[Aib^1,3,M]PTH(1-14) (peptide sequence Aib-V-Aib-E-I-Q-L-M-H-Q-Har-A-K-W-NH₂) was kindly provided by T. Gardella (Massachusetts General Hospital and Harvard Medical School, Boston); [Aib^1,3,M] refers to the combined modifications of Aib^1,3, Gln-10, Har¹¹, Ala-12, and Trp-14 on human PTH(1-14), where Aib is the α-amino-isobutyric acid and Har is homoarginine (8).

Molecular Biology and Cell Culture. Construction of the N-terminally GFP-tagged PTHR (^GFPN-PTHR) was performed following a PCR strategy using pfu DNA polymerase (Stratagene). cDNA encoding the N-terminal fragment of human PTHR (corresponding to amino acids 1-60) was amplified with primers containing restriction sites for EcoRI (5′) and BamHI (3′). EGFP's cDNA from Clontech was amplified without stop codon with primers containing restriction sites for BamHI (5′) and KpnI (3′). Both PCR products were ligated together after restriction enzymatic digestion with BamHI and subcloned into a previously described pCMV vector encoding the hemagglutinin (HA)-tagged human PTHR cDNA (9) using the restriction sites EcoRI and KpnI. In the resultant construct, ^GFPN-PTHR, exon 2, and the HA tag at the N terminus of the PTHR (residues 61-101) were replaced with the EGFP sequence preceded by the linker Arg-Leu-Ile-Ser-Gly-Ser. Note that deletion of exon 2 in PTHR does not alter the receptor pharmacology (10). The construct verified by restriction enzyme mapping and sequencing was subcloned into pcDNA3 (Invitrogen) for transient and stable expression in human embryonic kidney (HEK)293 cells. N-terminally GFP-tagged β₂-adrenergic receptor construct (^GFPN-β₂AR) in pcDNA3 was kindly provided by U. Zabel (University of Würzburg) and stably transfected into HEK293 cells for control experiments.

HEK293 cells were cultivated and transfected as described (11). After selection with 400 μg/ml G418 (Merck), stable cell clones showing membrane expression of the GFP-tagged receptors were selected. The HEK293 cell line stably expressing the WT PTHR was kindly provided by E. Blind (University of Würzburg) (12).

Confocal Microscopy. HEK293 cells stably expressing ^GFPN-PTHR or WT PTHR were plated onto poly-d-lysine-coated glass coverslips. Cells were incubated with 1 μM PTH(1-34) or 1 μM PTH(1-34)^TMR for different times, washed with PBS, and fixed for 10 min in 4% paraformaldehyde. Fixed cells were observed with an oil immersion ×63 Plan-Neofluar objective in a Leica TCS laser scanning microscope.

Pharmacology. Ligand binding and measurement of cAMP were performed as described (11) with minimal modifications. For binding studies, stably transfected cells grown in 24-well plates were incubated in DHB buffer (serum-free DMEM containing 20 mM Hepes and 1% BSA) for 1 h at 0°C, followed by a 1.5-h incubation with the same buffer containing ¹²⁵I-PTH(1-34) (100,000 cpm per well) as a radioligand with or without varying concentrations of unlabeled peptides. Cells were washed three times with iced PBS and extracted with 0.8 M NaOH, and cell-associated ¹²⁵I-PTH(1-34) was counted.

For cAMP assays, cells grown in 12-well plates were washed twice with Hepes buffer (137 mM NaCl/5 mM KCl/1 mM CaCl₂/1 mM MgCl₂/20 mM Hepes, pH 7.4) and incubated with isobutylmethylxanthine (0.5 mM) in the same buffer for 5 min at 37°C. Cells were stimulated with varying concentrations of PTH(1-34) or PTH(1-34)^TMR at 37°C for 15 additional min. Cellular cAMP was extracted and measured by RIA (Immunotech, Luminy, France; Beckman Coulter). Competition binding studies and concentration-response data were analyzed with the program prism 4.0 (GraphPad, San Diego).

