Kinetic and Thermodynamic Insights on Agonist Interactions with the Parathyroid Hormone Receptor-1 from a New NanoBRET assay

Abstract

Polypeptides that activate the parathyroid hormone receptor-1 (PTHR1) are important in human physiology and medicine. Most previous studies of peptide binding to this receptor have involved displacement of a radiolabeled ligand. We report a new assay format based on bioluminescence resonance energy transfer (BRET). Fusion of a NanoLuc luciferase (nLuc) unit to the N-terminus of the PTHR1 allows direct detection of binding by an agonist peptide bearing a tetramethylrhodamine (TMR) unit. Affinity measurements from the BRET assay align well with results previously obtained via radioligand displacement. The BRET assay offers substantial operational benefits relative to affinity measurements involving radioactive compounds. The convenience of the new assay allowed us to explore several questions raised by earlier reports. For example, we show that although the first two residues of PTH(1-34) (the drug teriparatide) are critical for PTHR1 activation, these two residues contribute little or nothing to affinity. Comparisons among the well-studied agonists PTH(1-34), PTHrP(1-34) and “long-acting PTH” (LA-PTH) reveal that the high affinity of LA-PTH arises largely from a diminished rate constant for dissociation relative to the other two. A D-peptide recently reported to be comparable to PTH(1-34) as an agonist of the PTHR1 was found not to bind detectably to the receptor and to be a very weak agonist.

INTRODUCTION

The parathyroid hormone receptor-1 (PTHR1) is a family B G protein-coupled receptor (GPCR) that is targeted by several polypeptide drugs.^1–3 Activation of the PTHR1 helps to regulate blood concentrations of calcium and phosphate ions, and this receptor plays important roles in bone physiology. The PTHR1 has two natural agonists, parathyroid hormone (PTH) and parathyroid hormone-related protein (PTHrP). PTH(1-34), a synthetic N-terminal fragment of PTH that matches the activity of the full-length hormone, is used to treat osteoporosis.⁴ Abaloparatide (ABL), a synthetic analogue of PTHrP(1-34), is used clinically for the same purpose.⁵ Full-length recombinant PTH is used to treat hypoparathyroidism.⁶ Understanding how agonists bind to and activate the PTHR1 represents a fundamental challenge in terms of biological signal transduction; such understanding could provide a basis for future therapeutic development. Here we describe a new type of assay for characterizing agonist-PTHR1 interactions, and we show how this new tool can be used to address fundamental questions regarding activation of this receptor.

As with other GPCRs, agonist binding to the PTHR1 alters the conformational equilibrium of the receptor.^7,8 This structural change is registered by cytoplasmic proteins that engage the inward-facing surface of the receptor, including G proteins, β-arrestins and GPCR kinases (GRKs). Family B GPCRs feature a large extracellular domain (ECD) that engages the C-terminal portion of the long polypeptide agonist.^8–12 The N-terminal region of the agonist binds into the core of the family B GPCR transmembrane domain (TMD), which contains the seven-helix bundle tertiary structure that is characteristic of all GPCRs.^13,14

Strategies to monitor GPCR activation are valuable for characterizing relationships between agonist structure and function. Many cell-based assay formats for GPCR activation feature a convenient optical readout. Agonists of the PTHR1 and other receptors that activate G_S, for example, are often characterized by their ability to stimulate production of the intracellular second messenger cAMP, which can be detected via luminescence generated by cAMP-binding proteins such as GloSensor.¹⁵ Direct recruitment of a partner protein (G protein, β-arrestin, GRK) to the intracellular surface of a GPCR can be detected via bioluminescence resonance energy transfer (BRET) if the receptor and partner are fused to complementary BRET partners, such as a luciferase donor and a fluorescent protein acceptor.^16–18 Other optical readouts, such as fluorescence resonance energy transfer (FRET), have been harnessed for this purpose as well.^19,20

Relative affinities among agonists for a given receptor need not correlate with relative agonist potencies; therefore, it is valuable to measure agonist binding directly, as a complement to activity measurements. Agonist binding to cognate GPCRs is commonly characterized with radioactive ligands, either through direct binding or competition with a radiolabeled tracer.^21–27 Handling radioactive materials is cumbersome, however, which has prompted a search for more convenient binding assays. Agonist binding to a few GPCRs has recently been characterized via BRET by using a receptor construct bearing a NanoLuc luciferase (nLuc) domain at the N-terminus in conjunction with ligands covalently attached to a complementary fluorophore, such as tetramethylrhodamine (TMR).^28,29 This approach does not require radiolabeled compounds and involves less sample manipulation relative to radioligand displacement assays. In addition, the BRET assay format allows real-time measurements. Here we describe application this strategy to the PTHR1 (Figure 1A).

Figure 1.

(A) Cartoon depiction of the NanoBRET binding assay. Nanoluciferase (nLuc) is attached to the extracellular side of the PTHR1, which allows energy transfer to a BRET acceptor that is covalently linked to the bound peptide. (B) Stimulation of intracellular cAMP production by nLuc-PTHR1-expressing cells activated by either PTH(1-34) or PTH(1-34)K35^TMR. The luminescence response of vehicle control is normalized to 0%, while maximal response of unlabeled PTH(1-34) is normalized to 100%. Each datapoint represents mean ± s.e.m from n ≥ 3 independent experiments. (C) Sequences of PTHR1 agonists. All peptides used in this work were prepared as C-terminal amides. (D) Representative BRET data for PTH(1-34)K35^TMR binding to the nLuc-PTHR1 at different peptide concentrations measured over 5 min. Each trace shows the result of a single measurement. Each value shown in Table 1 was derived from n ≥ 3 independent experiments.

RESULTS AND DISCUSSION

An nLuc-PTHR1 fusion protein enables measurement of binding kinetics and affinities, providing results that are consistent with previous reports involving other assay formats.

DNA coding for the human PTHR1 with nLuc fused to the N-terminus was transiently expressed in GS22 cells. These HEK293-derived cells stably express the GloSensor F-22 construct (Promega), which allows detection of intracellular production of cAMP. PTH(1-34) caused potent stimulation of cAMP production (EC₅₀ = 0.1 nM) in these cells (Figure 1B; sequence in Figure 1C), and this activity was comparable to the activity observed for PTH(1-34) with cells transiently expressing the native human PTHR1 (EC₅₀ = 0.3 nM). This agonist peptide and all others discussed below were synthesized as C-terminal amides. This comparison indicates that the N-terminal nLuc domain does not interfere with receptor activation, at least as measured by cAMP production.

