LA-PTH is a long-acting parathyroid hormone (PTH) peptide analog in pre-clinical development for hypoparathyroidism (HP). Like native PTH, LA-PTH contains a methionine at position 8 that is predicted to be critical for function. We assessed the impact of methionine oxidation on the functional properties of LA-PTH and control PTH ligands. Oxidation of PTH(1-34) resulted in marked (~20-fold) reductions in binding affinity on the PTH receptor-1 (PTHR1) in cell membranes, similarly diminished potency for cAMP signaling in osteoblastic cell lines (SaOS-2 and UMR106), and impaired efficacy for raising blood calcium in mice. Surprisingly, oxidation of LA-PTH resulted in little or no change in these functional responses. The signaling potency of oxidized-LA-PTH was, however, reduced ~40-fold compared to LA-PTH in cells expressing a PTHR1 construct that lacks the N-terminal extracellular domain (ECD). Molecular modeling revealed that while Met8 of both LA-PTH and PTH(1-34) is situated within the orthosteric ligand-binding pocket of the receptor’s transmembrane domain bundle (TMD), the Met8 sidechain position is shifted for the two ligands such that upon Met8 oxidation of PTH(1-34) steric clashes occur that are not seen with oxidized LA-PTH. The findings suggest that LA-PTH and PTH(1-34) engage the receptor differently in the Met8-interaction environment of the TMD bundle, and that this interaction environment can be allosterically influenced by the ECD component of the ligand-receptor complex. The findings should be useful for the future development of novel PTH-based peptide therapeutics for diseases of bone and mineral ion metabolism.

Peptide synthesis and purification

Peptides were based on the human PTH or PTHrP peptide sequence, except for the rat PTH(1-34) analog used as a radioligand.  Peptides were C-terminal amides except for LA-PTH which was C-terminal carboxyl.  Peptides were synthesized by conventional, solid-phase chemistry utilizing Fmoc-protected amino acid precursors.  Non-conventional amino acids used included nor-leucine (Nle), alpha-aminoisobutyric acid (Aib), homoarginine (hAr), methionine sulfoxide (Met-SO) and methionine sulfphone (Met-SO₂).  Peptide products were cleaved from the resin and deprotected utilizing Reagent K (82.5% TFA, 5% phenol, 5% H₂O, 5% thioanisole, 2.5% ethanedithiol) for two hours.  Peptide products were purified on a prep‐C18 column using reverse phase‐HPLC (solvent A: 0.1% TFA in water, solvent B: 0.1% TFA in acetonitrile).  Purity was assessed by narrow-bore RP-HPLC (C-18, 2.1 x 150mm column, flow rate 0.4ml/min, 5-95% gradient over nine minutes; solvent A: 0.06% TFA in water, solvent B: 0.05% TFA in 80% acetonitrile).  Masses were measured by MALDI‐TOF mass spectrometry and the observed masses -- either monoisotopic [M+H]⁺, m/z or average [M+H]⁺, m/z), confirmed identity.  Peptide stock solutions were prepared in 10 mM acetic acid at peptide concentrations of 1.0 or 2.0 mM, and stored as aliquots at -80˚C.  Peptide radioligands, ¹²⁵I-PTH(1-34) (¹²⁵I-[Nle^8,21,Tyr³⁴]-ratPTH(1-34)NH₂) and ¹²⁵I-M-PTH(1-15) (Aib-Val-Aib-Glu-Ile-Gln-Leu-Nle-His-Gln-hAr-Lys-Trp-Tyr.NH₂) were prepared by chloramine-T-based radioiodination followed by reversed-phase HPLC purification.

Post-synthetic oxidation was performed by incubating peptides (1.0 mM) in 1.0% H₂0₂ for 16 hours at room temperature.  The solutions were then diluted in dH₂0, lyophilized, and the peptide was reconstituted in 10 mM acetic acid to a stock peptide concentration of 1.0 mM.  Mass spectrometry confirmed 100% conversion of methionines to methionine-sulfoxide (Met-SO), as indicated by an increase of +16 mass units with LA-PTH and M-PTH(1-14), corresponding to oxidation of the single methionine at position 8, and +32 mass units with PTH(1-34), corresponding to oxidation of Met8 and Met18.  

Cell lines

            Ligand functional properties were evaluated in cell lines stably transfected to express the glosensor cAMP reporter (plasmid p22F, Promega Corp.) and expressing the PTHR1 either by stable transfection (GP-2.3 and GD-5y cells) or endogenously (SGS-72 and UGS-56 cells) (26,27).  GP-2.3 cells were derived from HEK-293 cells (ATCC CRL-1573) by stable transfection with glosensor and a plasmid encoding the intact hPTHR1.  GD-5y cells were derived from HEK-293 cells by stable transfection with glosensor and a plasmid encoding PTHR1-delNt^yfp, in which the receptor’s ECD is replaced by YFP (27).  SGS-72 cells were derived from the human osteoblastic cell line SaOS-2 (ATCC HTB-85) by stable transfection with glosensor.  UGS-56 cells were derived from the rat osteoblastic cell line UMR-106 (ATCC CRL-1661) by stable transfection with glosensor.

