New Insights into G-Protein-Receptor Coupling
Published: June 10, 2011
The combination of double electron-electron spin resonance EPR, biochemical studies, and molecular modeling reveal the intricacies of G-protein function.
A large number of hormones and neurotransmitters act by binding to and activating a G protein-coupled plasma membrane receptor. Formed as a complex of three subunits, Gα, Gβ, and Gγ, G proteins transmit the signal from an activated (ligand bound) receptor to an intracellular effector molecule, usually an enzyme. The mechanism by which G proteins communicate receptor signaling involves the interconversion of the Gα subunit between four different binding states (Figure 1). In the absence of an activated receptor, the Gα subunit binds GDP in its nucleotide binding site. This inactive complex (designated Gα(GDP)) has a high affinity for the combined, membrane-associated Gβγ subunit. Binding to this subunit forms the complete heterotrimeric G protein complex (Gα(GDP)βγ), and places the Gα subunit in close proximity to the receptor. Ligand binding to the receptor leads to receptor activation and formation of a complex between the receptor and the G protein. Then, a conformational change in the Gα subunit is accompanied by release of GDP to form the empty complex (Gα(0)βγ). Subsequent binding of GTP to the nucleotide binding site of Gα activates the subunit (Gα(GTP)), which dissociates from the complex. Subsequent binding of Gα(GTP) to a target effector enzyme modulates the enzyme’s activity. Slow hydrolysis of GTP to GDP returns Gα to its inactive state (Gα(GDP)) to recapitulate the cycle.
Figure 1. Different binding modes of Gα. In its inactive form, the Gα subunit (dark blue) binds GDP and then forms a heterotrimeric complex with Gβγ (dark pink). Interaction with an activated receptor (gray) leads to reduced affinity for and release of GDP to form the empty complex (light blue). GTP then binds to Gα, yielding the activated form (dark orange), which dissociates from the receptor complex to interact with target proteins. Reproduced from Van Eps, et al. (2011) Proc. Natl. Acad. Sci. U.S.A., published online May 23, DOI: 10.1073/pnas. 1105810108. Copyright 2011, Van Eps, et al.
Although the first Gα crystal structure was published nearly two decades ago, and we have a clear understanding of how this subunit cycles through its various binding states, the mechanism by which Gα interacts with an activated receptor is only poorly understood. Now, a team of researchers, including Vanderbilt Institute of Chemical Biology members Heidi Hamm and Jens Meiler along with Wayne Hubbell at the University of California at Los Angeles, combine double electron-electron resonance (DEER) EPR spectroscopy, biochemical techniques, and state-of-the-art molecular modeling to provide new insight into this important question [Van Eps et al. (2011) Proc. Natl. Acad. Sci. U.S.A., published online May 23, DOI: 10.1073/pnas.1105810108].
The crystal structure of the Gα subunit, published by Hamm with Paul Sigler, revealed two major domains, a nucleotide binding domain and a helical domain. The N-terminus of the protein is myristoylated, for anchoring to the membrane, while the α5 helix of the C-terminus interacts with the activated receptor (Figure 2). The crystal structure publication predicted that interaction with an activated receptor must trigger some movement of the nucleotide binding and helical domains that allows release of GDP from its binding site. To explore this possibility, the investigators first constructed a mutant protein by removing all reactive cysteine residues from a Gα subunit of the Gαi class. They verified that this protein retained the capacity to bind and exchange GDP and to respond to an activated receptor. They then selected pairs of amino acids that were fairly close to each other but located in different domains (one in the nucleotide binding domain and one in the helical domain). For each pair, the selected amino acids were mutated to cysteine, and then the protein was incubated with S-(1-oxy-2,2,5,5-tetramethylpyrroline-3-methyl-methanethiosufonate) to introduce a disulfide-linked spin label (Figure 3) at each cysteine residue. Five such proteins were constructed, each bearing a different pair of spin labeled amino acids. In each case, the investigators confirmed the ability of the protein to bind and exchange GDP and to respond to an activated receptor (rhodopsin that had been exposed to light).