To monitor agonist-receptor association, we treated cells expressing the nLuc-PTHR1 with NaN₃ for 30 min before introducing TMR-labeled agonist and luciferase substrate (h-coelenterazine). Azide pretreatment was used to halt cellular metabolism and prevent the internalization of receptor-agonist complexes, which normally occurs shortly after agonist engagement. The lack of internalization should ensure that ligand-receptor interaction occurs solely on the cell surface, and that the ligand concentration and the total receptor density remain largely unaffected for the duration of the experiment. Ligand-receptor binding led to a rapid rise in the BRET signal, which reached a plateau within five minutes (Figure 1D). The plateau level varied with agonist concentration. These data were fitted to a pseudo-first order binding model (see equations 1 and 2 in the Methods section; peptide in vast excess) to generate rate constants k_on and k_off. The dissociation constant, K_D, was calculated from these rate constants, k_off/k_on (Table 1).

Table 1.

Kinetic constants and dissociation constants of various PTHR1 binding ligands as measured by either direct binding or competition binding.

Direct Binding	kon (× 105 s−1M−1)	koff (s−1)	KD (nM)
PTH(1-34)-35KTMR	4.4 ± 0.8	0.0051 ± 0.0006	12 ± 2
PTH(1-34)-N33KTMR	4.0 ± 0.6	0.0043 ± 0.002	11 ± 4
PTHrP(1-34)-35KTMR	2.1 ± 0.6	0.010 ± 0.003	48 ± 20
PTHrP(1-34)-13KTMR	2.4 ± 0.7	0.011 ± 0.001	53 ± 20
LA-PTH(1-34)-35KTMR	2.3 ± 0.4	0.00056 ± 0.0002	2.4 ± 0.8
M-PTH(1-14)-15KTMR	0.77 ± 0.2	0.024 ± 0.001	310 ± 80
Competition
PTH(1-34)	2.5 ± 1	0.0038 ± 0.001	16 ± 2
PTHrP(1-34)	2.4 ± 0.6	0.024 ± 0.004	73 ± 40
nLuc-ΔECD-PTHR1
LA-PTH(1-34)-35KTMR	1.3 ± 0.1	0.0097 ± 0.003	72 ± 20
M-PTH(1-14)-15KTMR	1.8 ± 0.5	0.018 ± 0.003	106 ± 20

The kinetics and affinity parameters of various ligands measured by different experimental approaches. Direct binding Measurements on nLuc-PTHR1 were obtained by treating HEK293 cells stably expressing nLuc-PTHR1 with TMR-labeled ligands in the presence of the substrate, coelenterazine-h, and monitoring the BRET response. Competition Monitored BRET response with varying concentrations of the unlabeled analyte with fixed concentration of the tracer, PTH(1-34)K35^TMR. nLuc-ΔECD-PTHR1 Binding of LA-PTH(1-34)K35^TMR measured with fitting curves into one-phase association model. Binding of M-PTH(1-14)K15^TMR measured by fitting the curves into one-phase association then dissociation model. Each value is derived as mean ± SD from n ≥ 3 independent experiments.

We initially placed the TMR label at the C-terminus of PTH(1-34) or PTHrP(1-34) by appending a Lys residue at position 35 and linking the fluorophore to the Lys side chain. Varying the position of the (TMR)-Lys residue, to position 33 in PTH(1-34) (Figure S1) or to position 13 in PTHrP(1-34) (Figure S2, S3), did not cause a significant change in the kinetic parameters (Table 1). These observations suggest that our strategy characterizes contacts between the peptide agonist and the PTHR1 with minimal influence from the fluorophore attached to the peptide. The data indicate that PTH(1-34) has modestly higher affinity (~4-fold) for the PTHR1 relative to PTHrP(1-34).

The relative affinities of human PTH(1-34) and PTHrP(1-34) for the human PTHR1 emerging from our studies are consistent with earlier measurements involving radiotracer displacement. At least three studies have made this comparison based on binding to receptors on whole cells. Gardella et al.²¹ compared human PTH(1-34) and PTHrP(1-34) binding to the rat PTHR1 expressed in two cell types; for both ROS 17/2.8 cells and AR-C40 cells, IC₅₀ for PTH(1-34) was approximately two-fold smaller than IC₅₀ for PTHrP(1-34). Comparison of PTH(1-34) and PTHrP(1-36) for binding to COS-7 cells expressing the human PTHR1 suggested approximately seven-fold smaller IC₅₀ for PTH(1-34).²² A comparison of human PTH(1-34) and PTHrP(1-36) for binding to the rat PTHR1 on UMR 106 cells yielded indistinguishable IC₅₀ values.³⁰

Two studies have employed radiotracer displacement to compare the binding of PTH(1-34) and PTHrP(1-36) to distinct forms of the human PTHR1 in isolated cell membranes. In one version of this assay, the intracellular surface of the PTHR1 is bound to the G_S heterotrimer. This form of the receptor is designated the RG state. In the other version of this assay, the PTHR1 is functionally uncoupled from the G protein. This form of the receptor is designated the R⁰ state. Using membranes prepared from COS-7 cells that transiently overexpressed the human PTHR1, Dean et al. reported that human PTH(1-34) and PTHrP(1-36) have similar IC₅₀ values for the RG state, but that for the R⁰ state PTH(1-34) displays a ~4-fold lower IC₅₀ relative to PTHrP(1-36).²³ A similar comparison was subsequently made by Hattersley et al. with membranes from GP2.3 cells, which stably express the human PTHR1.²⁴ Again, human PTH(1-34) and PTHrP(1-36) displayed similar IC₅₀ values in the RG state assay, but PTH(1-34) bound with higher affinity to the R⁰ state, as indicated by a ~9-fold lower IC₅₀ relative to PTHrP(1-36). Comparison of our results to these precedents suggests that our cell-based assay may reflect the R⁰ state of the PTHR1, as has been predicted for cells that overexpress only the receptor but not G proteins.²⁵