Competitive binding assays

Ligand binding to the PTHR1 the R⁰ and RG conformations was assessed by competition methods (28) using membranes prepared from GP-2.3 cells (HEK-293/glosensor/hPTHR1).  Binding reactions (100 ul) were assembled at room temperature in 96-well vacuum filtration plates (0.65 mM Hydrophilic Low Protein Binding DV membrane, Millipore Corp., reference number MSDVN6510) in membrane binding buffer (20 mM HEPES, pH 7.4, 0.1 M NaCl, 3 mM MgSO₄, 20% glycerol, 3 mg/ml bovine serum albumin, protease inhibitors (Sigma-Aldrich Inc., reference number P8849). R⁰ reactions contained GTPgs (1x10^-5 M) and used ¹²⁵I-PTH(1-34) as the tracer radioligand, and RG reactions used ¹²⁵I-M-PTH(1-15) as the tracer radioligand. After adding radioligand (~20,000 cpm) and unlabeled ligand at varying concentrations, the reactions were initiated by adding cell membranes to a membrane protein concentration of 25 ng/ul (RG) or 50 ng/ul (R⁰).  Reactions were incubated for 90 minutes and terminated by rapid vacuum filtration followed by rinsing of the filter membrane with 200 ul buffer, after which the filters were removed and counted for gamma irradiation.  Values of non-specific binding, maximum binding, and the 50% inhibition constant (IC₅₀), were determined by curve-fitting the data using PRISM 8.0 software to a sigmoidal dose-response model with variable slope and constrained to shared minimum and maximum values. 

cAMP signaling assays

PTH peptide-induced cAMP signaling responses were assessed in intact cells utilizing the glosensor cAMP reporter.  Cells were seeded in white 96 well plates and utilized for assay 24-48 hours (GP2.3 cells) or 3-5 days (SGS-72 and UGS-56 cells) after reaching confluency.  At the time of assay, the cells were pre-loaded with luciferin by replacing the culture media with CO₂-independent media (Gibco Life Sciences; reference number 18045-088) containing 0.7% bovine serum albumin and 0.5 mM d-luciferin (Biotium, Hayward CA, USA, reference number 10101), and incubating the plates for 30 minutes at room temperature.  Ligands were then added at varying concentrations, and luminescence (as counts per second, cps), was recorded at two-minute intervals in a Perkin Elmer Envision plate reader for 18-30 minutes (Ligand-on phase).  For wash-out assays, the plates were removed from the instrument after the ligand-on phase, the cells were rinsed twice with media to remove free PTH peptide, new media containing d-luciferin (0.5 mM) was added, and luminescence was recorded at two minute intervals for 90 minutes (Wash-out phase).  The peak luminescence signal, which occurred ~10-15 minutes after ligand addition, or the area-under-the-curve (AUC) of the time vs. luminescence response curves obtained for each ligand dose was used to generate ligand dose-response curves.  The AUC values obtained in each assay with UGS-56 cells were normalized to the maximum AUC response observed in that assay for PTH(1-34) (Ligand-on phase) or LA-PTH (Wash-out phase).  Curves were fit to ligand dose-response data using PRISM 8.0 software and a sigmoidal dose-response model with variable slope and constrained to shared minimum and maximum values.  Data were compiled from three or more independent experiments, each in duplicate, and are reported as means±SEM. 

Calcemic Responses in mice

Calcemic responses to PTH peptides were assessed in CD1 female mice (aged 9 weeks).  Mice were treated in accordance with the ethical guidelines adopted by the Massachusetts General Hospital.  Mice (n = 5 per group) were injected subcutaneously with vehicle (10 mM citric acid/150 mM NaCl/0.05% Tween-80, pH 5.0) containing PTH(1-34) or Ox-PTH(1-34) at a dose of 40 nmol/kg body weight, or an LA-PTH analog at a dose of 2 nmol/kg body weight.  The doses used for the LA-PTH analogs were selected to be 20-fold lower than that used for the PTH(1-34) peptides based on prior studies that demonstrate a more robust and prolonged blood calcium response to injected LA-PTH as compared to PTH(1-34) in mice (5,29).  Just prior to injection (t = 0) and at 1, 2, 4 and 8 hours after injection, blood was collected from a 1 mm tail vein incision into a heparinized capillary tube (Multi-cap-S, Siemens Healthcare Diagnostics Inc, reference number 05656514) and immediately analyzed for pH-adjusted ionized calcium using a Siemens RapidLab 348 Ca2+/pH analyzer. 

Molecular models were generated using the cryo-EM structure of the LA-PTH•PTHR1•Gas/bg complex (PDB:6nbf) (11) and the X-ray crystal structure of the PTHR1 in complex with an engineered (e)PTH(1-34) analog (PDB:6fj3) (10).  To derive a PTH(1-34)-bound PTHR1 model from the ePTH-bound PTHR1 X-ray crystal structure, the eight engineered mutations in the ePTH(1-34) peptide were changed back to the endogenous PTH residues (Acp1àSer, Aib3àSer, Gln10àAsn, homoArg11àLeu, Ala12àGly, Trp14àHis, norLeu18àMet and Tyr34àPhe), and the two thermostabilizing mutations in the PTHR1 were mutated back to the wild type residues (K240àM and Y191àC). To assess differences in binding modes used by LA-PTH and PTH in the models, the sidechains of all amino acids in the ligand and those in the receptor within 4Å of the peptide were optimized by energy minimization using Monte Carlo simulation in the ICM Molsoft Pro software package (v3.8) (30). Model images were generated using ICM Molsoft Pro as well as Pymol software (version 1.8.6.2). 

Graphical data are provided as .pzfx files that were generated using Graphpad Prism 8.0 for Mac OS. A molecular model is provided as a .pse package file that was generated in MacPymol vers. 1.8.6.2. These files can be accessed using those two applications.