Skip to main content
NCBI home page
Biophysical Journal logoLink to Biophysical Journal
. 2011 Nov 16;101(10):2389–2398. doi: 10.1016/j.bpj.2011.10.015

GluN1-Specific Redox Effects on the Kinetic Mechanism of NMDA Receptor Activation

Iehab Talukder †,▵, Rashek Kazi †,▵, Lonnie P Wollmuth †,‡,∗
PMCID: PMC3218337  PMID: 22098737

Abstract

NMDA receptors are glutamate-activated ion channel complexes central to the functioning of the mammalian nervous system. Opening of the NMDA receptor ion channel pore is initiated by agonist-induced conformational changes in the extracellular ligand-binding domain (LBD) but the dynamic mechanism of this process remains unresolved. We studied how a disulfide bond in the obligatory GluN1 subunit—the sole site of redox modulation in NMDA receptors—controls this activation gating mechanism. This disulfide bond is located in the hinge region of the LBD, and presumably constrains agonist-induced cleft closure of the clamshell-like LBD. Elimination of this bond, by either DTT-mediated reduction or mutagenesis, enhances gating efficiency such that pore opening now occurs with higher frequency and longer duration. The most prominent effect was to shift opening modes to long duration openings reminiscent of a high Po gating mode that the NMDA receptor exhibits under ambient oxidizing conditions. In terms of preopen gating steps, elimination of this bond has effects only on the fast gating step consistent with this step being GluN1-specific and reflecting GluN1 gating movements immediately before channel opening. Overall, our results suggest that the dynamics of the GluN1 LBD have strong effects on late pore opening steps including regulating the duration of pore opening. This redox-mediated gating modulation could be an underlying mechanism of NMDA receptor malfunction in redox-dependent disease states and presents a potential target of pharmacologic action.

Introduction

NMDA receptors are a subtype of glutamate-activated ion channels that mediate rapid excitatory neurotransmission in the mammalian central nervous system. NMDA receptors are involved in a number of higher order brain functions including learning and memory as well as neurodevelopment (1). Aberrant activity of NMDA receptors can trigger and/or complicate both acute (e.g., stroke and seizure) and chronic (e.g., Parkinson's and Alzheimer's diseases) neurological and psychiatric (e.g., Schizophrenia) disorders (2). NMDA receptor function is regulated by a host of endogenous substances, including the coagonist glycine (or D-serine in certain brain regions), and monovalent (H+) and divalent (Ca2+, Mg2+, Zn2+) cations (3, 4). Further, NMDA receptor function is responsive to the local redox conditions, as both endogenous and exogenous reducing agents potentiate its activity (5, 6). This may be an important exacerbating factor in certain diseases states, e.g., status epilepticus and hypoxic ischemic insults, that are accompanied by an imbalance in the redox state of the brain (7, 8).

Redox modulation of NMDA receptors affects the efficiency with which ligand binding at an extracellular domain (ligand-binding domain or LBD) couples to opening of the ion channel pore (5). Reducing agents potentiate this coupling efficiency with an observed increase in the frequency and duration of channel openings (9, 10, 11, 12). Chimeric and mutagenesis studies as well as structural data have identified a redox-sensitive disulfide bond (C726 and C780) between a highly conserved pair of cysteines in the LBD to be the primary site of redox modulation (11, 13, 14). Interestingly, in an intact tetrameric NMDA receptor containing two glycine-binding GluN1 and two glutamate-binding GluN2A subunits, it is the GluN1-specific disulfide bond that mediates almost the entire redox modulation (12, 13) (Fig. 1 A). Despite the importance of the redox state to functionality, the effect of this GluN1-specific disulfide bond on the activation gating mechanism in NMDA receptors is unknown. Given that the disulfide bond is GluN1-specific, understanding its effects on the activation gating mechanism will further delineate subunit-specific mechanisms of gating in NMDA receptors.

Figure 1.

Figure 1

Open in a new tab

Redox-sensitive disulfide bond between C726 and C780 in the GluN1 ligand-binding domain (LBD) affects NMDA receptor function. (A) Backbone structure of an isolated GluN1 LBD in a glycine-bound conformation (PDB accession code No. 1PB7) (14) and an associated iGluR transmembrane domain (GluA2, subunit A, PDB accession code No. 3KG2) (44). The LBD is composed of polypeptide segments S1 (light gray) and S2 (dark gray). Glycine molecule (in ball configuration and in white) is depicted. The artificial glycine-threonine segment connecting S1 and S2 is also shown (white). The endogenous disulfide bond between C726 and C780 is also labeled. (Gray) Transmembrane α-helical segments M1, M3, and M4. These transmembrane segments are physically connected to the LBD by three 12–20 amino-acid-long linkers, S1-M1, M3-S2, and S2-M4 (dashed lines). Numbering is for the mature protein. (B) Representative membrane currents (holding potential, −60 mV) in Xenopus oocytes injected with NMDA receptor subunits, either GluN1/GluN2A (top) or GluN1(C780S)/GluN2A (bottom). Receptors were initially exposed to 15-s pulses of the coagonists glycine (20 μM) and glutamate (200 μM) (Gly/Glu, thin lines) to elicit NMDA receptor-mediated currents. A continuous application of DTT (4 mM, gray bar) was then started and currents were elicited three more times in the presence of DTT. (C) Mean % potentiation (mean ± SE, n ≥ 4) of current amplitudes in the presence of DTT calculated as 100 × (Ipost − Ipre)/Ipre, where Ipre and Ipost are the average current amplitudes, typically of 3–5 glutamate-activated currents, before and during DTT application, respectively. Only WT GluN1/GluN2A showed a significant DTT-induced change (potentiation) of current amplitudes. (Black and white bars) Value statistically significant from zero. (Asterisk) P < 0.05 (Student's t-test).

