Glutamate is the main neurotransmitter molecule in the brain responsible for communicating excitatory signals between brain cells. This communication is mediated through glutamate-activated receptor proteins embedded in the membranes of brain cells. One of these receptors, metabotropic glutamate receptor 5 (mGlu5), is crucial for learning and memory, and is an attractive target for the treatment of several psychiatric and neurological disorders. Writing in Nature, Koehl et al.1 report the first essentially full-length structures of mGlu5, which they obtained by using a combination of X-ray crystallography and cryo-electron microscopy (cryo-EM). These structures reveal how activation of the receptor alters its multidomain structure to initiate cell signalling.
G-protein-coupled receptors (GPCRs) are the largest superfamily of receptors found in cell membranes, and are also the largest group of drug targets2. All GPCRs have seven α-helices that span the cell membrane, collectively known as the 7TM domain. Class C GPCRs — the group to which mGlu5 belongs — differ from other types in that they must form a dimeric complex to function, and because they have a large extracellular amino terminus.
The binding site for the naturally occurring activators (agonists) of class C GPCRs is found in the N terminus, and is referred to as the Venus flytrap (VFT) domain because it is formed by two lobes. Most of these receptors also contain a domain that is rich in cysteine amino-acid residues, and this links the VFT to the 7TM domain. How the binding of an agonist in the VFT domain transmits a signal over a long distance (more than 120 ångströms) within the mGlu5 dimer to promote the active conformation of the 7TM domain has been unknown.
Until now, that is. Koehl et al. describe two X-ray crystal structures of dimeric mGlu5 VFT domains in complex with a positive allosteric modulator (a molecule, in this case a small antibody known as a nanobody, that stabilizes the binding of agonists) — one with and one without a synthetic agonist. These structures show that the binding of the agonist causes the two lobes of the mGlu5 VFT domain to close. The closed conformation strongly resembles that observed in the crystal structure of the mGlu5 VFT domain bound to its natural agonist, glutamate3.
The authors then obtained cryo-EM structures of dimeric mGlu5 that incorporate all of the receptor’s major domains, apart from the intracellular carboxy terminus. In these structures, the agonist-free VFT domains adopt a similar conformation to that seen in the equivalent X-ray crystal structure. The cysteine-rich domain forms a stalk that holds the VFT more than 55 Å above the 7TM domain, and each of the two 7TM domains are separated by more than 20 Å (Fig. 1).
Koehl et al. used two methods to reconstitute purified, full-length mGlu5 for their cryo-EM experiments, each generating a different receptor conformation for the mGlu5 dimer. The images of mGlu5 produced using both methods suggest that the agonist-free receptor has minimal or weak interactions between the cysteine-rich and transmembrane domains across the dimer. This agrees with the results of a previous biophysical study4 of the isolated 7TM domains of mGlu5 and of the related mGlu2 receptor. However, it contrasts with experiments5 in which proximal amino-acid residues across the dimer in agonist-free mGlu2 were identified on the basis of whether covalent crosslinks could be formed between those residues. Koehl and co-workers also used cryo-EM to visualize the structure of full-length mGlu5 in which the VFT domains were bound to an agonist, using positive allosteric modulators (a nanobody bound to the VFTs, and a small molecule bound to the 7TM domain) to stabilize this conformation of the receptor.
It was already known that the mGlu2 receptor must form a dimeric structure to enable its activation by glutamate, stimulating coupling of the receptor to a G protein4 and thereby triggering signalling in the cell. A comparison of the full-length mGlu5 structures in which the VFT domains are agonist-bound (activated) and agonist-free (inactivated) provides insight into how the activation of dimeric class C GPCRs transmits structural changes throughout the entire protein.
The cryo-EM structures show that closure of the VFT domain brings the cysteine-rich stalk and the 7TM domain of each subunit in the dimer closer together (Fig. 1), with the 7TM domains rotating, such that one of the α-helices (known as TM6) of each mGlu5 monomer forms a new interface between the two monomers. It is worth noting that the 7TM domains in the full-length receptors were reconstituted in different media (a nanodisc of lipids or a micelle of detergents) for each structure, and this might have influenced the relative orientation and proximity of the 7TM domains within the dimer. However, the authors carried out crosslinking experiments that provide more evidence of the structural changes proposed to occur on receptor activation, and further support comes from previously published studies5–7 of other class C GPCRs.
Koehl et al. also carried out experiments to examine the effects of mutations to mGlu5 on its activation mechanism. Their results suggest that an interaction between the cysteine-rich stalk and a region of the 7TM domain known as the second extracellular loop (ECL2) governs activation by agonists that bind to the VFT domain, but not activation by agonists that bind to the 7TM domain. The conformation of this loop modelled by the authors is similar to that observed in the X-ray structure8 of the 7TM domain of the related mGlu1 receptor in complex with an inhibitor. ECL2 is known to influence activation states in other GPCRs9, but Koehl and colleagues’ study provides the first indication that it also has a key role in mediating interdomain communication in class C GPCRs. However, the relatively low resolution of the new structures prohibits meaningful comparisons of the interactions between ECL2 and the cysteine-rich stalk in the inactive and active conformations of mGlu5. Whether these regions constitute targets suitable for drug discovery also remains an open question.
The resolution of the 7TM domains in both conformations is also insufficient to visualize the small-molecule inhibitors or activators that were used in the purification and reconstitution of the agonist-free and agonist-bound receptor structures, respectively. It remains to be seen how the structures of binding pockets in the 7TM domain change in the presence of inhibitors or activators. Indeed, the resolution of the 7TM domains is lower than those previously obtained10–12 for structures of the 7TM domains of mGlu5 bound to inhibitors. The structure of receptors in complex with an intracellular effector (such as a G protein) will be required to stabilize, and therefore visualize at high resolution, active 7TM conformations, and to understand how allosteric modulators binding to the 7TM domain alter the activation states, to enable the structure-guided design of drugs that target glutamate receptors.
Koehl and colleagues’ structures reveal the large-scale conformational changes that occur in a dimeric, full-length, class C GPCR when an agonist binds to the N terminus. The strategies used to stabilize full-length proteins will inform efforts to obtain the structures of other class C GPCRs, including receptors for ions and the inhibitory neurotransmitter GABA, as well as for receptors involved in taste. An appreciation of how these dimeric, multidomain receptors are organized should inform our understanding of how receptor complexes composed of two or more different class C GPCRs, or from different GPCR classes, are formed and activated. These structures might also guide future protein engineering of class C GPCRs to enable the identification of pockets that can be targeted by drugs, and might ultimately open up avenues of research for structure-guided drug discovery.
Nature 566, 42-43 (2019)