Molecular Mechanisms of PTH/PTHrP Class B GPCR Signaling and Pharmacological Implications

Abstract

The classical paradigm of G protein-coupled receptor (GPCR) signaling via G proteins is grounded in a view that downstream responses are relatively transient and confined to the cell surface, but this notion has been revised in recent years following the identification of several receptors that engage in sustained signaling responses from subcellular compartments following internalization of the ligand–receptor complex. This phenomenon was initially discovered for the parathyroid hormone (PTH) type 1 receptor (PTH₁R), a vital GPCR for maintaining normal calcium and phosphate levels in the body with the paradoxical ability to build or break down bone in response to PTH binding. The diverse biological processes regulated by this receptor are thought to depend on its capacity to mediate diverse modes of cyclic adenosine monophosphate (cAMP) signaling. These include transient signaling at the plasma membrane and sustained signaling from internalized PTH₁R within early endosomes mediated by PTH. Here we discuss recent structural, cell signaling, and in vivo studies that unveil potential pharmacological outputs of the spatial versus temporal dimension of PTH₁R signaling via cAMP. Notably, the combination of molecular dynamics simulations and elastic network model–based methods revealed how precise modulation of PTH signaling responses is achieved through structure-encoded allosteric coupling within the receptor and between the peptide hormone binding site and the G protein coupling interface. The implications of recent findings are now being explored for addressing key questions on how location bias in receptor signaling contributes to pharmacological functions, and how to drug a difficult target such as the PTH₁R toward discovering nonpeptidic small molecule candidates for the treatment of metabolic bone and mineral diseases.

Graphical Abstract

Essential Points.

• Molecular dynamics (MD) simulations linked to the anisotropic network model (ANM)–based method address the key question of how to identify druggable allosteric sites in the parathyroid hormone (PTH) type 1 receptor (PTH₁R) and small molecules that can execute allosteric modulation of receptor signaling and function

• Structure-encoded allosteric coupling between PTH₁R and PTH residues helps design synthetic PTH analogs with the desired bias to demonstrate that physiological responses to PTH depend on the subcellular location of receptor signaling via cyclic adenosine monophosphate (cAMP).

• PTH₁R-generated endosomal cAMP signaling is thought to ensure vitamin D and bone homeostasis whereas the PTHR-generated plasma membrane cAMP maintains phosphate ion homeostasis.

The Role of PTH₁R in Human Health

The parathyroid hormone (PTH) type 1 receptor (hereafter referred to as PTH₁R) is a G protein–coupled receptor (GPCR) and the canonical cell surface receptor for PTH and PTH-related protein (PTHrP) (1). Secreted as an 84 amino acid peptide, PTH is the quintessential regulator of ionized serum calcium (Ca²⁺) and phosphate (PO₄^3–) levels in the body (Fig. 1A). In response to hypocalcemic conditions when the Ca²⁺ concentration fall below the homeostatic set point in blood, or approximately 1.2 ± 0.1 mM, PTH is secreted from the parathyroid glands and acts in an endocrine fashion by exerting effects on specific bone and kidney cells to adjust the balance of calcium and phosphate ions, and regulate bone remodeling (2, 3). In bone, PTH-mediated activation of the PTH₁R expressed by osteoblasts and osteocytes ultimately promotes bone matrix resorption and liberation of Ca²⁺ into the systemic circulation (4–7). In the kidneys, PTH increases Ca²⁺ reabsorption by distal tubule cells by modulating expression and activity of Ca²⁺ transport proteins (8, 9). Concomitant with this effect of PTH is a decrease in inorganic phosphate reabsorption in proximal tubule cells via reduction of cell surface sodium–dependent phosphate transporters (10–12), which ensures that Ca²⁺ remains in the ionized state. PTH also acts on renal proximal tubule cells to increase circulating levels of active vitamin D3 (VitD) via upregulation of the 25-hydroxyvitamin D3 1α-hydroxylase, and active VitD subsequently serves to increase absorption of Ca²⁺ from the gastrointestinal tract. PTHrP is secreted as a 139, 141, or 171 amino acid protein that is capable of inducing analogous effects on bone and kidney cells observed for PTH, which is not surprising given the shared amino acid sequence homology within the N-terminal region of these peptides that has been shown to be critical for activity (13–15). While it was originally associated as a factor responsible for the hypercalcemia frequently observed in cancer patients, PTHrP is now recognized to play a role in the growth and development of a variety of tissues, including bone (16), mammary glands (17), and teeth (18). PTHrP mostly acts in a paracrine manner on numerous tissues, including bone, growth plate, and mammary gland during development (15, 19). While native PTH is constituted of 84 amino acids and PTHrP consists of splice variants of 139, 141, and 173 amino acids, synthetic or recombinant N-terminal fragments PTH_1-34 (also known as teriparatide) and PTHrP_1-36 maintain full activity and are primarily used in research studies (14, 20–22) (Fig. 1B).

Figure 1.

Functions and sequence of PTH. (A) Main PTH functions in the body. (B) Sequence alignment of PTH₁R peptide ligands. Residues conserved among all the listed PTH₁R peptide ligands are in bold. Residues critical for receptor activation and residues critical for receptor binding are colored. X is α-methylalanine in ABL.

Abnormalities in PTH₁R signaling cause a variety of diseases, previously reviewed in detail (23). In brief, the lack of a functional PTH₁R, such as through expression of homozygous inactivating PTH₁R mutants, causes fetal mortality in Blomstrand chondrodysplasia (24, 25). Heterozygous expression of inactive PTH₁R can cause enchondromatosis/Ollier disease, in which patients develop benign cartilaginous tumors (26), or primary failure of tooth eruption (27, 28). Conversely, heterozygous constitutively active PTH₁R variants with a single mutation in the transmembrane (TM) helices 2 (H223R), 6 (T410P), or 7 (I458R) cause Jansen metaphyseal chondrodysplasia, a very rare disease characterized by short-limbed dwarfism and hypercalcemia (29–31).

Anomalous hormone levels also trigger severe diseases associated with PTH₁R functions. PTHrP hypersecretion from tumoral cells induces excessive PTH₁R activity, leading to humoral hypercalcemia of malignancy encountered in ≈ 25% of cancer patients (32). Oversecretion or undersecretion of PTH causes hyperparathyroidism and hypoparathyroidism, respectively. Hyperparathyroidism treatment typically involves surgical removal of the parathyroid glands (33) or pharmaceutical therapy by calcimimetics (allosteric activators of the calcium-sensing receptor, a class C GPCR) to reduce PTH secretion by parathyroid glands. Patient intolerance and the risk of hypocalcemia associated with parathyroidectomy, however, limit these approaches. Current treatment of hypoparathyroidism includes oral calcium, active VitD, and daily subcutaneous administration of recombinant human PTH_1-34 or PTH (34).

