Energetics of the Activated Receptor-G Protein Interaction
By: Carol A. Rouzer, VICB Communications
Published: December 13, 2013
A model of the interaction between activated rhodopsin and Gi reveals the conformational changes in the Gα subunit that lead to GDP release.
Heterotrimeric G proteins play a critical role in the transduction of signals from the very large class of seven transmembrane G protein-coupled receptors (GPCRs). Comprising α, β, and γ subunits, the G protein binds to an agonist-activated receptor, and this interaction triggers exchange of GDP for GTP in the GTPase domain of the G protein’s α subunit. The GTP-bound α subunit dissociates from the βγ subunits, and these now separate species activate a range of effectors, most often enzymes or ion channels. Over time, the GTPase activity of the α subunit hydrolyzes the bound GTP to GDP, ending the subunit’s ability to modulate the activity of effector proteins. The GDP-bound α subunit then reunites with βγ, preparing it to interact with another activated receptor protein (Figure 1). Numerous laboratories have worked to identify the exact mechanism by which binding of the Gα subunit to an agonist-bound receptor results in GDP-GTP exchange. A crystal structure of the agonist-activated β2 adrenergic receptor bound to Gs (β2AR-Gs)(S. G. F. Rasmussen, et al. (2011) Nature, 477, 549) has contributed substantially to our understanding of the process. However, the crystal structure provides only a static picture of the receptor-G protein interaction. This led Vanderbilt Institute of Chemical Biology members Jens Meiler and Heidi Hamm and their laboratories to combine the crystal structure with experimental data to develop a dynamic model of the G-protein-receptor interaction [N. S. Alexander, et al. (2013) Nat. Struct. Mol. Biol., published online December 1, DOI: 10.1038/nsmb.2705, copyright (2013)].
Figure 1. Mechanism G protein-mediated signal transduction. Binding of an agonist to a G protein-coupled receptor leads to a conformational change that allows the α subunit of the G protein to bind to the receptor. This induces a conformational change in the G protein that promotes exchange of GTP for GDP in the α subunit. The GTP-bound G protein separates into α and βγ subunits, which activate effector molecules (here shown as adenylate cyclase and a calcium channel). Hydrolysis of GTP to GDP by the GTPase activity of the α subunit inactivates its effector modulating activity and promotes reassociation with the βγ subunits to restart the cycle. Image reprinted by permission from Macmillan Publishers Ltd: [S. G. F. Rasmussen, et al. (2011) Nature, 477, 549, copyright (2011)].
The investigators used the β2AR-Gs crystal structure data as a framework to create a model of activated rhodopsin (a well-characterized GPCR) interacting with Gi (R*-Gi). Similarly, they used crystal structure data for Gt (D. G. Lambright, et al. (1996) Nature, 379, 311) to construct a model of free Gi. Prior experiments using double electron-electron resonance (DEER) to measure the distances between key residues in free and receptor-bound Gi informed the model creation process, which was carried out using the Rosetta software suite.
The Gα subunit comprises a GTPase domain and a helical domain, with GDP bound between them. The GTPase domain contains a six-stranded β-sheet surrounded by six α-helices, which are numbered α1 through α5 plus αG. A single large α-helix (αA), surrounded by five shorter helices (αB through αF) constitute the helical domain. An N-terminal helix (αN) points away from the Gα subunit (Figure 2). The researchers carefully compared their models of R*-Gi and free Gi to identify key interactions within Gi and between Gi and activated rhodopsin (Figure 3). The results indicated that binding of the Gα subunit to the receptor occurred through an interaction of the Gα subunit’s C-terminal α5 helix and the receptor’s transmembrane helices that required a 5.7 Å translation and 63o rotation of α5. This movement induced conformational changes in the loop between α5 and the β6 strand, the α1 and αG helices, and the β-strands of the GTPase domain of Gα. The changes were accompanied by a weakening of the interactions between the GTPase and helical domains, leading to an opening of the interface between the two domains and the release of GDP.
Figure 2. Ribbon diagram of the crystal structure of the heterotrimeric G protein transducin (Gt). The α subunit (green) comprises a GTPase domain and a helical domain. The GTPase domain is composed of a six-stranded beta sheet (β1 through β6) surrounded by six α-helices (α1 through α5 plus αG). The helical domain contains a long central helix αA surrounded by five shorter helices (αB through αF). The N-terminal helix (αN) projects away from the α subunit. The β and γ subunits are shown in yellow and red, respectively. Image reprinted by permission from Macmillan Publishers Ltd: [D. G. Lambright, et al. (1996) Nature, 379, 311, copyright (1996)].