In ionotropic glutamate receptors, gating is initiated by the conformational change in the extracellular ligand-binding domain (LBD) induced by agonist binding. This conformational change is propagated to the ion channel-forming transmembrane domain (TMD), which ultimately promotes opening of the ion channel pore. Structurally, an individual LBD, composed of discontinuous polypeptide segments S1 and S2, adopts a clamshell-like structure (Fig. 1 A) (15, 16). Initial ligand interactions occur inside the cleft of this clamshell primarily with the membrane-distal lobe. Subsequently, the membrane-proximal lobe closes in on the ligand, creating additional high affinity ligand-protein interactions and trapping the ligand inside the closed cleft (1, 17, 18) (Fig. 1 A). This process leads to the generation of several Kcals of energy, the majority of which are thought to be transferred to the TMD through three LBD-TMD linkers promoting channel opening (19).

Here, we use chemical modulation and mutagenesis to define how the C726-C780 disulfide bond of the GluN1 LBD influences the activation gating mechanism of NMDA receptors. Consistent with previous reports, DTT-mediated reduction of this disulfide bond in GluN1/GluN2A receptors elicits a 25% potentiation of macroscopic current amplitudes that is manifested by increased frequency and longer duration channel openings at the single-channel level. We found that the disulfide bond plays a critical role in determining the overall distribution of open gating modes in GluN1/GluN2A NMDA receptors. Surprisingly though, we find that this disulfide bond has a localized effect on the kinetic mechanism of activation, being tightly integrated with a single preopen gating step already thought to be defined by GluN1-specific movements. Combining these two results, we provide significant functional evidence that this GluN1 hinge-region disulfide bond energetically destabilizes the closed cleft conformation of the LBD and confers substantial control to NMDA receptor gating.

Materials and Methods

Mutagenesis and expression

Xenopus oocytes were prepared, coinjected with rat GluN1 (GluN1a, accession No. P35439) and GluN2A (accession No. Q00959) cRNA, and maintained as previously described in Sobolevsky et al. (20). All numbering is for the mature protein using signal peptides of lengths of 18 (GluN1) and 19 (GluN2A) amino acids. A serine substitution (C780S) in GluN1 was generated using PCR-based methods (20). For mammalian cell expression, human embryonic kidney 293 (HEK 293) cells were cotransfected using Fugene 6 (Roche Applied Science, Roche Diagnostics, Indianapolis, IN) with cDNA for GluN1 and GluN2A, as well as a vector for enhanced green fluorescent protein (pEGFP-Cl; Clontech, Mountain View, CA), at a ratio of (in μg) 1:1:1.

Macroscopic current recordings

Xenopus oocytes membrane currents were recorded at room temperature (20–23°C) using a two-microelectrode voltage-clamp (TEV-200A; DAGAN, Minneapolis, MN) (20). The external solution consisted of 115 mM NaCl, 2.5 mM KCl, 0.18 mM BaCl2, 5 mM HEPES, and 100 μM ethylenediaminetetraacetic acid (EDTA) (pH 7.2, NaOH). Glycine (20 μM) and glutamate (200 μM), and the reducing agent dithiothreitol (DTT, 4 mM) were applied with the bath solution.

Single-channel recordings and analysis

Single-channel recordings were made at steady state using the cell-attached configuration on HEK 293 cells as described in detail previously by Talukder et al. (21). Briefly, currents were acquired using an Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA) filtered at 10 kHz (four-pole Bessel filter) and digitized at 40 kHz (model No. ITC-16 DA/AD converter interfaced with the software PatchMaster; both by HEKA Elektronik, Lambrecht, Germany). The bath solution consisted of 150 mM NaCl, 2.5 mM KCl, and 10 mM HEPES (pH 8, NaOH) (22). Recording electrodes were filled with an external (pipette) solution consisting of the bath solution supplemented with 1 mM EDTA, 0.1 mM glycine, and 1 mM glutamate. A voltage of +100 mV was applied through the recording pipette to elicit inward currents. For experiments assessing the effect of DTT on NMDA receptor single-channel activity, we treated the NMDA receptor-transfected HEK cells with 4 mM DTT both in our bath solution and in our patch pipette (containing the agonists).

All single-channel analysis was done using QuB (http://www.qub.buffalo.edu). Patches analyzed in this article contained a single NMDA receptor. For GluN1/GluN2A or GluN1(C780S)/GluN2A receptors with or without DTT, the recordings consisted of long clusters of activity separated by seconds-long periods of zero-activity, making it straightforward to detect more than one channel in the patch as simultaneous openings. In these cases, given the high Po (0.5–0.98) of GluN2A containing receptors and the minutes-long duration (with 10,000–320,000 events) of recordings without any apparent multiple openings, we are highly confident that these recordings certainly contained only a single channel in the patch (23).

Records cleaned of long periods of high noise (21) were idealized using the SKM algorithm after filtering to 12 kHz with a Gaussian digital filter. A conservative dead-time of 0.15 ms was imposed across all recording files. The idealization protocol may have missed very fast events. For our analysis, we did not correct for such missed events and assume that they were largely equal across different experimental conditions.

Kinetic analysis was performed using the maximum interval likelihood (MIL) algorithm in QuB. State models with increasing open and closed states were constructed and fitted to the recordings until log-likelihood (LL) values improved by <10 LL units/added state. We used a linear fully liganded state model containing three closed states, two desensitized states, and two-to-four open states (see Fig. 5) of NMDA receptor gating (24). The open-time components, comprising one common short duration (O1) and up to three long-duration (O2-4) intervals, arise from modal gating of NMDA receptors (25). Modal gating was observed for wild-type (WT) and mutant receptors with or without DTT (see Table 2), though not all modes (low, medium, high) were observed for all patches in either WT or under various experimental conditions. Time constants and the relative areas of each component, the transition rate constants, as well as mean closed time (MCT) and mean open time (MOT), were averaged for each receptor without and with DTT pretreatment and compared to each other.

Figure 5.