While continuous infusion of PTH was found to promote net bone catabolism (2, 35), intermittent PTH_1-34 injections promote net bone anabolism and are used to treat osteoporosis (36–38). A synthetic analog of PTHrP_1-34, abaloparatide (ABL) (Fig. 1B) was recently developed to treat postmenopausal women with osteoporosis (39). Recent studies have reported that ABL can be more effective than teriparatide at promoting bone formation while reducing the hypercalcemic effect (39–42). The molecular basis for the differing capacity of these ligands to engage a calcemic response in human is unknown but may be related to differences in their binding mode to the PTH₁R. Studies in cultured cells expressing the recombinant PTH₁R show that PTH_1-34 and ABL stabilize different PTH₁R conformations (21, 43). This ligand-dependent PTH₁R conformation selectivity is thought to cause distinct residence time (1/k_off where k_off is the dissociation rate constant) of ligands to the receptor (prolonged for PTH_1-34) (21), which in turn alters the spatiotemporal mode of PTH₁R signaling, an emerging paradigm discussed in the following paragraphs.

The PTHR Signaling Paradigm

GPCR Signaling

All GPCRs share a common structural architecture: a hallmark transmembrane domain (TMD) consisting of 7 membrane spanning α-helices connected by 3 extracellular loops and 3 intracellular loops, N-terminal (extracellular) and C-terminal (cytoplasmic) portions that differ among GPCR families. Based on sequence and functional analyses, human GPCRs are classified into 5 families: glutamate, rhodopsin, adhesion, frizzled, and secretin families, as defined by the International Union of Basic and Clinical Pharmacology (IUPHAR) “GRAFS” classification (44, 45).

Most prominent members of the glutamate receptor family (formerly class C) include metabotropic glutamate receptors, Ca²⁺-sensing receptor, γ-aminobutyric acid B receptors, and type I taste receptors. The N-terminus of these GPCRs is large and contains the ligand recognition domain composed of 2 lobes that form a cavity in which ligands bind, altogether reminiscent of the Venus flytrap and therefore often referred to as a Venus flytrap domain. The Venus flytrap domain may also accommodate allosteric binding sites and is connected to the TMD through a cysteine-rich domain, which also plays a role in receptor activation.

The rhodopsin family (also known as class A GPCRs) has the greatest number of members, which typically exhibit small N-terminal regions (46, 47). It comprises more than 700 receptors that share high sequence similarity (48) and several local structural characteristics with rhodopsin, including an NPXXY motif on TM helix 7 (TM7) and a DRY motif between TM3 and intracellular loop 2 (ICL2) (49). Ligand binding for receptors in this class typically takes place within a cavity formed between TM helices. Notable GPCRs in this class that hold significant physiological relevance include muscarinic acetylcholine receptors, dopamine receptors, adrenergic receptors, opioid receptors, adenosine receptors, and histamine receptors.

The adhesion family has recently attracted much attention due to several unique features. With a sole exception of ADGRA1 (GPR123), these receptors typically possess extremely large extracellular N-terminal domains that contain a highly conserved GPCR autoproteolysis-inducing domain, within which the conserved GPCR proteolysis site is located. Autoproteolysis typically takes place in endoplasmic reticulum and generates 2 fragments, N-terminal fragment and C-terminal fragment, that remain connected via noncovalent interactions. For most adhesion GPCRs, the N-terminal portion of the C-terminal fragment, spanning about 15 amino acid residues, comprises a tethered agonist peptide termed Stachel sequence. In addition to these characteristic adhesion GPCR features, the N-termini of some of these receptors also contain an abundance of serine and threonine residues that constitute sites for O-glycosylation, which act as mucin-like domains to create rigid structures that protrude from the cell surface and are important in cell–cell adhesion.

The frizzled/taste 2 receptor family is far less understood than other classes, and their similarity is largely confined to results from phylogenetic analysis. Only recently have frizzled receptors been established as true GPCRs following reports that binding of Wnt ligand can induce G protein coupling. Frizzled receptors are characterized by long N-termini with conserved cysteine residues that are considered critical for Wnt binding.

The PTH₁R belongs to the secretin family (or class B GPCR) characterized by a long N-terminus (or extracellular domain, ECD) that adopts a conserved fold composed of 2 major elements—an N-terminal α-helix and 2 antiparallel β-sheets—that are connected by several loops and stabilized by 3 disulfide bridges. Class B GPCRs bind relatively large peptide hormone ligands that act most often in an endocrine manner rather than a paracrine fashion, with PTHrP as an exception. There are currently 15 identified members of this family in humans (50).

In the conventional paradigm for understanding GPCR signaling, hormones transmit signals into cells through a series of biophysical and biochemical events that are initiated at the plasma membrane by ligand binding to a receptor (51, 52). This first event stabilizes the active receptor conformation that couples to and activates heterotrimeric G proteins. G protein activation involves release of the guanosine-5′-diphosphate (GDP) molecule bound to the Gα subunit of the G protein and the binding of a guanosine-5′-triphosphate (GTP); in this way, the activated GPCR acts as a guanine exchange factor. Since the concentration of GTP is several fold higher than that of GDP in cells, GTP readily binds to nucleotide-free Gα (53). GTP binding triggers the dissociation or rearrangement of Gα from the Gβγ heterodimer thus permitting Gα-GTP and Gβγ to independently regulate activities of membrane-bound enzymes and ion channels, respectively. There are 4 main classes of Gα subunits that have been identified based on sequence similarity (54): G_s, which stimulates adenylyl cyclase activity to increase intracellular cyclic adenosine monophosphate (cAMP) levels; G_i/o, which inhibits adenylyl cyclase activity; G_q/11, which leads to an increase in intracellular Ca²⁺ levels via activation of phospholipase C (PLC) β; and G_12/13, which regulates Rho guanine nucleotide exchange factors.

Each G protein subunit contains distinct domains that play specific roles during GPCR signaling. The Gα subunit consists of GTPase and flexible α-helical domains (55). In the presence of nucleotides, the GTPase domain binds and hydrolyses GTP to GDP, while the helical domain forms a lid over the nucleotide binding site (53, 56). There are 5 subtypes of Gβ and 12 subtypes of Gγ subunits expressed in humans (57). The Gβ subunit adopts a β-propeller structure with 7 blades along with a long α-helical N-terminus (55), whereas Gγ consists of 2 α-helical segments, which strongly bind to the N-terminal helix and to the blades 1, 4-7 of Gβ. While the specific actions of each subtype combination are under investigation, the Gβ₁γ₂ heterodimer is ubiquitously expressed and regulates many effectors, including several adenylyl cyclase subtypes, ion channels, and phosphoinositide 3-kinase gamma (PI3Kγ) (57, 58). Within seconds after agonist binding, the activated receptor interacts with GPCR kinases that initiate receptor desensitization by (1) preventing G protein coupling (59), and (2) receptor phosphorylation that engages or stabilizes receptor association with β-arrestins (β-arrestin 1 and β-arrestin 2) to promote internalization of the receptor–arrestin complex into endosomes (60–63). Such receptor desensitization leads to the termination of G protein–mediated downstream signaling and reassociation of subunits into the heterotrimeric complex (64). Following receptor internalization to endosomes, the agonist dissociation ensues, and the receptor can be either targeted to lysosomes for degradation (referred to as receptor downregulation) or dephosphorylated and recycled back to the plasma membrane for a new cycle of activation and signaling (referred to as receptor resensitization) (65).

