Abstract
Kinetics of NMDA receptor (NMDAR) ion channel opening and closing contribute to their unique role in synaptic signaling. Agonist binding generates free energy to open a canonical gate at the M3 helix bundle crossing. Single channel activity is characterized by clusters, or periods of rapid opening and closing, that are separated by long silent periods. A conserved glycine in the outer most transmembrane helices, the M4 helices, regulates NMDAR function. Here we find that the GluN1 glycine mainly regulates single channel events within a cluster, whereas the GluN2 glycine mainly regulates entry and exit from clusters. Molecular dynamics simulations suggest that, whereas the GluN2 M4 (along with GluN2 pre-M1) regulates the gate at the M3 helix bundle crossing, the GluN1 glycine regulates a ‘gate’ at the M2 loop. Subsequent functional experiments support this interpretation. Thus, the distinct kinetics of NMDARs are mediated by two gates that are under subunit-specific regulation.
Similar content being viewed by others
Introduction
Ion channels exist in two general conformations: a closed, nonconducting state and an open, conducting state, where the channel forms a water-filled pore or permeation pathway that certain ions can cross and impact membrane excitability. ‘Gates’ are barriers that occlude the flux of ions in the non-conducting or closed state. They can arise from a variety of mechanisms1, from a physical barrier where the pore is too narrow for ions to cross in its closed state to free energy barriers such as “dewetting” pores, where changes in hydrophobicity prevent charged ion entry2,3,4,5,6.
AMPA (AMPAR) and NMDA (NMDAR) receptors are glutamate-gated ion channels or ionotropic glutamate receptors (iGluRs) that participate in numerous brain functions7. These receptors are highly modular, layered proteins (Fig. 1a) and are members of the pore-loop family of ion channels reflecting that their channel pore is formed in part by a non-membrane spanning pore loop, referred to as the M2 pore loop (Fig. 1b)7,8,9,10,11. In homology to other pore-loop channels such as K+, Na+, and Trp channels, the remainder of the permeation pathway is formed by a transmembrane helix (i.e., M1) N-terminal to the pore loop, but mainly by an “inner” helix (i.e., M3) C-terminal to the pore loop (Fig. 1b)7,12. For other pore-loop channels, gates have been identified at the bundle crossing formed by the inner helices13 as well as at the pore loop4,5,14. Interestingly, some K+ channels may have gates at the bundle crossing and the pore loop4.
In iGluRs, the M3 helices form a canonical gate at the bundle crossing8,15,16,17,18. Binding of agonists, glutamate in the case of AMPARs and glutamate and glycine in the case of NMDARs, to the extracellular ligand-binding domain (LBD) (Fig. 1a) induces conformational changes that are transduced to this M3 bundle crossing, in part via the LILI motif (Fig. 1b)18, leading to channel opening19,20,21,22,23,24,25,26. Some studies have suggested a ‘gate’ at the M2 loop27,28 and mutations in the M2 loop can alter receptor gating29,30,31. Still, the significance of the M2 loop to receptor gating is unknown.
NMDARs are heterotetramers composed of two obligatory GluN1 and typically some combination of two GluN2(A-D) subunits. Surrounding the pore domain (M1-M3) (Fig. 1b) are eukaryotic specific M4 helices (Fig. 1c). The GluN1 and GluN2 M4 helices regulate the core gating machinery32,33,34,35,36. Critical in the M4 helices are highly conserved glycines positioned near the top of the M4s37. A variety of disease-associated missense mutations have been identified at or near these conserved glycines38,39,40, and even the most conservative amino-acid substitution, glycine-to-alanine (G-to-A), dramatically alters receptor gating. Notably, the G-to-A substitution in the GluN1 but not GluN2 alters channel gating at least in part via constriction of the M2 pore loop which modifies Ca2+ permeation37. Additionally, studies of nearby missense mutations in the GluN2 M4 suggest that the M4 may act through the pore lining GluN1 M3 helices38.
Here, we investigate the roles of the GluN1 and GluN2A M4 helices in regulating NMDAR function by focusing on the conserved glycines. The G-to-A substitution in either the GluN1 or GluN2A subunits dramatically curtailed channel opening37. However, we find that they have distinct effects on the overall pattern of single channel activity. The GluN2A G-to-A, which our molecular dynamics (MD) simulations indicate primarily alters the M3 gate, affects entry and exit from prolonged periods of activity known as “clusters”. On the other hand, the GluN1 G-to-A, which our MD simulations show impedes ion flux through the M2 constriction, affects single channel opening within a cluster. These data suggest that the M3 gate mediates the longer opening that reflect entry into clusters, while the M2 gate mediates, at least in part, the faster gating activity within clusters. Reactivity of MTS reagents within the vestibule separating the M3 gate and the M2 loop further confirms the presence of a gate at the M2 constriction. Thus, our results support a two-gate model of NMDAR activity with these gates regulated by specific subunits and controlling different aspects of NMDAR activity.
Results
A G-to-A substitution (or G-A for short) at a conserved glycine in the GluN1a but not GluN2 M4 helices affects Ca2+ permeation by altering the M2 loop constriction37, suggesting that the M2 loop may function as a ‘gate’. To address this notion, we characterized more extensively the single channel properties of the G-to-A substitutions at the conserved glycines in GluN1 and GluN2A.
M4 helices regulate distinct components of NMDAR gating
We recorded in the on-cell configuration single channel activity of wild-type GluN1a/GluN2A, GluN1a(G815A)/GluN2A, and GluN1a/ GluN2A(G819A) for long durations, typically 20 min or longer (Fig. 2a and Supplementary Fig. 1a). For these experiments, we used saturating concentrations of co-agonists and recorded in the absence of divalents to enhance resolution (see “Methods” section). Both G-to-A substitutions dramatically reduce the equilibrium open probability (eq. Popen) to the same extent (Fig. 2b & Supplementary Table 1). The eq. Popen is the fraction of time channels are open during the entire recording including during long-lived closed states. However, these constructs have differential effects on mean closed (MCT) and open (MOT) times with GluN1(G815A) (GluN1 G-A) notably reducing MOT (Supplementary Fig. 1b). To understand this difference, we characterized the effect of the G-As on the pattern of the single channel activity including clusters and interclusters (Fig. 2c–d and Supplementary Fig. 1c–e and Table 2).
