Computational analysis of interactions between AMPA receptor and con-ikot-ikot conotoxin

Abstract

AMPA receptors (AMPARs) are a subtype of ionotropic glutamate receptors (iGluRs) found in excitatory synapses in the central nervous system, where they mediate the majority of excitatory synaptic transmission. AMPARs are encoded by 4 different genes, GRIA1–4, whose protein products assemble into functional tetramers. In vivo, AMPARs can be present as homomers, diheteromers and triheteromers. Each of these AMPAR subtypes has specific kinetics, expression profile and hence a distinct physiological role. Synapses involved in information storage change their AMPAR subunit composition as well as the total number of AMPARs. This phenomenon is known as synaptic plasticity and is widely recognised as the molecular mechanism underlying the processes of memory formation and learning. However, to date, there are no known compounds that specifically recognise one AMPAR subunit composition over the other.

Con-ikot-ikot (CII), a naturally occurring conotoxin found in the venom of a marine cone snail Conus striatus, specifically binds to AMPARs with EC50 of ∼5 nM and exhibits unique potential to serve as a scaffold to generate novel, heteromer-specific AMPAR binders and modulators. Its unique binding site, localised within the AMPAR extracellular domains, where it makes contacts with all 4 subunits, might provide a starting point for design of heteromerspecific binders. At the same time, its small size, minimal perturbation of AMPAR size (about 1 nm increase in receptor height), 1:1 binding stoichiometry and the fact that it leaves AMPAR extracellular domains open for physiological interactions make it an attractive, novel labelling tool for fluorescent studies of AMPARs, including super-resolution imaging (SRI).

SRI was essential in demonstrating activity-dependent changes in AMPAR numbers within synapses. In these experiments, AMPARs are labelled with a fluorescent tag either by genetic fusions (e.g. GFP or Halo tag) or by fluorescently labelled antibodies. While overexpression of genetic fusions can lead to non-physiological conditions, antibodies are large (10–15 nm in size) and almost all bind to the surface of AMPAR extracellular domains. This could significantly hinder free movement of AMPARs into and out of the already crowded and narrow synaptic space, where AMPAR extracellular domains interact with other synaptic and presynaptic proteins. CII toxin has the potential to overcome all these limitations.

Here, I explore the interactions between CII toxin and AMPARs using computational approaches followed by experimental validation of the results. I have performed a series of molecular dynamics simulations and developed NAppEd – a novel approach to analyse the resulting trajectories based on residue interaction network analysis. Using NAppEd in combination with computational alanine scanning, I have confirmed residues known to be important for the interaction between the toxin and AMPAR and identified new interactions. Importantly, this includes a network of novel interactions between the toxin and the AMPAR amino-terminal domains (ATDs). This interaction network is of particular importance as ATD sequence varies between AMPAR subunit types. My analysis identifies CII as the only known non-antibody AMPAR binder that interacts with AMPAR ATDs. The results of computational analyses were validated by mutating CII tyrosine 54 into alanine (Y54A), identified as one of the strongest interactors with AMPARs by NAppEd. The single-point mutant resulted in a right-shift in CII dose response curve in electrophysiology experiments, validating the computational findings.

These results, along with additional computational analyses showing minimal effects of fluorescent labelling on CII binding to AMPARs, confirm the unique potential of the CII as a novel fluorescent probe and as a scaffold for design of AMPAR-heteromer specific binders for investigation of AMPARs.