Paradigm Shift in PTH₁R Signaling via G Proteins

The PTH₁R predominantly activates G_s and G_q/11 in response to PTH or PTHrP binding. G_s activates adenylyl cyclases resulting in the production of cAMP and protein kinase A (PKA) activation (1, 66), whereas the G_q/11 (Gα_q, Gα₁₁) subfamily activates phospholipase Cβ (PLCβ), which then cleaves phosphatidylinositol (4, 5)-bisphosphate to generate diacylglycerol and inositol (1,4,5)-trisphosphate (IP3). IP3 activates calcium channels at the endoplasmic reticulum, releasing stored Ca²⁺ into the cytosol, while diacylglycerol activates protein kinase C) The PTH₁R can also activate the G_12/13 subfamily (Gα₁₂, Gα₁₃), which engages the phospholipase D/RhoA signaling pathway (67, 68), and Gi (69).

Back in the 2000s, it was well established that PTH_1-34 infusion stimulates significant increases in serum Ca²⁺, serum vitamin D in its active form (ie, 1,25-dihydroxy vitamin D3), urinary phosphate ion (PO₄^3–), and bone resorption. However, it was unclear why these physiological parameters were not increased at a comparable dose of PTHrP_1-36 infusion given that PTH and PTHrP share similar pharmacological properties as they relate to binding affinity and stimulation of cAMP production (Fig. 2A and 2B) (70).The development and application of quantitative Förster resonance energy transfer (FRET)-based approaches to determine kinetics and rate-limiting reactions for each step along the PTH₁R signaling cascade at the single-cell level (reviewed in (71)) unveiled signaling differences between PTH_1-34 and PTHrP_1-36 and a new mode of GPCR signaling via cAMP. Ligand–PTH₁R interaction and cAMP production measured by FRET showed that PTHrP_1-36 binding is fully reversible and induces transient cAMP after ligand washout, whereas PTH_1-34 remains bound and induces sustained cAMP generation (Fig. 2C). Knowing that PTH induces rapid PTH₁R internalization (72–74), the protracted cAMP response is inconsistent with the conventional model of GPCR desensitization as G protein-dependent signaling by activated receptor is quickly terminated following receptor phosphorylation by GPCR kinases and the recruitment of β-arrestins, which both sterically occlude G protein binding and promote endocytosis of ligand-bound GPCRs. Thus, signaling responses by GPCRs were initially considered transient in nature and confined to the plasma membrane. However, this model has been revised in recent years after the observation that blocking PTH₁R internalization prevents the sustained cAMP response (75, 76) (Fig. 3A and 3B) and that PTH mediates sustained cAMP signaling from early endosomes following internalization of a ternary PTH₁R–Gβγ−β-arrestin complex (Fig. 3C) (77). This unexpected aspect of PTH₁R signaling is due to the apparent ligand-dependent nature of endosomal signaling by this receptor and will be discussed in the next section. Notably, the production of 2 distinct pools of cAMP has been verified for several other GPCRs in class A and class B (76, 78, 79) and reviewed in (80) and is now considered an integral part of GPCR signaling (81).

Figure 2.

Pharmacological properties of PTH₁R. (A, B) Comparison between the competitive binding isotherms (A) and concentration–effect curves for cAMP accumulation in HEK293 cells expressing the recombinant human PTH₁R. (C) Recording real-time binding of peptide ligands to PTH₁R in a single cell. Ligand and receptor interaction is measured by FRET between green fluorescent protein (GFP)-tagged PTH₁R and tetramethylrhodamine (TMR)-labeled peptide ligands. Changes of GFP emission due to a FRET increase in response to rapid superfusion of ligand–TMR (horizontal black bar) are shown. (D) cAMP production over 35 minutes measured by FRET changes from a single HEK293 cells stably expressing PTH₁R and transiently expressing the cytoplasmic cAMP FRET sensor epac1-CFP/YFP and in the absence of phosphodiesterase (PDE) inhibitors. Cells were continuously perfused with control buffer or peptide ligands (horizontal black bar). Images show the propagation of the cAMP production represented as pseudocolored FRET ratio before (green) and after (red) stimulation of a single cell with PTH_1-34.

Figure 3.

Modes of PTH₁R signaling via cAMP. (A) Time courses of cAMP production induced by PTH_1-34 in single cell (here a mouse osteoblast) without (control) or with expression of the recombinant dominant-negative dynamin mutant (DynK44A) that blocks receptor internalization. (B) Confocal images representing a 3D view of HEK-293 cells expressing PTH₁R N-terminally tagged with GFP (PTHR^GFP) with TMR-labeled PTH_1-34 (PTH^TMR). In control cells, PTH₁R^GFP and PTH^TMR colocalized (colored spots) in endosomes; in cells coexpressing Dyn-K44A, the ligand–receptor complex remains at the cell surface. (C) The 1st pool of cAMP production takes place at the cell membrane following PTH binding to PTH₁R (step 1). This response is short-lived due to rapid receptor desensitization via PTHR phosphorylation by GPCR kinases, followed by recruitment of β-arrestins (β-arrs) (step 2) driving receptor internalization and its redistribution in early endosomes (step 3). PTH interacts tightly with PTH₁R in a conformation-dependent fashion and induces the 2nd pool cAMP production from endosomes (step 4). Endosomal cAMP production is prolonged until endosomal acidification induces the release of PTH from the receptor. Created with BioRender.