Wild-type GluN1a/GluN2A shows a characteristic pattern of single channel activity: periods of high activity referred to as clusters (bursts, superclusters) (gray bars), during which the channel rapidly transitions between open and closed states, interspersed by periods of inactivity, which we refer to as interclusters (Fig. 2a, upper trace)41,42. Under our conditions, wild-type GluN1a/GluN2A showed a cluster duration around 12.3 sec (12.3 ± 1.0 s, n = 16; mean ± SEM, n = number of on-cell patches) and an intercluster duration around 2.3 s (2.3 ± 0.3 s; Supplementary Table 2).
While the G-to-A substitutions in GluN1a and GluN2A have the same effects on eq. Popen, they have different effects on cluster Popen (Fig. 2c), which is significantly more reduced in GluN1a G-A than in GluN2A G-A with this effect related to a severely reduced MOT in GluN1 G-A (Supplementary Fig. 1c). Further, the G-As alter the cluster and intercluster pattern but do so in distinct ways with GluN2A G-A significantly reducing cluster length and extending intercluster length (Supplementary Fig. 1d, e). To contrast these differences, we characterized the probability of being in a cluster during the entire recording period (Pcluster) (Fig. 2d), which showed no difference between wild-type and GluN1 G-A but with GluN2A G-A being significantly reduced relative to both. Hence, the GluN2A M4 but not the GluN1a M4 regulates the cluster probability.
NMDAR single channel activity is characterized by five closed states42,43,44,45, which we found to hold for wild-type as well as both G-to-A constructs (Supplementary Figure 2 & Table 3). Cluster analysis depends on the definition of a Tcrit, which is a cut-off between the third (C3) and fourth (C4) closed kinetic states that minimizes false events46 (see “Methods” section). To analyze the effects of the G-to-A constructs on the gating pattern independent of defining Tcrit, we compared the mean duration of the two longest closed states, C4 and C5 (Supplementary Fig. 1f, g). Consistent with the differential roles of the M4 helices in regulating the gating (cluster) pattern of NMDARs, the mean C4 and C5 durations for GluN2A G-A were consistently longer relative to both wild-type and GluN1 G-A. This difference is further revealed when the mean closed durations of the G-to-A substitutions are compared as ratios relative to wild-type, either for C4 (Fig. 2e) or C5 (Fig. 2f). In contrast, the shortest-lived closed state, C1, was predominantly altered by GluN1 G-to-A (Fig. 2g and Supplementary Fig. 1h), whereas C2 (Fig. 2h) and C3 (Fig. 2i) were equally affected.
These results suggest that the M4 helices in GluN1 and GluN2A subunits contribute differently to the kinetics of NMDAR gating. The GluN2A M4s predominantly regulate a ‘cluster’ gate, dictating entry into and exit from long-lived closed states (Fig. 2d). In contrast, the GluN1 M4s predominantly regulate channel activity ‘within’ a cluster, impacting mainly the duration of open time (Supplementary Fig. 1b) and hence cluster Popen (Fig. 2c) as well as compressing the shortest-lived closed state (Fig. 2g).
MD simulations suggest that the M2 pore loop acts as a ‘gate’
The GluN1 M4 but not the GluN2 M4 regulates the pore radius of the M2 pore loop and consequently the magnitude of Ca2+ permeation37. Given this association of the GluN1 M4 helix with the M2 pore loop, we carried out MD simulations on an ‘open’ state homology model for GluN1/GluN2B37. In part due to the uncertainty of the wild-type open state structure, we are mainly interested in outcomes for the G-to-A substitutions relative to our wild-type and refer to this approach as comparative MD simulations. Although the MD simulations were run on GluN1/GluN2B constructs, we note that GluN2B G-A and GluN2A G-A have the same relative effect on single channel activity37. With the LBD-TMD linkers ‘locked’ in a presumed open conformation, we ran 15 replicate simulations for 500 nanoseconds each and repeated the simulations after introducing G-A in either GluN1 or GluN2B (Fig. 3a–e and Supplementary Figs. 3 and 4). Achieving the open conformation for the LBD-TMD linkers requires significant rearrangements of the extracellular domain47,48, which we did not explicitly model here. The ions in the simulations were monovalent (with Na+ as permeant ions), matching conditions in our single channel recordings.
We characterized the effect of the G-to-A substitutions on pore radius, focusing on the M3 helix bundle crossing, encompassing positions +2, +6, and +10, and the M2 pore loop (Fig. 1b and Supplementary Fig. 3a, b). The G-to-A substitutions in GluN1 or GluN2 did not affect pore radius at the M3 crossing (Fig. 3b, c, upper panels), reflecting that throughout the simulations the LBD-TMD linkers are locked in an open conformation. In contrast and consistent with previous simulations37, the G-to-A substitution in GluN1, but not in GluN2, reduced the pore dimensions at the M2 loop (Fig. 3b, c, lower panels).
To address how these changes in pore radius might affect ion flow, we tracked the movement of Na+ ions across the pore during the simulations (Fig. 3d and Supplementary Fig. 4). We defined three compartments along the permeation pathway (Fig. 3a, inset): extracellular to the +10 position in the M3 helix; the vestibule between the M3 helix bundle crossing and the M2 pore loop; and intracellular, below the M2 loop. We classified ‘permeation events’ as full transitions either through the M3 bundle crossing (red arrows) or through the M2 loop (green arrows) (Fig. 3a, inset). At the start of each simulation, a pool of ions was available for permeation in the extracellular (+30 Å) or intracellular (−30 Å) space or in the vestibule. Ions once inside the vestibule could leave by traversing the M2 loop (downward green arrows; Fig. 3d) but more frequently by traversing the M3 bundle crossing (upward red arrows). Ions could also traverse from the extracellular (downward red arrows) or intracellular (upward green arrows) space to the vestibule (Fig. 3d and Supplementary Fig. 4).
We characterized M2 crossings as a fraction of the total number of crossings (Fig. 3e, upper panel). Notably, M2 crossings represented a significant fraction of the total crossings for both wild-type, at 39% of total crossings, and GluN2A G-A, at 20% of total crossings. In contrast, M2 crossings were rare in GluN1 G-A, at only 3.7% of total crossings.