Figure 5

Open in a new tab

GluN1-specific C726-C780 disulfide bond exclusively influences late gating transitions leading to pore opening. (A and B) Sequential state model (24) of NMDA receptor activation with the rate constants (s−1) of transitions averaged from fits of individual single-channel recordings. Significant differences are indicated with asterisks (P < 0.05, Student's t-test). Models were generated with the maximum number of open states best fit to that particular recording. Table S2 shows mean (mean ± SE) values for these kinetic rate constants including those for GluN1(C780S)/GluN2A + DTT. (C) Free-energy landscape plotted with respect to C3. The off-pathway steps to and from C4 and C5 are excluded. For clarity, only the O4 open state is shown. The three traces (solid line, GluN1/GluN2A; shaded line, GluN1/GluN2A + DTT; and dashed line, GluN1(C780S)/GluN2A) are horizontally offset for clarity. (D) Simulated macroscopic currents, obtained by using the software QuB and based on models shown in panel A, for 500 receptors elicited by a 1 s pulse of agonists. The kinetic models and rate constants from panel A are used for simulations. Two consecutive ligand-binding steps (solid currents, GluN1/GluN2A and shaded currents, GluN1/GluN2A + DTT) were connected to the C3 gating step, with glutamate binding and unbinding constants of 1.7 × 107 M−1 s−1 and 60 s−1, respectively.

Table 2.

GluN1 C726-C780 disulfide bond affects multiple-channel open-time components

N1/N2A N1/N2A +DTT N1(C780S)/N2A N1(C780S)/N2A +DTT
No. of recordings 8 7 6 6
τO1, ms 0.21 ± 0.04 0.39 ± 0.25 0.15 ± 0.02 0.25 ± 0.07
aO1, % 1.7 ± 0.2 5.8 ± 2.9 3.2 ± 0.4∗ 3.2 ± 0.5∗
τO2, ms 4.5 ± 0.8 (8) 3.4 ± 1.0 (6) 3.1 ± 0.8 (6) 6.0 ± 2.0 (4)
aO2, % 37 ± 7.3 12 ± 4.1∗ 9.1 ± 2.8∗ 39 ± 20
τO3, ms 10 ± 1.3 (8) 16 ± 2.0∗ (7) 15 ± 1.9∗ (6) 23 ± 1.5∗ (6)
aO3, % 60 ± 6.8 53 ± 6.6 60 ± 7.2 44 ± 9.3
τO4, ms 43 ± 15 (6) 37 ± 3.4 (7) 33 ± 3.0 (6) 44 ± 2.9 (4)
aO4, % 2.0 ± 0.9 31 ± 9∗ 28 ± 6.5∗ 41 ± 10∗
Open in a new tab

Mean values (mean ± SE) for the durations and their relative occupancies of up to four fitted exponential components in the open time histograms. Idealization and maximum interval likelihood (MIL) fitting with five closed and two-to-four open states was done with the software QuB. The four open components arise from modal gating, with each mode containing a common short duration (O1) and one of three long duration (O2, O3, or O4) openings. All recordings displayed at least two modes and most displayed all four. The number of recordings observed for each mode is shown in parentheses.

∗

P < 0.05, relative to GluN1/GluN2A (Student's t-test).

Macroscopic current simulations

Simulations were made using QuB software from 500 channels, each with a 7.5 pA current amplitude. The simulated pulse consisted of a fully liganded square pulse lasting 1 s during a 5-s recording (250-ms prepulse). Kinetic models and rate constants used for each simulation are given in Fig. 5 D. Two consecutive ligand binding steps were connected to the C3 gating step, with glutamate binding and unbinding constants of 1.7 × 107 M−1 s−1 and 60 s−1, respectively (26, 27). Peak (Ipk) and steady-state (Iss) currents were measured. Percent of desensitization was calculated as 100 × (Ipk – Iss)/Ipk. The current trace during deactivation was fitted with a single exponential function, with τdeact = time constant of deactivation.

Statistics

Data analysis was done using the softwares IgorPro (WaveMetrics, Lake Oswego, OR), QuB, and Microsoft Excel (Microsoft, Redmond, WA). Results are presented as mean ± SE. A Student's t-test was used to test for significance with the reference in all instances WT GluN1/GluN2A (24). Significance was defined at P < 0.05.

Results

Redox-dependent potentiation of NMDA receptor macroscopic currents is mediated by a GluN1-specific C726-C780 disulfide bond

Reduction of NMDA receptors by extracellularly applied reducing agents such as DTT potentiates agonist-activated macroscopic currents (12) via reduction of an endogenous disulfide bond between the cysteine pair C726 and C780 in the GluN1 LBD (13, 14) (Fig. 1 A). We reproduced this redox effect on GluN1/GluN2A NMDA receptors expressed in Xenopus oocytes in the absence of any other modulatory divalents (100 μM EDTA) (Fig. 1, B and C). Extracellular application of DTT (4 mM) potentiates macroscopic agonist-activated (20 μM glycine and 200 μM glutamate) currents carried by WT GluN1/GluN2A receptors (27 ± 5% potentiation, n = 8; mean ± SE, n = number of recordings) (Fig. 1 B, top and C). In NMDA receptors lacking the GluN1 C726-C780 disulfide bond (serine substitution at residue C780, hereby referred to as GluN1(C780S)), this DTT-induced potentiation was abolished (−3 ± 1% inhibition, n = 8) (Fig. 1 B, bottom, and C).

DTT exposure to GluN1/GluN2A receptors increases the efficiency of channel openings

To address the kinetic mechanisms underlying this redox-dependent potentiation, we recorded the activity of single GluN1/GluN2A NMDA receptors in the cell-attached configuration. NMDA receptors were activated with saturating agonist concentrations (0.1 mM glycine and 1 mM glutamate) either in the absence (Fig. 2 A, top) or continuous presence (Fig. 2 A, bottom) of 4 mM DTT (see Materials and Methods). We made long-duration recordings with at least 10,000 events for greater statistical power (23). In the absence of divalent cations (1 mM EDTA) and proton inhibition (pH 8) (28, 29), GluN1/GluN2A typically undergo unitary-level openings that are clustered together, with each cluster separated from the next by sustained periods of inactivity (Fig. 2 A, top).

Figure 2.

Figure 2

Open in a new tab

In the presence of DTT, a single WT GluN1/GluN2A receptor undergoes high-frequency and long-duration openings. (A) Representative recordings of two different on-cell patches containing single GluN1/GluN2A receptors either in the absence (top) or presence (bottom) of DTT (4 mM). Recordings were done under steady-state conditions (0.1 mM glycine and 1 mM glutamate) (digitized at 40.0 kHz, filtered at 1 kHz) from transiently transfected HEK cells. For each, the bottom trace is an expanded view (filtered at 5 kHz) of the respective boxed regions. (B) Open-time duration of the same two single channel patches (top and bottom) shown in panel A. For these patches, the open-time duration histograms were well fitted with four exponential components. (Insets) Time constants and relative areas of the exponential components. (C) Closed-time duration histograms of the same two single channel patches (top and bottom) shown in panel A. The closed-time duration histograms were well fitted with five exponential components. (Insets) Time constants and relative areas of the exponential components.

