High-affinity zinc inhibition of NMDA NR1-NR2A receptors

Abstract

Micromolar concentrations of extracellular Zn2+ are known to antagonize native NMDA receptors via a dual mechanism involving both a voltage-independent and a voltage-dependent inhibition. We have tried to evaluate the relative importance of these two effects and their subunit specificity on recombinant NMDA receptors expressed in HEK 293 cells and Xenopus oocytes. The comparison of NR1a-NR2A and NR1a-NR2B receptors shows that the voltage-dependent inhibition is similar in both types of receptors but that the voltage-independent inhibition occurs at much lower Zn2+ concentrations in NR1a-NR2A receptors (IC50 in the nanomolar range) than in NR1a-NR2B receptors (IC50 in the micromolar range). The high affinity of the effect observed with NR1a-NR2A receptors was found to be attributable mostly to the slow dissociation of Zn2+ from its binding site. By analyzing the effects of Zn2+ on varied combinations of NR1 (NR1a or NR1b) and NR2 (NR2A, NR2B, NR2C), we show that both the NR1 and the NR2 subunits contribute to the voltage-independent Zn2+ inhibition. We have observed further that under control conditions, i.e., in zero nominal Zn2+ solutions, the addition of low concentrations of heavy metal chelators markedly potentiates the responses of NR1a-NR2A receptors, but not of NR1a-NR2B receptors. This result suggests that traces of a heavy metal (probably Zn2+) contaminate standard solutions and tonically inhibit NR1a-NR2A receptors. Chelation of a contaminant metal also could account for the rapid NR2A subunit-specific potentiations produced by reducing compounds like DTT or glutathione.

Fig. 1.

NR1a–NR2A responses are potentiated by the heavy metal chelator TPEN. Recombinant NMDA receptors expressed in HEK 293 cells (A, B) or in Xenopusoocytes (C, D) were activated by applying saturating concentrations of glutamate (100 μm) and glycine (100 μm). The responses were compared before (control) and after (TPEN) addition of 1 μm TPEN to the external solution.A, B, HEK cells. Glutamate was applied on a background of glycine for 2 sec every 10 sec on cells held at −50 mV. The external Ca²⁺ concentration was 1 mm. A, TPEN potentiates NR1a–NR2A responses. Each trace is the average of five records.B, NR1a–NR2B responses are not affected by 1 μm TPEN. Each trace is the average of fifteen records. C, D,Xenopus oocytes. Voltage ramps from −70 to +50 mV were applied in the absence (control) or presence of 1 μmTPEN. Current–voltage curves corresponding to the leak currents were substracted from those obtained during steady applications of 100 μm glutamate and 100 μm glycine. The only external divalent cation was Ba²⁺ (0.3 mm). C, TPEN potentiates NR1a–NR2A responses over the whole voltage range.D, NR1a–NR2B currents recorded in the absence or in the presence of TPEN are superimposed.

Fig. 3.