A potential complication is that, while random, more ions entered the vestibule from the extracellular space than from the intracellular space (Supplementary Fig. 4), thereby raising the number of M3 crossings. To address this concern, we analyzed those events where the ion started in the vestibule, with the assumption that all other things being equal there is a 50% chance of going either outward across the M3 gate or inward across the M2 loop (Fig. 3e, lower panel). For wild-type, 52% of total crossings were via the M2 loop (12 out of 23 total crossings), whereas for the GluN2A G-A 14% were via the M2 (2 out of 14 total crossing). In contrast, despite having a highest number of starting ions in the vestibule, 32 total, none of these ions crossed the M2 loop in the GluN1 G-A mutant, with all crossings at the M3 gate (Fig. 3e, lower panel).
For the simulations reported in Fig. 3, no transmembrane potential was present. We carried out additional simulations applying a −300 mV transmembrane potential (16 replicates of 250-ns duration for each construct). As expected, the transmembrane potential increases the number of crossings (M3 or M2) for all constructs. Importantly, the percentage of M2 crossings relative to the total number of crossings in GluN1 G-A is still strongly attenuated (15.7%, 16 out of 102 total crossings) compared to either wild-type (46.6%, 41 out of 88) or GluN2 G-A (36.4%, 24 out of 66).
In summary, these MD simulations are consistent with the idea that the M2 loop can act as a ‘gate’, controlling the flux of ions across the channel and, moreover, that the GluN1 M4 regulates this M2 gate. A further notable outcome is that the M2 gate can close while the M3 gate is in a fully open conformation (Fig. 3c). Hence, the M2 gate can transition between the presumed open and closed conformation while the M3 gate is open.
The GluN2 M4 helix regulates the M3 gate
For the MD simulations in Fig. 3b–e, we restrained the LBD-TMD linkers in a presumed open conformation, which prevented any major change in the M3 bundle crossing (Fig. 3b, c, upper panels). To assess how the M4 helices might impact the M3 gate, we released the restraints on the LBD-TMD linkers and carried out MD simulations for 3000 ns and averaged the pore radii over the last 1500 ns for each construct (Fig. 3f, g).
The starting pore radii for all three constructs before restraint release were comparable and we use their average as reference (thin green line, Fig. 3f). After the LBD-TMD linker restraints were removed, the most dramatic change was a collapse of the M3 gate (+2 to +10) for GluN2B G-A. While this effect was most pronounced around the +2 position, we averaged the pore radii from +2 to +6 (Fig. 3g, upper panel): the GluN2B G-A showed a strong reduction in pore radius while there was no difference between wild-type and GluN1 G-A. In contrast, averaging across the M2 pore loop did not reveal a significant difference between the constructs (Fig. 3g, lower panel), although wild-type did show some collapse around −12 Å. Hence, the GluN2 M4 impacts the M3 gate whereas GluN1 M4 does not.
Locking the M3 gate ‘open’ suggests two distinct gates
Studying a potential ‘gate’ at the M2 loop is challenging since it will be intertwined with the external M3 gate, which if closed would preclude access to this internal gate. To circumvent this problem, we took advantage of a tyrosine (Y) substitution at an alanine adjacent to the Lurcher position in GluN2A (A650Y) that dramatically enhances receptor gating49. Since this position is part of the M3 gate7,8, we assume that A650Y locks the M3 gate in an ‘open’ conformation, comparable to our MD simulations where the LBD-TMD linkers locked the channel open (Fig. 3b–e). Consistent with this idea, GluN1a/GluN2A(A650Y) channels showed extremely high eq. Popen, about 0.96 (0.96 ± 0.01, n = 9) compared to 0.68 for wild-type and long MOTs, about 40 ms (40 ± 5 ms) compared to 6.4 ms for wild-type (Fig. 4a, upper panel, 4b and Supplementary Fig. 5 and Table 5). GluN1a/GluN2A(A650Y) did visit the closed state, but these transitions were typically extremely brief C1 and C2 closures (Supplementary Figs. 6 and 7 and Table 6).
To assay the impact of the GluN1a and GluN2A M4 helices on this A650Y-induced gating pattern, we introduced the G-As in the GluN2A(A650Y) background (Fig. 4a–d; Supplementary Figs. 5–7 and Tables 5–8). The effects of these substitutions in the A650Y background were largely comparable to those in the wild-type background (Fig. 2a–i), including significant reduction in eq. Popen (Fig. 4b and Supplementary Fig. 5b and Table 5), greater reduction in cluster Popen by GluN1a G-A (Fig. 4c), and greater reduction in Pcluster by GluN2A G-A (Fig. 4d, Supplementary Figs. 5c–e and 6, and Table 6). These comparable outcomes support the idea of the distinct roles of the GluN1 and GluN2 M4s in receptor function.
MTS modification rates support two gates
A salient effect of the G-to-A substitutions in the A650Y background was to induce a significant reduction in eq. Popen (Fig. 4b), with this effect much greater for GluN1 G-A. We hypothesize that these decreased open probabilities reflect increased closures of the M3 gate (by GluN2A G-A) and the M2 gate (by GluN1 G-A). To test this hypothesis, we quantified the rate of reactivity of the MTS reagent aminoethyl-methanethiosulfonate (MTSEA) with a cysteine substituted in the external vestibule, GluN1(V644C), located between the M3 gate and the M2 loop (Fig. 5a), in the GluN2A(A650Y) background. Given this positioning of the introduced cysteine, closures of the external or upstream M3 gate would slow the rate of reactivity whereas closures of the downstream M2 gate would have no effect on the reactivity rate.
Consistent with our hypothesis, the rate of MTS reactivity with GluN1 G-A was indistinguishable from the control A650Y background (Fig. 5b–d) despite GluN1 G-A showing a stronger decrease in eq. Popen for constructs either lacking cysteines (Fig. 4b) or with cysteines (Fig. 5e and Supplementary Table 9). In contrast, GluN2A G-A showed a significantly reduced MTS reactivity rate (Fig. 5b–d), again consistent with the idea that it is affecting the M3 gate.
To reveal possible correlations between MTSmodification rates and other parameters including eq. Popen (left panel) and probability of being in a cluster, Pcluster (right panel), which we assume is a reporter of the M3 gate (see Fig. 2d), we normalized these parameters by the wild-type values (Fig. 5f). MTSmodification rates and eq. Popen in GluN1 G-A were incongruent, with a reduction in eq. Popen despite no change in MTSmodification rate (Fig. 5f, left panel). Thus, the divergence in GluN1 G-A eq. Popen and MTSmodification rate strongly supports the hypothesis that the M2 loop can function as a ‘gate’. Furthermore, reductions in Pcluster were congruent with MTSmodification rates, consistent with the assumption that this parameter is a reporter of the M3 gate (Fig. 5f, right panel). The congruence between Pcluster and MTS rate further supports the hypothesis that the M3 gate regulates entry and exit from clusters.