When exposed to DTT, the basic architecture of GluN1/GluN2A single-channel openings is noticeably different, with the receptors opening to the same unitary current level but staying open for extended periods of time (Fig. 2 A, bottom). The composite channel open-time histogram fitted with the sum of multiple exponentials shows this single-channel current phenotype with more openings shifted toward longer durations (Fig. 2 B). Additionally, subtle changes are also evident in the channel closed-time histogram, especially for the short-duration closures (<1 log millisecond range) (Fig. 2 C).

As summarized in Table 1, quantification of these single-channel parameters reveals that DTT-induced reduction of GluN1/GluN2A receptors exclusively alters gating by causing a 22% increase in equilibrium channel open probability (eq. Po) (GluN1/GluN2A versus GluN1/GluN2A + DTT; 0.71 ± 0.04, n = 8 vs. 0.88 ± 0.04, n = 7) without affecting single channel conductance (as gauged by the unitary current amplitude, I) (−7.6 ± 0.3 vs. −7.5 ± 0.2 pA). The higher Po is manifested chiefly by an increased mean open time (MOT) (8.1 ± 0.9 vs. 20 ± 4 ms). Mean closed time (MCT), however, is unchanged (3.0 ± 0.04 vs. 2.5 ± 0.9 ms).

Table 1.

Absence of the C726-C780 disulfide bond in GluN1 LBD increases NMDA receptor single-channel activity

No. I (pA) eq. Po MCT (ms) MOT (ms)
N1/N2A 8 −7.6 ± 0.3 0.71 ± 0.04 3.0 ± 0.4 8.1 ± 0.9
N1/N2A + DTT 7 −7.5 ± 0.2 0.88 ± 0.04∗ 2.5 ± 0.9 20 ± 3.8∗
N1(C780S)/N2A 6 −6.9 ± 0.4 0.86 ± 0.03∗ 2.8 ± 0.6 19 ± 2.2∗
N1(C780S)/N2A + DTT 6 −8.1 ± 0.5 0.83 ± 0.05 4.3 ± 1.6 24 ± 4.4∗
Open in a new tab

Mean values (mean ± SE) for single channel current amplitudes (I), equilibrium open probability (eq. Po), mean closed time (MCT), and mean open time (MOT). Single-channel recordings were done in the cell-attached mode under steady-state conditions at pH 8 and in 1 mM EDTA (see Materials and Methods). Idealization and maximum interval likelihood (MIL) fitting with five closed and two-four open states was done with the software QuB. The value eq. Po is defined as the fractional occupancy of the open states in the MIL fitted single-channel recordings.

∗

P < 0.05, relative to GluN1/GluN2A (Student's t-test).

DTT-induced effects on single-channel activity are mediated by the GluN1-specific C726-C780 disulfide bond

GluN1(C780S)/GluN2A receptors are not potentiated by DTT (Fig. 1). We therefore performed single-channel recordings of GluN1(C780S)/GluN2A (Fig. 3), under the same conditions as WT receptors to determine whether its activity was similar to that in WT receptors exposed to DTT. Single-channel currents of these mutant receptors, either in the absence (Fig. 3, top current traces and histograms) or presence (Fig. 3, bottom current traces and histograms) of DTT, are phenotypically indistinguishable from those of DTT-exposed WT receptors (compare to Fig. 2, bottom current traces and histograms). Likewise, Po (0.86 ± 0.03) and MOT (19 ± 2) of GluN1(C780S)/GluN2A receptors (n = 6) are high, reminiscent of DTT-exposed WT receptors (Table 1). DTT exposure to GluN1(C780S)/GluN2A receptors (n = 6) does not cause any additional changes to the measured single-channel parameters (Table 2).

Figure 3.

Figure 3

Open in a new tab

GluN1(C780S)/GluN2A receptors display high single-channel activity like DTT-exposed WT receptors. (A–C) Same as Fig. 3. Note that, like DTT-exposed WT receptors (Fig. 2B), a single GluN1(C780S)/GluN2A receptor undergoes high-frequency and long-duration openings independent of DTT.

Thus, the redox effects on GluN1/GluN2A receptors are mediated solely by the C726-C780 disulfide bond in GluN1 (11). In the absence of this disulfide bond, either by its DTT-induced reduction (Fig. 2) or by preventing its formation with C780S substitution (Fig. 3), NMDA receptors undergo a more efficient activation gating mechanism.

The GluN1-specific C726-C780 disulfide bond regulates the distribution of open gating states

To further explore the gating properties defined by the GluN1-specific C726-C780 disulfide bond, we analyzed the individual exponential components in the composite open-time histograms of the single-channel recordings. NMDA receptors display modal gating, wherein channel openings consist of a common short duration (O1) and one of three long duration (O2, O3 or O4) components in each mode (O1 – O2, O1 – O3, or O1 – O4) (25). All of our single-channel recordings displayed at least three open-time components and most (20 of 27) had four (Table 2). Furthermore, O1, the shortest-duration open-time component common to all modes in either GluN1/GluN2A or GluN1(C780S)/GluN2A receptors, was not consistently affected by the redox status of the C726-C780 disulfide bond (Table 2). Thus, in the absence of the GluN1-specific C726-C780 disulfide bond, either DTT-exposed GluN1/GluN2A or GluN1(C780S)/GluN2A receptors, the general features of modal gating—transitions between three open modes—persisted.