Microscopic FRET Measurements. FRET experiments were performed as described (7) with some modifications. In brief, cells plated on poly-d-lysine-coated glass coverslips maintained in Hepes buffer containing 0.1% (wt/vol) BSA were placed at room temperature on a Zeiss inverted microscope (Axiovert135) equipped with an oil immersion ×100 Plan-Neofluar objective and a dual emission photometric system (TILL Photonics, Planegg, Germany). Cells were excited with light from a polychrome IV (TILL Photonics). The illumination time was typically set to 25 ms applied with a frequency of 40 Hz. FRET was monitored as the decrease emission of GFP (F_GFP) from ^GFPN-PTHR in the presence of PTH(1-34)^TMR, recorded at 520 ± 20 nm (beam splitter 530 DCLP) upon excitation at 480 ± 20 nm (beam splitter 498 DCLP). Note that during the recording of binding events in real time, PTH(1-34)^TMR was continuously superfused, and the presence of TMR fluorescence in the FRET channel (emission at 580 nm) caused by the direct excitation of TMR [i.e., bound and unbound PTH(1-34)^TMR] at 480 nm does not permit the differentiation of the TMR emission caused by FRET or direct excitation. Because the diminution of GFP fluorescence was caused by FRET between GFP and TMR (see below) we used the decline of GFP fluorescence as readout for the binding event. The spillover of TMR into the 520-nm channel (2%) was negligible at PTH(1-34)^TMR concentrations <3 μM and subtracted from the signal of the 520-nm channel at concentrations of 3 μM or higher to give a corrected F_GFP. FRET between ^GFPN-PTHR and PTH(1-34)^TMR was also determined by donor (GFP) recovery after acceptor (TMR) photobleaching after three consecutive cycles of 6 min of continuous illumination at 550 nm (orange glass 530 nm, beam splitter 555 DCLP, emission filter HQ605/75).

Recording of Ligand-Induced Changes in FRET. A single cell expressing ^GFPN-PTHR was continuously superfused with Hepes/BSA buffer and PTH(1-34)^TMR at different concentrations with a computer-assisted solenoid valve rapid superfusion device (ALA-VM8, ALA Scientific Instruments, Westbury, NY, solution exchange 5-10 ms, ref. 7). Signals detected by avalanche photodiodes were digitalized with an AD converter (Digidata1322A, Axon Instruments, Union City, CA) and stored on a personal computer by using clampex 9.0 software (Axon Instruments). The change in the fluorescence emission of GFP as a function of time caused by FRET was analyzed by using nonlinear regression to one- and two-exponential models, where A₁ and A₂ are the amplitudes of the two phases, A₀ is the baseline at t = ∞, k₁ and k₂ are the rate constants (s^-1) for the two phases, and t is the time. The best fits were judged by the analysis of residuals (i.e., differences between the experimental data and the calculated curve fits). Changes in emission caused by photobleaching were systematically subtracted. Data were analyzed with origin 6.1 software (Microcal, Amherst, MA).

First, confocal microscopy showed that ^GFPN-PTHR was well expressed at the plasma membrane and underwent time-dependent internalization upon incubation with PTH(1-34) (Fig. 1B Top). PTH(1-34)^TMR specifically decorated the cell surface of cells expressing PTHR and internalized with kinetics similar to that observed for ^GFPN-PTHR in response to PTH(1-34) (Fig. 1B Middle and Bottom).

Second, binding competition experiments showed that the affinity of PTH(1-34)^TMR (K_i = 47.5 ± 1.1 nM) was close to that of PTH(1-34) for ^GFPN-PTHR (K_i = 22.3 ± 1.1 nM) (Fig. 2A and Table 1, which is published as supporting information on the PNAS web site) and approximately 1 order of magnitude lower than that of PTH(1-34) for the WT PTHR (K_i = 2.4 ± 0.1 nM) (7). The presence of GFP in the N-terminal domain of the receptor is presumably responsible for the loss in binding affinity.