Discussion
Based on functional experiments and molecular dynamics simulations, we derive two general conclusions about NMDARs (Fig. 6). First, the gating pattern of NMDARs – entry and exit from long- or short-lived closed states – reflects the activity of two gates that directly control the flux of ions through the channel as opposed to modulating this process: a primary ‘gate’ at the M3 helix bundle crossing that controls entry and exit predominantly from long-lived closed states (C4 & C5), and a secondary gate at the M2 pore loop that may regulate mainly short-lived closed states (C1–C3). Our second major conclusion is that these two gates are regulated by the M4 helices of different subunits, with the GluN1 M4 strongly regulating the M2 gate and the GluN2 M4 strongly regulating the M3 gate. These gates are most likely coupled and therefore these actions are not exclusive.
These dual gates lead to a fundamental gating schema. In this schema, entry into and exit from clusters is mediated by an M3 gate that requires a significant amount energy to open, presumably reflecting a required displacement or rearrangement of the GluN2 pre-M1 and M435,50,51,52 (Fig. 6b, c). Notably, the opening of the M3 gate would mediate the initial step in channel activity. Once the M3 helices have rearranged out of a constrained conformation (pre-M1 and M4 displaced), we envision that both the M3 low-energy gate (where outer structures are already displaced) and the M2 gate mediate short-lived closed states (C1, C2, C3) within a cluster (Fig. 6d, e). Although present experiments cannot assign which states are mediated by the M3 or M2 gates, disruption of the M2 gate mostly affects the short-lived states, C1 and C2, while disruption of the M3 gate has the opposite effect, most severely affecting the longest closed states, C4 and C5 (Supplementary Tables 3 and 7).
Our results reveal that the M3 bundle and the M2 loop can function as ‘gates’, regulating the flux of ions through the pore. Nevertheless, there remain many unresolved issues. While our experiments support the idea that the M3 gate mainly regulates C4 and C5 and the M2 gate mainly regulates the short-lived closed states (C1, C2, and C3), this is by no means absolute since both gates impacts all closed states to some extent. Indeed, it is very likely that the two gates are energetically linked, with the status of the M3 gate, whether opened or closed, impacting the M2 gate and vice-versa. In this scenario, the status of the M2 gate would be one factor that might impact the duration of clusters (Supplementary Fig. 1d), though this idea needs to be directly tested. In addition, the time course of NMDAR activation is too fast to be solely mediated by the slow C4 and C5 and must involve faster opening events. Whether this reflects a fast component of the activation that is independent of cluster activity is unknown. Finally, the C4 and C5 components are associated with receptor desensitization. Hence, the GluN2 M4 as well as the S2-M4 linker is associated in some fashion with receptor desensitization; however, the mechanisms underlying this association are unknown.
At present we do not know the nature of either gate, whether it is physical occlusion or some other energetic constraint. Given the tight crossing, the M3 gate presumably functions by physical occlusion8,17. However, it has been shown that such narrow regions do not necessarily preclude ion flux3,53. Although evidence from our MD simulations and macroscopic current recordings suggest that reductions in the pore diameter of the M2 gate correlate with reduced Ca2+ permeation37, this pore loop site is evolutionarily related to the gate in K+ channels7, which have been shown to occlude ions due to changes in hydrophobicity, or “dewetting”2. Thus, it is possible that M3 and M2 gates act via different mechanisms. Further studies are needed to better understand how each gate occludes ion flow.
Subunit-specific regulation (GluN1 vs GluN2) in NMDARs has been identified previously16,18,43,54. Here, we showed that the M3 and M2 gates are regulated in a subunit-specific manner, and therefore may be targets for endogenous and pharmacological modulation. For example, zinc’s modulatory “rolling” mediated actions on the ATD and LBD that act on the GluN2A M455 may act on the M3 gate. Similarly, post translational modifications in the intracellular GluN2 CTD56,57, which is directly attached to the GluN2 M4, also likely act on the M3 gate. Conversely, modulation of NMDAR activity through the GluN1 CTD, notably the calcium-dependent inhibition of prebound calmodulin58 is likely mediated through the M2 gate. Thus, the present work strongly suggests that modulation in NMDARs is achieved in a gate specific manner.
Though it was previously imagined that a single gate generates the distribution of closed and open states seen in MCT and MOT histograms of NMDAR single channels, the introduction of a second gate implies that some of these states are the product of the two gates. Assuming the two gates can enter different structural arrangements partially independent of each other, the number of states observed may be a product of the number of distinct M3 and M2 states. Thus, the presence of two gates may reflect a need for the channel to have more complex control over gating speeds to mediate the wide variety of signaling events that NMDARs control. However challenging isolation of the separate activities of these two gates may prove to channel physiologists, such work is necessary to determine what physiology and pathophysiology each distinct gate mediates. Indeed, pharmacologic tools designed with gate specificity in mind could provide great therapeutic advancements in patients affected by the myriad of NMDAR-associated pathologies59.
Methods
Molecular biology and cell culture
For all experiments, we used rat GluN1 (GluN1a) (NCBI Protein database accession No. P35439) or GluN2A (Q00959) subunits. Numbering includes the signal peptide (GluN1, 18 residues; GluN2A, 19 residues). QuickChange site-directed mutagenesis kits (Agilent) with XL1-Blue super-competent cells were used to generate amino acid substitutions.
Human embryonic kidney 293 (HEK 293) cells (CRL-1573) are obtained from ATCC® approximately every two years. Multiple aliquots of these stocks are made and frozen down in liquid nitrogen. Approximately once per month, we thaw one of these aliquots and use for the subsequent time period. Cells derived from these aliquots display highly consistent morphology and growth patterns. If we have concerns about our stocks of HEK 293 cells, we obtain new stocks from ATCC®.