The effects on channel gating by removal of the disulfide bond were generally concentrated on the longer duration open-time components. Strikingly, the exposure of WT GluN1/GluN2A receptors to DTT increased the fraction of channel openings to O4 by ∼16-fold (aO4) (GluN1/GluN2A versus GluN1/GluN2A + DTT; 2 ± 0.9 vs. 31 ± 9%) (Table 2). This was apparently accomplished by shifting the channel openings away from O2 and O3 and into O4, with significant decreases in the fraction of O2 openings (aO2, by 37 ± 7 vs. 12 ± 4%) (Table 2). Furthermore, the duration of O3 was slightly but significantly increased (τO3, by 10 ± 1.3 vs. 16 ± 2 ms) (Table 2). The same changes to channel openings are fully recapitulated in GluN1(C780S)/GluN2A receptors under both reducing and nonreducing conditions (Table 2). In summary, whereas O4 normally represents a small fraction of NMDA receptors openings (∼2%), disruption of the GluN1-specific disulfide bond drastically increases the receptor preference for this long-duration open state. These shifts to long-duration openings were stable over time (see Fig. S1 in the Supporting Material).

Disruption of the GluN1-specific C726-C780 disulfide bond exclusively affects the shortest duration channel-closed state

Although removal of the GluN1 C726-C780 disulfide bond does not affect the overall mean closed time (Table 1), it elicits changes in the distribution of discrete closed-time events (Figs. 2 B and 3 B). To define these changes, we analyzed the individual exponential components that comprised the closed-time duration histograms. Consistent with NMDA receptor gating behavior, all of our single-channel recordings were fitted best with five exponential closed time components: three shorter intracluster (C1, C2, and C3) and two longer intercluster (C4 and C5) components (24, 29, 30, 31). Of these, the most notable changes in the absence of the GluN1 disulfide bond were a shift away from C2 to the shortest duration component (C1) (Fig. 4). The duration of C1 was slightly but significantly increased (τ1) (GluN1/GluN2A versus GluN1/GluN2A + DTT; 0.13 ± 0.002 vs. 0.17 ± 0.01 ms) (Fig. 4 A). More prominently though, the fractional occupancy of C1 was increased 2.5-fold (a1) (14 ± 1 vs. 40 ± 3%), likely by shifting it away from C2 (a2) (54 ± 6 vs. 35 ± 2%) (Fig. 4 B; see Table S1 in the Supporting Material). These specific changes in closed-time intervals were recapitulated in GluN1(C780S)/GluN2A receptors under the two tested redox conditions (Fig. 4). Therefore, the C726-C780 disulfide bond helps define GluN1 specific preopen gating behavior.

Figure 4.

Figure 4

Open in a new tab

GluN1/GluN2A receptors with no C726-C780 disulfide bond, display localized changes to a single closed-time component. (Top) Mean fold-change in duration (mean ± SE) of closed-time components as determined by MIL fitting of single-channel recordings. For WT GluN1/GluN2A, the time constants were: τ1 0.13 ± 0.002 ms, τ2 1.1 ± 0.1 ms, τ3 2.5 ± 0.3 ms, τ4 31 ± 6 ms, and τ5 1100 ± 140 ms. Average values for all constructs and experiment conditions are shown in Table S1 in the Supporting Material. (Bottom) Mean relative areas (mean ± SE) of the closed-time components. Note values of a4 and a5 are displayed with an expanded y axis (right). Significant differences are indicated (asterisks) (P < 0.05, Student's t-test).

Redox effects mediated by the GluN1-specific C726-C780 disulfide bond influence a fast preopen gating step

A kinetic mechanism of NMDA receptor activation, that arranges the multiple open- and closed- time components as sequential gating reactions, has been described (24, 29, 30, 31). The mechanism, C3 – C2 – C1 – O1, captures as-of-yet undefined conformational events initiating with ligand-bound cleft closure and terminating in channel pore opening, with the two longest closed-time components, C4 and C5, representing off-pathway microscopic desensitized states (Fig. 5) (22, 27, 32).

Within this activation mechanism, disulfide bond reduction locally affects only the late gating transitions leading to pore opening (C2 – C1 – O1). Both the forward and reverse rates for the C2 – C1 transition were significantly slowed (GluN1/GluN2A versus GluN1/GluN2A + DTT) (kC2→C1: 1900 ± 140 vs. 1100 ± 60 s−1; kC1→C2: 3600 ± 130 vs. 1700 ± 120 s−1) (Fig. 5 A; and see Table S2). Furthermore, the forward rate for the C1 – O transition was significantly faster (kC1→O1) (3400 ± 190 vs. 4400 ± 280 s−1) (Fig. 5 A). Biochemical or mutagenic disulfide bond elimination showed comparable differences in the kinetic activation mechanism (Fig. 5 B). In the context of the thermodynamics of the activation gating mechanism, the specific changes caused by either DTT-induced reduction of or eliminating the C726-C780 disulfide bond lowers the energetic barrier between the C2 and C1 states, thus biasing the receptor gating preference toward pore opening (Fig. 5 C). Further, there was a strong bias, once in the open state, toward longer openings (Fig. 5, A and B; and see Table S2). From a thermodynamic perspective, eliminating the disulfide bond greatly lowers the energy barrier for entry into O4 (Fig. 5 C).

To validate the kinetic alterations brought about by disruption of the C726-C780 disulfide bond, we simulated macroscopic currents (Fig. 5 D). The simulated macroscopic currents based on the kinetic models of WT NMDA receptors with or without the intact disulfide bond reliably reproduce the experimentally observed DTT-mediated changes in macroscopic currents (12, 33, 34). First, at steady state, the simulated macroscopic currents were potentiated 27% by DTT, like the DTT-mediated potentiation of macroscopic currents in Xenopus oocytes (Fig. 1, B and C). Second, the simulated currents showed a ∼42% reduction in desensitization (100 × (Ipk – Iss)/Ipk, see Materials and Methods) caused by DTT (12). Third, the DTT-based simulated macroscopic currents reduced the rate of deactivation with a twofold increase in the time constant of deactivation (Fig. 5 D, inset) (12, 34).