Long-acting PTH (LA-PTH) is a designed agonist containing PTH-derived residues in the N-terminal region and PTHrP-derived residues in the C-terminal region.²⁶ LA-PTH induces PTHR1 activation of very long duration in cellular assays and in animals.¹⁴ Binding measurements involving radiotracer displacement indicated that the affinity of LA-PTH for the RG state of the PTHR1 is similar to the RG affinities of PTH(1-34) and PTHrP(1-36), while the affinity of LA-PTH for the R⁰ state is higher relative to the affinities of PTH(1-34) and PTHrP(1-36).²⁴ We conducted binding measurements with LA-PTH bearing a TMR unit on a lysine residue at the C-terminus (Figure S3). k_on for this peptide was similar to that of PTH(1-34) or PTHrP(1-34), but k_off for LA-PTH was an order of magnitude lower than k_off for the other two peptides (Table 1). Thus, K_D for LA-PTH was smaller than for PTH(1-34) or PTHrP(1-34), which is consistent with the reported trend in IC₅₀ values for the R⁰ state.²⁴

Ferrandon et al. conducted single-cell FRET studies of agonist binding to a GFP-PTHR1 fusion protein using peptides bearing a TMR fluorophore at position 13.³¹ These measurements, which have a much shorter time scale relative to our BRET measurements, supported a two-state binding mechanism. In this mechanism, the C-terminal portion of the agonist associates with the receptor ECD in the initial step. In a second, slower step, the N-terminal portion of the agonist engages the receptor TMD. For the first step, PTH(1-34) and PTHrP(1-36) displayed k_on values similar to those we measured, but k_off was ~50- to 100-fold faster for this step than for the interaction characterized in our assay. The two agonists differed dramatically in the second step, according to the single-cell FRET study, with PTHrP(1-36) displaying a k_off value comparable to the value we measured, but little dissociation observed for PTH(1-34).³¹ This latter behavior was attributed to rapid internalization of the PTHR1-PTH(1-34) complex, which continues to stimulate cAMP production from endosomes. In addition to the shorter time scale, the measurements of Ferrandon et al. differ from our BRET-based measurements in that our studies employ cells pretreated with NaN₃ to prevent receptor internalization, while internalization mechanisms were active under the conditions used by Ferrandon et al.³¹

Overall, the comparisons with previous studies summarized above suggest that the results provided by the new BRET-based assay are consonant with results for agonist-PTHR1 association obtained by other methods. The BRET-based assay can be performed with a conventional luminescence plate reader, making this assay more easily implemented than the assays previously employed to characterize ligand-PTHR1 binding.

Competition assay format.

For comparative studies of multiple peptides, it is convenient to conduct binding assays that involve displacement of a single labeled tracer peptide (Figure 2A). To evaluate the nLuc-PTHR1 construct in competition mode, we used 50 nM PTH(1-34)-K35^TMR as the tracer and unlabeled competitors at variable concentration, and monitored tracer binding over time. As the concentration of the competitor peptide was increased, the degree of tracer binding diminished. Figure 2B shows data for tracer binding in the presence of varying concentrations of PTH(1-34), and Figure 2C shows analogous data with PTHrP(1-34). PTH(1-34) has a smaller dissociation rate constant compared to PTHrP(1-34), and therefore it was necessary to increase the measurement time relative to that with PTHrP(1-34) to determine accurate values. The k_on, k_off and K_D values derived from these data were similar to those measured with TMR-bearing peptides in the direct-binding format (Table 1).

Figure 2.

(A) Cartoon depiction of the competition mode of NanoBRET binding kinetics measurements. Varying concentrations of the peptide of interest and a fixed concentration of the tracer were introduced to cells expressing nLuc-PTHR1. The decrease in the BRET response was used to calculate the kinetic properties and the affinities of the competitors. (B) Representative BRET data generated using 50 nM PTH(1-34)K35^TMR and varying concentrations of PTH(1-34). Each trace shows the result of a single measurement. (C) Representative BRET data generated using 50 nM PTH(1-34)K35^TMR and varying concentrations of PTHrP(1-34). Each trace shows the result of a single measurement. (D) Competition equilibrium binding of PTH(1-34), LA-PTH, and ABL in the presence of 30 nM PTH(1-34)K35^TMR on nLuc-PTHR1 expressed on metabolically poisoned cells (0.02% NaN₃ (w/v)). Each datapoint represents mean ± s.e.m from n ≥ 3 independent experiments.

We used equilibrium competition binding experiments with 30 nM PTH(1-34)-K35^TMR as the tracer to evaluate the designed agonist LA-PTH. These measurements suggested IC₅₀ of 4.1 ± 1 nM for LA-PTH, compared to 16 ± 10 nM for PTH. Thus, the difference in IC₅₀ values between PTH and LA-PTH measured in the competition binding assay (~4-fold) is comparable to the difference in K_D between these two peptides measured in the direct-binding assay (~5-fold; Table 1). Another PTHR1 agonist, ABL, which is as potent as PTH(1-34) in terms of stimulating cAMP production²⁴, failed to fully inhibit PTH(1-34)K35^TMR binding even at 1 μM ABL (Figure 2D). This observation is consistent with a previous report that ABL has diminished affinity for the R⁰ state²⁴ and our observation that ABL(1-34)K35^TMR failed to produce a significant BRET response with nLuc-PTHR1, in contrast to PTH(1-34)K35^TMR.

Garton et al. reported that a peptide comprised entirely of D-amino acid residues displayed potency comparable to that of PTH(1-34) (L-amino acid residues) in activating the PTHR1.³² In the original publication, it was not clear whether the C-terminal amide or C-terminal acid form of this peptide was evaluated, so we examined both forms (Figure S4). In our equilibrium binding competition BRET assay, no displacement of PTH(1-34)K35^TMR by either version of the D-peptide could be detected (Figure S4). The ability of the D-peptide C-terminal amide to stimulate cAMP production via the PTHR1 was extremely weak, with estimated EC₅₀ at least 10,000-fold higher than EC₅₀ for PTH(1-34). The D-peptide C-terminal acid was an even weaker agonist.