For details on cell culture and transfection see34. Briefly, human embryonic kidney 293 (HEK 293) cells were grown in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 10% fetal bovine serum (FBS), for 24 h prior to transfection. Non-tagged cDNA constructs were co-transfected into HEK 293 cells along with a separate pEGFP-Cl vector at a ratio of 4.5:4.5:1 (N1/N2/eGFP) for macroscopic recording, and at a ratio of 4.5:1:1 for single channel recording using X-tremeGene HP (Roche, 06-366). HEK 293 cells were bathed in a media containing the NMDAR competitive antagonist dl-2-amino-5-phosphentoic acid (APV, 100 µM, Tocris) and Mg2+ (100 µM). In some instances, we also used 1 mM Mg2+. Experiments were performed 18–48 h post-transfection.
Molecular modeling and MD simulations
The initial structure for the TMD plus LBD-TMD linker construct was built by homology modeling37. The sequences of the subunits were from Xenopus laevis GluN1/GluN2B, with two fragments comprising residues L541-P670 (M1-M3 containing) and R794-K841 (M4 containing) for GluN1 and two fragments comprising residues M537-K669 (M1-M3 containing) and G799-Q845 (M4 containing) for GluN2B (all residue numbering based on rat GluN1/GluN2B sequences). The template for homology modeling were the structure of the GluA2 AMPA receptor in the open state (Protein Data Bank entry 5WEO23). G-to-A substitutions were introduced at the conserved glycine positions (G815 in GluN1 and G820 in GluN2B) to generate the N1 G-A and N2B G-A constructs.
Simulations with linkers restrained to an open conformation
The CHARMM-GUI membrane builder server60 (https://www.charmm-gui.org/) was used to prepare the three constructs in a bilayer with 222 POPC lipids, solvated by 25,201 water molecules (including pore water). All the systems were neutralized using Na+ and Cl- ions and the final concentration of NaCl was 150 mM. The dimensions of the simulation box were 100.15 Å × 100.15 Å × 127.0 Å. The CHARMM-GUI six-segment protocol was followed to prepare the systems. Specifically, after energy minimization (5000 steps), six short segments of MD simulation were run. The first three segments were 25 ps each with a time step of 1 fs, and the second three segments were 100 ps each with a time step of 2 fs. Segments 1 and 2 were under constant NVT whereas segments 3–6 were under constant NPT. Harmonic restraints were imposed on two dihedral angles and all phosphorus atoms of POPC as well as on the backbone heavy atoms of the proteins. The restraint force constant for the dihedral angles was 250 kcal mol−1 rad−2 during the energy minimization and segment 1 of simulation but was reduced to 100, 50, 50, and 25 kcal mol−1 rad−2 in segments 2–5, respectively, and zeroed out in segment 6; the corresponding values for the restraint on the phosphorus atoms were 2.5, 2.5, 1.0, 0.5, and 0.1 kcal mol−1 Å−2. For the backbone heavy atoms of the proteins, the restraint force constant was 10.0, 5.0, 2.5, 1.0, 0.5, and 0.1 kcal mol−1 Å−2, respectively, in segments 1–6. Equilibration was performed for 12 ns at constant NPT, in which protein Cα atoms were restrained; the restraint force constant for all the Cα atoms except those on the linker tips was ramped down in the sequence of 5.0, 2.5, 1.0, 0.5, 0.25, and 0.1 kcal mol−1 Å−2 in 2-ns intervals; the restraint force constant on the linker tip Cα atoms was maintained at 5.0 kcal mol−1 Å−2. The simulations continued in the production run of 500 ns in 15 replicates for each construct, with only the Cα atoms of the linker tip residues (L541-K544, I667-P670, and R794-E797 in GluN1; M537-R540, L666-K669, and G799-H802 in GluN2B) restrained to the open conformation with a force constant of 5.0 kcal mol−1 Å−2.
The CHARMM-GUI six-segment preparation was run using Sander in Amber17, and the equilibration and production were run using pmemd.cuda on GPUs61. The force field was CHARMM3662 and the water model was TIP3P63. Bonds connected to hydrogens were constrained by the SHAKE algorithm64. The cutoff distance for nonbonded interactions was 8 Å; long-range electrostatic interactions were treated by the particle mesh Ewald method65. Temperature was kept at 300 K by the Langevin thermostat and pressure was kept at 1 bar by the Berendsen barostat66.
Simulations with a transmembrane potential
After the 12 ns equilibration, a second set of 16 replicate simulations was performed under a transmembrane potential of −300 mV67 for 250 ns at constant NVT. The settings of these simulations were otherwise identical to those without a membrane potential, except that the cutoff distance for nonbonded interactions was 10 Å. Trial simulations were also run under membrane potentials of −100 and −200 mV but were not further analyzed because the ion permeation events were relatively low and no clear distinction from simulations without membrane potential could be drawn.
Simulations of TMD closure
A third set of MD simulations mimicked the transition of the ion channel from the open state to closed state, by releasing the restraints on the linker tip residues. A snapshot from the first 100 ns of the simulations in the open state without membrane potential was selected for each construct, and simulations in four replicates were restarted without any restraints and run for 3000 ns using pmemd.cuda on GPUs.
MD analysis
Snapshots were saved in 100 ps intervals, resulting in 5000 and 2500 frames for each open simulation at 0 and −300 mV membrane potential, respectively, and 15,000 frames from the second half of each closure simulation. Pore radii along the pore (or z) axis were calculated every 10th saved frame using the HOLE (v2.2.005) program68, and averaged over the replicate simulations. The number of water molecules in the pore was obtained by counting the waters inside the channel from the M3 gate +10 position (V656 in N1 and I655 in N2B) to the M2 top residues (V613 in N1 and L612 in N2B).
Na+ ions that ever visited the pore (i.e., between M3 + 10 and M2 top) were tracked for crossings between three compartments along the permeation pathway. The z coordinate of each such ion was monitored as a function of simulation time. For example, the number of such ions was five in simulation 1 for the wild-type in the open state (z traces shown in Supplementary Fig. 4). A crossing from the vestibule (between M3 + 2 and M2 top) to the extracellular space) was counted only when the ion crossed M3 + 10 and diffused into the extracellular space.
Single-channel recordings and analysis
Single channel recordings were collected at room temperature (21–24 °C) using an Axopatch 200B (Molecular Devices), filtered at 10 kHz (four-pole Bessel filter), and digitized between 25 and 50 kHz (ITC-16 interfaced with PatchMaster, HEKA). Recording pipettes were pulled from thick-wall borosilicate capillary glass (Sutter Instruments).