Mechanisms of Endosomal PTH₁R Signaling via cAMP and its Regulation

Two distinct active PTH₁R conformations are thought to be responsible for the distinct duration of cAMP production (20, 71). The binding affinity of 1 of the conformations, referred to as the R⁰ conformation, is independent of G protein coupling. The R⁰ conformation is preferentially stabilized by PTH and maintains cAMP production when the receptor internalizes in early endosomes. The second conformation, referred to as the R^G conformation, corresponds to the classical G protein–dependent high-affinity receptor, which is indistinguishably stabilized by PTH or PTHrP_1-36. Accordingly, whereas the R^G conformation stabilized by PTHrP_1-36 is G protein dependent and is associated with transient cAMP responses from the plasma membrane, the R⁰ state is not altered by G protein coupling but nevertheless permits sustained cAMP generation following internalization of the receptor (20, 76, 82, 83). The C-terminal portion of PTH (ie, residues 15-34) plays a critical role in receptor selectivity and affinity by binding to the PTH₁R extracellular domain (PTH₁R^ECD) (84, 85). Structural studies revealed a perfect alignment for peptides containing the same C-terminal 15-34 amino acid residues, namely, ^CPTH-type including PTH and ePTH (a mimetic PTH agonist where PTH amino acid residues 1,3,10,11,12,14,18,34 are replaced with aminocyclopentane-1-carboxylic acid, α-aminoisobutyric acid, Gln, homoarginine, Ala, Trp, norleucine, and Tyr, respectively), or ^CPTHrP-type including PTHrP, but C-terminal tips of the ^CPTH-type and ^CPTHrP-type peptides diverge to opposite sides of PTHR^ECD (indicated by arrows in Fig. 4A). Such differential binding mode at PTH₁R^ECD likely translates to differential overall binding to PTH₁R and results in distinct signaling modes. The N-terminal portion, especially the most N-terminal residues, is critical for PTH₁R signaling, as deletion of PTH Ser1 and Val2 significantly reduces cAMP production (86, 87). Also, PTH_7-34 acts as a PTH₁R antagonist (88). From sequence analyses, PTH_1-34 shares several residues and chemical moieties that are not present in PTHrP_1-34 or ABL, which may be critical for stabilizing the R⁰ conformation and inducing sustained cAMP signaling in endosomes; these include Ala1, Ala3, Ile5, Met8, the amide group on residue 10 (Fig. 1B). Additionally, Arg11, Ala12, and Trp14 were initially identified in an alanine scanning to increase activity of a short N-terminal PTH fragment that was referred to as modified-PTH_1-14 (M-PTH_1-14) (89–92). Further optimization of the signaling properties of M-PTH_1-14 resulted in the chimeric peptide M-PTH_1-34, which binds the the R⁰ conformation of PTH₁R with high affinity and trigger prolonged elevation of serum Ca²⁺ (sCa) when injected subcutaneously into mice (82). Additional structure–activity studies to improve the solubility of this peptide by substitution of C-terminal hydrophobic PTHrP residues Leu18, Phe22, and His26 by Ala, Ala, and Lys, respectively, resulted in the [Ala^18,22, Lys²⁶]-M-PTH_1-14/PTHrP_15-36 peptide referred to as LA-PTH, which maintains a higher affinity binding to the PTHR R⁰ conformation than PTH, and sustains endosomal cAMP signaling at time points by which PTH has dissociated from the receptor (93, 94). Given the high affinity and residence time of LA-PTH (43, 83, 95) for PTH₁R, this ligand was used to solve high-resolution cryogenic electron microscopy (cryo-EM) structures of PTH₁R in complex with G_S. This part is discussed under “Structural Characterization of PTH₁R.”

Figure 4.

Structural characterization of PTH₁R activation. (A) Overlaid structures of PTH₁R^ECD in complex with distinct peptides, aligned by PTH₁R^ECD. PTH, red coil (PDB 3C4M); PTHrP, cyan coil, (PDB 3H3G); ePTH, green coil (PDB 6FJ3); LA-PTH, dark blue coil (PDB 6NBF). The N- and C-termini of the peptides are denoted by -NH₂ and -COOH, respectively. (B) Crystal structure of inactive state PTHR bound to ePTH (green coil, PDB 6FJ3). (C) Overlay of inactive (receptor in purple, ePTH in green, PDB 6FJ3) and fully activated (receptor in wheat, LA-PTH in dark blue, PDB 6NBF) PTH₁R structures, aligned by their TMDs. TM6 is in deep purple or orange in inactive and fully active states, respectively. (D) Cryo-EM structure of fully activated LA-PTH-bound (dark blue coil) PTH₁R (state 1) coupled to G_s heterotrimer (PDB 6NBF).

In the case of PTH₁R, a ternary complex composed of PTH-bound PTH₁R, β-arrestin, and Gβγ is required for endosomal cAMP production. The mechanism by which this ternary complex maintains production and elevated levels of cAMP was studied in cells expressing recombinant or native PTH₁R and proceeds as follows (Fig. 5A-5C). First, it promotes the activation cycle of G_s from endosomes (77). Second, it activates extracellular signal–regulated protein kinase 1/2 (ERK1/2) via β-arrestins, thus reducing the activity of the cAMP-specific phosphodiesterase PDE4 (96). Third, it stimulates the conditional activation of adenylate cyclase type 2 via Gβγ (94). Following the recognition of these functional roles, White et al (97) have demonstrated that G_q activation by PTH is determinant in the formation of the ternary PTH₁R–Gβγ−arrestin complex. Gβγ released upon G_q activation promotes the conversion of PI(4,5)P₂ to PI(34,5)P₃ by phosphoinositide 3-kinase β (PI3K_β), which in turn promotes β-arrestin recruitment to PTH₁R and formation of the ternary receptor complex (Fig. 5B, steps 3 and 4).

Figure 5.

Mechanism of endosomal PTH₁R signaling via cAMP. (A) PTH₁R activates heterotrimeric G_s and G_q proteins at the plasma membrane in response to PTH (steps 1 and 2). G_s activates adenylate cyclases (AC) leading to cAMP production and activation of cytosolic protein kinase A (PKA). The 1st pool of cAMP production is short due to the action of phosphodiesterases (PDE). G_q activates phospholipase C (PLC_β) that cleaves phosphatidylinositol (4,5)-bisphosphate (PI(4,5)P₂) into inositol (14,5)-trisphosphate (IP3). Free diffusing IP3 binds and activates IP3-gated Ca²⁺ channels, which in turn release stored Ca²⁺ into the cytosol. (B) Following Gq activation, Gβγ promotes PI3Kβ-dependent generation of PI(34,5)P₃ (step 3) at the plasma membrane that triggers β-arrestin (β-arr) recruitment to the PTHR (step 4) and the formation and internalization of the ternary PTH₁R–β-arr–Gβγ complex. Reassembly of the ternary PTH₁R complex with Gα_s is thought to be dependent on Gα_S diffusion. (C) PTH triggers sustained cAMP production after receptor internalization to endosomes via a β-arrestins (β-arr) and clathrin-coated pit-dependent pathway (2nd pool) (step 6). The sustained phase of endosomal cAMP production is due to the inhibitory action of extracellular signal-regulated protein kinase 1/2 (ERK1/2) on PDE4 (step 7). Endosome-generated cAMP can efficiently diffuse into the nucleus to activate nuclear PKA (step 8). (D) Endosomal cAMP production is prolonged until endosomal acidification caused by v-ATPase activity induces the release of PTH from the receptor (step 9) and assembly of the PTH₁R-retromer complex (step 10) allowing receptor transfer to the Golgi apparatus and receptor recycling. (Created with BioRender after adaptation from Sutkeviciute and Vilardaga (80).

Termination of the sustained cAMP response by the PTH₁R is predominantly controlled by mechanisms instilled within the endosomal trafficking pathway. Previous work has highlighted that the release of β-arrestins from the PTH₁R localized in endosomes is concomitant with association of the retromer complex with the receptor (96) (Fig. 5D). The retromer complex is comprised of 2 membrane-bound sorting nexins (SNX1 and SNX2) as well as a heterotrimer consisting of vesicle protein sorting (Vps) 26, 29, and 35, which collectively serve to regulate the sorting of cargo proteins from early endosomes to the trans-Golgi network (98, 99). Examination of this complex in the context of PTH₁R-mediated cAMP showed that overexpression of Vps subunits reduces the duration of cAMP generation, whereas siRNA targeting of the subunits prolonged cAMP signaling time courses (96). This process was recently shown to be dependent upon a negative feedback loop in which activated PKA increases the proton pump activity of endosomal membrane-bound v-ATPase, resulting in endosomal acidification (100). The decrease in endosomal pH then leads to the disassembly of the signaling PTH₁R–Gβγ–arrestin complex and concomitant assembly of inactive PTH₁R with the retromer, forming a complex that traffics to the Golgi apparatus (96, 100) and/or engages receptor recycling back to the plasma membrane (Fig. 5D) (100).

