17% of the complexes were unstable and decayed rapidly (t 1/2 Ͻ 1 min), whereas the remaining 83% were stable and decayed much more slowly (t 1/2 ϳ 4 h).   36 ]PTHrP(1-36)NH 2 (C) were prebound to the human PTHR in membranes prepared from HKRK-B7 cells for 90 min; then dissociation was initiated (t ϭ 0) by the addition of the homologous unlabeled analog (5 ϫ 10 Ϫ7 M), added either alone (F) or with GTP␥S (5 ϫ 10 Ϫ5 M, E). At each time point, aliquots were withdrawn and immediately processed by rapid vacuum filtration to separate bound from free radioactivity. Nonspecific binding was determined in tubes containing the homologous unlabeled ligand (5 ϫ 10 Ϫ7 M) during both the preincubation and dissociation phases. The specifically bound radioactivity (SB) at each time point is expressed as a percentage of the specific binding observed at t ϭ 0. Shown are aggregate data from four (A), five (B), or three (C) experiments. For each tracer radioligand, the respective values (means Ϯ SEM) of total radioactivity (counts per minute) added, total radioactivity bound at t ϭ 0, and nonspecifically bound radioactivity (averaged over the time course of each experiment) were 26 79% remained stable (t 1/2 ϭ ϳ2 h). These findings agree closely with previous dissociation studies performed on this radioligand, and highlight the capacity of PTH  to bind to a high affinity PTHR conformation (R 0 ) that is functionally uncoupled from heterotrimeric G proteins by GTP␥S (15,16). Biphasic kinetics were also observed for [ 125 I]PTHrP ; however, where the complexes were again mostly stable in the absence of GTP␥S (68% decayed with a t 1/2 of ϳ3 h), most became unstable upon addition of GTP␥S (72% decayed with a t 1/2 of ϳ1 min; Fig. 1B The divergent residues at position 5 in PTH and PTHrP (Ile and His, respectively) have been shown to play important roles in PTHR-binding affinity (7,20) and PTHR-subtype selectivity (21,22). We therefore examined the receptor-dissociation properties of [ 125 I]Ile 5 -PTHrP , to see whether the His 5 3Ile substitution altered complex stability. This radioligand dissociated from the receptor slowly and with monophasic kinetics, both in the presence and absence of GTP␥S (t 1/2 Ն 2 h; Fig. 1C). Thus, the His 5 3Ile substitution markedly enhanced the stability with which PTHrP bound to the PTHR, in the G protein-coupled, and especially the G protein-uncoupled state.

Effects of GTP␥S on Equilibrium Binding
We further assessed the effects of GTP␥S on the binding of the above radioligands to the PTHR under approximate equilibrium conditions. The radioligands were thus incubated with cell membranes for 90 min in the absence or presence of varying concentrations of GTP␥S. Figure 2A shows that the binding of We also assessed binding to the rat PTHR using membranes prepared from the rat osteoblastic cell line ROS17/2.8 (endogenous PTHR expression). As seen with the human PTHR ( Fig. 2A), the binding of [ 125 I]Ile 5 -PTHrP  to the rat PTHR was again largely insensitive to GTP␥S (Fig. 2B). The binding of [ 125 I]PTH  to the rat PTHR appeared more sensitive to GTP␥S than was its binding to the human PTHR (Fig. 2 (Fig. 2B). These equilibrium binding data thus further suggest that although there are clearly differences among the different species of receptor, PTH(1-34) and Ile 5 -PTHrP(1-36) bind more strongly to the G protein-uncoupled conformation of the PTHR, R 0 , than do PTHrP  or [Aib 1,3 ,M]PTH (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15) and that the latter two peptides bind preferentially to the G protein-coupled conformation, RG.