On-cell single channel recordings were recorded at steady state. Recording pipettes were fire-polished to final resistances ranging from 10 to 30 MΩs when measured in the bath solution (with an applied positive pipette pressure of ~200 mbar). Seal resistances were between 1 and 15 GΩ. The pipette solution contained (in mM): 150 NaCl, 2.5 KCl, 10 HEPES, 0.05 EDTA, 1 glutamate, and 0.1 glycine, pH 8.0 (NaOH). The high pH and EDTA were used to minimize proton and divalent mediated inhibitory effects, respectively44. To elicit inward current amplitudes, we held the electrode voltage at +100 mV. The recordings were ~4–60 min in duration to provide a substantial number of events for analysis. In general, we did not include on-cell patches in analysis unless it contained a minimum of 5000 (typically >50,000) events.
For on-cell single channel analysis, single channel records were idealized in QuB using the segmental k-means (SKM) algorithm with a dead time of 20 μs. Closed and open state fits were performed using the maximum interval likelihood (MIL) algorithm in QuB. Kinetic models of NMDAR gating contain approximately 5 closed states and 3–4 open states43,44. For each individual record, state models with increasing closed (3–6) and open (2–4) states were constructed and fitted to the recordings until log-likelihood (LL) values improved by less than 10 LL units/added state or if the next added state showed 0% occupancy69. Clusters were identified by a critical time (Tcrit), which was defined by minimizing the false event ratio between the third (C3) and fourth (C4) closed kinetic state46,70. To verify single channels in patches, especially for those constructs with a low equilibrium Popen, we used statistical approaches71.
Parameters derived include probability of being open (Popen), mean open (MOT) and mean closed (MCT) times. These measurements were made either for the entire recording period, which is referred to as ‘equilibrium’ (e.g., equilibrium Popen), or for just events just during clusters (e.g., cluster Popen). We also derived the probability of being in a cluster (Pcluster = mean cluster length/(mean cluster length + mean intercluster length)).
Macroscopic current recordings
Macroscopic currents in the whole-cell mode, isolated from HEK 293 cells, were recorded at room temperature using an Axopatch 200B (Molecular Devices), filtered at 2.8 kHz (four-pole Bessel filter), and digitized at 10 kHz (ITC-16 interfaced with PatchMaster, HEKA). Patch microelectrodes were filled with an intracellular solution (in mM): 140 KCl, 10 HEPES, 1 BAPTA, 4 Mg2+-ATP, 0.3 Na+-GTP, pH 7.3 (KOH), 297 mOsm (sucrose). The extracellular solution consisted of (mM): 150 NaCl, 2.5 KCl, and 10 HEPES, 0.05 EDTA, pH 8.0 (NaOH). Pipettes had resistances of 2–6 MΩs when filled with the pipette solution and measured in the standard Na+ external solution. Ca2+ was omitted from the extracellular solution to prevent run-down over time. Currents were measured within 15 min of going whole-cell.
External solutions were applied using a piezo-driven double barrel application system. The open tip response (10-90% rise time) of the application system was between 400 and 600 μs. For display, NMDAR currents were digitally refiltered at 500 Hz and resampled at 1 kHz.
MTS modification reaction rates
To assay the effects of mutations on gating configuration, we calculated the rate of inhibition in agonist-activated macroscopic currents during MTS reagent application. MTS reagents (Toronto Research Chemicals, Toronto, Ontario, Canada) were made as stocks in powder form every day at room temperature and mixed into the recording solution within a few minutes before use. Modification rates were determined at −60 mV by continuous application of MTS reagents in the presence of glycine and glutamate and changes in current amplitudes, which were fitted with a single-exponential function to obtain the time constant (τ). The apparent second-order rate constant for MTS modification rate (k), was related to τ by:
where [MTS] is the concentration of the MTS reagent.
Statistics
Data analysis was performed using IgorPro 7, QuB, and Excel. All average values are presented as mean ± SEM. The number of replicates is indicated in the figure legend or in a table associated with the figure. In instances to determine if outcomes were statistically different from wild-type, we used an unpaired two-tailed Student’s t-test to test for significant differences. For multiple comparisons, we used an analysis of variance (ANOVA), followed by the Tukey test. Unless otherwise noted, statistical significance was set at p < 0.05. Sample sizes reflects those from prior publications.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The data that support this study are available from the corresponding authors upon request. Raw datasets generated during the current study as well as beginning and ending pdb of the MD simulations are available at the Open Science Framework repository [10.17605/OSF.IO/Q5B48]. The source data underlying Figs. 2–5 have been provided as a Source Data file, which includes all numbers derived from the analysis of the raw current records and statistics. Previously published PDB files can be accessed under accession codes 6WHR and 5WEO. Rat GluN1a (NCBI Protein database accession No. P35439) and GluN2A (Q00959) were used in this study. Source data are provided with this paper.
References
Hille, B. Ion Channels Of Excitable Membranes. 3rd edn (Sinauer Associates, Inc., 2001).
Jensen, M. O. et al. Principles of conduction and hydrophobic gating in K+ channels. Proc. Natl Acad. Sci. USA 107, 5833–5838 (2010).
Aryal, P., Sansom, M. S. & Tucker, S. J. Hydrophobic gating in ion channels. J. Mol. Biol. 427, 121–130 (2015).
Posson, D. J., Rusinova, R., Andersen, O. S. & Nimigean, C. M. Calcium ions open a selectivity filter gate during activation of the MthK potassium channel. Nat. Commun. 6, 8342 (2015).
Jia, Z., Yazdani, M., Zhang, G., Cui, J. & Chen, J. Hydrophobic gating in BK channels. Nat. Commun. 9, 3408 (2018).
Black, K. A., Jin, R., He, S. & Gulbis, J. M. Changing perspectives on how the permeation pathway through potassium channels is regulated. J. Physiol. 599, 1961–1976 (2021).
Hansen, K. B. et al. Structure, function, and pharmacology of glutamate receptor ion channels. Pharmacol. Rev. 73, 298–487 (2021).
Sobolevsky, A. I., Rosconi, M. P. & Gouaux, E. X-ray structure, symmetry and mechanism of an AMPA-subtype glutamate receptor. Nature 462, 745–756 (2009).
Karakas, E. & Furukawa, H. Crystal structure of a heterotetrameric NMDA receptor ion channel. Science 344, 992–997 (2014).
Lee, C. H. et al. NMDA receptor structures reveal subunit arrangement and pore architecture. Nature 511, 191–197 (2014).
Meyerson, J. R. et al. Structural mechanism of glutamate receptor activation and desensitization. Nature 514, 328–334 (2014).