Structural Characterization of PTH₁R

As for many class B GPCRs, the crystal structure of the isolated PTH₁R^ECD was first resolved either in the absence or presence of ligands (84, 85, 101). These structures demonstrate direct interaction between the C-terminal portion of peptide ligands and the receptor ECD and provide snapshots of the first step in peptide ligand binding to its receptor (Fig. 4A). High resolution structures of PTH₁R were then solved by different approaches concomitantly. An inactive signaling conformation of PTH₁R was crystallized with a mimetic PTH agonist ePTH (102) (Fig. 4B). The receptor is maintained in an inactive state via a dense interaction network (cyan dashes in Fig. 4B) sealing TM6 to the core of the receptor (see detailed in a close-up view, where interacting residues are shown in stick representation). The conserved PxxG motive in the middle of TM6 (highlighted in red with P415^6.42b, using the class B GPCR numbering in superscript, in Fig. 4B) faces outward and away from inactive-state stabilizing network; however, these observations need confirmation given that stabilizing mutations and fusion protein inserted in ICL3 of the receptor construct used for crystallization likely contribute to the inactive receptor state. Three fully active conformations (referred to as states 1-3) of LA-PTH-bound PTH₁R in complex with G_s were determined by cryo-EM (87) (Fig. 6). LA-PTH interactions with PTH₁R involve a continuous α-helix connecting the PTH₁R^ECD and the TMD (PTH₁R^TMD), as seen in other published structures of class B GPCR–G_s complexes (103–106). Specifically, the C-terminal residues 18-34 of LA-PTH are dissociated from the PTH₁R^ECD in the state 3 structure (Fig. 6), suggesting that residues at the C-terminus of the peptide can undergo many dissociation–rebinding cycles while in complex with PTH₁R. Contacts between peptide residues 1-14 and the receptor TMD are likely critical for maintaining active peptide–receptor complex during dissociation and possibly reassociation events of the C-terminal peptide part.

Figure 6.

Structural dynamics of LA-PTH–PTH₁R complex. Overlaid 3 distinct conformational states of LA-PTH–PTH₁R complex obtained from cryo-EM data analysis, aligned by TMDs. PTH₁R and LA-PTH are shown as different shades of purple and orange, respectively. The oscillation of the PTH₁R^ECD is evident (close-up view), while the TMD core shows higher degree of stability, likely due to tight binding of ^NLA-PTH. Notably, state 3 captures the event of partial dissociation of ^CLA-PTH from PTH₁R^ECD (arrow).

From these data, a hypothetical model for sustained endosomal signaling by LA-PTH can be proposed in which the large number of contacts between LA-PTH and PTH₁R^TMD (shown as cyan dashed lines in Fig. 4D, left close-up view) reduce ligand dissociation at endosomal pH, despite dissociation of the peptide C-terminal part. In contrast, the destabilization of PTH–PTH₁R at endosomal pH leads to the release of PTH from the receptor and termination of signaling (100) (step 9 in Fig. 5D).

To date, structures of nearly all class B GPCRs in active G_s-coupled states have been solved (107–117) and display a common structural feature—a sharp outward kink in the middle region of TM6, which opens the receptor cytosolic interface for coupling to G proteins or β-arrestins. Such hallmark feature is also present in active state PTH₁R structures (87) (Fig. 4C and 4D) and is the most pronounced structural rearrangement occurring when the receptor switches from the inactive to its fully active state as highlighted in Fig. 4C. In particular, the P415^6.42b of the conserved class B GPCR PxxG motif faces away from TMD core in an inactive state but is twisted inward in the active state and engages in interactions with Q451^7.45b and V455^7.49b on TM7 (see close-up view in Fig. 4C), overall resulting in partial unwinding of a portion of TM6 above (C-terminal to) the PxxG motif and a large outward movement of the lower part of TM6 N-terminal to the PxxG motif (red continuous arrow in Fig. 4C) (112). This results in formation of a sharp kink in the middle of TM6 of fully active PTH₁R and a large opening of the receptor's cytosolic core for G protein binding. In addition, PTH₁R^ECD may also undergo rearrangement of its position with respect to TMD to adapt the peptide ligand into TMD for receptor activation (top red arrow in Fig. 4C); the type, extent, and stability of such motion might be dependent on the type of peptide ligand agonist.

Other structural rearrangements upon receptor activation also include minor reorganization in the extracellular portions of TM1 and TM7 move toward TM6 in the active state receptor. The other TM helices also exhibit shifts, mostly outward from the receptor core upon activation. These structural rearrangements may contribute to the outward TM6 kink (112). Comparing active state cryo-EM structures with the inactive state crystal structure shows that the first 3 residues of LA-PTH push against TM6, and ligand residue Glu4 stabilizes an extensive polar core network within the receptor (Fig. 4D, left close-up), both of which promote the outward TM6 kink characteristic of receptor activation. The large outward movement of TM6 is required and presumably induced and stabilized by Gα_S subunit binding via interactions of its C-terminal α5-helix (highlighted green in Fig. 4D) with the receptor's cytosolic core. Such profound movement of TM6 to open cytosolic core of class B GPCRs likely is not induced by agonist peptide binding alone. Recent cryo-EM studies of calcitonin gene–related peptide receptor in apo and CGRP-bound states revealed that CGRP-bound and apo receptor structures are highly similar (118). Agonist binding increases structural dynamics at lower halves of TM5 and TM6 in CGRP-bound receptor indicating that agonist binding destabilizes inactive-state interaction networks, and this is required for effective G protein coupling eventually leading to a sharp kink formation in TM6 receptors and full opening of the cytosolic core. Such activation mechanism is likely common among all class B GPCRs.

PTH Binding to PTHR: A Structure-Encoded Allosteric Mechanism

The binding of PTH to PTH₁R proceeds via a 2-step association mechanism (119). The molecular model of this 2-step reaction is displayed in Fig. 7. In the first step, the C-terminal part of the ligand (residues 16-34) rapidly binds to the receptor N-extracellular domain (PTH₁R^ECD) following bimolecular reaction kinetics, defined by k_obs = k_off + k_on[L], where k_obs is the measured rate constant (s^–1), and k_on and k_off are the association and dissociation rate constants, respectively, and [L] is the ligand concentration. In the second and slower step, the N-terminal part of the ligand (residues 1-15) binds to the receptor TMD (PTH₁R^TMD) via a complex reaction involving a bimolecular interaction coupled to conformational changes in both peptide hormone and receptor. This reaction follows a hyperbolic dependence on hormone concentration.

Figure 7.

