Distinct functional and pharmacological properties of Triheteromeric GluN1/GluN2A/GluN2B NMDA receptors

Abstract

NMDA receptors are tetrameric ligand-gated ion channels comprised of GluN1, GluN2, and GluN3 subunits. Two different GluN2 subunits have been identified in most NMDA receptor-expressing cells, and the majority of native receptors are triheteromers containing two GluN1 and two different GluN2. In contrast to diheteromeric NMDA receptors, little is known about the function of triheteromers. We developed a method to provide selective cell-surface expression of recombinant GluN1/GluN2A/GluN2B triheteromers and compared properties of these receptors with those of GluN1/GluN2A and GluN1/GluN2B diheteromers. We show that glutamate deactivation of triheteromers is distinct from those of GluN1/GluN2A and GluN1/GluN2B and reveal modulation of triheteromers by subunit-selective antagonists ifenprodil, CP-101,606, TCN-201, and extracellular Zn(2+). Furthermore, kinetic measurements suggest variation in the ifenprodil binding site of triheteromers compared to GluN1/GluN2B diheteromers. This work provides insight into the distinct properties of GluN1/GluN2A/GluN2B triheteromers, which are presumably the most abundant NMDA receptors in the adult forebrain.

Figure 1

Control of NMDA receptor composition using engineered peptide tags. (A) Linear representations of polypeptide chains show the amino terminal domain (ATD), S1 and S2 segments of the agonist binding domain, transmembrane helices (M1, M3, and M4), and the reentrant pore loop (M2) of GluN2A, GluN2B, GluN2A with C1- or C2-tags fused to the intracellular C-terminus (2A_CX), and chimeric GluN2B subunits with the C-terminal domain replaced by that of C1- or C2-tagged GluN2A (2B_ACX). (B) Co-expression of GluN1 (omitted from cartoon for clarity) with two different GluN2 subunits in heterologous systems would normally generate three populations of functional NMDA receptors. However, fusing the C1 and C2 tags to the intracellular C-termini of GluN2 subunits should not allow cell-surface expression of receptors that contain two C1-tagged or two C2-tagged GluN2 subunits. (C) Representative two-electrode voltage-clamp recordings of responses from recombinant NMDA receptors expressed in Xenopus oocytes activated by 100 μM glutamate plus 50 μM glycine. GluN1 was co-expressed with either 2A_C1 and mutated 2B_AC2 (R519K + T691I in 2B_AC2(RK+TI)), mutated 2A_C1 (R518K + T690I in 2A_C1(RK+TI)) and 2B_AC2, or 2A_C1 and 2B_AC2. The RK+TI mutations abolish binding of glutamate and render any NMDA receptor containing this subunit non-functional. (D) Co-expression of 2A_C1 and 2B_AC2 (as well as GluN1) produced robust currents responses that increased following cRNA injection (black), whereas current responses from 2A_C1/2A_C1 and 2B_AC2/2B_AC2 receptors that may have escaped ER retention remained small. Data are mean ± SEM and each data point is from 5 batches of oocytes with 6 oocytes for each batch (N = 30). (E) The sum of the fractional currents assessed using the RK+TI mutations provided an estimate of the percent “escape” current in oocytes co-expressing 2A_C1 and 2B_AC2, 2A_C1 and 2A_C2, or 2B_AC1 and 2B_AC2. See also Figures S1–S3 and Table S1.

Figure 2

Inhibition of triheteromeric receptors by GluN2A-selective antagonists. (A) TCN-201 data were generated for wild type receptors (wt 2A/2A and wt 2B/2B) and receptors containing C1- and C2-tagged GluN2 subunits expressed in Xenopus oocytes using two-electrode voltage-clamp recordings. Current responses were activated by 100 μM glutamate plus 3 μM glycine. Data are mean ± SEM from 8–10 oocytes. (B) Data for inhibition by extracellular Zn²⁺ were generated for the NMDA receptor subtypes. Current responses were activated by 50 μM glutamate plus 50 μM glycine. Data are from 6–12 oocytes. (C–E) Inhibition of wild type GluN1/GluN2A (wt 2A/2A) and GluN1/GluN2A/GluN2B triheteromers (2A_C1/2B_AC2) by extracellular Zn²⁺ was compared at pH 7.8, 7.3, and 6.8. Data are from 6–28 oocytes. (F,G) The representative recordings show inhibition by extracellular Zn²⁺ at physiological pH 7.3. (H) Data for inhibition by extracellular Zn²⁺ at physiological pH 7.3 of receptors with a mutation in a single or both Zn²⁺ binding sites (H128S in GluN2A and H127A in GluN2B). The blue dashed line represents the fit of actual data from 2A_C1/2B_AC2 shown in panel D. Data are from 6–11 oocytes. See also Table 1 and Figure S4.