Competition Analysis of R 0 and RG Binding Affinity
We then used competition methods to analyze the relative affinities with which PTH and PTHrP ligands bind to the RG and R 0 conformations of the PTHR. To assess binding to RG, we used [ 125 I][Aib 1,3 ,M]PTH(1-15) as a tracer radioligand (binds predominantly to RG, see above) and membranes prepared from COS-7 cells cotransfected with the hPTHR and a negativedominant G␣ S subunit (G␣ S ND), which enriches for RG, vs. R or R 0 , receptor conformations (16,23,24). To assess binding to R 0 , we used [ 125 I]PTH(1-34) as a radioligand (binds predominantly to R 0 ) and membranes prepared from COS-7 cells transfected with the hPTHR alone; we also added GTP␥S (1 ϫ 10 Ϫ5 M) to the binding reactions so as to functionally uncouple receptor-heterotrimeric G protein complexes and thus enrich for the R 0 (and R) conformations vs. RG. Comparison of the apparent affinities with which an unlabeled PTH or PTHrP ligand bound to the PTHR in these two assay formats would thus provide an assessment of the selectivity with which that ligand bound to the RG vs. R 0 conformation. Figure 3A shows that PTH  bound to the R 0 conformation with a 4-fold weaker affinity than it did to the RG conformation (IC 50 ϭ 4.0 vs. 0.91 nM; P ϭ 0.001; see R 0 :RG ratio of Table 1). PTHrP  bound to R 0 with a 66-fold weaker affinity than it did to RG (28 vs. 0.42 nM; P ϭ 0.04; Fig. 3B). PTHrP(1-36) was thus approximately 16-fold more selective for the RG conformation vs. R 0 than was PTH(1-34). Reciprocal exchange of residue 5 in these ligands reversed the pattern of conformational selectivity. Thus, His 5 -PTH(1-34) bound to R 0 with a 670-fold weaker affinity than it did to RG (P ϭ 0.01), whereas Ile 5 -PTHrP  bound to R 0 with only a 3-fold weaker affinity than it did to RG (P ϭ 0.0004; Fig. 3, C and D, and Table 1). The His 5 -PTH(1-34) analog was therefore approximately 220-fold more selective for the RG conformation than was the Ile 5 -PTHrP(1-36) analog.
We also assessed the effects of the Ile 5 3His substitution on the binding affinity of human PTH(1-34)NH 2 and rat PTH(1-34)NH 2 peptides that lacked the methionine 8   bind well to the RG receptor conformation of the PTHR, PTH(1-34) binds with higher affinity to the R 0 conformation than does PTHrP(1-36) and that residue 5 in the ligand plays a key role in modulating this conformational selectivity. We note, however, that residues C-terminal of position 15 in PTH(1-34) are also likely to contribute to the capacity of the ligand to bind strongly to R 0 , because  [Aib 1,3 ,M]PTH(1-15), which contains isoleucine at position 5, bound only weakly to R 0 , while exhibiting strong affinity for RG (supplemental Fig. 1C; Table 1).

Direct Recording of PTHR Activation
The fluorescent resonance energy transfer (FRET) approach has recently been used to assess, in real time and in intact cells, the processes of ligand binding and receptor activation used by the PTHR (8,25). We thus used this approach as an independent means to compare the time courses by which PTH and PTHrP ligands interact with the PTHR. The approach exploits an intramolecular FRET signal that occurs in a PTHR construct, PTHR-CFP IC3 /YFP CT , that contains cyanfluorescent protein (CFP) in the third intracellular loop and yellow-fluorescent protein (YFP) in the C-tail (25). The FRET signal produced by PTHR-CFP IC3 /YFP CT occurs in the basal state and diminishes upon agonist binding, presumably due to conformational changes associated with receptor activation (25).