Modified assay for binding to a version of the PTHR1 lacking the extracellular domain.

Gardella et al. demonstrated that an N-terminally truncated version of the PTHR1 lacking most of the extracellular domain is competent for signal transduction.²⁷ PTH(1-34) is a weak agonist of the truncated PTHR1, but other designed peptides, including LA-PTH and M-PTH(1-14), potently activate the truncated receptor. M-PTH(1-14) differs only modestly from the first 14 residues of LA-PTH (Figure 3A). M-PTH(1-14) contains aminoisobutyric acid at positions 1 and 3, which are Ala in LA-PTH. M-PTH(1-14) contains homoarginine at position 11, while LA-PTH contains arginine at this position.

Figure 3.

(A) Sequences of LA-PTH(1-34) and M-PTH(1-14). (B) Cartoon depiction of BRET between fluorescently labeled ligand and nLuc-ΔECD-PTHR1 after binding. (C) Representative BRET data for LA-PTH(1-34)K35^TMR binding to nLuc-ΔECD-PTHR1. Each trace shows the result of a single measurement. (D) Representative BRET data for association then dissociation of M-PTH(1-14)K15^TMR on nLuc-ΔECD-PTHR1. Each trace shows the result of a single measurement.

We constructed a fusion protein containing the truncated PTHR1 preceded by a NanoLuc luciferase (nLuc) (nLuc-ΔECD-PTHR1) (Figure 3B). HEK293H cells transiently transfected with this receptor were used to conduct BRET-based binding studies as described above. Binding of TMR-labeled derivatives of PTH(1-34) and PTHrP(1-34) could not be reliably detected (data not shown). In each case, a BRET signal was observed at peptide concentrations of 200 nM or higher, but control studies showed that at these TMR-peptide concentrations, BRET or radiative energy transfer can arise without specific binding to the receptor. In contrast, binding could be measured for derivatives of LA-PTH and M-PTH(1-14) bearing a TMR unit (Table 1). For LA-PTH, data for peptide association could be fitted to a pseudo-first order binding model as described above to generate k_on and k_off. For binding of M-PTH(1-14)K15^TMR, however, we employed an association-then-dissociation approach. This modification in approach was necessary because although M-PTH(1-14)K15^TMR showed detectable BRET response at the highest concentration employed (100 nM), this peptide failed to produce responses at lower concentrations; the original approach requires data sets obtained over a range of concentrations. k_off was calculated from the dissociation curve based on a one-phase exponential decay model, and this rate constant was then used to fit the association curve and calculate the k_on value. k_on measured for LA-PTH with the nLuc-ΔECD-PTHR1 was indistinguishable from k_on measured for this peptide with the full-length nLuc-PTHR1, but k_off measured for the truncated receptor was considerably larger relative to k_off for the full-length receptor. The similarity of the k_on values suggests that this rate constant is primarily determined by interaction between LA-PTH and the receptor TMD, even when the ECD is present. This conclusion is consistent with the recent cryo-EM structure for the complex between LA-PTH and the PTHR1.¹³ Three bound states were resolved among the particles, all with approximately the same interactions between the N-terminal portion of LA-PTH and the receptor TMD. However, significant variations were observed in the relationship between the C-terminal portion of LA-PTH and the ECD, with one of the three states showing no interaction in this region of the complex. The difference in k_off values at the full-length receptor for LA-PTH vs. M-PTH(1-14) suggests that interactions between the C-terminal portion of LA-PTH and the ECD contribute significantly to the overall affinity of the receptor for LA-PTH.

Sequence variants of known PTHR1 agonists.

We selected several peptides derived from PTH(1-34) and/or PTHrP(1-36) for evaluation with the new assay based on previous reports documenting effects of sequence modifications on agonist activity and/or PTHR1 binding. The goals of these studies were: (1) to enhance our understanding of the relationship between data generated by the new assay and previous measurements, and (2) to provide new insights on previously described agonists.

N-terminal truncations of PTH(1-34).

Early studies with PTH(1-38) showed that bone anabolic effects in rats were diminished upon removal of the N-terminal residue (to generate PTH(2-38)), and these effects were lost altogether with a further truncation (to generate PTH(3-38)).³³ We observed that PTH(2-34) was moderately less potent than PTH(1-34), as indicated by a ~10-fold increase in EC₅₀ for PTH(2-34) relative to PTH(1-34), while PTH(3-34) showed only minimal activity (Figure 4A). Derivatives of these N-terminally truncated peptides containing Lys at the C-terminus with TMR on the side chain were very similar to the peptides lacking TMR in terms of their ability to stimulate intracellular cAMP production (Figure S7). In studies with the nLuc-PTHR1, k_on values for PTH(2-34)K35^TMR and PTH(3-34)K35^TMR (Figure 4B,C) were indistinguishable from that of the full-length analogue, PTH(1-34)K35^TMR (Table 2). The k_off values were indistinguishable for the full-length peptide and PTH(2-34)K35^TMR, while k_off was modestly larger for PTH(3-34)K35^TMR (Table 2). These results show that although the N-terminal residues of PTH are critical in terms of receptor activation, these two residues contribute little or nothing to peptide affinity for the PTHR1. The recent cryo-EM structure of the complex between LA-PTH and the PTHR1 suggests that contacts between the N-terminus of the peptide and residues in receptor TM helix 6 promote adoption of the activated conformation of the PTHR1.¹³ It has been proposed that these contacts are repulsive,³⁴ a view that seems to be consistent with our binding data.

Figure 4.

(A) Stimulation of intracellular cAMP production by cells expressing WT-PTHR1 and Glosensor construct activated by PTH(1-34) or its N-terminally truncated analogue PTH(2-34) or PTH(3-34). The luminescence response of vehicle control is normalized to 0%, while maximal response of unlabeled PTH(1-34) is normalized to 100%. Each datapoint represents mean ± s.e.m from n ≥ 3 independent experiments. (B) Representative BRET data for binding of PTH(2-34)K35^TMR to nLuc-PTHR1. (C) Representative BRET data for binding of PTH(3-34)K35^TMR to nLuc-PTHR1. Each trace shows the result of a single measurement.

Table 2.