Expanded 2-step model of PTH binding to PTH₁R. Schematic representation and structural model where PTH₁R is colored wheat, with ECL2 highlighted in dark blue. PTH is colored red, with residues His9, Asn16, and Ser17 displayed as cyan sticks. In the fast first step (step 1), binding of the C-terminal part of PTH to PTH₁R^ECD (step 1a) rigidifies the C-terminal helix as well as residues 6-15 within the N-terminal part of PTH (step 1b). These conformational changes may expand the small N-terminal helix from residues 6-9 to 6-15. In the slow second step (step 2), the N-terminal part of PTH is inserted into PTH₁R^TMD (step 2a). Interactions between the N-terminal portion of PTH and PTH₁R^TMD and ECL2 residues, promote the formation of a continuous PTH helix (step 2b). Adapted from Fig. 2 in Clark et al (120).

NMR experiments titrating unlabeled purified PTH₁R^ECD into a sample of ¹⁵N-labeled PTH_1-34 provided further insights on the molecular mechanism of this complex reaction (120). In brief, the binding of the PTH C-terminal part to PTH₁R^ECD triggers a distinct and more structured conformation in the N-terminal region of PTH_1-34 independent of the interaction with PTH₁R^TMD, which in turn primes this N-PTH region for interaction with PTH₁R^TMD and resulting in a continuous PTH helix conformation. Such placement of the PTH N-terminal part, especially of residues 1-4, stabilizes the active receptor conformation, given that N-terminal PTH residues 1-4 are essential to trigger activation of G_s proteins and cAMP production (87).

Toward elucidating the allosteric changes in PTH₁R structure triggered by PTH binding, we examined the spectrum of conformational motions, or thermal fluctuations (also called Brownian motions), accessible to the PTH–PTH₁R complex embedded in a lipid bilayer (120). The spectrum comprises a wide range of relaxational motions, from low-frequency modes that cooperatively engage the entire structure to those in the high-frequency regime that are highly localized (121). Several studies have demonstrated that the low-frequency modes of motions, robustly defined by the overall architecture, represent cooperative changes in structure that are often required for biological function, including allosteric activation (122). The mode spectrum can be uniquely determined for each structure by normal mode analysis. We analyzed that of the PTH–PTH₁R complex using an extension of the anisotropic network model (123) (ANM) that takes into account the constraints exerted by the lipid bilayer implemented in the DynOmics server (124), and we focused on the most cooperative movements predicted for the complex. Our analysis revealed the predisposition of the complex to a cooperative motion, which enables the “opening” of the intracellularly exposed ends of TM5 and TM6 and the connecting intracellular loop ICL3 (Fig. 8A and 8B), reminiscent of the activated state of class B GPCRs (see Fig. 4D), suggesting that the conformational transition triggered by ligand binding can readily induce a structural change at the cytoplasmic ends of TM5 and TM6.

Figure 8.

PTH₁R collective dynamics, inter-residue cross-correlations, and allosteric communication. (A, B) Collective motions accessible to PTH-bound PTH₁R in the presence of lipid bilayer (not shown). A and B show the respective side and bottom views of 2 conformers visited during fluctuations of the complex along ANM mode 14 that cooperatively engages the entire complex (126). Here, PTH is green, PTH₁R residues G357-L481, orange. Note the large swinging movements at the cytoplasmic ends of TM5-TM6 with respect to helix 8, facilitated by a disruption at TM6. Corresponding animation are in ref. (126). (C) Cross-correlation map between residue motions presented in A and B. The color code refers to the correlation cosines between the directions of motion of all residue pairs (scale bar on the right). Red and blue regions indicate the regions engaged in strongly correlated (coupled, same direction) and anticorrelated (coupled, opposite direction) movements. (D) PTH₁R–PTH complex color coded by the type and strength of correlations of all movements of the residues with the TM5-TM6 cytoplasmic end (derived from the column corresponding to T392 at the ICL3 in C; see the vertical box). Red and blue residues undergo coupled same and opposite direction fluctuations with respect to T392 (or ICL3). G357, T392, and F447 are displayed in space filling. PTH V1-H14 (enclosed in a semitransparent orange ellipse, residues shown in sticks) is strongly correlated with PTHR G357-F447 (red) and the N-terminal (extracellularly exposed) part of TM1, and anticorrelated with F447-L481 (blue). Adapted from Clark et al (120).

To visualize more clearly the coupling between the PTH binding site and the cytoplasmic regions that undergo large rearrangements, we analyzed the cross-correlations between (PTH and PTH₁R) residue motions in this mode. The results are presented as a heat map in Fig. 8C. The red blocks indicate the structural elements that undergo highly correlated movements, and the blue blocks indicate the pairs of structural elements that undergo anticorrelated (opposite direction) fluctuations. We focused on the type and strength of coupling between the PTH N-terminal residues S1-H14 inserted into the PTH₁R^TMD (Fig. 8D; orange ellipse) and the ICL3 connecting TM5 and TM6, using T392 located at the cytosolic tip of TM6, as a reference point for probing conformational rearrangements associated with receptor activation. The heat map clearly showed that PTH V1-H14 was strongly correlated with TM5-ICL3-TM6, as well as the extracellularly exposed region of TM1. Furthermore, examination of the correlation between T392/ICL3 and all PTH₁R residues (vertical rectangular box in Fig. 8C and ribbon diagram in Fig. 8D color-coded by these correlations) showed that T392 belongs to a highly coherent block encompassing the substructure from approximately the mid-part of TM5 all the way to its cytoplasmic end, the entire TM6 and the N-terminal half of TM7 (up to F447), in strong support of the coupling of this entire substructure to PTH S1-H14. This highly coupled region is naturally expected to allosterically transmit the perturbations at the PTH binding site to the TM5-ICL3-TM6 cytoplasmic region.

