Structural Basis for Negative Allosteric Modulation of GluN2A-Containing NMDA Receptors

Abstract

NMDA receptors mediate excitatory synaptic transmission and regulate synaptic plasticity in the central nervous system, but their dysregulation is also implicated in numerous brain disorders. Here, we describe GluN2A-selective negative allosteric modulators (NAMs) that inhibit NMDA receptors by stabilizing the apo state of the GluN1 ligand-binding domain (LBD), which is incapable of triggering channel gating. We describe structural determinants of NAM binding in crystal structures of the GluN1/2A LBD heterodimer, and analyses of NAM-bound LBD structures corresponding to active and inhibited receptor states reveal a molecular switch in the modulatory binding site that mediate the allosteric inhibition. NAM binding causes displacement of a valine in GluN2A and the resulting steric effects can be mitigated by the transition from glycine bound to apo state of the GluN1 LBD. This work provides mechanistic insight to allosteric NMDA receptor inhibition, thereby facilitating the development of novel classes NMDA receptor modulators as therapeutic agents.

Figure 1. Glycine-sensitivity of inhibition by GluN2A-selective NAMs

A) Chemical structures of GluN2A-selective negative allosteric modulators (NAMs). B) Representative recordings showing inhibition of diheteromeric GluN1/2A and triheteromeric GluN1/2A/2B receptors by NAMs. Responses are activated by 100 μM glutamate in the continuous presence of 3 μM glycine and recorded using two-electrode voltage-clamp electrophysiology. Triheteromeric GluN1/2A_C1/2B_AC2 receptors are expressed without diheteromeric GluN1/2A and GluN1/2B receptors using engineered C1 and C2 tags (Hansen et al., 2014). Horizontal and vertical scale bars represent 1 min and 100 nA, respectively. C-D) NAM concentration-inhibition data at GluN1/2A receptors comprised of two GluN1 and two GluN2A subunits and GluN1/2A/2B receptors comprised of two GluN1 and two different GluN2 subunits (GluN2A and GluN2B). Inhibition is measured for responses activated by 100 μM glutamate in the continuous presence of 1, 3, 30, or 300 μM glycine. Data are mean ± SEM from 5-14 oocytes. See also Figures S1 and S2; Table S1.

Figure 2. Structures of NAM-bound GluN1/2A LBD heterodimers

A) Structure of the NMDA receptor composed of two GluN1 and two GluN2 subunits (PDB ID 4PE5; (Karakas and Furukawa, 2014). One GluN1/GluN2 LBD heterodimer is highlighted in orange and yellow. B) Crystal structure of the soluble glycine/glutamate-bound GluN1/2A LBD heterodimer in complex with TCN-201. Expression and purification of soluble NMDA receptor LBDs is made possible by deleting the amino-terminal domain (ATD) and replacing the transmembrane domain (TMD) with a Gly-Thr dipeptide linker. C) NAM ligands from crystal structures of glycine/glutamate-bound GluN1/2A LBD heterodimers in complex with TCN-201 (2.28 Å), MPX-004 (2.52 Å), and MPX-007 (2.11 Å). The NAM ligands could be unambiguously fitted to their electron densities shown as grey mesh (mF_o-DF_c omit map contoured at 2.5σ). D) Chemical structures of the NAMs, here exemplified by TCN-201, contain a halogenated ring A, a variable ring B, which is phenyl in TCN-201 and pyrazine in MPX-004/MPX-007, and a variable head group with a ring C that is phenyl in TCN-201 and thiazole in MPX-004/MPX-007. E) Overlay of TCN-201 (grey), MPX-004 (cyan), and MPX-007 (salmon) from their respective GluN1/GluN2A LBD structures. F-H) Views of the NAM binding site in structures of the glycine/glutamate-bound GluN1/2A LBD heterodimer in complex with TCN-201 (grey; F), MPX-004 (cyan, G), and MPX-007 (salmon; H). The NAM binding site is formed by residues from β-sheet 10 and α-helices J and K in GluN2A (yellow) and β-sheets 10 and 14 in GluN1 (orange). See also Figure S4 and Table S2.

Figure 3. Analysis of NAM-mediated structural changes in GluN1/2A LBD structures

A) Structural alignment of the NAM binding site in the GluN1/2A LBD heterodimer without bound NAM (yellow and orange) and in complex with TCN-201 (grey), MPX-004 (cyan), and MPX-007 (salmon). B) Difference distance maps showing pairs of Cα atoms that are less than 15 Å apart in the NAM-lacking structure and change by more than 0.5 Å in NAM-bound structures. The difference distance ∣ΔD_ij∣ is the absolute value of the distance measured in the NAM-lacking structure minus the distance measured in the NAM-bound structure. See also Figure S5.