The kinetic constants and dissociation constants of two N-terminally truncated analogues of PTH(1-34).

	kon (× 105 s−1M−1)	koff (s−1)	KD (nM)
PTH[1-34]-35KTMR	4.4 ± 0.8	0.0051 ± 0.006	12 ± 2
PTH[2-34]-35KTMR	2.5 ± 1.7	0.0053 ± 0.004	21 ± 6
PTH[3-34]-35KTMR	3.4 ± 0.8	0.0094 ± 0.003	27 ± 12

Comparison of binding data for PTH(1-34)K35^TMR and the N-terminally truncated analogues. The data shown for PTH(1-34)K35^TMR are identical to those shown for this peptide in Table 1. Each value is derived as mean ± SD from n ≥ 3 independent experiments.

Variations at position 5 of PTH and PTHrP.

There are many differences between the sequences of PTH(1-34) and PTHrP(1-34), but the difference at position 5 (Ile in PTH vs. His in PTHrP) is known to have a particularly large effect in terms of affinity for the receptor and signaling profile.^21,22 Comparisons involving PTH(1-34) and the I5H variant, and PTHrP(1-36) and the H5I variant, revealed that neither change caused a substantial alteration in affinity for the RG state of the PTHR1.²³ In contrast, these changes had large effects on affinity for the R⁰ state, with Ile at position 5 being favored in both the PTH and PTHrP contexts. Thus, native PTH(1-34) displayed ~20-fold lower IC₅₀ relative to the I5H variant in the competition binding assay for the R⁰ state, and the H5I variant of PTHrP(1-36) displayed a ~10-fold lower IC₅₀ relative to the version with the native His. These binding assays were conducted in DPBS buffer at pH 7.4.²³

We used the BRET assay (direct binding mode) to compare PTH(1-34)K35^TMR with the I5H variant, and to compare PTHrP(1-34)K35^TMR with the H5I variant (Table 3). Trends from measurements conducted at pH 7.5 were qualitatively consistent with those reported for R⁰ state binding based on radioligand displacement.²³ For PTH(1-34), the native Ile5 led to ~6-fold lower K_D relative to non-native His5, while for PTHrP(1-36), the non-native Ile5 led to ~5-fold lower K_D relative to native His5. The major kinetic effect of Ile vs. His at position 5 was manifested in k_off, which was an order of magnitude larger with His in both cases. k_on was larger with His as well, but only slightly.

Table 3.

The effect of pH on the kinetic constants and the dissociation constants of PTH(1-34) and PTHrP(1-36), as well as their analogues with substitution at the 5^th position (Figure S5).

	pH	kon (× 105 s−1M−1)	koff (s−1)	KD (nM)
PTH	7.5	4.8 ± 0.5	0.0049 ± 0.002	10 ± 3
	6.5	4.2 ± 0.8	0.0049 ± 0.001	12 ± 3
	5.5	9.8 ± 4*	0.039 ± 0.006*	39 ± 6*
PTH(I5H)	7.5	7.6 ± 1	0.049 ± 0.01	64 ± 10
	6.5	6.8 ± 3*	0.071 ± 0.04*	104 ± 50*
	5.5	4.4 ± 2*	0.088 ± 0.04*	200 ± 110*
PTHrP	7.5	3.1 ± 0.5	0.021 ± 0.006	68 ± 20
	6.5	3.4 ± 0.2	0.049 ± 0.02	140 ± 50
	5.5	2.2 ± 0.5*	0.071 ± 0.04*	330 ± 160*
PTHrP(H5I)	7.5	2.9 ± 0.2	0.0035 ± 0.0005	12 ± 2
	6.5	2.7 ± 0.3	0.0034 ± 0.0006	12 ± 2
	5.5	1.5 ± 0.3*	0.071 ± 0.02*	48 ± 10*

Comparison of binding data for PTH(1-34)K35^TMR, PTHrP(1-36)K35^TMR, and analogues modified at position 5 at pH 7.5, 6.5, and 5.5. Each buffer contained 137 mM NaCl, 1 mM CaCl₂, 0.5 mM MgCl₂, 0.02% NaN₃ (w/v), and 0.1% BSA (w/v), and pH was maintained using HEPES (7.5), MES (6.5), or acetate (5.5). Data showing BRET response over time are provided in Figure S5. Each value is derived as mean ± SD from n ≥ 3 independent experiments.

^*p<0.05, relative to pH 7.5 data, analyzed by T-test

The pH 7.5 parameters in Table 3 for PTH(1-34)K35^TMR and PTHrP(1-36)K35^TMR differ slightly from those in Table 1. These small differences may arise because of differences in media for the two assays. The assays summarized in Table 1 were conducted in Dulbecco’s phosphate-buffered saline. We measured the pH of this medium to be in the range 7.0–7.3, which is slightly lower than the HEPES buffer used for the pH 7.5 measurements summarized in Table 1.

The complex formed between the PTHR1 and PTH(1-34) remains intact and continues to stimulate cAMP production after internalization.³¹ In contrast, internalization of the complex formed between PTHrP(1-36) and PTHR1 causes dissociation of the peptide and halts cAMP production.³¹ The divergent behaviors of these two agonists may arise from changes in pH encountered by the peptide-receptor complex as it is trafficked from the cell surface to endosomes.³⁵ The shift from neutral pH at the cell surface to the mildly acidic environment of a late endosome (pH ~ 5.5) should enhance protonation of His side chains. Protonation of His5 of PTHrP(1-36), which is buried in the receptor TMD core, could cause a substantial decline in affinity for the PTHR1. In contrast, protonation at position 5 of PTH(1-34) is not possible because the Ile side chain does not contain a basic site. Thus, the difference between PTH and PTHrP at position 5 could play a critical role in the divergent spatiotemporal signaling profiles of these agonists.