PTH₁R Druggability

Overall, ANM analysis (or its extension ESSA (125) implemented in the interface ProDy) unveils the structural elements that mediate the allosteric signaling properties of target proteins. Recent application to PTH₁R, in conjunction with druggability simulations, pointed to an allosteric and druggable region in the vicinity of the extracellular ends of TM1 and TM2 (126). In particular, the TM1 residues E180 and R181 were distinguished by their high affinity for binding drug-like probe molecules. Pharmacophore modeling using snapshots from druggability simulations, followed by virtual screening of libraries of small molecules led to the identification of a series of small molecules. One of these small molecules, Pitt12, was experimentally verified to act as a negative allosteric modulator of PTH₁R (126). The binding pose of Pitt12 onto LA-PTH-bound PTH₁R observed in molecular dynamics (MD) simulations, as illustrated in Fig. 9A, displays the particular atoms engaged in highly stabilizing interactions with selected E180 and R181 atoms (Fig. 9B). Notably, a recently resolved positive allosteric modulator of the glucagon-like peptide 1 (GLP-1) receptor (GLP₁R), another Class B GPCR mentioned above, also binds the same site, but this modulator (LSN3160440) inserts slightly deeper into the TMD. The structural alignment of the allosteric modulator binding sites for Pitt12-bound PTH₁R and LSN3160440-bound GLP₁R (Fig. 9C) supports the view that the interfacial region between the N-terminal region of TM1 and the bound peptide or substrate (here LA-PTH and GLP-1) harbors a site with a high propensity to bind allosteric small molecules that might be a common allosteric druggable site among all class B GPCRs. The allosteric effect of ligand binding to this site is consistent with the highly cooperative interaction between the PTH N-terminal residues, the cytoplasm-exposed halves of TM5 and TM6 and connecting loop ICL3, and the N-terminal extracellularly exposed end of TM1 noted in Fig. 8D. Although Pitt12 and LSN3160440 share a similar binding site and pose in their respective receptors, these small allosteric molecules produce opposite modulation of receptor signaling via G_S/cAMP. In the case of LSN3160440, it acts as a positive allosteric modulator by stabilizing the interaction of GLP-1_9-36 for the active GLP₁R conformation (127). In the case of Pitt12, MD simulations predict that it operates by shifting the position of the PTH_1-34 peptide's N-terminal tip resulting in an outward displacement of TM5 and TM6 helices; this conformational rearrangement would be consistent with the negative allosteric modulator activity of Pitt12 in reducing PTH₁R coupling to G protein. Another nonpeptidic molecule, PCO371, acting as agonist for the PTH₁R with a µM potency for the stimulation of cAMP in cells expressing the recombinant at PTH₁R has been previously discovered (127). The binding site is not resolved and initial phase 1 clinical trials to test this molecule as a potential orally available drug candidate for the treatment of hypoparathyroidism (hypocalcemia) have been terminated (128, 129). Acting as a negative allosteric modulator, Pitt 12 is under optimization for improving its efficacy and potency as a potential candidate for future treatment of bone and mineral diseases linked to PTH₁R hyperactivity due to hypersecretion of PTH or PTHrP encountered in hyperparathyroidism or in hypercalcemia of malignancy, respectively.

Figure 9.

Interaction of a negative allosteric modulator with PTH₁R. (A) Results from MD simulations of the binding of an allosteric modulator, named Pitt12, to LA-PTH–bound PTH₁R (PDB 6NBF) (87). The binding pose and structure of Pitt12 were deduced from druggability simulations followed by pharmacophore modeling using Pharmmaker and virtual screening of libraries of small molecules (126). PTH₁R is shown in the wheat ribbon diagram and LA-PTH in green. E180 and R181 (orange sticks) were identified as 2 critical residues that coordinate the small molecule. Pitt12 is shown in teal space filling. (B) 2D representations of Pitt12. Atoms highlighted in colored spheres (at the top) interact with the E180 and R181 atoms shown by the same color in the diagram at the bottom. (C) Comparison of the computationally predicted binding pose of Pitt12 onto PTH₁R, and experimentally resolved (127) binding pose of a positive allosteric modulator, LSN3160440, onto GLP-1R (PDB 6VCB) (127). GLP-1R is shown in light gray, GLP-1 (structurally aligned against LA-PTH) is in dark gray, and LSN3160440 is in yellow. A and B are adapted from Sutkeviciute et al (126).

Location Bias in PTH₁R Signaling and Pharmacological Implications

Sustained endosomal cAMP production has likely important pharmacological and physiological implications, as injection of LA-PTH into mice and monkeys increases serum calcium and decreases serum phosphate more efficiently than does PTH_1-34 (83, 95). Conversely, a PTH mutant (PTH-R25C) found in patients with a severe form of chronic hypocalcemia cannot stimulate endosomal PTH₁R signaling via cAMP in cells and fails to elevate blood Ca²⁺ elevation in animals (21, 130) (Fig. 10A). The recent (2021) design of a novel PTH_1-34 variant contributed to better understand the functional role of location bias (plasma membrane vs endosomes) in PTH₁R signaling independently of the duration of cAMP. The ANM analysis of inter-residues cross-correlations in the PTH-bound receptor (Fig. 8C) let to conceive a PTH analog via epimerization of Leu 7 (PTH^7d) with the desired bias property to sustain cAMP production exclusively from the plasma membrane (131). Findings from studies in cells and mice comparing the effects of the 2 biased PTH agonists, PTH^7d and LA-PTH, which induce a similar elevation and duration of cAMP production but generated in separate cell locations—plasma membrane (for PTH^7d) vs endosomes (for LA-PTH)—helped examine the role of location bias in cAMP production for prime pharmacological functions of the PTH₁R (131) (Fig. 10B). In mice, LA-PTH and PTH7^d injections led to a similar reduction in serum phosphate ion due to inhibition of renal phosphate reabsorption (a major physiological role of PTH acting via PTH₁R to maintain plasma PO₄^3– level constant); however, LA-PTH induces significantly more elevation of serum vitamin D and Ca²⁺ than PTH^7d (Fig. 9B). In polarized Mardin–Darby canine kidney cells, the LA-PTH–generated endosomal cAMP pool ensures nuclear cAMP accumulation and nuclear PKA activity leading to production of the 25-hydroxyvitamin D 1-α-hydroxylase (1α[OH]ase), the rate-limiting enzyme catalyzing the formation of active vitamin D. Conversely, the PTH^7d-generated plasma membrane cAMP pool does not lead to nuclear PKA activity and elevation of the 1α[OH]ase. These observations support possible functional implications of distinct locations of cAMP generation in the control of mineral-ion balance and vitamin D levels.

Figure 10.

Pharmacological actions of biased PTH ligands in mice. (A) Pharmacological actions of LA-PTH and PTH-R25C in cell, mice and monkeys from studies reported in (95, 131). (B) LA-PTH and PTH^7d are biased PTH agonists inducing a similar sustained cAMP production that originates from distinct cell locations: plasma membrane for PTH^7d and endosomes for LA-PTH. When injected into mice, both ligands induce a similar reduction in serum phosphate ion; however, LA-PTH mediates a remarkable increase in serum levels of calcium ion and vitamin D.

PTH₁R Coupling to β-Arrestins

Binding Mode of Arrestin to GPCRs

Arrestins engage with activated GPCR at 2 main sites: (1) a positively charged groove located in the N-terminal domain of arrestin binds to the phosphorylated C-terminal tail (C-tail) of the receptor; (2) several regions, including a finger loop, interact with the receptor core (132, 133). Binding of phosphorylated receptor C-tail to arrestin displaces the arrestin C-tail, disrupting a 3-element interaction and a polar core that stabilize the inactive arrestin conformation (134). These changes promote an active conformation in which the N- and C-terminal domains are rotated by 20°. The arrestin finger loop adopts a helical conformation upon engagement with the receptor core (135). Residues in the finger loop, lariat loop, and elsewhere on arrestin form hydrophobic and polar interactions with the receptor core. In comparison with an arrestin structure bound to a synthetic receptor phosphopeptide, the structure of rhodopsin (the light-sensing GPCR)-bound arrestin (135) exhibits unique loop conformations and a small change in the interdomain twist. While binding to the receptor C-tail was previously considered a prerequisite for engagement with the receptor core (132), recent research suggests that arrestin can also engage with the receptor core independent of receptor C-tail engagement (136, 137). Extensive MD simulations of rhodopsin–arrestin-1 complex determined that interactions between arrestin loops, not including the finger loop, with receptor intracellular loops (ICLs) 2 and 3 are the main mediators of arrestin activation via the receptor core. Furthermore, both receptor core and phosphorylated C-tail binding trigger arrestin activation by promoting distinct conformations of arrestin C-loop and lariat loop. While the molecular mechanism of arrestin C-tail displacement is not clear, the displaced C-tail binds clathrin and clathrin adaptor AP2, permitting receptor internalization (62, 137, 138).