Duration of cAMP Signaling Capacity
The data so far support the notion that certain PTH and PTHrP ligands can bind to a G protein-uncoupled form of the PTHR, R 0 , with high affinity and thereby form a stable LR 0 complex. We next explored the hypothesis that this LR 0 complex could, over time, isomerize to LRG, and thus become active in terms of  cell signaling. Because this hypothesis predicts that a ligand that can bind stably to R 0 would have the capacity to produce a more prolonged signaling response than would a ligand that binds only weakly to R 0 , we assessed the capacities of the ligands to stimulate cAMP production in PTHR-expressing cells at times after initially binding to the receptor. We thus treated cells with ligand at a relatively high concentration (100 or 300 nM) for 10 min, rinsed the cells thoroughly to remove unbound ligand, incubated the cells in buffer alone for various times, removed the buffer, and replaced it with a buffer containing the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX), and after a 5-min incubation in IBMX, lysed the cells and measured intracellular cAMP. By this delayed cAMP protocol, only the intracellular cAMP generated during the 5-min IBMX phase would accumulate in the cells to measurable levels.
The experiments of Fig. 5A compare the time courses of the cAMP responses produced by PTHrP(1-36) and Ile 5 -PTHrP(1-36) in HKRK-B7 cells. Immediately after the wash-out step (t ϭ 0), cells treated with either ligand produced approximately the same, near-maximal amount of cAMP (ϳ100-fold above basal). Two hours later, the cells treated with Ile 5 -PTHrP(1-36) were still producing a cAMP response, which was 50% of the initial response (Fig.  5A). In contrast, the cells treated with PTHrP(1-36)   , were also observed in ROS17/2.8 cells, which express the PTHR endogenously and at a relatively low level (70,000 per cell vs. 950,000 per cell for HKRK-B7 cells), as well as in HKRK-B64 cells, which stably express the hPTHR, also at a relatively low level (90,000 per cell) (supplemental Fig. 2, A and B). Figure 5B compares the capacity of PTH or PTHrP ligands to produce a sustained cAMP signaling responses in HKRK-B64 cells at the 60-min time point. In these experiments, the cAMP response observed at 60 min after ligand washout is expressed as a percentile of the maximal cAMP response observed for that ligand (determined by treating cells concomitantly with that ligand and IBMX for 10 min and omitting the washout phase), as shown in the figure inset. At 60 min after washout, the cAMP responses generated by PTH(1-34) and Ile 5 -PTHrP(1-36) were 47 and 40% of their corresponding maximal responses, respectively, whereas those of In conventional cAMP dose-response assays performed in HKRK-B64 cells, little or no difference in potency or efficacy was observed for the analogs (supplemental Fig. 3A), whereas in ROS17/2.8 cells, a modest 4.5-fold reduction in cAMP signaling potency was observed for His 5 -PTH(1-34), as compared with PTH(1-34) (supplemental Fig. 3B). No difference in inositol triphosphate signaling potency was observed for the analogs in COS-7 cells transiently transfected to express the hPTHR (supplemental Fig. 3C).

