Research Article

Structure and organization of heteromeric AMPA-type glutamate receptors

See allHide authors and affiliations

Science  29 Apr 2016:
Vol. 352, Issue 6285, aad3873
DOI: 10.1126/science.aad3873

You are currently viewing the abstract.

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution

Signaling at the synapse

Neurons signal to each other at synapses using neurotransmitters. Glutamate is a key neurotransmitter, and AMPA-type glutamate receptors (AMPARs) mediate rapid responses to glutamate release. These receptors mainly occur as heteromers comprising GluA1-4 subunits. Herguedas et al. used electron microscopy and x-ray crystallography to determine the structure of GluA2/3 and GluA2/4 heteromers. The structures differ from those determined previously for GluA2 homomers but emphasize how signals may be transmitted through these dynamic receptors.

Science, this issue p. 10.1126/science.aad3873

Structured Abstract


Ionotropic glutamate receptors (iGluRs) are cation channels that mediate signal transmission by depolarizing the postsynaptic membrane in response to L-glutamate release from the presynaptic neuron. There are three major iGluR subtypes, and together their activity is pivotal to learning and memory. The first subtype, the AMPA-type receptor (AMPAR), initiates signaling and activates the second subtype, the N-methyl-d-aspartate receptor (NMDAR), thereby admitting Ca2+ ions that drive synaptic plasticity and ultimately enabling learning. The rapid submillisecond response of the AMPAR permits accurate transmission of high-frequency presynaptic impulses. In the brain, AMPARs preferentially exist as heterotetramers composed of the GluA1 to GluA4 subunits in various combinations. Their signaling properties are dominated by the GluA2 subunit, which is present in the majority of AMPARs, yet the architecture and subunit arrangement of GluA2-containing AMPAR heteromers has thus far been elusive.


So far, AMPAR structures have been limited to GluA2 homomers. We used a combination of x-ray crystallography and cryogenic electron microscopy (cryo-EM) to obtain structural information for AMPAR heteromers. We first solved crystal structures of the N-terminal domain (NTD) layer, which constitutes the more sequence-diverse upper half of the receptor. Our structures of the GluA2/3 and GluA2/4 NTDs provide atomic resolution of the interface that initiates subunit assembly and reveal tetrameric arrangements in both of these heteromeric combinations that are distinct from known homomers. Cysteine cross-linking confirmed this architecture in full-length receptors and permitted us to determine cryo-EM structures of an intact GluA2/3 heteromer, which is a prominent variety in neural tissue.


In both GluA2/3 and GluA2/4 NTD crystal structures, the four subunits alternate around a central axis to create a compact O shape that deviates from the loose N shape commonly found in GluA2 homomers. Simulations suggest that the receptor can transition between these two states. Structure-guided cysteine mutants allowed us to trap receptors in the O state and facilitated further analysis by cryo-EM. The GluA2/3 receptor was captured in a ligand-free state and resulted in two models, determined at 8.2 and 10.3 Å. Both deviate from existing GluA2 homomer structures. Associated with the compact conformation of the NTD layer was a substantial vertical compression of the receptor heteromer, forming distinct interlayer interfaces. Contact points between the NTD and the ligand-binding domain (LBD) suggest that the NTD can contribute to signaling. We also observed a rearrangement of the gating machinery at the level of the LBD layer, and our results provide a structural snapshot of a desensitized receptor in the absence of agonist. Our data indicate that subunit placement in AMPAR tetramers is not strict, unlike in NMDARs, a finding that may extend the functional spectrum of heteromeric AMPARs.


The GluA2/3 structure and our simulations emphasize that AMPARs are dynamic complexes. The observed rearrangements of the extracellular region are expected to be of functional consequence in synaptic environments, where the NTD engages various interaction partners. Transitioning between N and O states could break existing contacts and make new ones, thereby affecting AMPAR clustering, a prerequisite for synaptic plasticity. Moreover, the vertical compression observed in GluA2/3 would permit crosstalk between the NTD and LBD, thus incorporating the NTD in allosteric regulation, as has been observed in NMDARs. The sequence-diverse AMPAR NTD, like that of NMDARs, may therefore offer a target for drug development—one that is as yet unexplored.

Structure of a GluA2/3 AMPAR heteromer.

A GluA2/3 heteromer (left) is shown with a GluA2 homomer (right) for comparison. The top panel shows the NTD layers of the compact O-shaped GluA2/3 and the loose N-shaped GluA2 [Protein Data Bank identifier (PDB ID) 3H5V], viewed from the top. The bottom panel shows full-length cryo-EM structures of GluA2/3 (ligand-free) and a GluA2 homomer (antagonist-bound; PDB ID 4UQJ). The difference in vertical packing between the NTD and LBD is indicated by arrows.


AMPA-type glutamate receptors (AMPARs), which are central mediators of rapid neurotransmission and synaptic plasticity, predominantly exist as heteromers of the subunits GluA1 to GluA4. Here we report the first AMPAR heteromer structures, which deviate substantially from existing GluA2 homomer structures. Crystal structures of the GluA2/3 and GluA2/4 N-terminal domains reveal a novel compact conformation with an alternating arrangement of the four subunits around a central axis. This organization is confirmed by cysteine cross-linking in full-length receptors, and it permitted us to determine the structure of an intact GluA2/3 receptor by cryogenic electron microscopy. Two models in the ligand-free state, at resolutions of 8.25 and 10.3 angstroms, exhibit substantial vertical compression and close associations between domain layers, reminiscent of N-methyl-d-aspartate receptors. Model 1 resembles a resting state and model 2 a desensitized state, thus providing snapshots of gating transitions in the nominal absence of ligand. Our data reveal organizational features of heteromeric AMPARs and provide a framework to decipher AMPAR architecture and signaling.

View Full Text