Third, the ability of PTH(1-34)^TMR to stimulate maximal cAMP responses in cells expressing the WT PTHR or ^GFPN-PTHR was indistinguishable from that of PTH(1-34) (Fig. 2 B and C and Table 1), although hormone potencies were reduced for cells expressing ^GFPN-PTHR (half-maximally effective concentrations, EC₅₀ ranged between 0.10 and 0.20 nM for PTHR and between 4 and 6 nM for ^GFPN-PTHR) (compare Fig. 2 B and C and Table 1). Taken together, these data indicate the ^GFPN-PTHR has, in comparison to the WT PTHR, slight altered binding and activation properties and thus is well suited for studying the association of hormone and receptor.

High Resolution of Binding Kinetics. Kinetics of binding of PTH(1-34)^TMR to ^GFPN-PTHR were recorded under a fluorescence microscope at the single-cell level. Fig. 1A illustrates the principle of the experiment. After excitation at 480 nm the maximal fluorescence of GFP was recorded at 520 nm and rapid (<10 ms) addition of PTH(1-34)^TMR (1 μM) led to a strong and fast decrease in GFP emission (Fig. 3A Left): the latter decreased by 15% with a half-time t_1/2 = 700 ± 20 ms (n = 14). Photo-destruction (i.e., photobleaching) of TMR verified that the decrease in GFP fluorescence was caused by FRET: when GFP (the donor) and TMR (the acceptor) exhibited FRET, then photobleaching of TMR by direct excitation at 550 nm should increase GFP emission at 515 nm. Fig. 3B Left shows that this was exactly the case: after formation of the complex ^GFPN-PTHR/PTH(1-34)^TMR photobleaching of TMR led to recovery of the GFP emission. Control experiments where TMR was not photobleached demonstrated that the fluorescence emission of GFP and TMR moieties was stable over the time frame of the experiments (Fig. 3B Right).

In the following experiments we proved that the decrease in GFP emission corresponds to a specific interaction between PTH(1-34)^TMR and ^GFPN-PTHR. First, we investigated the effect of PTH(1-34)^TMR on ^GFPN-β₂AR, which was also well expressed at the cell surface and functionally active but that does not bind PTH(1-34) nor PTH(1-34)^TMR (data not shown). As shown in Fig. 3A Center, the addition of PTH(1-34)^TMR (1 μM) had no effect on the fluorescence intensity of GFP because of the absence of FRET between PTH(1-34)^TMR and ^GFPN-β₂AR. The second experiment showed that sequential addition of PTH(1-34) (1 μM) followed by PTH(1-34)^TMR (1 μM) to ^GFPN-PTHR almost completely prevented the FRET signal, indicating specificity of the response and that little, if any, agonist dissociation occurs during the experiments (Fig. 3A Right). These data demonstrated that the change in GFP emission caused by FRET can be used to specifically record the association between ^GFPN-PTHR and PTH(1-34)^TMR.

Two-Step Binding Mechanism to Switch On PTHR. Recording the temporal resolution of the FRET signal from a single cell with increasing concentrations of PTH(1-34)^TMR allows the kinetic analysis of the interaction of hormone and receptor (Fig. 4A). Binding responses were well fitted to a two-exponential model (Fig. 4B). The biphasic nature of the signal suggests that the association of PTH(1-34) to PTHR involves two binding events. At a maximal concentration of PTH(1-34)^TMR (15 μM) the signal rapidly reached equilibrium (t_1/2 ≈ 120 ms), and the binding responses followed the sum of two exponential components with time constants τ₁ = 140 ± 8 ms (n = 14) for the fast step and τ₂ = 1.2 ± 0.1 s (n = 14) for the slower step.

Analysis of the first fast phase showed that the observed rate constant linearly depends on hormone concentration (Fig. 5A). The strict linearity of the rate constant implies that the first step reflects a simple bimolecular interaction between the hormone (H) and its receptor (R) described by the following reaction, where k₁ and k_-1 represent the binding and dissociation rate constants, respectively, and the time course of the interaction is defined by under pseudofirst-order conditions (14). The slope of versus the concentration of hormone gave an association rate constant k₁ = (0.7 ± 0.1) × 10⁶ M^-1·s^-1, which is too low to be under diffusion control (10⁸ to 10⁹ M^-1·s^-1) but was within the range for interactions between protein—protein or peptide-receptor (10⁵ to 10⁶ M^-1·s^-1) (15, 16). The intercept with the ordinate axis gave a dissociation rate constant of k_-1 = 0.04 ± 0.01 s^-1, reflecting the slow dissociation of the bound hormone. The ratio (k_-1/k₁) yielded a dissociation constant (K_D) of 50 nM, which is consistent with the value obtained from radioligand binding analysis (K_i = 47.5 nM).