DISCUSSION
The studies presented here suggest intriguing differences in the mechanisms by which PTH and PTHrP interact with the PTH/PTHrP receptor, in that they point to differences in the capacities of the ligands to bind to and stabilize distinct receptor conformations. Our kinetic and equilibrium binding assays performed in cell membranes in the presence of GTP␥S lead us to our main conclusion that whereas PTH  and PTHrP  bind with similar affinities to the G protein-coupled PTHR conformation, RG, PTH  binds with greater affinity to the G protein-uncoupled conformation, R 0 , defined here as a receptor conformation that can bind ligand with high affinity in the presence of GTP␥S (15)(16)(17).
In considering the potential biological implications for such differences in the capacity of PTH or PTHrP ligands to bind to different PTHR conformations, and in particular, R 0 , we hypothesized that R 0 , although not coupled to G protein, is a preactive state that is primed to interact efficiently with a G protein as it encounters one. Thus, the LR 0 complex, upon engaging a G protein, can isomerize to LRG and so become signaling competent. This hypothesis, in turn, predicts that a ligand that can bind stably to R 0 will have the potential to produce a biological signal at longer times after it initially binds to the receptor than will a ligand that binds only weakly to R 0 . We performed the delayed cAMP time-course assays of Fig. 5 and supplemental Fig. 2 as means to test these hypotheses. The results showed that ligands that exhibited relatively high affinities for R 0 in the binding assays, PTH  and Ile 5 -PTHrP , produced greater cAMP responses at times after initial binding than did ligands that bound more weakly to R 0 , PTHrP(1-36) and His 5 -PTH . The results thus support the hypothesis that PTH and PTHrP ligands exhibit different conformational selectivities for the PTHR and, more generally, support the notion that the capacity of a ligand to bind to R 0 can determine, in part, the overall signaling capacity of that ligand in target cells.
The capacity to produce a signaling response at times after initial binding of a ligand to its receptor could be biologically relevant in situations where the cognate G proteins are in low abundance, relative to the receptor, due, for example, to differences in expression levels and/or subcellular compartmentalization. It is also possible that formation of a stable LR 0 complex could enable multiple (catalytic) rounds of G protein activation (26,27) and thereby contribute to signal amplification. In any event, our present data suggest that receptor conformational selectivity can be a factor that contributes to the functionality of a PTH receptor ligand, and this might be a property of the PTHR that can be exploited in ligand-design efforts. In this regard, we have recently developed a PTH analog, [Ala 1,12 ,Aib 3 ,Gln 10 ,Har 11 ,Trp 14 ,Arg 19 ]hPTH(1-28)NH 2 , that binds to the R 0 conformation of the PTHR with considerably higher (ϳ80-fold on the rat PTHR) affinity than does hPTH(1-34)NH 2 and, when injected into mice, produces biological responses (increases in serum calcium and suppression of serum phosphate) that are significantly more prolonged than those produced by PTH(1-34) (28,29). These findings, which are not based on differing serum concentrations of the ligands, based on pharmacokinetic data, strongly suggest that the capacity of a ligand to bind to the R 0 conformation of the PTHR can contribute importantly to the biological response profile of that ligand.
Our acute dose-response signaling assays detected little if any difference in the potencies with which PTH  and PTHrP(1-36) ligands stimulate cAMP or inositol phosphate (IP) accumulation (supplemental Fig. 3; Table 2), results that, by themselves, are consistent with the view that the two ligands interact with the PTHR via largely similar mechanisms. The timedelayed cAMP assays of Fig. 5 (and supplemental Fig.  2) brought out previously unappreciated differences in the signaling properties of the two ligands, evident as differences in the signal output at times after initial binding of ligand to the receptor. Although these findings are consistent with a model involving altered selectivities for different PTHR conformations, we cannot, at present, exclude a possible role for differences in receptor desensitization mechanisms (30)(31)(32)(33). The relationship between receptor conformational selectivity and such internalization/desensitization processes will be an interesting and important matter to explore in future studies.
The capacity of ligand to bind stably to LR 0 might also facilitate coupling to secondary G proteins that presumably have lower affinity for the ligand-receptor complex than does the primary G protein. For the PTHR, this could involve coupling to G proteins of the G␣ q/11 , G␣ i/o , or G␣ 12/13 subclasses, which have been shown to be activated by the PTHR in response to PTH . More studies are needed to assess the relationship between PTHR ligand conformational selectivity and activation of these other G protein signaling pathways. It is also interesting to note that some capacity to form a stable LR 0 complex might be an intrinsic property of the class B GPCRs, most if not all of which use a two-site ligand-binding mechanism involving interactions to both the N and J receptor domains. Thus, several others of these receptors, including the receptors for calcitonin (34), CRH (17), and glucagon (35) have been shown to form a stable complex with their cognate peptide ligand in the presence of a nonhydrolyzable guanine nucleotide analog.
We do not know whether differing capacities to bind to the R 0 state of the receptor explain any of the differences in pharmacological or physiological effects attributed to the two ligands, PTH and PTHrP. It has been of interest to speculate how two physiologically different systems, one paracrine/autocrine (PTHrP) and the other endocrine (PTH) can effectively use one common receptor, widely present in cells, yet produce different effects. PTHrP is produced locally with high regional concentrations likely (36), whereas PTH is secreted into the circulation. Stabilization or formation of distinctive PTHR conformations that differ in terms of duration of signaling could be one factor that underlies, in part, the different modes of action. For example, a shorter duration of signaling might be a useful mechanism for a paracrine factor involved in the timing of cell differentiation programs. PTHrP  peptide, in the limited human studies reported so far, does seem to differ from PTH  in the extent of hypercalcemia induced after a single sc injection (14), stimulation of 1,25-(OH) 2 vitamin D 3 production after iv infusion (13), and stimulation of bone resorption after several months of daily sc administration (12). The differences in the capacities of PTH and PTHrP ligands to bind to the R 0 conformation of the PTHR described herein could potentially explain some of these differing pharmacological properties of the ligands and may be of value to explore therapeutically. Biopolymer Core facility, as described (38). The human PTH(1-34) (free carboxy terminus) used in FRET analyses was purchased from Bachem California (Torrance, CA). Peptide quality was verified by analytical HPLC, matrixassisted laser desorption/ionization mass spectrometry and amino acid analysis, and peptide concentrations of stock solutions were established by amino acid analysis. Radiolabeled peptide variants were prepared by the oxidative chloramine-T procedure using Na 125 I (specific activity, 2200 Ci/mmol; PerkinElmer/NEN Life Science Products, Boston, MA) and were purified by reversed-phase HPLC.