Concentration dependence of the Zn²⁺ inhibition of NR1a–NR2A and NR1a–NR2B receptors. Recombinant NMDA receptors were expressed inXenopus oocytes, and dose–response curves were constructed from I–V curves obtained as in Figure 2. In each experiment currents were expressed as a fraction of the current recorded in the presence of a Zn²⁺ chelator (“0” Zn²⁺). Thecurves in A and Brepresent least-squares fits to the data points with the two-binding-site isotherm y = 1 − ((a/(1 + IC₅₀₍₁₎/[Zn²⁺])) + (b/(1 + IC₅₀₍₂₎/[Zn²⁺]))), in whichy is the relative current, and a andb are the respective weights of each isotherm. Thecurves in C and Drepresent least-squares fits to the data points with the single-binding-site isotherm y = 1 − (a/(1 + (IC₅₀/[Zn²⁺])ⁿ)), in which y is the relative current, n is the Hill coefficient, and a is a weight factor.A, B, Dual antagonism by Zn²⁺ of NR1a–NR2A responses recorded at negative potentials (−60 mV). A, Zn²⁺concentrations correspond to free Zn²⁺concentrations in solutions buffered with tricine (10 mm; see Fig. 2 and Materials and Methods). Data are from seven cells, with each point being the mean of three to four values. The estimated IC₅₀₍₁₎, IC₅₀₍₂₎,a, and b are 17 nm, 26 μm, 0.75, and 0.25, respectively. B, Zn²⁺ concentrations correspond to added Zn²⁺ concentrations corrected for an assumed 10 nm contaminating Zn²⁺. TPEN (1 μm; n = 13) or DTPA (2 μm; n = 4) were used for reference (“0” Zn²⁺ concentration). Data are from 17 cells, each point being the value obtained from 1 cell (30 and 300 nm added Zn²⁺) or the mean value obtained from 2 cells (5 and 10 nmand 3 μm), 3 cells (30 and 300 μm), 4 cells (1 and 100 nm), 9 cells (10 and 100 μm), or 17 cells (1 μm). The estimated IC₅₀₍₁₎, IC₅₀₍₂₎,a, and b are 6 nm, 32 μm, 0.78, and 0.22, respectively. C, A single-binding-site isotherm is sufficient to describe the Zn²⁺ inhibition of NR1a–NR2B responses recorded at negative potentials (−60 mV). Zn²⁺ concentrations correspond to added Zn²⁺ concentrations with no correction for Zn²⁺ contamination. TPEN (1 μm) was used for the “0”Zn²⁺ solution. Data are from five cells, with each point being the mean of three to five values. The value ofa was fixed to 1. The estimated IC₅₀ andn are 490 nm and 0.78, respectively.D, At positive potentials the low-affinity Zn²⁺ inhibition of NR1a–NR2A responses is absent. Zn²⁺ concentrations were corrected by assuming 10 nm contaminating Zn²⁺. TPEN (1 μm; n = 13) or DTPA (2 μm; n = 4) were used for the“0” Zn²⁺ concentration. Data are from 17 cells, each point being the value for 1 cell (5, 30, and 300 nm added Zn²⁺) or the mean value for 2 cells (10 nm and 3 and 30 μm), 3 cells (1 nm and 300 μm), 4 cells (100 nm), 8 cells (10 and 100 μm), or 17 cells (1 μm). The value of n was fixed to 1. The estimated values of the IC₅₀ and of a are 5 nm and 0.79, respectively.

Fig. 2.

Nanomolar external Zn²⁺concentrations selectively inhibit NR1a–NR2A responses. Leak-substracted NMDA currents were recorded at different concentrations of external Zn²⁺ during voltage ramps from −70 to +50 mV applied in Xenopus oocytes expressing NR1a–NR2A or NR1a–NR2B receptors. Glutamate and glycine were applied at saturating concentrations (100 μm each). The “0” Zn²⁺ concentration refers to a solution containing a Zn²⁺ chelator with no added Zn²⁺ (see Results). A, NR1a–NR2A responses are inhibited by Zn²⁺concentrations of a few nanomolars. The inhibition is voltage-independent. At higher Zn²⁺ concentrations the inhibition saturates at ∼75% of the response in the“0” solution. The indicated Zn²⁺concentrations correspond to calculated free Zn²⁺concentrations in solutions buffered with 10 mm tricine (see Materials and Methods). The “0” solution contained 10 mm tricine and no added Zn²⁺. B, An additional voltage-dependent inhibition is produced by micromolar concentrations of Zn²⁺. The responses are from a different oocyte. The indicated Zn²⁺ concentrations correspond to nominal values. The “0” Zn²⁺solution contained 1 μm TPEN. C, Zn²⁺ antagonism of NR1a–NR2B responses is of lower affinity, is total, and is mainly voltage-independent. The indicated Zn²⁺ concentrations correspond to nominal values. The “0” Zn²⁺ solution contained 1 μm TPEN. D, Expanding the current scale at a high Zn²⁺ concentration (30 μm) reveals a voltage-dependent component of the inhibition of NR1a–NR2B responses. Shown is the same cell as in C. For clarity, the data were fit with a third-order polynomial.

Fig. 5.