Tikhonov, D. B. & Zhorov, B. S. The pore domain in glutamate-gated ion channels: Structure, drug binding and similarity with potassium channels. Biochim. Biophys. Acta Biomembr. 1862, 183401 (2020).
Johnson, J. P. Jr. & Zagotta, W. N. Rotational movement during cyclic nucleotide-gated channel opening. Nature 412, 917–921 (2001).
Heer, F. T., Posson, D. J., Wojtas-Niziurski, W., Nimigean, C. M. & Berneche, S. Mechanism of activation at the selectivity filter of the KcsA K(+) channel. Elife 6, e25844 (2017).
Zuo, J. et al. Neurodegeneration in Lurcher mice caused by mutation in delta2 glutamate receptor gene. Nature 388, 769–773 (1997).
Jones, K. S., VanDongen, H. M. & VanDongen, A. M. The NMDA receptor M3 segment is a conserved transduction element coupling ligand binding to channel opening. J. Neurosci. 22, 2044–2053 (2002).
Chang, H. R. & Kuo, C. C. The activation gate and gating mechanism of the NMDA receptor. J. Neurosci. 28, 1546–1556 (2008).
Ladislav, M. et al. The LILI motif of M3-S2 linkers is a component of the NMDA receptor channel gate. Front. Mol. Neurosci. 11, 113 (2018).
Armstrong, N. & Gouaux, E. Mechanisms for activation and antagonism of an AMPA-sensitive glutamate receptor: crystal structures of the GluR2 ligand binding core. Neuron 28, 165–181 (2000).
Zhang, W., Cho, Y., Lolis, E. & Howe, J. R. Structural and single-channel results indicate that the rates of ligand binding domain closing and opening directly impact AMPA receptor gating. J. Neurosci. 28, 932–943 (2008).
Blanke, M. L. & VanDongen, A. M. The NR1 M3 domain mediates allosteric coupling in the N-methyl-D-aspartate receptor. Mol. Pharmacol. 74, 454–465 (2008).
Kazi, R., Dai, J., Sweeney, C., Zhou, H. X. & Wollmuth, L. P. Mechanical coupling maintains the fidelity of NMDA receptor-mediated currents. Nat. Neurosci. 17, 914–922 (2014).
Twomey, E. C., Yelshanskaya, M. V., Grassucci, R. A., Frank, J. & Sobolevsky, A. I. Channel opening and gating mechanism in AMPA-subtype glutamate receptors. Nature 549, 60–65 (2017).
Chen, S. et al. Activation and desensitization mechanism of AMPA receptor-TARP complex by cryo-EM. Cell 170, 1234–1246.e1214 (2017).
Chou, T. H., Tajima, N., Romero-Hernandez, A. & Furukawa, H. Structural basis of functional transitions in mammalian NMDA receptors. Cell 182, 357–371.e13 (2020).
Yelshanskaya, M. V., Patel, D. S., Kottke, C. M., Kurnikova, M. G. & Sobolevsky, A. I. Opening of glutamate receptor channel to subconductance levels. Nature 605, 172–178 (2022).
Beck, C., Wollmuth, L. P., Seeburg, P. H., Sakmann, B. & Kuner, T. NMDAR channel segments forming the extracellular vestibule inferred from the accessibility of substituted cysteines. Neuron 22, 559–570 (1999).
Sobolevsky, A. I., Beck, C. & Wollmuth, L. P. Molecular rearrangements of the extracellular vestibule in NMDAR channels during gating. Neuron 33, 75–85 (2002).
Chen, N., Li, B., Murphy, T. H. & Raymond, L. A. Site within N-methyl-D-aspartate receptor pore modulates channel gating. Mol. Pharmacol. 65, 157–164 (2004).
Brown, P. M., Aurousseau, M. R., Musgaard, M., Biggin, P. C. & Bowie, D. Kainate receptor pore-forming and auxiliary subunits regulate channel block by a novel mechanism. J. Physiol. 594, 1821–1840 (2016).
Li, J. et al. De novo GRIN variants in NMDA receptor M2 channel pore-forming loop are associated with neurological diseases. Hum. Mutat. 40, 2393–2413 (2019).
Talukder, I., Borker, P. & Wollmuth, L. P. Specific sites within the ligand-binding domain and ion channel linkers modulate NMDA receptor gating. J. Neurosci. 30, 11792–11804 (2010).
Ren, H., Zhao, Y., Dwyer, D. S. & Peoples, R. W. Interactions among positions in the third and fourth membrane-associated domains at the intersubunit interface of the N-methyl-D-aspartate receptor forming sites of alcohol action. J. Biol. Chem. 287, 27302–27312 (2012).
Amin, J. B. et al. Divergent roles of a peripheral transmembrane segment in AMPA and NMDA receptors. J. Gen. Physiol. 149, 661–680 (2017).
Gibb, A. J. et al. A structurally derived model of subunit-dependent NMDA receptor function. J. Physiol. 596, 4057–4089 (2018).
Vyklicky, V. et al. Surface expression, function, and pharmacology of disease-associated mutations in the membrane domain of the human GluN2B subunit. Front. Mol. Neurosci. 11, 110 (2018).
Amin, J. B., Leng, X., Gochman, A., Zhou, H. X. & Wollmuth, L. P. A conserved glycine harboring disease-associated mutations permits NMDA receptor slow deactivation and high Ca(2+) permeability. Nat. Commun. 9, 3748 (2018).
Chen, W. et al. Functional evaluation of a de novo GRIN2A mutation identified in a patient with profound global developmental delay and refractory epilepsy. Mol. Pharmacol. 91, 317–330 (2017).
Perszyk, R. E. et al. Hodgkin-Huxley-Katz Prize Lecture: genetic and pharmacological control of glutamate receptor channel through a highly conserved gating motif. J. Physiol. 598, 3071–3083 (2020).
Amin, J. B., Moody, G. R. & Wollmuth, L. P. From bedside-to-bench: What disease-associated variants are teaching us about the NMDA receptor. J. Physiol. 599, 397–416 (2021).
Gibb, A. J. & Colquhoun, D. Glutamate activation of a single NMDA receptor-channel produces a cluster of channel openings. Proc. Biol. Sci. 243, 39–45 (1991).
Gibb, A. J. & Colquhoun, D. Activation of N-methyl-D-aspartate receptors by L-glutamate in cells dissociated from adult rat hippocampus. J. Physiol. Lond. 456, 143–179 (1992).