We evaluated binding to the nLuc-PTHR1 as a function of pH (7.5 vs. 6.5 vs. 5.5) for PTH(1-34)K35^TMR, PTHrP(1-36)K35^TMR and the TMR-bearing analogues modified at position 5 to test the hypothesis in the preceding paragraph, i.e., to determine whether a drop in pH causes a larger decline in affinity for the peptides with His at position 5 relative to those with Ile at this position (Table 3). For all four peptides, affinity was modestly lower at pH 5.5 relative to pH 7.5. However, the effect of diminished pH was similar for peptides containing Ile5 and those containing His5. Thus, at pH 5.5 relative to pH 7.5, K_D was ~4-fold larger for PTH(1-34), and K_D was ~5-fold larger for PTHrP(1-36). A comparable change was observed for I5H variant of PTH(1-34). Among the two kinetic parameters, the larger changes were observed in k_off, and pH-dependent variation seemed larger for the peptides containing Ile relative to those containing His at position 5. Overall, these data indicate that protonation of the His residue near the N-terminus of PTHrP(1-36) does not constitute a switch that dramatically lowers the affinity of this agonist for the PTHR1 within endosomes compared to receptor at the cell surface.

Both PTHrP(1-34) and PTHrP(1-36) have been subjects of previous studies. To ask whether this small difference in length affects agonism trends, we compared the H5I variants of PTHrP(1-36) and PTHrP(1-34) by measuring affinities for the nLuc-PTHR1 at different pH. Values obtained for PTHrP(1-34)-His5-Ile-K35^TMR, K_D = 5.0 nM at pH 7.5, K_D = 5.7 nM at pH 6.5, and K_D = 56 nM at pH 5.5, are similar to those obtained for PTHrP(1-36)-His5-Ile-K35^TMR (Table 3).

PTH/PTHrP hybrids.

Early studies evaluated synthetic hybrids of PTH(1-34) and PTHrP(1-34) in terms of binding to and activation of the PTHR1.^21,22 Binding was assessed via radioligand displacement assays involving whole cells. These studies were motivated by the hypothesis that the N-terminal and C-terminal portions of these peptides likely engaged distinct portions of the receptor, the TMD and ECD, respectively. This hypothesis was subsequently validated by structural analysis of receptor-peptide complexes.^{13,14,36–38}

The hybrid PTH(1-14)/PTHrP(15-34) was reported to be indistinguishable from PTH(1-34) in competition binding measurements with the rat PTHR1 expressed in ROS17/2.8 or AR-C40 cells,²¹ but a subsequent comparison involving the human PTHR1 expressed in COS-7 cells indicated weaker binding (~12-fold higher IC₅₀) for the hybrid peptide relative to PTH(1-34).²² The complementary hybrid, PTHrP(1-14)/PTH(15-34), in contrast, bound much more weakly than PTH(1-34) in all three comparisons. IC₅₀ for the second hybrid was ~1000-fold larger in the study with ROS17/2.8 or AR-C40 cells and ~200-fold larger in the study with COS-7 cells. The dramatic difference between the two hybrids in terms of PTHR1 affinity motivated the development of LA-PTH, the C-terminal portion of which is very similar to PTHrP(15-36), and the N-terminal portion of which is derived from PTH(1-14).²⁶ In the COS-7 cell study, agonist activities were evaluated.²² Despite the large differences in binding, as reflected in IC₅₀ values, PTH(1-34), PTHrP(1-36) and the two hybrid peptides were very similar in terms of their ability to stimulate cAMP production, as indicated by EC₅₀ and E_max values.

These precedents motivated us to evaluate two sets of PTH/PTHrP hybrid peptides, with varying sequence proportions drawn from the two prototype peptides (Table 4). One set contained PTH residues at the N-terminus and PTHrP residues at the C-terminus; the other set had the opposite arrangement. Each peptide had Lys at position 35, with TMR on the side chain. All of these peptides were potent as agonists: EC₅₀ for cAMP production varied between 0.1 and 1.0 nM, and each hybrid matched E_max achieved with PTH(1-34) (Table 4). This behavior is consistent with the activities reported for hybrid peptides with COS-7 cells.²²

Table 4.

The kinetic constants, dissociation constants, and the activity profiles of various PTH/PTHrP hybrids.

Kinetic Profiles			cAMP formation
	kon (× 105 s−1 M−1)	koff, (s−1)	KD (nM)	EC50 (nM)	%Emax
PTH[1-12] PTHrP[13-36]	8.0 ± 0.5*	0.0035 ± 0.001*	4.4 ± 1*	0.55 ± 0.1	107 ± 5
PTH[1-14] PTHrP[15-36]	3.5 ± 0.4*	0.0022 ± 0.0009*	6.1 ±3*	0.89 ± 0.2	106 ± 1
PTH[1-16] PTHrP[17-36]	5.3 ± 0.3*	0.0017 ± 0.001*	3.4 ± 2*	0.79 ± 0.1	103 ± 4
PTH[1-18] PTHrP[19-36]	3.2 ± 0.7*	0.0079 ± 0.001*	26 ± 4*	0.99 ± 0.2	106 ± 1
PTHrP[1-12] PTH[13-34]	5.9 ± 1*	0.029 ± 0.009*	49 ± 20*	0.15 ± 0.02	100 ± 2
PTHrP[1-14] PTH[15-34]	6.8 ± 2*	0.043 ± 0.01*	63 ± 20*	0.39 ± 0.1	97 ± 6
PTHrP[1-16] PTH[17-34]	4.3 ± 0.6*	0.041 ± 0.01*	95 ± 30*	0.48 ± 0.2	105 ± 1
PTHrP[1-18] PTH[19-34]	4.8 ± 0.8*	0.026 ± 0.006*	53 ± 10*	0.16 ± 0.04	99 ± 5

Comparisons of binding data and the activity profiles for PTH/PTHrP hybrid peptides. Each value is derived as mean ± SD from n ≥ 3 independent experiments.

^*p<0.05, relative to PTH(1-34)-35K^TMR data in Table 1, analyzed by T-test. Data are provided in Figures S6 and S7.

