|
Figure 1. The M3 transmembrane segment is the major channel-gating element in iGluRs. (A) Backbone structure of two glutamate receptor subunits (GluA2cryst; subunits B and C, Protein Data Bank accession no. 3KG2), each harboring an extracellular LBD (in gray) comprised of polypeptide segments S1 and S2, transmembrane segments M3 and M4, and their associated linkers M3–S2 and S2–M4. For clarity, the M1 transmembrane segment and the M2 pore loop are not shown. GluN1 is assumed to adopt the A/C conformation (red) and GluN2 the B/D conformation (blue) (Sobolevsky et al., 2009). The dashed line indicates point of view (looking down the ion channel) shown in B. Red and blue squares depict regions of intrasubunit cross-linking of the M3–S2 and S2–M4 linkers in GluN1 and GluN2A, respectively. Although there are limitations in comparing NMDA receptors to the AMPA receptor structure and the derived open-state structural model, for example, domain arrangements may be different (Stroebel et al., 2011), we use this information to illustrate general features of gating in iGluRs, specifically for the TMD. (B) Presumed gating movements of the M3 transmembrane segment leading to pore opening. (Left) Tetrameric arrangement of M3/M3–S2 and M4/S2–M4 adopting the A/C (~GluN1; red) and B/D (~GluN2A; blue) conformations in an antagonist-bound channel closed state (Sobolevsky et al., 2009). The M3 transmembrane helices line the channel pore (depicted with a dot), whereas the external helices surrounding this core are the M4 segments. (Right) Reorientation of the M3 segments in the channel open state as predicted from superposition of the GluA2cryst on the closed KcsA and the open Shaker K+ channels (Sobolevsky et al., 2009). In the present study, we restrict these gating rearrangements of M3 through intrasubunit cross-linking of the M3–S2 and S2–M4 linkers in GluN1 (A, red box) or GluN2A (A, blue box).
|
|
Figure 2. DTT-induced potentiation of macroscopic currents in NMDA receptors containing intrasubunit double-cysteine substitutions in GluN1 or GluN2A. (A and B) NMDA receptors with intrasubunit GluN1- or GluN2A-specific double-cysteine substitutions were assayed for DTT-induced changes in macroscopic current amplitudes using two-microelectrode voltage clamp in Xenopus oocytes. Double-cysteine–substituted GluN1 (A) or GluN2A (B) subunits were coexpressed with WT GluN2A or GluN1 subunits, respectively. (Left) Schematic representation of regions around M3–S2 and S2–M4 linkers. Positions substituted with cysteine are indicated with a “C” and numbered next to the endogenous residue. Tested pairs of cysteines are shown with a connecting line. Darker lines indicate pairs that showed significant DTT-induced current potentiation relative to GluN1/GluN2A and hence can presumably spontaneously cross-link. Numbering is for the mature protein. Proximal parts of S2 and the hydrophobic segments M3 and M4 are colored as magenta and gray, respectively. Boxed regions around the hydrophobic segments represent the α-helical extent of the transmembrane segments in an AMPA receptor structure (Sobolevsky et al., 2009). (Right) Mean percent change (±SEM; n ≥ 4) of current amplitudes after DTT. In the recording protocol for the GluN1 double-cysteine substitutions, (A) DTT was applied continuously in the presence and absence of agonists for at least 2 min (raw recordings not depicted). The recording protocol for the GluN2A double-cysteine substitutions (B) was identical to those in C. Filled bars indicate values significantly different from those of WT receptors (P < 0.05). Our experiments focused on GluN1(R645C,S784C)/GluN2A and GluN1/GluN2A(Q642C,K785C) receptors. (C) Representative membrane currents (holding potential, −60 mV) in Xenopus oocytes injected with WT GluN1/GluN2A, GluN1(R645C,S784C)/GluN2A, or GluN1/GluN2A(Q642C,K785C) receptors. Hereafter, GluN1(R645C,S784C) and GluN2A(Q642C,K785C) are referred to as GluN1(C,C) and GluN2A(C,C), respectively. Currents were elicited by coapplication of 20 µM glycine and 200 µM glutamate (thin black lines). 4 mM DTT (2 min; gray bars), applied in the presence of competitive antagonists DCKA (10 µM) and APV (100 µM) (open boxes), strongly potentiated subsequent current amplitudes in the double-cysteine–substituted receptors. (D) Mean percent change (±SEM; n ≥ 4) of current amplitudes after DTT. Filled bars indicate values significantly different from those of WT receptors (P < 0.05). (E) Western blot analysis of membrane proteins purified from Xenopus oocytes under nonreducing conditions. Formation of intersubunit cross-linking, either homomeric or heteromeric, was assayed with anti-GluN1 (top) or anti-GluN2A (bottom) antibodies. The “+ Control” is GluN1(N521C,L777C)/GluN2A(E516C,L780C) receptors that form intersubunit dimers (Furukawa et al., 2005). Expected molecular weights are: monomeric GluN1 (114 kD) and GluN2A (173 kD), homodimeric GluN1 (228 kD) and GluN2A (346 kD), and heterodimeric GluN1/GluN2A (287 kD). Other than the monomeric bands, no apparent homomeric or heteromeric dimer bands were detected for GluN1/GluN2A or either of the double-cysteine–substituted receptors, although dimers were present for the “+ Control” (n = 4).