Figure 3

Inhibition of triheteromeric receptors by GluN2B-selective antagonists. (A) The representative two-electrode-voltage-clamp recordings show ifenprodil inhibition of 2B_AC1/2B_AC2 diheteromers and 2A_C1/2B_AC2 triheteromers expressed in Xenopus oocytes. Current responses were activated by 100 μM glutamate plus 50 μM glycine. (B) Data for ifenprodil inhibition were generated for the NMDA receptor subtypes. Data are mean ± SEM from 6–16 oocytes. (C) Data for CP-101,606 inhibition were generated for the NMDA receptor subtypes. Data are from 7–10 oocytes. (D,E) Data for ifenprodil and CP-101,606 inhibition were generated in the continuous presence of 300 nM extracellular Zn²⁺. The blue dashed lines represents fits of actual data from 2A_C1/2B_AC2 (control) generated in the absence of extracellular Zn²⁺ as shown in panels B and C. Data are from 6–8 oocytes. See also Table 1 and Figures S4–S5.

Figure 4

Deactivation time course of triheteromers and rate of ifenprodil inhibition. (A) The overlay compares representative whole-cell patch-clamp recordings of current responses from wild type GluN1/GluN2B (wt 2B/2B) expressed in HEK293 cells. The receptors are activated by 100 μM glutamate in the continuous presence of 100 μM glycine. The time course of ifenprodil inhibition of steady-state responses can be described using mono-exponential fits (black lines). (B) Representative response from GluN1/GluN2A/GluN2B triheteromers (2A_C1/2B_AC2) activated by 100 μM glutamate in the continuous presence of 100 μM glycine and inhibited by ifenprodil. (C) The rate of ifenprodil inhibition (1/τ_inhibition) was linearly related to the concentration of ifenprodil, and the slope of the linear fit provides the rate of ifenprodil binding (k_on), which was 0.263 μM⁻¹·s⁻¹ for wild type 2B/2B (N = 6) and 0.099 μM⁻¹·s⁻¹ for 2A_C1/2B_AC2 triheteromers (N = 6). Data are mean ± SEM. (D) The overlay compares representative whole-cell patch-clamp recordings of responses normalized to the peak amplitude. The diheteromeric wt 2A/2A and 2B/2B and triheteromeric 2A_C1/2B_AC2 receptors were expressed in HEK293 cells and activated by brief pulses (5 ms) of 1 mM glutamate in the continuous presence of 100 μM glycine. (E) The graph plots the weighted time constants for deactivation following removal of glutamate from individual cells (circles) as well as the mean (horizontal line) for wild type 2A/2A (N = 11), 2A_C1/2B_AC2 triheteromers (N = 17), and wild type 2B/2B (N = 15). * denotes significantly different (unpaired t test; two-tailed, P < 0.05). See also Table 2, Figure S6, and Table S2.

Figure 5

Effects of extracellular Zn²⁺ and ifenprodil on deactivation time course. (A) The overlay compares representative whole-cell patch-clamp recordings of current responses normalized to the peak amplitude from wild type 2A/2A (left), 2A_C1/2B_AC2 triheteromers (middle), and wild type 2B/2B (right) expressed in HEK293 cells. The receptors were activated by brief pulses (5 ms) of 1 mM glutamate in the absence (control; black line) or presence of 1 μM ifenprodil (red line) or 300 nM extracellular Zn²⁺ (blue line). Glycine (100 μM) was present at all times. (B) The bar graph shows the percent inhibition of the peak amplitude by ifenprodil or extracellular Zn²⁺ relative to control. Each bar represents mean ± SEM from 6 cells. (C–E) The graphs plot the weighted time constants of individual cells expressing wild type 2A/2A, 2B/2B, and 2A_C1/2B_AC2 triheteromers. Responses to brief (5 ms) glutamate applications in absence of antagonists (control) were compared to responses in the presence of either 1 μM ifenprodil or 300 nM extracellular Zn²⁺ obtained from the same cell (i.e. paired recordings). * denotes significantly different (paired t test; two-tailed, P < 0.05), and ns indicates not significant. See also Table 2.