A Model for Predicting Cation Selectivity and Permeability in AMPA and NMDA Receptors Based on Receptor Subunit Composition

Abstract

Glutamatergic AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) and NMDA (N-methyl-D-aspartate) receptors are implicated in diverse functions ranging from synaptic plasticity to cell death. They are heterotetrameric proteins whose subunits are derived from multiple distinct gene families. The subunit composition of these receptors determines their permeability to monovalent and/or divalent cations, but it is not entirely clear how this selectivity arises in native and recombinantly-expressed receptor populations. By analyzing the sequence of amino acids lining the selectivity filters within the pore forming membrane helices (M2) of these subunits and by correlating subunit stoichiometry of these receptors with their ability to permeate Na+ and/or Ca2+, we propose here a mathematical model for predicting cation selectivity and permeability in these receptors. The model proposed is based on principles of charge attractivity and charge neutralization within the pore forming region of these receptors; it accurately predicts and reconciles experimental data across various platforms including Ca2+ permeability of GluA2-lacking AMPARs and ion selectivity within GluN3-containing di- and tri-heteromeric NMDARs. Additionally, the model provides insights into biophysical mechanisms regulating cation selectivity and permeability of these receptors and the role of various subunits in these processes.

Visualizing the triheteromeric N-methyl-D-aspartate receptor subunit composition

“…The discovery and cloning of the GluN3 subunits (GluN3A-B), the final members of the NMDAR family, augmented the functional diversity of conventional glutamate-activated GluN1/GluN2(A-D) containing NMDARs by introducing unconventional glycine-activated GluN1/GluN3(A-B) NMDARs that were originally thought to form relatively Ca 2+ -impermeable cation channels and be expressed presynaptically ( Das et al, 1998 ; Chatterton et al, 2002 ; Perez-Otano and Rodriguez-Moreno, 2019 ; Crawley et al, 2022 ). We showed previously that the GluN3 subunit can combine with GluN1 and GluN2 (A and/or B) to make glutamate activated t -NMDARs which are distinguishable from GluN2-containing d -NMDARs electrophysiologically (have excitatory postsynaptic currents with markedly different current-voltage relationships), have reduced affinity for Mg 2+ , and increased selectivity for Ca 2+ over Na + , making them highly Ca 2+ permeable ( Pilli and Kumar, 2012 , 2014 ; Beesley et al, 2019 , 2020b ; Kumar and Kumar, 2021 ). These receptors are blocked by the pan-NMDAR antagonist D-(-)-2-Amino-5-phosphonopentanoic acid (D-AP5) and by D-serine, a potential gliotransmitter and a co-agonist of conventional NMDARs ( Kumar, 2016 ; Beesley et al, 2019 , 2020a ).…”

“…An upshot of this work is the possibility of visually analyzing pathology underlying neurodegenerative disorders like temporal lobe epilepsy (TLE) from the synaptic/receptor perspective. For example, we have previously shown how t -NMDARs, by virtue of their increased selectivity for Ca 2+ render neurons vulnerable to excitotoxic damage and contribute to the pathology (vulnerability and pattern of neuronal loss) and by extension to the pathophysiology (Ca 2+ -induced excitotoxicity) underlying TLE ( Beesley et al, 2020a ; Kumar and Kumar, 2021 ). By assaying the spatial expression of their subunit proteins (GluN1, GluN2A, GluN2B, and GluN3A) using area-specific tissue analysis (ASTA), a novel methodology for harvesting brain chads from hard-to-reach regions within brain slices for Western blotting, we recently showed that GluN3A was expressed in a gradient along the mid-lateral extent of layer three MEA and along the CA1-subicular axis in the hippocampus, unlike GluN1 and GluN2A which were uniformly distributed.…”

“…Triheteromeric NMDARs, by contrast, contain three different types of subunits (e.g., GluN1-2A-1-2B), and include receptors that are composed of one or more subunits from each of the three gene families, designated t -NMDARs ( Kumar, 2016 ) (e.g., GluN1-2A-3A-2A). We showed previously that GluN3-containing t -NMDARs in the brain can be distinguished from GluN2-containing d -NMDARs electrophysiologically, have reduced affinity for Mg 2+ and increased selectivity for Ca 2+ over Na + , making them highly Ca 2+ permeable ( Pilli and Kumar, 2012 ; Beesley et al, 2020b ; Kumar and Kumar, 2021 ). These receptors are blocked by the pan-NMDAR antagonist D-(-)-2-Amino-5-phosphonopentanoic acid (D-AP5) and by D-serine, a potential gliotransmitter and a co-agonist of conventional NMDARs ( Kumar, 2016 ; Beesley et al, 2019 , 2020a ).…”