Molecular recognition of parathyroid hormone by its G protein-coupled receptor

Abstract

Parathyroid hormone (PTH) is central to calcium homeostasis and bone maintenance in vertebrates, and as such it has been used for treating osteoporosis. It acts primarily by binding to its receptor, PTH1R, a member of the class B G protein-coupled receptor (GPCR) family that also includes receptors for glucagon, calcitonin, and other therapeutically important peptide hormones. Despite considerable interest and much research, determining the structure of the receptor-hormone complex has been hindered by difficulties in purifying the receptor and obtaining diffraction-quality crystals. Here, we present a method for expression and purification of the extracellular domain (ECD) of human PTH1R engineered as a maltose-binding protein (MBP) fusion that readily crystallizes. The 1.95-A structure of PTH bound to the MBP-PTH1R-ECD fusion reveals that PTH docks as an amphipathic helix into a central hydrophobic groove formed by a three-layer alpha-beta-betaalpha fold of the PTH1R ECD, resembling a hot dog in a bun. Conservation in the ECD scaffold and the helical structure of peptide hormones emphasizes this hot dog model as a general mechanism of hormone recognition common to class B GPCRs. Our findings reveal critical insights into PTH actions and provide a rational template for drug design that targets this hormone signaling pathway.

Fig. 1.

Purification and function of the MBP-PTH1R-ECD fusion. (A) Overview of the method, in which an MBP-ECD fusion protein is coexpressed with DsbC in the oxidizing cytoplasm of an E. coli trxB gor host. Affinity purification of the fusion protein yields multiple protein species, e.g.: (i) multimers linked by intermolecular disulfide bonds; (ii) misfolded monomer; (iii) monomeric, correctly folded protein. The mixture is subjected to in vitro disulfide shuffling in a redox buffer in the presence of DsbC to increase the percentage of the total protein present as species iii, which can be purified to homogeneity. (B) Gel analysis of MBP-PTH1R-ECD-H₆ at various stages of purification. Lanes 1, 3, and 4 show samples from affinity purification via the His₆ and MBP tags, with lane 4 containing 20 mM DTT. Lane 5 shows a disulfide shuffling reaction incubated overnight at 20°C in the presence of 1 mM each GSH and GSSG, and DsbC. Lanes 2 and 6 show the final purified sample after removing DsbC and remaining misfolded fusion protein. (C) AlphaScreen assay for the association of biotin-PTH-(7–34)-NH₂ with the purified MBP-PTH1R-ECD at 20°C. The reaction mixtures contained 15 μg/ml each streptavidin-coated donor beads and Ni-chelate-coated acceptor beads, and protein and peptide as indicated. The data represent the average of duplicate samples. (D) AlphaScreen assay for the ability of PTH(15–34)-NH₂ to compete with the interaction of biotin-PTH(7–34)-NH₂ and MBP-PTH1R-ECD-His₆. The data represent the average of three independent experiments, and the curve yielded an IC₅₀ value of 3.0 μM. (E) Isothermal titration calorimetry analysis of PTH-(15–34)-NH₂ binding to the MBP-PTH1R-ECD at 27°C. (Upper) Raw data after subtraction of the peptide heats of dilution. (Lower) Enthalphy vs. the PTH to MBP-PTH1R-ECD-His₆ ratio. The best fit of the data yielded a K_d value of 0.98 μM; number of binding sites, 0.985; ΔH = −20,190 calories per mol; and ΔS = −39.8 calories per mol per degree.

Fig. 2.

Structure of the PTH–PTH1R ECD complex. (A and B) Two views of the complex with the ECD in green, PTH in yellow, and the disulfide bonds depicted as sticks. The dashed red line designates the break in the chain caused by the disordered loop between residues 57 and 105. For clarity, MBP is not shown in this and all subsequent figures. All structural figures were prepared with PyMol (55). (C) Overview of the three-layer α–β–βα structure of the PTH1R ECD with PTH removed. (D) Residues that stabilize the ECD fold. The carbon atoms are colored based on amino acid sequence conservation among the 15 human class B GPCR ECDs. Cyan indicates residues that are invariant, magenta indicates residues that have conservative substitutions, and green indicates residues that are not conserved. Hydrogen bonds are depicted as red dashes, and the red sphere is a water molecule. (E) Sequence alignment of human class B GPCR ECDs with the secondary structural elements noted over the sequences. Invariant residues are shown in white lettering on a red background, and conservative substitutions are shown in red lettering.

Fig. 3.

Structural and biochemical analysis of the PTH–ECD interaction. (A) Detail of the PTH–ECD interface. The PTH backbone is shown as a yellow coil and selected side chains as sticks. The ECD is shown as a green ribbon diagram covered by a transparent molecular surface. ECD atoms within 8 Å of PTH are shown in stick representation. Hydrogen bonds are indicated by red dashes, and the red sphere is a water molecule. (B) Molecular surface of the ECD showing the hydrophobic patch that contacts W23, L24, L28, and V31 of PTH. Carbon atoms are colored gray, nitrogen atoms blue, and oxygen atoms red. (C) The σA weighted 2F_o − F_c electron density map for PTH shown as a blue mesh contoured at 1 σ. (D) Interactions between PTH and the ECD where ECD residues are boxed. The lines indicate hydrophobic or van der Wall contacts, and the arrows indicate hydrogen bonds from donors to acceptors. (E) AlphaScreen assay for the ability of alanine scan mutants of PTH-(15–34)-NH₂ (100 μM) to compete with the binding of the biotin-PTH-(7–34)-NH₂ (40 nM) to the MBP-PTH1R-ECD (40 nM). The data represent the average of duplicate samples. (F) Sequence comparison of human PTH, PTHrP, and bovine TIP39.

Fig. 4.

Comparison of the PTH1R ECD–PTH complex with the mouse CRFR2β ECD–astressin and human GIPR ECD–GIP complexes. (A) Structural alignment between the PTH1R ECD–PTH complex and the NMR solution structure of the mouse CRFR2β ECD bound to astressin (PDB code 2JND). C^α backbone traces are shown. The structures were aligned with PyMol based on the ECD structures. The PTH1R ECD is green and PTH is yellow; the CRFR2β ECD is colored red and astressin blue. (B) Alignment with the crystal structure of the GIPR ECD–GIP complex (PDB code 2QKH). The color scheme is as in A.