|
|
Figure 3. A subset of cell surface NMDA receptors with double-cysteine–substituted GluN1 or GluN2A is resistant to pore block by MK801. To optimize current amplitudes in these experiments, we injected WT GluN1/GluN2A at 1 ng/µL and GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) at 10 ng/µL into Xenopus oocytes. (A and B) Example recordings depicting steady-state MK801 inhibition of NMDA receptor–mediated macroscopic currents. MK801 (open bar; 1–2 µM; 1 min), applied in the presence of agonists (thin lines), inhibited current amplitudes for GluN1/GluN2A (A), GluN1(C,C)/GluN2A (B), and GluN1/GluN2A(C,C) (not depicted) receptors. Subsequent application of DTT (filled bar) in the channel closed state (as in Fig. 2 C) significantly potentiated current amplitudes of the double-cysteine–substituted receptor (B) relative to WT GluN1/GluN2A (A). (C) Mean percent change (±SEM; n ≥ 4) in current amplitudes either immediately after MK801 (MK801) or after MK801, but with an intervening treatment by DTT in the presence of antagonists (DTT) or antagonists alone (antag.). For DTT and antagonist-alone treatments, percent change was calculated relative to the current amplitudes preceding these treatments but after MK801 block. Negative and positive values represent current inhibition and potentiation, respectively. Filled bars indicate values significantly different from those of WT GluN1/GluN2A receptors (P < 0.05). (D and E) 25 nM MK801 was applied in the presence of agonists until steady-state current inhibition was reached. (E, right) For GluN1(C,C)/GluN2A, MK801 was also applied to DTT-potentiated currents. Single- (gray dashed lines) and biexponential (green dashed lines) fits to MK801-mediated current inhibition are shown, as well as the residuals (Ires) to the two fits. For GluN1/GluN2A, single-exponential fits were sufficient to describe the time course of MK801-mediated current inhibition, whereas for GluN1(C,C)/GluN2A, biexponential fits were required, as determined by qualitative minimization of Ires. (F) Normalized MK801-mediated inhibition of currents with the overlayed best fits (dashed lines). (G) Averaged time constants of single-exponential fits (τ), as well as the fast (τf) and slow (τs) components of the biexponential fits (±SEM) to MK801-mediated current inhibition. (H) Averaged percentage of area (±SEM) occupied by each component of the exponential fits.
|
|
Figure 4. NMDA receptors with intrasubunit GluN1 or GluN2A disulfide cross-links show dramatically reduced single-channel activity that can be reversed by DTT. (A–C) Recordings of on-cell patches containing single GluN1/GluN2A (A), GluN1(C,C)/GluN2A (B), or GluN1/GluN2A(C,C) (C) receptors under steady-state conditions (0.1 mM glycine and 1 mM glutamate; digitized at 50 kHz, filtered at 1 kHz) from transiently transfected HEK cells. For recordings shown on the right, cells were exposed to 4 mM DTT in the presence of antagonists DCKA (10 µM) and APV (100 µM) before forming the on-cell patches. For each, the bottom trace is an expanded view (filtered at 5 kHz) of the respective boxed regions. The left and right recordings are from different on-cell patches.
|
|
Figure 5. Exponential fitting of composite histograms identify five closed and two to four open-time components. (A–C) Shut- (top) and open- (bottom) time–duration histograms of the same single-channel patches shown in Fig. 4. The shut-time–duration histograms were well fitted with five exponential components, whereas the open-time–duration histograms were well fitted with two to four exponential components. The time constants and relative areas of the exponential components are given in the insets.