Figure 2. Model based on the crystal structure of Gαi, showing the nucleotide domain (gray), the helical domain (green), GDP (magenta), the Cterminal α5 helix (yellow), and the paired spin label sites (connected with dotted lines). Reproduced from Van Eps, et al. (2011) Proc. Natl. Acad. Sci. U.S.A., published online May 23, DOI:10.1073/pnas.1105810108. Copyright 2011, Van Eps, et al.
The presence of the spin labels allowed the investigators to use DEER to measure the distance between each pair of cysteine residues. They carried out these measurements for each labeled protein in all four binding states, Gα(GDP), Gα(GDP)βγ, activated receptor-bound Gα(0)βγ, and Gα(GTP). The data revealed that the distance between the pairs of spin-labeled amino acids remained similar for all of the states of the protein with one exception. In the case of the empty state (Gα(0)βγ), the two labeled amino acids were farther apart (by up to 20 Å) and exhibited much greater positional flexibility than was observed in any of the other states.
Figure 3. Structure of the spin label bound to cysteine via a disulfide bridge.
The DEER results suggested that a shift from Gα(GDP)βγ to Gα(0)βγ was associated with a substantial conformational change that moved the nucleotide domain and the helical domain apart. To visualize how this might occur, the investigators constructed a molecular model based on two previously published X-ray crystal structures, one of Gαi(GDP)βγ and the other of opsin (the activated form of rhodopsin) bound to the α5 domain of a Gα subunit from the Gαt class. Models constructed from the crystal structure data were merged to position the G protein in correct orientation relative to the receptor. The investigators then refined the model to place the protein complex in correct orientation relative to the membrane. With this accomplished, the structure could then be altered to match the amino acid side chain distances measured for the Gα(0)βγ structure as determined from the DEER experiments. Of 140,000 different modeled structures, 1,000 met a set of predefined criteria based on the DEER data. Cluster analysis of these 1,000 structures provided a unified model that represented the best overall fit of the experimental data. The results (Figure 4) revealed that conversion of Gα(GDP)βγ to Gα(0)βγ led to a translation of 8 Å and a rotation of 29o between the nucleotide binding and the helical domains. This conformational change provided an opening for egress of GDP from the nucleotide binding site.
Figure 4. Molecular model of Gαi, showing the nucleotide domain (gray), the helical domain (green), GDP (magenta), the C-terminal α5 helix (yellow), and Gβγ (tan and black). The structure on the left shows the G protein complex in the presence of an inactive receptor (orange). On the right, the activated receptor binds the G protein, triggering a major conformational change in Gαi, leading to the release of GDP. Reproduced from Van Eps, et al. (2011) Proc. Natl. Acad. Sci. U.S.A., published online May 23, DOI: 10.1073/pnas.1105810108. Copyright 2011, Van Eps, et al.
Previous data from the Hamm and Hubbell laboratories using a mutant Gα containing a flexible loop in the C-terminal α5 helix suggested that the α5 helix plays an important role in promoting the conformational change that occurs during the transition from the Gα(GDP)βγ to the Gα(0)βγ state. Indeed, placement of a flexible linker composed of five glycine residues between amino acids 343 and 344 of that helix eliminated receptor-mediated GDP exchange. Presumably, the flexible helix prevents the movement of the α5 helix that occurs upon receptor binding from being transferred to the remainder of the protein. DEER analysis showed that Gαi proteins containing the flexible linker exist permanently in an equilibrium between two conformations similar to those observed for Gα(GDP)βγ and Gα(0)βγ, and that receptor binding no longer affects the position of this equilibrium. Chemical cross-linking of the amino acids at positions 90 and 238, across the helical and nucleotide binding domains, prevented the domain rearrangement and also eliminated receptor-mediated GDP exchange. This finding provided key evidence that GDP release requires the opening between the two domains.
Together the data form the framework for understanding the mechanism by which receptor interactions mediate nucleotide exchange in the G protein Gα subunit, a process that is key to the function of the vast array (estimated 800 human genes) of G-protein coupled receptors.