NMDA receptors: linking physiological output to biophysical operation

Abstract

NMDA receptors are preeminent neurotransmitter-gated channels in the CNS, which respond to glutamate in a manner that integrates multiple external and internal cues. They belong to the ionotropic glutamate receptor family and fulfil unique and crucial roles in neuronal development and function. These roles depend on characteristic response kinetics, which reflect the operation of the receptors. Here, we review biologically salient features of the NMDA receptor signal and its mechanistic origins. Knowledge of distinctive NMDA receptor biophysical properties, their structural determinants and physiological roles is necessary to understand the physiological and neurotoxic actions of glutamate and to design effective therapeutics.

Figure 1. iGluR family members have similar structures, but distinctive output

(A) Structural models of two iGluR representatives reveal overall similarity between NMDA and AMPA receptor architectures; both have large ectodomains composed of N-terminal (NTD) and ligand binding (LBD) domains, a short transmembrane domain (TMD), and cytoplasmic C-terminal domains (CTD), the latter being structurally unresolved. (B) Stereotypical map and electrical responses of central excitatory synapses. Top, glutamate molecules released from presynaptic vesicles diffuse across the synaptic cleft, bind to post-synaptic iGluRs to produce the EPSC, and to extrasynaptic glutamate receptors and transporters (EAAT). Bottom left, the typical EPSC has two components, which can be separated pharmacologically into AMPA (red) and NMDA (blue) currents, each with characteristic time-dependent amplitudes. The NMDA receptor-mediated component is visibly longer; its decay kinetics set the EPSC decay timecourse. Bottom right, simultaneous recording of NMDAR-EPSC and corresponding rise in Ca²⁺ fluorescence in a single synaptic spine^,. (C) NMDAR-EPSCs differ in maximal amplitude and kinetics across synapse development: hippocampal synapse at P10 compared with P30 (top)^,; cellular type: cerebellar compared with spinal synapse (middle)^,; and synaptic state: spinal synapse before and after induced neuropathic pain (bottom). All traces represent simulations based on values from the cited reports.

Figure 2. Observable features of the NMDA receptor output

(A) Macroscopic currents elicited with brief synaptic-like stimuli (1 ms, 1 mM Glu) are described by their peak amplitude (I_pk) and decay kinetics (τ_d). Decay kinetics are quantified by fitting declining mono-exponential (τ_d) or bi-exponential functions (τ_f, τ_s) to the declining phase of the current. (B) Macroscopic currents elicited with long nonsynaptic-like stimuli (>2 s, 1 mM Glu) are described by peak (I_pk) and steady-state (I_ss) current amplitudes measured directly from the record; desensitization kinetics, determined by fitting declining mono-exponential functions (τ_D) to the response between I_pk and I_ss; and desensitization extent, calculated as the I_ss/I_pk ratio. (C) Microscopic currents are recorded as downward deflections from a zero current baseline (top) and expanded (below) to allow direct measurement of current amplitude distributions (i) (bottom right) and duration for each opening (t_o) and closure (t_c); responses elicited with brief synaptic-like stimuli (1 ms, 1 mM Glu) also reveal the receptor’s latency to the first opening (t₁). (D) Complex probability distributions of open (t_o) and closed (t_c) durations (thick lines) observed in each record yield quantitative information about channel output, including the number and lifetime of kinetic components determined by fitting multiple exponential functions to the data (thin lines).

Figure 3. Models of NMDA receptor operation

(A) Minimal model (left) required to describe macroscopic currents includes two sequential glutamate binding steps to resting (C^U) and monoliganded (C^M) receptors, followed by one-step isomerization reactions: gating (C—O) and desensitization (C—D). Minimal model (right) required to describe one-channel currents also includes two sequential glutamate binding steps, followed by a complex gating sequence (C-C-C-O-O)^–, and two desensitization steps (C—D)^,. (B) The conceptual model fits well (simu, purple) the rise and decay of recorded synaptic-like macroscopic current (patch, black); but not the distribution of closures observed in one-channel currents (patch, black). The statistical model (model, blue) estimates well both the synaptic-like current and the single-channel event distributions. Both models predict mono-exponential decay for the synaptic-like current and are therefore too simple to fully account for NMDAR-EPSC. (C) Traces were simulated in response with the ‘average’ statistical models reported for GluN1/GluN2A, GluN1/GluN2B, GluN1/GluN2C, and GluN1/GluN2D receptors; they predict substantial differences in peak open probabilities, overall kinetics, and charge transferred in response to synaptic-like (left) and non-synaptic-like (right) glutamate transients (thick black lines).

Figure 4. Insights from statistical models

(A) Tiered model accounts for modal gating, defined as spontaneous transitions between patterns of activity with distinguishable rates. Periods of low, medium, and high gating have been identified statistically and separated in one-channel records to estimate mode-specific rate constants^,,,,; (B) Synaptic-like unitary responses (top, black) recorded successively from the same GluN1/GluN2A receptor can be grouped by their pattern of opening consistent with modal gating. The sum currents for each kinetic group have distinct peak amplitudes and decay kinetics. (C) Statistical models derived from modes observed in GluN1/GluN2A and GluN1/GluN2B single-receptor recordings predict distinct peak open probabilities, and decay times that span several orders of magnitude. Therefore, the complex decay kinetics observed for NMDAR-EPSCs and for synaptic-like NMDA receptor currents may reflect molecular (subtype) as well as kinetic (modes) heterogeneities of the activated receptors.

Box 1

Recording synaptic NMDA receptor signals

Box 2

Recording non-synaptic NMDA receptor signals

Box 3

Conceptual models of ligand-gated ion channels