Cell Culture
Cells were cultured at 37 C in a humidified atmosphere containing 5% CO 2 in DMEM supplemented with 10% fetal bovine serum (HyClone, Logan UT), 100 U/ml penicillin G, and 100 g/ml streptomycin sulfate (Invitrogen Corp., Carlsbad, CA). The PTHR-expressing cell lines used were HKRK-B7, HKRK-B64, ROS 17/2.8, and HEK-PTHR-CFP IC3 /YFP CT . The HKRK-B7 and HKRK-B64 lines are derivatives of the porcine kidney cell line LLC-PK 1 and are stably transfected to express the human PTHR at approximate surface densities of 950,000 and 90,000 PTH-binding sites per cell, respectively (39). ROS 17/2.8 cells are rat osteosarcoma cells (40) and express the endogenous rat PTHR at an approximate surface density of 70,000 PTH-binding sites per cell (41). HEK-PTHR-CFP IC3 /YFP CT cells are derived from HEK-293 cells and stably express PTHR-CFP IC3 /YFP CT , a human PTHR construct previously called PTHR-Cam (25), that contains CFP inserted at Gly 395 in the third intracellular loop and YFP inserted into the C-terminal tail. Cells were propagated in T75 flasks and divided into 24-well plates for assays with intact cells, sixwell plates for membrane preparations, or onto glass coverslips for FRET studies. COS-7 cells were transiently transfected using Fugene-6 (Roche Diagnostics, Indianapolis, IN) and CsCl-purified plasmid DNA (3 l Fugene-6 per 1 g DNA). The wells of six-and 24-well plates were transfected with 1 g and 250 ng DNA per well, respectively. Cells were transfected with the PTHR alone or cotransfected with the PTHR and a negative-dominant G␣ S subunit, G␣ S ND. This G␣ S ND subunit binds more effectively, but unproductively, to receptors than does wild-type G␣ S (24) and thus enhances binding of [ 125 I][Aib 1,3 ,M]PTH(1-15)NH 2 radioligand to the PTHR in an RG conformation (see below) (23).

Binding Studies
Binding studies were performed using cell membranes as described (16). Reactions were incubated at room temperature in membrane assay buffer [ Reactions contained a total membrane protein concentration of 20-100 g/ml and a total radioactivity concentration of approximately 150,000 cpm/ml. Unlabeled peptide ligands and/or GTP␥S (Sigma-Aldrich) were added to the reactions as indicated. At the end of the reaction, bound and free radioligand were separated by vacuum filtration using a 96-well vacuum filter plate and vacuum filter apparatus (Multi-Screen system with Durapore HV, 0.65-m filters; Millipore Corp., Milford, MA); the air-dried filters were then detached from the plate and counted for ␥-radioactivity using a ␥-counter.

Radioligand Dissociation
These studies were performed as bulk reactions in 15-ml roundbottom polystyrene snap-cap tubes (Falcon) (total reaction volume ϭ 5.0 ml). Membranes and radioligand were preincubated for 90 min to allow complex formation; the dissociation phase was then initiated by the addition of an excess of the unlabeled analog of the radioligand (5 ϫ 10 Ϫ7 M final concentration), with or without GTP␥S (5 ϫ 10 Ϫ5 M). Immediately before this addition (t ϭ 0), and at successive time points thereafter, 0.2-ml aliquots (ϳ30,000 cpm) were withdrawn and immediately processed by vacuum filtration, as described above. Nonspecific binding was determined in parallel reaction tubes containing the unlabeled analog (5 ϫ 10 Ϫ7 M) in both the preincubation and dissociation phases. The specifically bound radioactivity at each time point was calculated as a percent of the radioactivity specifically bound at t ϭ 0.

Equilibrium Competition Binding and GTP␥S Inhibition
Binding reactions performed with [ 125 I][Aib 1,3 ,M]PTH(1-15) radioligand were assembled and incubated in the wells of the 96-well, Multi-Screen vacuum filtration plates. Membranes, tracer radioligand, and various concentrations of unlabeled ligands and/or GTP␥S were incubated in the wells for 90 min, following which, the reaction plates were processed by rapid vacuum filtration, as described above. Binding reactions performed with [ 125 I]PTH(1-34) radioligand were assembled and incubated in 96-well polystyrene microtiter plates (Falcon, total reaction volume ϭ 230 l) and at the end of the incubation were transferred to the Multi-Screen vacuum filtration plates and processed, as described above. This transfer minimized nonspecific binding of [ 125 I]PTH  to the Multi-Screen filter membranes. For both radioligands, the nonspecific binding was determined in reactions containing a saturating concentration of the unlabeled analog of the radioligand. The specifically bound radioactivity was calculated as a percentage of the radioactivity specifically bound in the absence of a competing ligand or GTP␥S.
To assess the binding of unlabeled peptide ligands to the G protein-uncoupled PTHR conformation (R 0 ), we used membranes prepared from COS-7 cells transiently transfected with the PTHR, [ 125 I]PTH(1-34), as a tracer radioligand and added GTP␥S to the binding reactions (1 ϫ 10 Ϫ5 M final concentration). This binding format is based on the premise that [ 125 I]PTH(1-34) binds predominantly to the R 0 conformation of the PTHR and that this conformation is enriched in the membranes, relative to RG, by the presence of GTP␥S (15,16). A similar approach, using radiolabeled peptide agonist and GTP␥S, was used by Hoare and colleagues (17) to assess binding to the R 0 conformation of the corticotropinreleasing factor receptor-1. To assess binding to the G protein-coupled conformation (RG), we used membranes prepared from cells cotransfected with the PTHR and a negative-dominant G␣ S subunit (G␣ S ND), and we used [ 125 I][Aib 1,3 ,M]PTH(1-15) as a tracer radioligand. This binding format is based on the premise that [ 125 I][Aib 1,3 ,M]PTH(1-15) binds predominantly to the RG conformation of the PTHR and that this conformation is enriched in the membranes, relative to R or R 0 , by the presence of G␣ S ND (7,23,24). We note that binding of a ligand to any low-affinity PTHR conformation (R) will not be detectable in these assays, given the low concentrations (ϳ25 pM) of tracer radioligands used.