Banke, T. G. & Traynelis, S. F. Activation of NR1/NR2B NMDA receptors. Nat. Neurosci. 6, 144–152 (2003).
Popescu, G. & Auerbach, A. Modal gating of NMDA receptors and the shape of their synaptic response. Nat. Neurosci. 6, 476–483 (2003).
Schorge, S., Elenes, S. & Colquhoun, D. Maximum likelihood fitting of single channel NMDA activity with a mechanism composed of independent dimers of subunits. J. Physiol. 569, 395–418 (2005).
Magleby, K. L. & Pallotta, B. S. Burst kinetics of single calcium-activated potassium channels in cultured rat muscle. J. Physiol. 344, 605–623 (1983).
Dong, H. & Zhou, H. X. Atomistic mechanism for the activation and desensitization of an AMPA-subtype glutamate receptor. Nat. Commun. 2, 354 (2011).
Cerny, J., Bozikova, P., Balik, A., Marques, S. M. & Vyklicky, L. NMDA receptor opening and closing-transitions of a molecular machine revealed by molecular dynamics. Biomolecules 9, 546 (2019).
Murthy, S. E., Shogan, T., Page, J. C., Kasperek, E. M. & Popescu, G. K. Probing the activation sequence of NMDA receptors with lurcher mutations. J. Gen. Physiol. 140, 267–277 (2012).
Ogden, K. K. et al. Molecular mechanism of disease-associated mutations in the pre-M1 helix of NMDA receptors and potential rescue pharmacology. PLoS Genet. 13, e1006536 (2017).
McDaniel, M. J. et al. NMDA receptor channel gating control by the pre-M1 helix. J. Gen. Physiol. 152, e201912362 (2020).
Amin, J. B., Gochman, A., He, M., Certain, N. & Wollmuth, L. P. NMDA receptors require multiple pre-opening gating steps for efficient synaptic activity. Neuron 109, 488–501.e484 (2021).
Black, K. A. et al. A constricted opening in Kir channels does not impede potassium conduction. Nat. Commun. 11, 3024 (2020).
Sobolevsky, A. I., Prodromou, M. L., Yelshansky, M. V. & Wollmuth, L. P. Subunit-specific contribution of pore-forming domains to NMDA receptor channel structure and gating. J. Gen. Physiol. 129, 509–525 (2007).
Esmenjaud, J. B. et al. An inter-dimer allosteric switch controls NMDA receptor activity. EMBO J. 38, e99894 (2019).
Salter, M. W. & Kalia, L. V. Src kinases: a hub for NMDA receptor regulation. Nat. Rev. Neurosci. 5, 317–328 (2004).
Vieira, M., Yong, X. L. H., Roche, K. W. & Anggono, V. Regulation of NMDA glutamate receptor functions by the GluN2 subunits. J. Neurochem. 154, 121–143 (2020).
Iacobucci, G. J. & Popescu, G. K. Resident calmodulin primes NMDA receptors for Ca(2+)-dependent inactivation. Biophys. J. 113, 2236–2248 (2017).
Perszyk, R. E. et al. Biased modulators of NMDA receptors control channel opening and ion selectivity. Nat. Chem. Biol. 16, 188–196 (2020).
Jo, S., Kim, T., Iyer, V. G. & Im, W. CHARMM-GUI: A web-based graphical user interface for CHARMM. J. Comput. Chem. 29, 1859–1865 (2008).
Salomon-Ferrer, R., Gotz, A. W., Poole, D., Le Grand, S. & Walker, R. C. Routine microsecond molecular dynamics simulations with AMBER on GPUs. 2. Explicit solvent particle mesh ewald. J. Chem. Theory Comput. 9, 3878–3888 (2013).
Klauda, J. B. et al. Update of the CHARMM all-atom additive force field for lipids: validation on six lipid types. J. Phys. Chem. B 114, 7830–7843 (2010).
Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W. & Klein, M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926–935 (1983).
Ryckaert, J.-P., Ciccotti, G. & Berendsen, H. J. C. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327–341 (1977).
Essmann, U. et al. A smooth particle mesh Ewald method. J. Chem. Phys. 103, 8577 (1998).
Berendsen, H. J. C., Postma, J. P. M., Gunsteren, W. F. v., DiNola, A. & Haak, J. R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 81, 3684–3690 (1984).
Aksimentiev, A. & Schulten, K. Imaging alpha-hemolysin with molecular dynamics: ionic conductance, osmotic permeability, and the electrostatic potential map. Biophys. J. 88, 3745–3761 (2005).
Smart, O. S., Neduvelil, J. G., Wang, X., Wallace, B. A. & Sansom, M. S. HOLE: a program for the analysis of the pore dimensions of ion channel structural models. J. Mol. Graphics 14, 354–360 (1996).
Talukder, I. & Wollmuth, L. P. Local constraints in either the GluN1 or GluN2 subunit equally impair NMDA receptor pore opening. J. Gen. Physiol. 138, 179–194 (2011).
Alsaloum, M., Kazi, R., Gan, Q., Amin, J. & Wollmuth, L. P. A molecular determinant of subtype-specific desensitization in ionotropic glutamate receptors. J. Neurosci. 36, 2617–2622 (2016).
Kazi, R. et al. Asynchronous movements prior to pore opening in NMDA receptors. J. Neurosci. 33, 12052–12066 (2013).
Acknowledgements
We thank Donna Schmidt for technical assistance. This work was supported by NIH Grant R01 NS088479 (L.P.W. and H-X.Z.) and an SBU-BNL SEED Grant (LPW).
Author information
Authors and Affiliations
Contributions
J.B.A., M.H., R.P., H-X.Z., and L.P.W. designed research; M.H, J.B.A., and L.P.W. carried out and/or analyzed the functional experiments including single channel recordings, cluster analysis, and MTS applications; R.P., X.L., and H-X.Z., performed the computational studies including modeling and molecular dynamics simulations; J.B.A., M.H., R.P., H-X.Z., and L.P.W. wrote the paper including approving the final version.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Jiri Cerny and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Amin, J.B., He, M., Prasad, R. et al. Two gates mediate NMDA receptor activity and are under subunit-specific regulation. Nat Commun 14, 1623 (2023). https://doi.org/10.1038/s41467-023-37260-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-023-37260-y
- Springer Nature Limited