Discussion

In glutamate-activated ion channels, the energy driving gating of the ion channel pore originates in the ligand-binding domain (LBD) and propagates along structural linkers to the transmembrane domain (TMD). Working with NMDA receptors, we studied the gating effect of a single disulfide bond (C726-C780) located at the hinge region of the LBD in the obligatory GluN1 subunit (14). Eliminating this GluN1-specific C726-C780 disulfide bond by either DTT-mediated reduction (Fig. 2) or mutagenesis (Fig. 3) results in more efficient gating, with the probability of pore opening significantly increased (Table 1) (6). In terms of a kinetic mechanism of activation gating, eliminating this disulfide bond had two notable effects. The first effect was a dramatic increase in the occurrence of long open events (O4) (Table 2). Because these long open events are a feature of modal gating in NMDA receptors (24, 25), the GluN1-specific disulfide bond and its spontaneous breakage may represent one component mediating modal gating in NMDA receptors. The second major effect occurred in the preopen gating steps where eliminating this disulfide bond exclusively affected a fast kinetic step (C2 – C1). This result is consistent with other work suggesting that this fast step is GluN1-specific (30, 35). Overall, our results suggest that the dynamics of the GluN1 LBD, influenced by the redox state, regulate components of modal gating and late gating steps in channel opening/closing.

NMDA receptors display modal gating consisting of three open-state modes (low, medium, high) (24, 25). In any given mode the channel opens to a common short duration (τO1 ∼ 0.2 ms) and one of three long duration (τO2 ∼ 4.5 ms, τO3 ∼ 10 ms, and τO4 ∼ 43 ms) states (Table 2). In our hands, WT GluN1/GluN2A receptors displayed all three gating modes with O2 (low mode) and O3 (medium mode) comprising the overwhelming majority (∼97%) of openings (Table 2). Although modal gating was intact in the absence of the GluN1 C726-C780 disulfide bond, elimination of the disulfide bond increased the duration of O3 (∼1.5-fold), but more prominently the occupancy of O4 (∼16-fold) at the expense mainly of O2 occupancies. Hence, this GluN1-specific disulfide bond might contribute to the distribution of modal gating states, with limited openings to O4 under ambient oxidizing conditions reflecting transient breakage of this disulfide bond. Still, whether breakage of this endogenous disulfide bond represents a component of modal gating especially in native or synaptic NMDA receptors (32) will need to be directly tested. We anticipate that additional structural elements and mechanisms must also contribute to modal gating because it persisted even with the elimination of this disulfide bond (Table 2).

NMDA receptor activation gating proceeds through a core kinetic mechanism of sequential transitions across at least three preopen states, C3 → C2 → C1 → O1, as well as two off-pathway desensitized states, C4 and C5 (Fig. 5, A and B) (24, 29, 30, 31). This core gating mechanism has been postulated to incorporate a slow (C3 – C2, τs ∼ 13 ms) GluN2-specific and a fast (C2 – C1, τf ∼ 0.7 ms) GluN1-specific preopen gating step, followed by a very fast (C1 – O, τvf < 0.1 ms) concerted pore opening (22, 30, 31). Elimination of the GluN1-specific disulfide bond affected the forward and reverse rates associated with the fast C2 – C1 preopen gating step (Fig. 5, A–C) consistent with this fast step being GluN1-specific. Additionally, although the relationship between kinetic entry/exit rates and the time constants/areas of the corresponding exponential components are complex, we detected redox-dependent changes in the duration (τ1 ∼ 0.13 ms) and occupancy of a single exponential component, reflecting the observed changes in rates associated with C1 (Fig. 4). (Note that τ1 ≠ τf due to different single channel data collection and analysis regimes, e.g., filtering and digitization frequencies, tcrit, etc.) Thus, the C726-C780 disulfide bond modulates GluN1 conformations involved in specific kinetically defined preopen gating steps.

The conformational dynamics of agonist binding in ionotropic glutamate receptors encompasses two general steps: a binding step termed “docking”, and a clamshell-like closing step termed “locking” where the cleft surrounds agonist preventing its exit. This “locking” step presumably provides the majority of free energy required for pore opening (36, 37). Based on the findings that eliminating this disulfide bond increases agonist efficacy (13) and that it is located at the back, “hinge-region” of the LBD, it has been proposed that this disulfide bond normally hinders cleft closure or “locking” of the LBD (14). By hindering the “locking” process, the disulfide bond presents an energetic barrier to the closed-cleft state and hence destabilizes the conformational changes inducing pore opening. Thus, in the absence of this disulfide bond, we saw a decline in the energetic barrier in the C2 – C1 – O transition (Fig. 5 C), promoting channel open states. In addition, Kussius and Popescu (35) trapped the GluN1 LBD in a closed-cleft conformation using introduced cysteines in the mouth of the cleft and observed a similar kinetic disruption, specifically changes in the fast C2 – C1 transition.

Nevertheless, these manipulations, whether trapping the closed-cleft conformation via disulfides or eliminating the hinge-region disulfide bond, are not completely equivalent. Unlike their study, we found a robust increase in Po (Table 1) arising from a redistribution of open states (Table 2). We also found that the direction of changes in the C2 – C1 preopen steps was not identical: “locking” cleft closure enhanced the C2 – C1 forward transition, whereas elimination of the hinge disulfide bond slowed the C2 – C1 reverse transition. The basis for these differences is unknown, but might reflect, for example, that elimination of the hinge disulfide bond allows for additional twisting of LBD helices (38) and/or the hinge disulfide bond might have additional interactions with the LBD-TMD linkers or with the TMD themselves. In any case, additional experiments will be needed to define the specific structural effect of removing the GluN1-specific C726-C780 disulfide bond.

Energetically, eliminating the hinge disulfide bond appears to stabilize LBD conformations that favor the open pore conformation. Interestingly, if the sole energetic effect of the hinge disulfide bond is to destabilize cleft closure, this would suggest that the dynamics of the GluN1 LBD plays a central role in regulating late gating steps including the stability of the open state (17, 35, 39, 40, 41, 42, 43). Thus, the preference for the specific gating modes displayed by a functional receptor might arise from structural dynamics within the LBD.