Figure 4. Relationship between NAM inhibition and the GluN1 LBD conformation

A) Cartoon illustrating conformations of the GluN1/2A LBD heterodimer with the NAM binding site in the subunit interface. In this cartoon, glutamate is continuously bound to the GluN2A LBD and glycine must induce a closed GluN1 LBD conformation for NMDA receptor activation. The cartoon also depicts the engineered disulfide bond introduced by N499C + Q686C mutations in GluN1 (GluN1-CC). B) Representative two-electrode voltage-clamp recordings of responses from GluN1-CC/2A receptors with the GluN1 LBD locked in the closed, active conformation. Reponses are activated by 100 μM glutamate alone, followed by co-application of glycine (300 μM), TCN-201 (5 μM), MPX-004 (10 μM), or MPX-007 (30 μM). Horizontal and vertical scale bars represent 30 sec and 50 nA, respectively. C) Summary of the effects of glycine (300 μM), the competitive glycine-site antagonist 7CKA (10 μM), TCN-201 (5 μM), MPX-004 (10 μM), and MPX-007 (30 μM) on glutamate-activated responses from GluN1-CC/2A receptors. Data are mean ± SEM from 5-6 oocytes. * indicates significantly different from all other groups (P < 0.05; one-way ANOVA with Tukey-Kramer posttest). D-E) Representative recordings of responses from GluN1-CC/2A receptors activated by 100 μM glutamate alone. Co-application of DTT (4 mM) plus the competitive antagonist 7CKA (10 μM) or MPX-007 (30 μM) diminished glutamate-activated responses. Horizontal and vertical scale bars represent 1 min and 100 nA, respectively. F) Summary of the effects of DTT alone (4 mM) and co-application of DTT plus 7CKA (10 μM), TCN-201 (5 μM), MPX-004 (10 μM), or MPX-007 (30 μM) on glutamate-activated responses from GluN1-CC/2A receptors. Data are mean ± SEM from 5-24 oocytes See also Figure S6.

Figure 5. Crystal structures of conformational states in the NAM inhibition cycle

A) Cartoon illustrating conformations of the GluN1/2A LBD heterodimer in the NAM inhibition cycle. In this cartoon, glutamate is continuously bound to the GluN2A LBD and the competitive glycine-site antagonist DCKA stabilizes the open (i.e. glycine-lacking) state of the GluN1 LBD. PDB IDs of structures corresponding to the four conformations are listed. B) Overlay of DCKA/glutamate-bound (PDB ID 5JTY; yellow and orange) and glycine/glutamate-bound (PDB ID 5I59; grey) structures in complex with MPX-007, illustrating the difference in GluN1 LBD domain closure (indicated by the arrow). C) Crystal structures of the GluN1/2A LBD heterodimer corresponding to conformational states in the NAM inhibition cycle. Arrows indicate transitions between states and structural changes.

Figure 6. Structural changes in the NAM binding site during allosteric inhibition

A) Structural changes in modulatory binding site upon NAM binding to the glycine/glutamate-bound GluN1/2A LBD heterodimer. Binding of NAM (i.e. MPX-007) displaces GluN2A V783 and this “push” is accompanied by a steric effect on GluN1 F754. B) Structural changes upon glycine unbinding from the NAM-bound GluN1/2A LBD heterodimer. Glycine unbinding and opening of the GluN1 LBD (here stabilized by DCKA) enable a conformational change of GluN1 F754, removing steric interaction with GluN2A V783. C) Structural changes upon NAM unbinding from the DCKA/glutamate-bound GluN1/2A LBD heterodimer. NAM unbinding enables a relaxation of residues in the GluN1 hinge region. D) Structural changes upon glycine binding to the NAM-lacking GluN1/2A LBD heterodimer. Glycine binding and closure of the GluN1 LBD reengages the steric interaction between GluN2A V783 and GluN1 F754, and places residues in the GluN1 hinge region in a conformation permissive for NAM binding. See also Figure S7.

Figure 7. Mutational analyses of the influence of GluN2A V783 and GluN1 F754 on glycine potency

A) View of the NAM binding site in the glycine/glutamate-bound GluN1/2A LBD heterodimer in complex with MPX-007. The van der Waals radius of the methyl group of MPX-007 is shown as transparent sphere. This methyl group contacts the backbone carbonyl of GluN2A P527 and the side chain of GluN2A V783. There is also a nonpolar interaction between side chains of GluN2A V783 and GluN1 F754. B) Glycine concentration-response data for wild type and mutated GluN1/2A receptors. Responses are recorded in the presence of 100 μM glutamate using two-electrode voltage-clamp electrophysiology. Data are mean ± SEM from 9-16 oocytes. See also Table S3.

Figure 8. Evaluation of GluN1 Y535 and F754 in allosteric inhibition of glycine binding

A) Operational model for TCN-201 inhibition by allosteric modulation of glycine binding without changing agonist efficacy (E). A is the agonist glycine, B is the NAM, and R is the receptor. Agonist binding affinity (K_A) is changed by an allosteric constant α upon NAM binding, and NAM binding affinity (K_B) is changed by α upon agonist binding. The equation for the dose ratio DR, the ratio of agonist EC₅₀ values in presence and absence of modulator, is shown. B) Glycine concentration-response data for GluN1/2A in the absence or presence of TCN-201 are analyzed by simultaneously fitting all data to both the dose ratio DR equation and the Hill equation using global nonlinear regression, yielding the glycine EC₅₀, TCN-201 binding affinity (K_B), and allosteric constant α that describe all the experimental data. C) DR values derived from fitting individual concentration-response data shown in B) are plotted as a function of TCN-201 concentration in the Schild plot, which also illustrates the effects of changing the allosteric constant α, but with constant K_B (27 nM). The black line is generated using the K_B and α obtained by fitting all data in B) using global nonlinear regression (i.e. the black line is not obtained by fitting directly to DR values). Competitive antagonists have α = 0 (dashed line). D-G) Schild analysis of TCN-201 inhibition for mutated GluN1/2A receptors. Glycine concentration-response data are analyzed using global nonlinear regression (left side) and the resulting binding affinity (K_B) and allosteric constant (α) are used to generate the black line in the Schild plot (right side) with the corresponding DR values. Dashed red line indicates the Schild plot for wild type GluN1/2A receptors. See also Figures S1 and S8; Table S3.