FRET
FRET analyses using HEK-PTHR-CFP IC3 /YFP CT cells were performed as described (25). With PTHR-CFP IC3 /YFP CT , excitation of the CFP donor with UV light ( max.ex. ϭ 436 nm; max.em. ϭ 480 nm) produces an intramolecular FRET to the YFP acceptor, resulting in emission from that YFP ( max.ex. ϭ 480 nm, max.em. ϭ 535 nm). This FRET response is observable as a decrease in intensity of CFP light emission at 480 nm and an increase in intensity of YFP light emission at 535 nm. The FRET signal occurs in the ground-state receptor and decreases upon binding of a PTH agonist ligand (25). PTH ligands were applied to and washed from the cells using a computer-assisted, solenoid valve-controlled, rapid superfusion device (ALA Scientific Instruments, Westbury, NY). Solution-exchange times were 5-10 msec. Fluorescence was monitored using a Zeiss inverted microscope equipped with a ϫ100 objective and a dual-emission photometric system (Till Photonics, Planegg, Germany), coupled to an avalanche photodiode detection system and an analog-digital converter (Axon Instruments, Foster City, CA). The FRET signal detected upon excitation at 436 nm was calculated as a normalized FRET ratio: F YFP(535nm) /F CFP(480nm) ,where F YFP(535nm) is the emission at 535 nm, corrected for spillover of the CFP signal into the YFP channel, and F CFP(480nm) is the emission at 480 nm, corrected for spillover (minimal) of the YFP emission into the CFP channel. Changes in fluorescence emissions due to photobleaching were subtracted.

Stimulation of Intracellular cAMP and Inositol Phosphate
Intracellular cAMP levels were measured by RIA (38). The capacity of a ligand to produce a delayed cAMP response in cells was assessed as follows (33,42). The cells in 24-well plates were rinsed in binding buffer [50 mM Tris-HCl (pH 7.7), 100 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 5% heat-inactivated horse serum, 0.5% heat-inactivated fetal bovine serum] and then incubated in binding buffer with or without a peptide ligand (1 ϫ 10 Ϫ7 or 3 ϫ 10 Ϫ7 M) for 10 min at room temperature. The cells were then washed with three changes of binding buffer and incubated further in binding buffer for varying times (1-120 min). Then, the buffer was replaced by binding buffer containing IBMX (2 mM), and after an additional 5-min incubation, the intracellular cAMP was quantified. By this approach, only the cAMP produced during the final IBMX stage of the incubation is measurable, because that produced before IBMX addition is degraded by cellular phosphodiesterases.
The stimulation of intracellular IPs was measured in transiently transfected COS-7 cells that were prelabeled (16 h) with [ 3 H]myo-D-inositol (2 Ci/ml) (38). Cells were treated with ligand in DMEM containing fetal bovine serum (10%) and LiCl (30 mM) for 30 min. The cells were then lysed with ice-cold, trichloroacetic acid (5%), and IPs were extracted from the acid-lysates by ion-exchange filtration.

Data Calculations
Data were processed using Microsoft Excel and GraphPad Prism 4.0 software packages. Dissociation time-course data were analyzed using a biexponential decay equation, except when an F test analysis indicated a monoexponential equation provided a better fit (P␣ Ͼ 0.02). Data from equilibrium binding, cAMP, and IP dose-response assays were analyzed using a sigmoid dose-response equation with variable slope, which yielded values of EC 50 , IC 50 (concentration of ligand producing the half-maximal effect), and E max (maximal cAMP or IP response). Paired data sets were statistically compared using the Student's t test (two tailed) assuming unequal variances for the two sets.