Because the energy provided by LBD cleft closure mediates NMDAR gating, stabilizing the closed cleft state by disulfide bond reduction can readily account for the observed changes in gating. Overall, our results suggest that under endogenous oxidative conditions, energy provided by LBD cleft closure is dissipated through structural components distant from the ion channel pore. This disulfide bond likely represents an “energy sink” draining the energy necessary for stabilizing long-lived open states. By eliminating this sink, the energy of LBD cleft closure is more efficiently coupled to channel openings and long open-state modes. This result is promising in defining the structural elements modulating modal gating and would lead us to argue that modal gating is not solely due to intracellular interactions. Other unidentified extracellular components in both GluN1 and non-GluN1 subunits may also confer molecular control to gating in NMDA receptors.

Acknowledgments

We thank Dr. Stephen Traynelis and Quan Gan for helpful discussions and/or comments on the manuscript. We thank Janet Allopenna for technical assistance.

This work was supported by a National Institutes of Health RO1 grant (MH066892) from the National Institute of Mental Health (to L.P.W.) and an American Heart Association Predoctoral Fellowship (to I.T.).

Editor: Cynthia Czajkowski.

Footnotes

This is an Open Access article distributed under the terms of the Creative Commons-Attribution Noncommercial License (http://creativecommons.org/licenses/by-nc/2.0/), which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Supporting Material

Document S1. One figure and two tables
mmc1.pdf (157.6KB, pdf)