Figure 3. Diagrammatic representation of the energetics of key interactions within regions of the free Gα subunit (left) and between the Gα subunit and the receptor in the receptor-bound state (right). Gray indicates the GTPase domain, green the helical domain, yellow the α5 helix, and orange, the receptor transmembrane domain. The thickness of the arrows corresponds to the magnitude of the energy of interactions as indicated by the key. Energy values were calculated from the molecular model and are in Rosetta energy units. Image reprinted by permission from Macmillan Publishers Ltd: [N. S. Alexander, et al. (2013) Nat. Struct. Mol. Biol., published online December 1, DOI: 10.1038/nsmb.2705, copyright (2013)].
The β2AR-Gs crystal structure supports the conclusion that the Gα-receptor interaction leads to an opening of the interface between the GTPase and helical domains of Gα. However, the crystal structure shows a much larger movement of the helical domain than was predicted by the model or prior DEER data (Figure 4). The investigators used the rigid body docking protocols in the Rosetta program to explore a wide range of positions of the helical domain of receptor-bound Gα. This approach generated 739 models, some of which agreed with the β2AR-Gs crystal structure closely. However, when the researchers selected the nine models most consistent with the DEER data, none placed the Gα helical domain in the same position as in the β2AR-Gs crystal structure. The DEER data indicated that the helical domain is quite flexible, so the investigators concluded that the position of that domain in the β2AR-Gs crystal structure represents a poorly populated conformation that was stabilized during crystallization.
Figure 4. Ribbon diagrams of (a) the crystal structure of the β2AR-Gs complex, (b) the model of the R*-Gi complex, and (c) the model of the Gi heterotrimer derived from the transducin crystal structure (Figure 2). The receptor is shown in orange, the Gα helical domain in green, the Gα GTPase domain in gray, the Gα α5 helix in yellow, the Gβ subunit in light brown, and the Gγ subunit in black. Note the striking difference in the position of the Gα helical domain in structure (a) as compared to structure (b). Image reprinted by permission from Macmillan Publishers Ltd: [N. S. Alexander, et al. (2013) Nat. Struct. Mol. Biol., published online December 1, DOI: 10.1038/nsmb.2705, copyright (2013)].
The researchers compared the results of prior single-particle electron microscopy analysis of the β2AR-Gs complex to the structure predicted by their model and found good agreement. They also used continuous wave electron paramagnetic resonance spectroscopy (CW-EPR) and peptide amide hydrogen-deuterium exchange mass spectrometry data to calculate changes in the solvent accessibility of key residues in the free versus receptor-bound Gα subunit. The results of these calculations agreed well with predictions based on the model. To further test the model’s validity, they selected an amino acid in the GTPase domain and another in the helical domain of Gi. They expressed a Gi protein bearing mutations in both of these amino acids to cysteine and then attached a sulfur-accessible nitroxide probe to each cysteine. The altered protein allowed them to use DEER to assess the changes in the distance between the GTPase and helical domains upon interaction of Gα with the activated receptor. The results, using two different pairs of residues, were consistent with the helical domain movement predicted by the model. Placement of nitroxide probes at two sites, one in the GTPase domain and the other in the linker between the two domains, allowed the investigators to use EPR measurements to assess solvent accessibility at these two sites. Again, the results for the free and receptor-bound states of Gα agreed with model predictions. Finally, they used the model to select three residues that play an important role in interactions within the Gα subunit and one residue that is critical for the interaction between Gα and the receptor. They mutated each of these residues to cysteine and measured the effect of the mutation on GDP exchange. The results were in good agreement with predictions based on model calculations.
Together, the data suggest that the model of the Gi-activated rhodopsin interaction designed by the Meiler and Hamm groups provides a strong framework to explore the energetics that drive this dynamic process. The model not only agrees with available prior data, it has also successfully predicted the results of new experiments. The insights the model provides concerning the effects of receptor binding on Gα conformation (which are beautifully illustrated in a video available in the paper’s supporting information) will provide an invaluable guide to our future understanding of GPCR signaling.