Conformational states of the full-length glucagon receptor

Abstract

Class B G protein-coupled receptors are composed of an extracellular domain (ECD) and a seven-transmembrane (7TM) domain, and their signalling is regulated by peptide hormones. Using a hybrid structural biology approach together with the ECD and 7TM domain crystal structures of the glucagon receptor (GCGR), we examine the relationship between full-length receptor conformation and peptide ligand binding. Molecular dynamics (MD) and disulfide crosslinking studies suggest that apo-GCGR can adopt both an open and closed conformation associated with extensive contacts between the ECD and 7TM domain. The electron microscopy (EM) map of the full-length GCGR shows how a monoclonal antibody stabilizes the ECD and 7TM domain in an elongated conformation. Hydrogen/deuterium exchange (HDX) studies and MD simulations indicate that an open conformation is also stabilized by peptide ligand binding. The combined studies reveal the open/closed states of GCGR and suggest that glucagon binds to GCGR by a conformational selection mechanism.

Figure 1. EM analysis of glucagon receptor–mAb23 complex.

(a) Exemplary two-dimensional (2D) class average (sum of 343 individual particles) of negatively stained GCGR–mAb23 complex (left) and corresponding 3D surface representation of 3D map (∼30 Å resolution) determined using EM random conical tilt methods. 3D map is shown (right) in similar orientation to the 2D average (centre) and also rotated into an orientation convenient for comparison with the X-ray map. White scale bar, 50 nm. (b) Schematic interpretation of the domains in the EM map, rotated into an orientation convenient for comparison. (c) 3D envelope of the EM map (left), molecular model of the mAb23-bound full-length glucagon receptor structure based on mAb1-bound ECD (PDB code: 4ERS) and 7TM (PDB code: 4L6R) crystal structures (middle) and the molecular model fitted into the EM map (right). (d) View of panel c rotated 90° clockwise. Additional information on the EM maps is provided in Supplementary Fig. 1.

Figure 2. Stabilization of GCGR by peptide ligand in the HDX studies.

Changes to average percent deuterium are shown on the full-length GCGR model based on ECD and 7TM crystal structures. Dark blue regions of receptor indicate areas of large decreased exchange in the presence of the des-His¹-[Nle⁹-Ala¹¹-Ala¹⁶]-glucagon-NH₂ peptide ligand (depicted as green ribbon) and cyan indicates regions with smaller decreased exchange, while black indicates no significant change and white indicates regions where no peptide ions were detected using mass spectrometry. HDX plots for selected regions are shown around the structure. The data are shown as mean±s.d. of three independent experiments. Average percent deuterium values and percent deuterium values at 10 s are reported in Supplementary Fig. 4 and Table 1, respectively.

Figure 3. Motions of the ECD with respect to the 7TM domain in the simulations of glucagon-GCGR and apo-GCGR.

(a) Definitions of the Cartesian coordinate system, the polar angle (θ, the included angle between vector OC and axis z), the Azimuthal angle (φ, the included angle between the OC projection on the xy plane and axis x) and the distance between the COMs of the ECD and 7TM domain (d) for describing the motions of the ECD with respect to the 7TM domain in the average structure in the simulation on the glucagon-GCGR complex. (b) The probability map of the MD snapshots with θ and φ as coordinates. The probability of the most abundant conformation is set to 1 and the relative probabilities of other conformations with respect to this conformation are shown. There are two states with higher probabilities (with θ and φ in areas circled by dotted lines): the open state (marine, represented by conf_open) that can be stabilized by glucagon (semitransparent green cartoon) and the closed state (red, represented by conf_closed). (c–e) Time dependences of θ, φ and d in the MD simulations on apo-GCGR and glucagon-GCGR and indication of conf_open and conf_closed MD snapshots.

Figure 4. Comparison of open glucagon-bound and -closed apo-GCGR structures.

(a) Representative snapshots of glucagon-bound GCGR (conf_open, blue) and apo-GCGR (conf_closed, red) in the MD simulations (defined in Fig. 3). Regions investigated in HDX studies (see Fig. 2 and Table 1) are coloured dark (red/blue) and shown in more detail in (b) the top region of the stalk; (c) the N terminus of the ECD; (d) ECL1; and (e) ECL3. Residues involved in specific interactions or adopting different conformations in glucagon-bound and apo-GCGRs are depicted as sticks.

Figure 5. Conformational states of the ECL3 chimera in the MD simulation.

(a) Orientations of the ECD with respect to the 7TM domain in the simulation on the ECL3 chimeric apo-GCGR. Time dependences of the polar angel (θ, defined in Fig. 3) in the simulation are shown at the bottom and typical snapshots taken from specific periods of the trajectories are displayed at the top. Comparison of conformations of the TM1 stalk region (b) and interaction between ECD and ECL3 (c) in the closed-like structure in the ECL3 chimera (green) with the closed state in the wild-type apo-GCGR (red). The snapshot at 300 ns in the simulation on the chimera and the structure of conf_closed (Fig. 3) are used. Residues with different conformations in the two structures are depicted as sticks and C_α–C_α distance between R94 and E371 is labelled. For clarity, mutations are labelled in orange.

Figure 6. Transition between the two states and intervention by disulfide crosslinking studies.

(a) Modes 1 and 2 of the NMA on the open structure (top, orange arrow), and modes 3 and 4 of the NMA on the closed structure (bottom, yellow arrow) based on representative open state (conf_open) and closed state (conf_closed) structures (Fig. 3). The vectors representing both the amplitudes and directions of residues during the conformational changes are mapped on the ECD. (b) Selected residue pair H372^ECL3–H89^ECD for cysteine substitution is shown in the closed state of apo-GCGR, with the average Cβ–Cβ distance in the last 1,000 ns simulation. (c) Time dependences of the Cβ–Cβ distance between H372^ECL3 and H89^ECD in the MD simulations on glucagon-GCGR (blue) and apo-GCGR (red). (d) MS/MS spectra of the HCD fragmentation of the doubly charged disulfide-containing peptide are shown; b, y, B and Y indicate types of fragment ions. Graphical fragment map correlates the fragmentation ions to the peptide sequence in which the disulfide-linked cysteine residues C89 and C372 are shown in red. The top-right panel shows a LC-MS analysis-extracted ion chromatogram of GCGR from Spodoptera frugiperda (Sf9) cells with chymotrypsin and trypsin digestion, representing the doubly charged crosslinked peptide between YLPWHC(89)K and AFVTDEC(372)AQGTLR through a disulfide bond.

Figure 7. HDX studies of open peptide ligand-bound wild-type GCGR versus closed apo H89C/H372C mutant GCGR.

Changes to average percent deuterium of extracellular regions observed in HDX studies are shown on full-length models of peptide ligand-bound wild-type GCGR (left) and closed H89C/H372C mutant GCGR (right). The N-terminal region of the ECD (N-term) and TM1 stalk are depicted in dark blue, indicating large decreased exchange in the presence of the des-His¹-[Nle⁹-Ala¹¹-Ala¹⁶]-glucagon-NH₂ peptide ligand (green), while the smaller decreased exchange of the ECL1 region is coloured cyan. The ECD is coloured magenta, while residues C89 and C372 are coloured yellow. Differences in average percent deuterium values are reported in Supplementary Table 2.