|
|
Figure 6. NMDA receptors with intrasubunit GluN1 or GluN2A cross-links display largely symmetrical changes in multiple open- and closed-time components. (A, left) Mean fold change (±SEM) in duration of open-time components as determined by MIL fitting of single-channel recordings (see Materials and methods). Although single-channel records (in the cell-attached configuration) comprised two to four open-time components arising from modal gating behavior of NMDA receptors (Popescu and Auerbach, 2003), they were combined for simplicity into aggregates of two states, one short duration (for N1/N2A, 0.12 ± 0.01 ms) and one long duration (for N1/N2A, 8.71 ± 0.9 ms) (see Table II). (Right) Mean relative areas (±SEM) of the open-time components. (B, top) Mean fold change in duration (±SEM) of closed-time components as determined by MIL fitting of single-channel recordings. For WT GluN1/GluN2A, the time constants were (ms): τ1 0.14 ± 0.01, τ2 1.03 ± 0.13, τ3 4.16 ± 0.4, τ4 26.9 ± 5.6, and τ5 1,789 ± 366 (see Table II). (Bottom) Mean relative areas (±SEM) of the closed-time components. Significant differences (P < 0.05) relative to WT and between GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) are indicated with filled bars and asterisks, respectively.
|
|
Figure 7. Effects of the intrasubunit GluN1 or GluN2A cross-links on the kinetic mechanism of NMDA receptor activation. (A) Sequential-state model (Kussius and Popescu, 2009) of NMDA receptor activation with the rate constants (s−1) of transitions averaged from fits of individual single-channel recordings. Significant differences (P < 0.05) relative to WT are indicated with colored rate constants. Significant differences (P < 0.05) between GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) are indicated with asterisks. (B) Mean fold change (±SEM) in rate constants relative to GluN1/GluN2A. The left and right axes show the reverse and forward rate constants, respectively. Significant differences (P < 0.05) relative to WT and between GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) are indicated with filled bars and asterisks, respectively. (C) Free energy landscape plotted with respect to C2. The off-pathway steps to and from C4 and C5 are excluded. The three traces are horizontally offset for clarity.
|
|
Figure 8. After exposure to DTT, NMDA receptors composed of WT or double-cysteine–substituted subunits displayed comparable kinetic behavior. Kinetic analysis, as in Fig. 7, for single-channel patches exposed to DTT. (A) Sequential-state model (Kussius and Popescu, 2009) of NMDA receptor activation with the rate constants (s−1) of transitions averaged from fits of individual single-channel recordings. Significant differences (P < 0.05) relative to WT are indicated with colored rate constants. Significant differences (P < 0.05) between GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) are indicated with asterisks. (B) Mean fold change (±SEM) in rate constants relative to GluN1/GluN2A. The left and right axes show the reverse and forward rate constants, respectively. Significant differences (P < 0.05) relative to WT and between GluN1(C,C)/GluN2A and GluN1/GluN2A(C,C) are indicated with filled bars and asterisks, respectively. (C) Free energy landscape plotted with respect to C2. The off-pathway desensitization-related steps to and from C4 and C5 are excluded. The three traces are horizontally offset for clarity.
|
|
Figure 9. A model of the intrasubunit gating dynamics of M3–S2 and S2–M4 in the NMDA receptor. (A) Potential intrasubunit movements of the M3–S2 and S2–M4 linkers during pore opening. The channel closed conformations are from the antagonist-bound GluA2 (subunit C) crystal structure (Sobolevsky et al., 2009). The channel open conformations are based on homology modeling of the channel closed GluA2 structure with an open Shaker K+ ion channel and ligand-bound AMPA receptor LBD structure (Sobolevsky et al., 2009). Views from the top-down (top) and side (bottom) are shown. In this model, the centers of the M3–S2 linker (in the A/C subunit) and pre-M4 helix (dashed line) separate by ~9 Å in the transition from the closed- to the open-channel states. (B) Sequences of the M3–S2 and S2–M4 linkers in the GluA2, GluN1, and GluN2A subunits. Proximal parts of S2 are colored as magenta. In GluA2, the boxed region represents the transmembrane α-helical segments (in the A/C subunit), and the dashed boxed region depicts the pre-M4 helix. GluN1 is presumed to adopt the A/C and GluN2A the B/D conformations. Based on sequence alignment (Sobolevsky et al., 2009), the S2–M4 linkers in the GluN1 and GluN2A subunits have notable gaps, particularly in the presumed pre-M4 helix, complicating a direct comparison of NMDA receptor subunits to the GluA2 structure.
|