In contrast, the rate constant of the slow phase followed a hyperbolic dependence on hormone concentrations and reached a constant value at saturating concentrations (Fig. 5). This behavior is consistent with the following two-step reaction in which the fast binding step is followed by a slower conformational change (denoted by *) and where the time course is defined by (17). At low hormone concentrations, the rate constants increased linearly with hormone concentrations (Fig. 5B), indicating that the hormone binding to PTHR was the rate-limiting step as observed for the first binding step. But at high hormone concentrations the rate constants approached a maximal value, suggesting that a step other than the association of hormone receptor became rate-limiting. The time constant required for this step was ≈1 s (Fig. 5). This step is consistent with a conformational change of the bound agonist. Presumably the bound PTH(1-34) is reoriented to reach its final conformation.

The kinetic profile observed for the second binding step coincides with that recently reported for the activation switch of the PTHR (Fig. 5 and ref. 7), also recorded in a single cell with a similar FRET approach using the receptor sensor PTHR^CFP/YFP (Fig. 7). At high concentrations of hormone, rate constants governing the conformational change in the second binding step and in receptor activation were in the same range (n = 14) for binding and k_obs = 0.86-1.03 s^-1 (n = 10) for receptor activation (7). These kinetic similarities between the second association step and receptor activation suggest that these two events might be coupled.

To explore the possibility that the slow step of binding related to receptor activation we used [Aib^1,3, M]PTH(1-14) (see Materials and Methods), a conformationally constrained PTH(1-14)-derived agonist with increased α-helicity and similar binding and signaling properties as PTH(1-34) (8, 18). Extensive studies have demonstrated that [Aib^1,3, M]PTH(1-14) and other PTH(1-14) analogs (8, 19-21) mainly, if not exclusively, bind to the PTHR J-domain in contrast to PTH(1-34), which binds to both receptor's N- and J-domains (22). When [Aib^1,3, M]PTH(1-14) is bound, it prevents PTH analogs from occupying their binding site at the J-domain of the receptor (8, 18). In the next experiment, we examined whether the specific association of [Aib^1,3, M]PTH(1-14) to the receptor J-domain is able to differentiate the two kinetically distinct phases of PTH(1-34)^TMR binding to ^GFPN-PTHR. As we observed in Fig. 4B, PTH(1-34)^TMR (3 μM) association was described by the sum of two exponential phases with association rate constants (k_obs) of 2.30 ± 0.14 s^-1 and 0.38 ± 0.04 s^-1 (n = 12) for the first and second phases, respectively. The corresponding amplitudes (A₁ and A₂)for these latter phases represented 7 ± 0.6% and 5.5 ± 0.4% of the decline in FRET. However, Fig. 4C shows that simultaneous addition of PTH(1-34)^TMR (3 μM) and [Aib^1,3, M]PTH(1-14) (10 μM) eliminated the second phase, whereas the first phase remained mostly unchanged (k_obs = 2.78 ± 0.15 s^-1 and a corresponding amplitude of A = 6.5 ± 0.7%, n = 6). The subsequent removal of [Aib^1,3, M]PTH(1-14) restored the full response seen with PTH(1-34)^TMR alone (Fig. 4D). These observations are consistent with a model where the interaction of [Aib^1,3, M]PTH(1-14) to the receptor J-domain prevents the second PTH(1-34)^TMR-binding event to take place. It also implies that residues in the 15-34 portion of PTH interact with the receptor N-domain and dictate the initial binding step. Taken together, these data support a model where the slower second step in agonist binding is consistent with a conformational change of the bound PTH(1-34)^TMR accompanying the receptor switch.