PTH₁R–Arrestin Interaction

Part of the contact residues between PTH-activated PTH₁R and β-arrestin-1 (βarr1) have been identified using photo-crosslinking coupled to mass spectrometry experiments (120, 139). Crosslinks were identified between F75 of βarr1 and PTH₁R residues V384^5.64b and T392^5.71b, located at the cytoplasmic end of TM5 (120). These photo-crosslinks support a conformation in which the βarr1 finger loop is engaging with the receptor cytosolic core as observed in the crystal structure of rhodopsin bound to the murine visual arrestin (140).

To visualize these photo-crosslinks, models of PTH₁R bound to βarr1 were generated using MD snapshots of PTH₁R (120) and 3 structures of GPCR–arrestin complexes: rhodopsin–visual arrestin (PDB 5W0P, model 1) (135), M2 muscarinic receptor–βarr1 (PDB 6U1N, model 2) (141), and neurotensin receptor 1–βarr1 (PDB 6UP7, model 3) (142) (Fig. 11). PyMOL was used to perform structural, sequence-independent alignment of the TMD of the GPCR–βarr1 template structure to the PTH₁R^TMD (residues 180–460) from each snapshot. In the model of PTH₁R–βarr1 complex generated using rhodopsin–visual-arr (model 1) (135), C-edge loops are slightly embedded in the membrane (Fig. 11A), as observed in other GPCR–arrestin structures (135, 141–143). The βarr1 in model 3 clashes significantly with the membrane (Fig. 11C), and interactions between PTH₁R helix 8 residues and the membrane are eliminated (Fig. 11D and E). Model 1 satisfies distance restraints between βarr1 F75 and PTHR V384^5.64b/T392^5.7b that are well within photo-crosslinking restraints (Fig. 11A). The inward movement of receptor TM6 relative to the G_s-bound conformation has been reported essential for βarr coupling (144). The PTH₁R–βarr1 model permits this TM6 inward movement without clashing with βarr1. Interface in this PTH₁R–βarr1 complex (model 1) includes many contacts with ICL1 and ICL2, more so than in model 2 (Fig. 11B).

Figure 11.

Position of β-arrestin 1 relative to the membrane in PTH₁R–βarr1 models. In all panels, PTH is cyan, PTH₁R is dark green, and β-arrestin 1 is slate blue. The orange spheres represent membrane phosphates after 200 ns MD simulation. In A and B, yellow spheres are β-arrestin 1 C-edge loop residues that contact the membrane. (A) Model 1: PTH₁R–βarr1 model generated using the rhodopsin–arrestin structure (PDB 5W0P) as a template. (B) Model 2: PTH₁R–βarr1 model with M2R–βarr1 template structure (PDB 6U1N). (C-E) Model 3: PTH₁R–βarr1 model with NTSR1–βarr1 template structure (PDB 6UP7). (C) Top. Model 3 with same orientation as A and B. Bottom. Model 3 rotated 90° to highlight significant embedment of β-arr1 C-edge in the membrane. In D and E, PTH₁R alone and model 3 positioned in membrane (red spheres) by the OPM server. Helix 8 residues from PTH₁R alone interacting with the membrane in panel D are represented as lime sticks. Such interactions are eliminated in model 3.

Next Questions to Address

Ongoing basic research is providing important insights into structural mechanisms of PTH₁R signaling in response to PTH and PTHrP, including ligand binding, receptor activation, and coupling to G_s and β-arrestins. Since its initial discovery, significant advancements have been made toward elucidating molecular and cellular determinants that underlie PTH₁R endosomal cAMP. Even so, several outstanding questions remain. Although we made models of PTH₁R–βarr based on available structures, conformational heterogeneity in GPCR–arrestin complexes has been observed (142). Cryo-EM analysis of 2D class averages from purified PTH₁R–βarr could reveal possible conformations of this complex. Importantly, it is not clear what are the structural differences between PTH₁R–arrestin complexes at the plasma membrane versus those in endosomes (ie, supercomplex with both β-arrestin and G_s bound). What structural features contribute to PTH₁R coupling to G_q and G_12/13? Also, how does PTHrP/ABL stabilize unique conformations of PTH₁R that lead to transient cAMP production at the plasma membrane? Recent studies addressed in part this question by showing that PTHrP_1-36 does not mimic signaling mediated by the native PTHrP_1-141, which induces sustained cAMP at the plasma membrane due to impaired β-arrestin coupling (145, 146). These data motivate a reinterpretation of our prior understanding on how native PTHrP acts on the PTH₁R. Additional structural and cellular studies will also provide a better understanding of the differences in signaling observed between PTH₁R ligands. Do different ligands trigger distinctive membrane lipids dynamics that result in either receptor internalization or retention of receptors at the plasma membrane? Does the ternary PTH₁R–Gβγ–arrestin complex internalize into endosomes whose lipid composition differs when PTH₁R is activated by either PTH_1-34 or LA-PTH? If there is such a difference, does it contribute to the prolonged PTH₁R–LA-PTH interaction that engages sustained cAMP? Furthermore, research is needed to determine whether endosomal cAMP production directly regulates bone turnover. Answering these questions will help explain the differential effects of PTH₁R ligands on bone turnover, calcium levels, and vitamin D homeostasis and could serve as a foundation for the development of clinically relevant PTH analogs tailored to target specific aspects of PTH₁R signaling involved in bone and mineral diseases.

Abbreviations

ABL

abaloparatide

cAMP

cyclic adenosine monophosphate

ECD

extracellular domain

ENM

elastic network model

FRET

Förster resonance energy transfer

GDP

guanosine-5′-diphosphate

GLP-1

glucagon-like peptide 1

GLP₁R

GLP-1 receptor

GPCR

G protein–coupled receptor

GTP

guanosine-5′-triphosphate

ICL

intracellular loop

IP3

inositol (1,4,5)-trisphosphate

MD

molecular dynamics

PI3K

phosphoinositide 3-kinase

PLC

phospholipase C

PTH

parathyroid hormone

PTH₁R

PTH type 1 receptor

PTHR

PTH receptor

PTHrP

PTH-related protein

TM

transmembrane

TMD

transmembrane domain

VitD

vitamin D3

Funding

Redaction of this chapter was supported by the National Institute of Diabetes and Digestive and Kidney Disease (NIDDK), the National Institute of Dental and Craniofacial Research (NIDCR), and the National Institute for General Medical Sciences (NIGMS) of the US National Institutes of Health (NIH) under grant Awards Numbers R01DK116780, R01DK122259, and R21DE032478 (to J.-P.V.), and R01GM139297-01A1 (to I.B.)

Disclosure

I.S., J.Y.L., I.B. and J.-P.V. acknowledge potential competing financial interests for Small molecule allosteric modulators of class B GPCR and method to identify them.