Trends in PTHR1 affinity we measured for the hybrid peptides were qualitatively consistent with those observed in radioligand displacement assays,²² although the range of variation in our binding assay was much smaller than in the earlier studies.⁹ The hybrid peptides with N-terminal PTHrP sequence and C-terminal PTH sequence had K_D values about one order of magnitude higher than those with N-terminal PTH sequence and C-terminal PTHrP sequence. The point at which the sequence switch occurred had little effect on K_D (~2-fold variation) among the series spanned by PTHrP(1-12)/PTH(13-34) and PTHrP(1-18)/PTH(19-34). A similar trend was observed among PTH(1-12)/PTHrP(13-36), PTH(1-14)/PTHrP(15-36) and PTH(1-16)/PTHrP(17-36), but K_D was modestly higher for PTH(1-18)/PTHrP(19-36). k_on values were very similar among all of the hybrid peptides. In contrast, the series with N-terminal PTHrP sequence/C-terminal PTH showed significantly larger k_off values relative to the series with N-terminal PTH sequence/C-terminal PTHrP sequence.

Conclusions.

We have developed a convenient BRET-based assay for characterizing the binding of agonist peptides to the PTHR1, a target of multiple drugs in current clinical use. This assay format offers practical advantages relative to radioligand displacement. Multiple comparisons with earlier radioligand displacement studies suggest that the BRET assay provides generally consistent results, although some variations were evident.

The BRET assay can be conducted in a “direct binding” mode with agonists that bear a TMR appendage, which is straightforward to introduce in conjunction with solid-phase peptide synthesis. Addition of a TMR group represents a modest change in the context of a 34- to 36-residue peptide, which led us to hypothesize that this modification would not cause a profound alteration in peptide-receptor interactions if the TMR unit were strategically placed. Our data support this hypothesis. In addition, our data indicate that the receptor does not experience a major functional alteration as a result of the fused NanoLuc luciferase (nLuc). We note that the BRET-based strategy has been applied recently to family A GPCRs, the agonists of which are similar in size to the fluorophore that must be attached.^28,29,39

Use of peptides bearing a TMR group in the BRET assay allows direct measurement of the rate constants k_on and k_off for agonist-PTHR1 interaction, and K_D can be determined from those values. Alternatively, the BRET assay can be run in competition format, which accommodates unlabeled peptides and streamlines comparisons.

We have used the new assay to ask whether the Ile vs. His difference at position 5 of PTH and PTHrP underlies the difference in spatiotemporal signaling profiles between these two agonists. PTH continues to stimulate cAMP production after the agonist-receptor complex has been trafficked to endosomes, while PTHrP signaling is terminated upon endocytosis.³¹ There is great interest in PTHR1 signaling timeframe because the temporal pattern of agonist administration strongly influences the impact on bone growth.^40,41 Our data suggest that endosomal acidification does not lead to a large change in the relative affinities of PTH(1-34) and PTHrP(1-36), as would have been predicted if His protonation destabilized the bound form of PTHrP relative to the bound form of PTH.

The new assay provides insights that complement early studies of N-terminal truncation of PTH-derived peptides.³³ Removal of the first residue substantially impairs receptor activation, as manifested in stimulation of cAMP production, and removal of the first two residues further hinders agonist activity. Measurements from the BRET assay, however, show that PTH(2-34) and PTH(3-34) are comparable to PTH(1-34) in terms of affinity for the PTHR1. On the other hand, use of the BRET assay to examine a D-peptide³² that was reported to be comparable to PTH(1-34) in potency as an agonist for the PTHR1 indicates that the D-peptide binds only very weakly to the receptor. These observations prompted us to evaluate the D-peptide for the ability to stimulate cAMP production via the PTHR1; only very weak activity was detected, which is inconsistent with the earlier report.

The new BRET assay based on the nLuc-PTHR1 fusion should be useful to researchers who study signal transduction mediated by this medically important receptor. The construct will be available via Addgene, which should facilitate exploration by others.

Methods

Molecular Biology.

NanoLuc luciferase was fused to the N-terminus of the PTHR1 using Gibson assembly. Briefly, plasmids encoding nLuc-GLP-1R⁴² and Human PTHR1 in a pcDNA3.1+ vector (cDNA Resource Center, #PTHR100000) were linearized and amplified with appropriate overhangs to facilitate Gibson assembly using the Phusion HF polymerase PCR kit (Thermo) according to the manufacturer’s instructions. The PCR products were treated with DpnI (0.8 μl; Agilent) for 1 h at 37 °C, and DNA was isolated using concentrator spin columns (Zymo). Then, 0.03 pmol and 0.06 pmol of linearized DNA encoding receptor + vector and NLuc, respectively, were assembled using 10 μl of 2× master mix at 50 °C for 1 h according to manufacturer’s instructions (NEB). 5α Competent Escherichia coli (30 μl; NEB) was then transformed with 2 μl of the assembly reaction mixture by heat shocking for 30 s at 42 °C. These cells were then allowed to grow in SOC medium (0.5 ml) for 1 h at 37 °C with shaking. The cell suspension was spread onto LB agar plates with ampicillin (100 μg ml⁻¹), the plates were incubated overnight at 37 °C, and colonies were selected for plasmid Maxiprep (Qiagen) according to the manufacturer’s protocol. Site-directed mutagenesis using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent) according to the manufacturer’s instructions was used to install the desired stop codon. Sanger sequencing performed at the University of Wisconsin-Madison Biotechnology Center DNA Sequencing Facility was used to confirm the desired sequence.

Peptide synthesis.