References

  • 1.Traynelis S.F., Wollmuth L.P., et al. Dingledine R. Glutamate receptor ion channels: structure, regulation, and function. Pharmacol. Rev. 2010;62:405–496. doi: 10.1124/pr.109.002451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Kalia L.V., Kalia S.K., Salter M.W. NMDA receptors in clinical neurology: excitatory times ahead. Lancet Neurol. 2008;7:742–755. doi: 10.1016/S1474-4422(08)70165-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Cull-Candy S., Brickley S., Farrant M. NMDA receptor subunits: diversity, development and disease. Curr. Opin. Neurobiol. 2001;11:327–335. doi: 10.1016/s0959-4388(00)00215-4. [DOI] [PubMed] [Google Scholar]
  • 4.Paoletti P., Neyton J. NMDA receptor subunits: function and pharmacology. Curr. Opin. Pharmacol. 2007;7:39–47. doi: 10.1016/j.coph.2006.08.011. [DOI] [PubMed] [Google Scholar]
  • 5.Aizenman E., Lipton S.A., Loring R.H. Selective modulation of NMDA responses by reduction and oxidation. Neuron. 1989;2:1257–1263. doi: 10.1016/0896-6273(89)90310-3. [DOI] [PubMed] [Google Scholar]
  • 6.Tang L.H., Aizenman E. Allosteric modulation of the NMDA receptor by dihydrolipoic and lipoic acid in rat cortical neurons in vitro. Neuron. 1993;11:857–863. doi: 10.1016/0896-6273(93)90115-8. [DOI] [PubMed] [Google Scholar]
  • 7.Sanchez R.M., Wang C., et al. Jensen F.E. Novel role for the NMDA receptor redox modulatory site in the pathophysiology of seizures. J. Neurosci. 2000;20:2409–2417. doi: 10.1523/JNEUROSCI.20-06-02409.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Li X., Wallin C., et al. Sandberg M. Net efflux of cysteine, glutathione and related metabolites from rat hippocampal slices during oxygen/glucose deprivation: dependence on γ-glutamyl transpeptidase. Brain Res. 1999;815:81–88. doi: 10.1016/s0006-8993(98)01097-x. [DOI] [PubMed] [Google Scholar]
  • 9.Tang L.H., Aizenman E. The modulation of N-methyl-D-aspartate receptors by redox and alkylating reagents in rat cortical neurones in vitro. J. Physiol. 1993;465:303–323. doi: 10.1113/jphysiol.1993.sp019678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Brimecombe J.C., Boeckman F.A., Aizenman E. Functional consequences of NR2 subunit composition in single recombinant N-methyl-D-aspartate receptors. Proc. Natl. Acad. Sci. USA. 1997;94:11019–11024. doi: 10.1073/pnas.94.20.11019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Brimecombe J.C., Potthoff W.K., Aizenman E. A critical role of the N-methyl-D-aspartate (NMDA) receptor subunit (NR) 2A in the expression of redox sensitivity of NR1/NR2A recombinant NMDA receptors. J. Pharmacol. Exp. Ther. 1999;291:785–792. [PubMed] [Google Scholar]
  • 12.Köhr G., Eckardt S., et al. Seeburg P.H. NMDA receptor channels: subunit-specific potentiation by reducing agents. Neuron. 1994;12:1031–1040. doi: 10.1016/0896-6273(94)90311-5. [DOI] [PubMed] [Google Scholar]
  • 13.Sullivan J.M., Traynelis S.F., et al. Lipton S.A. Identification of two cysteine residues that are required for redox modulation of the NMDA subtype of glutamate receptor. Neuron. 1994;13:929–936. doi: 10.1016/0896-6273(94)90258-5. [DOI] [PubMed] [Google Scholar]
  • 14.Furukawa H., Gouaux E. Mechanisms of activation, inhibition and specificity: crystal structures of the NMDA receptor NR1 ligand-binding core. EMBO J. 2003;22:2873–2885. doi: 10.1093/emboj/cdg303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Oswald R.E., Ahmed A., et al. Loh A.P. Structure of glutamate receptors. Curr. Drug Targets. 2007;8:573–582. doi: 10.2174/138945007780618526. [DOI] [PubMed] [Google Scholar]
  • 16.Mayer M.L. Structure and mechanism of glutamate receptor ion channel assembly, activation and modulation. Curr. Opin. Neurobiol. 2011;21:283–290. doi: 10.1016/j.conb.2011.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Abele R., Keinanen K., Madden D.R. Agonist-induced isomerization in a glutamate receptor ligand-binding domain. A kinetic and mutagenetic analysis. J. Biol. Chem. 2000;275:21355–21363. doi: 10.1074/jbc.M909883199. [DOI] [PubMed] [Google Scholar]
  • 18.Cheng Q., Du M., et al. Jayaraman V. Evolution of glutamate interactions during binding to a glutamate receptor. Nat. Chem. Biol. 2005;1:329–332. doi: 10.1038/nchembio738. [DOI] [PubMed] [Google Scholar]
  • 19.Sun Y., Olson R., et al. Gouaux E. Mechanism of glutamate receptor desensitization. Nature. 2002;417:245–253. doi: 10.1038/417245a. [DOI] [PubMed] [Google Scholar]
  • 20.Sobolevsky A.I., Prodromou M.L., et al. Wollmuth L.P. Subunit-specific contribution of pore-forming domains to NMDA receptor channel structure and gating. J. Gen. Physiol. 2007;129:509–525. doi: 10.1085/jgp.200609718. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Talukder I., Borker P., Wollmuth L.P. Specific sites within the ligand-binding domain and ion channel linkers modulate NMDA receptor gating. J. Neurosci. 2010;30:11792–11804. doi: 10.1523/JNEUROSCI.5382-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Erreger K., Dravid S.M., et al. Traynelis S.F. Subunit-specific gating controls rat NR1/NR2A and NR1/NR2B NMDA channel kinetics and synaptic signaling profiles. J. Physiol. 2005;563:345–358. doi: 10.1113/jphysiol.2004.080028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Colquhoun D., Hawkes A.G. Stochastic properties of ion channel openings and bursts in a membrane patch that contains two channels: evidence concerning the number of channels present when a record containing only single openings is observed. Proc. R. Soc. Lond. B Biol. Sci. 1990;240:453–477. doi: 10.1098/rspb.1990.0048. [DOI] [PubMed] [Google Scholar]
  • 24.Kussius C.L., Popescu G.K. Kinetic basis of partial agonism at NMDA receptors. Nat. Neurosci. 2009;12:1114–1120. doi: 10.1038/nn.2361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Popescu G., Auerbach A. Modal gating of NMDA receptors and the shape of their synaptic response. Nat. Neurosci. 2003;6:476–483. doi: 10.1038/nn1044. [DOI] [PubMed] [Google Scholar]
  • 26.Popescu G., Robert A., et al. Auerbach A. Reaction mechanism determines NMDA receptor response to repetitive stimulation. Nature. 2004;430:790–793. doi: 10.1038/nature02775. [DOI] [PubMed] [Google Scholar]
  • 27.Kussius C.L., Kaur N., Popescu G.K. Pregnanolone sulfate promotes desensitization of activated NMDA receptors. J. Neurosci. 2009;29:6819–6827. doi: 10.1523/JNEUROSCI.0281-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Banke T.G., Dravid S.M., Traynelis S.F. Protons trap NR1/NR2B NMDA receptors in a nonconducting state. J. Neurosci. 2005;25:42–51. doi: 10.1523/JNEUROSCI.3154-04.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.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. 2005;569:395–418. doi: 10.1113/jphysiol.2005.095349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Banke T.G., Traynelis S.F. Activation of NR1/NR2B NMDA receptors. Nat. Neurosci. 2003;6:144–152. doi: 10.1038/nn1000. [DOI] [PubMed] [Google Scholar]
  • 31.Auerbach A., Zhou Y. Gating reaction mechanisms for NMDA receptor channels. J. Neurosci. 2005;25:7914–7923. doi: 10.1523/JNEUROSCI.1471-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Zhang W., Howe J.R., Popescu G.K. Distinct gating modes determine the biphasic relaxation of NMDA receptor currents. Nat. Neurosci. 2008;11:1373–1375. doi: 10.1038/nn.2214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Choi Y., Chen H.V., Lipton S.A. Three pairs of cysteine residues mediate both redox and zn2+ modulation of the NMDA receptor. J. Neurosci. 2001;21:392–400. doi: 10.1523/JNEUROSCI.21-02-00392.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Paoletti P., Ascher P., Neyton J. High-affinity zinc inhibition of NMDA NR1-NR2A receptors. J. Neurosci. 1997;17:5711–5725. doi: 10.1523/JNEUROSCI.17-15-05711.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Kussius C.L., Popescu G.K. NMDA receptors with locked glutamate-binding clefts open with high efficacy. J. Neurosci. 2010;30:12474–12479. doi: 10.1523/JNEUROSCI.3337-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lau A.Y., Roux B. The free energy landscapes governing conformational changes in a glutamate receptor ligand-binding domain. Structure. 2007;15:1203–1214. doi: 10.1016/j.str.2007.07.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Lau A.Y., Roux B. The hidden energetics of ligand binding and activation in a glutamate receptor. Nat. Struct. Mol. Biol. 2011;18:283–287. doi: 10.1038/nsmb.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Birdsey-Benson A., Gill A., et al. Madden D.R. Enhanced efficacy without further cleft closure: reevaluating twist as a source of agonist efficacy in AMPA receptors. J. Neurosci. 2010;30:1463–1470. doi: 10.1523/JNEUROSCI.4558-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Maier W., Schemm R., et al. Laube B. Disruption of interdomain interactions in the glutamate binding pocket affects differentially agonist affinity and efficacy of N-methyl-D-aspartate receptor activation. J. Biol. Chem. 2007;282:1863–1872. doi: 10.1074/jbc.M608156200. [DOI] [PubMed] [Google Scholar]
  • 40.Zhang W., Cho Y., et al. 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. 2008;28:932–943. doi: 10.1523/JNEUROSCI.3309-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Robert A., Armstrong N., et al. Howe J.R. AMPA receptor binding cleft mutations that alter affinity, efficacy, and recovery from desensitization. J. Neurosci. 2005;25:3752–3762. doi: 10.1523/JNEUROSCI.0188-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Weston M.C., Gertler C., et al. Rosenmund C. Interdomain interactions in AMPA and kainate receptors regulate affinity for glutamate. J. Neurosci. 2006;26:7650–7658. doi: 10.1523/JNEUROSCI.1519-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Fenwick M.K., Oswald R.E. On the mechanisms of alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptor binding to glutamate and kainate. J. Biol. Chem. 2010;285:12334–12343. doi: 10.1074/jbc.M109.086371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Sobolevsky A.I., Rosconi M.P., Gouaux E. X-ray structure, symmetry and mechanism of an AMPA-subtype glutamate receptor. Nature. 2009;462:745–756. doi: 10.1038/nature08624. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. One figure and two tables
mmc1.pdf (157.6KB, pdf)

Articles from Biophysical Journal are provided here courtesy of The Biophysical Society

RESOURCES