Peptides were prepared by microwave-assisted solid-phase peptide synthesis using either a CEM Mars microwave reactor (manual) or an automated CEM Liberty Blue reactor. Low-loading Rink Amide resin (CEM) was used as the solid support. For peptide coupling steps, 5 equiv. of Fmoc-amino acid, 10 equiv. of diisopropylcarbodiimide, and 5 equiv. of Oxyma with respect to the resin loading were used. NMP or DMF solutions of these compounds were added to the reaction vessel. The resulting slurry was heated to 90°C and stirred for 2 min, except for histidine, which was heated to 50°C for 10–30 min. After the heating, the resin was washed with DMF thrice, and the Fmoc was removed by adding 20% piperidine solution in DMF with 0.1% Oxyma, and heating to 80°C for 2 min. For orthogonal removal of the alloc protecting group on a lysine side chain, the resin was washed with methylene chloride thrice, and 10 equiv. of phenylsilane and 1 equiv. of tetrakis(triphenylphosphine)palladium(0) in ethylene dichloride were added to the reaction vessel. The resulting slurry was heated to 35°C for 10 min. This process was repeated twice. The resin was washed with methylene chloride thrice and dimethylformamide thrice. 4 equiv. 6-caboxytetramethylrhodamine, 4 equiv. of diisopropylcarbodiimide, and 4 equiv. of Oxyma in biotech-grade DMF were added, and the resulting mixture was heated to 75°C for 10 min. At the end of the solid-phase synthesis, the peptide was cleaved and deprotected in a solution containing 90% TFA, 5% thioanisole, 3% DODT and 2% anisole by incubating at room temperature for 4 h. The resulting solution was filtered from the resin, and product was precipitated by addition of 10-fold excess of −20 °C diethyl ether. Collected precipitate was washed twice with diethyl ether, dried, and dissolved in DMSO for further purification with a preparative reverse-phase HPLC system (Waters) using a C-18 column (12 mL/min flowrate, with gradients of H₂O + 0.1% TFA as solvent A and MeCN + 0.1% TFA as solvent B). The collected fractions were analyzed based on mass using MALDI-TOF-MS (Bruker UltraFlex) with CHCA as matrix. Isolated peptide was assessed for purity using a UPLC system (Waters) with a C-18 column (0.3 mL/min flowrate, with gradients of H₂O + 0.1% TFA as solvent A and MeCN + 0.1% TFA as solvent B) (Supplementary Information pages S2–S22). The concentration of peptide in stock solutions was determined using a NanoDrop One UV/Vis system (Thermo Scientific).

Cell Culture.

Cells were grown in Dulbecco’s Modified Eagle Medium (with D-glucose but lacking L-glutamine and sodium pyruvate) supplemented with 10% fetal bovine serum (Corning, Cat# 45000-734). Cells were incubated at 37°C with 5% CO₂. Dulbecco’s Phosphate Buffered Saline (DPBS; Gibco, Cat#14190235) and trypsin-EDTA (Corning, Cat# 45001-082) at 0.05% were used to detach cells for passaging and transfer to assay plates.

Transfection.

HEK293FT cells were seeded onto a 100 mm dish (Fisher, Cat# FB012924) and grown to 80–90% confluency as observed under a microscope. On the day of transfection, 10 μg of plasmid was mixed with 30 μL of FugeneHD (Promega, Cat# PAE2311) in Opti-MEM (Gibco, Cat# 31985062) to bring the final volume to 1 mL. The resulting mixture was gently mixed by flicking the Eppendorf tube,which was set to incubate at room temperature for 20–30 min. The medium from the 100 mm dish was aspirated, and the cells were washed with DPBS without calcium or magnesium. To the cells was added 4.5 mL of McCoy’s 5A supplemented with 10% FBS, and the cells were incubated at 37°C for 5 min. Following the incubation, 4.5 mL of DMEM with 10% FBS was added, and the transfection solution was added gently over the cells. The cells were left to incubate at 37°C for 24–36 h.

Glosensor Assay.

HEK293 cells stably expressing the WT-PTHR1 and the Glosensor construct (GP2.3) were grown in DMEM with 10% FBS. GP2.3 cells were harvested and seeded onto a 96-well assay plate (Costar, Cat# 29444-041), which was incubated at 37°C. On the day of the experiment, the medium was aspirated, and 90 μL of DPBS solution containing 500 μM of D-luciferin was added to each well. The cells were incubated at room temperature for 20 min. To the cells was added 10 μL of the peptide solution at various concentrations, and the luminescence was measured using a Biotek plate reader over 30 min. The luminescence at the 15-minute time point was used for activity analysis.

Kinetics Assay using BRET.

HEK293H cells stably expressing the nLuc-PTHR1 were generated using zeocine resistance as the selective marker. The cells were grown in DMEM supplemented with 10% FBS. Before the day of the experiment, the cells were seeded onto a white opaque 96-well assay plate at 5,000–10,000 cells per well. The cells were incubated at 37°C for 24 h. On the day of the experiment, the medium was aspirated, and 70 μL of DPBS supplemented with 0.02% NaN₃, 1 mM CaCl₂ and 0.5 mM MgCl₂ was added. The cells were incubated at room temperature for at least 30 min before the experiment. The labeled peptide solution was prepared in the same solution that also included 5 μM h-coelenterazine. For the experiment, the medium was aspirated from the well, 100 μL of the peptide solution with h-coelenterazine was added to the well, and the luminescence from 460 nm and 590 nm was measured over 5–10 min at 5-second intervals. The ratio of the luminescence from the 590 nm channel over that from the 460 nm channel with a 1,000x multiplier was used as the mBRET value for generating plots. For nLuc-ΔECD-PTHR1, HEK293FT cells were transiently transfected with the appropriate construct.

Competition Binding Kinetics.

The general procedure for preparation was the same as for the direct binding assay. For the assay buffer, 50 nM PTH(1-34)K35^TMR and three concentrations of each unlabeled ligand were prepared; each solution contained 5 μM h-coelenterazine. After ligand addition, the three wells (one for each unlabeled ligand concentration) were monitored at 9-second intervals.

Data Analysis.

For the direct binding kinetics experiments, the mBRET values were plotted against time in seconds, and fitted to equation 1. The extracted k_obs values were then plotted against the concentration of the tracer in equation 2 to calculate the slope and the Y-intercept, corresponding to k_on and k_off, respectively. The K_D values were calculated as the ratio of k_on and k_off (equation 3). For the association-then-dissociation kinetics experiment, the dissociation curve was fitted to a one-phase dissociation with Graphpad Prism software to measure k_off. Then the k_off value was used as a constraint for fitting the association curve to extract the k_on value. For the competition kinetics experiment, the curves generated were fitted into “Competition Kinetics” in Graphpad Prism software with the predetermined k_on, k_off and concentration values of the tracer to calculate the k_on and k_off values of the unlabeled peptide.

$$\text{Y}={\text{Y}}_{\text{t}(\text{inf})}\times \left(1-{\text{e}}^{-k_{obs}\text{t}}\right)$$ Equation 1.

$$k_{obs}=k_{on}[\text{L}]+k_{off}$$ Equation 2.

$${\text{K}}_{\text{D}}=\raisebox{1ex}{$k_{off}$}\!\left/ \!\raisebox{-1ex}{$k_{on}$}\right.$$ Equation 3.