Low-affinity voltage-dependent Zn²⁺ block of NMDA NR1a–NR2A responses.A, Voltage ramps from −100 to +50 mV were applied inXenopus oocytes expressing NR1a–NR2A receptors in the absence of Zn²⁺ (“0”Zn²⁺ solution containing 2 μm DTPA) or after the addition of 1, 10, 30, 100, or 300 μm added Zn²⁺. For clarity, leak-substractedI–V curves are shown on an expanded current scale, because the voltage-dependent block by Zn²⁺ appears in a concentration range in which ∼80% of the maximal response already is eliminated by the high-affinity voltage-independent Zn²⁺ inhibition (see Figs. 2, 3). The Zn²⁺ block increases with increasing Zn²⁺ concentrations and with hyperpolarization. However, even at the highest concentration of Zn²⁺tested (300 μm) and at the most negative potentials, an inward current can still be recorded. B, Voltage dependence of the low-affinity Zn²⁺ inhibition. Shown is the same cell as in A. The unblocked fraction was calculated by dividing, at each concentration of Zn²⁺, the NMDA current by the NMDA current recorded in “0” Zn²⁺ and by subsequently normalizing to 1 for a membrane potential of +50 mV. Data points that take artifactual values around the reversal potential have been omitted. C, Concentration dependence of the voltage-dependent block by Zn²⁺ at −100, −80, −60, −40, −20, 0, +20, and +40 mV. Data points were calculated from curves similar to those shown in B, obtained in a series of seven experiments. Each point corresponds to the mean value obtained from three to seven measurements. The data points at 0 mV were obtained by interpolation with a polynomial fit (see Materials and Methods). The lines drawn through the data points are least-squares fits of the single binding isotherm:y = y_max · (1 − (1/(1 +(IC₅₀/[Zn²⁺])ⁿ))), in which y is the relative voltage-dependent Zn²⁺ inhibition and n is the Hill coefficient. The weight factor y_max was introduced to eliminate the residual voltage dependence seen at 1 μm Zn²⁺ (see Results). The estimated IC₅₀ and n are, respectively, 22 μm and 0.9 at −100 mV, 31 μm and 0.9 at −80 mV, 41 μm and 0.9 at −60 mV, 100 μmand 0.9 at −40 mV, 346 μm and 0.9 at −20 mV, and 1162 μm and 0.8 at 0 mV. D, Voltage dependence of the low-affinity Zn²⁺ inhibition. The IC₅₀ values are those estimated by the fits shown inC. Note that, on this semi-log plot, a linear relation does not fit the data. The slope of the relation between the IC₅₀ and the voltage increases with depolarization and reaches a maximum between −40 and 0 mV. In this range the slope (e-fold for ∼16 mV, dotted line) is consistent with an apparent electrical depth of the Zn²⁺ binding site of 0.77. The dashed line is drawn according to the equation used by Christine and Choi (1990) for the fit of their single-channel data obtained in the range between −70 and −20 mV (an IC₅₀ at 0 mV of 909 μm and an electrical depth of 0.51).

Fig. 7.

Most native NMDA receptors are potentiated by TPEN. A, Comparison of the effects of TPEN (1 μm) on the peak amplitude of NMDA currents recorded in neurons and in HEK 293 cells expressing either NR1a–NR2A or NR1a–NR2B receptor subtypes. Native NMDA responses were elicited by a 2 sec pulse of NMDA (200 μm) on a background of glycine (10 μm). Recombinant NMDA responses were recorded with protocols identical to those shown in Figure 1. The holding potential was −50 mV. Each circle corresponds to the peak ratio (TPEN/control) obtained from one experiment. Thefilled and hatched circles correspond to the two separate experiments, which are illustrated inB. The potentiation of native NMDA responses by TPEN was highly variable, with peak ratios ranging from 1.0 (no potentiation) to a maximum of 1.6. The mean value was 1.25 ± 0.17 (n = 12). The mean peak ratios for NR1a–NR2A and NR1a–NR2B receptors were 2.9 ± 0.3 (n = 6) and 1.07 ± 0.04 (n = 6), respectively.B, Variability of the effect of TPEN on neuronal NMDA responses. Shown are superimposed traces recorded in the control solution and in the presence of 1 μm TPEN. Eachtrace is the average of three individual responses. Thetop panel shows an example with a marked potentiation by TPEN (the peak ratio TPEN/control of 1.41 corresponds to thefilled circle in A), whereas thebottom panel shows an example in which TPEN was ineffective (the peak ratio of 1.0 corresponds to the hatched circle in A).

Fig. 4.

Slow dissociation of Zn²⁺ from NR1a–NR2A receptors. Current relaxations that follow a Zn²⁺ concentration jump were analyzed in HEK 293 cells expressing NR1a–NR2A receptors. Each tracerepresents an individual response to a 14.5 sec pulse of glutamate (100 μm) on a background of glycine (100 μm). Tricine (10 mm) was present throughout the experiment. Zn²⁺ was applied for 3 sec during the pulse of glutamate once the response had reached a steady level. The Zn²⁺ concentrations as indicated in the figure correspond to the calculated free Zn²⁺ concentration in the tricine-buffered solutions (see Materials and Methods). The onset and offset of the inhibition by Zn²⁺ were fit by single exponentials (superimposed on the current traces) with time constants τ_on and τ_off, as indicated in the figure. Data were filtered at 100 Hz and sampled at 140 Hz. The holding potential was −50 mV.

Fig. 6.

A pore mutation selectively eliminates the low-affinity voltage-dependent Zn²⁺ inhibition.A, Comparison of the inhibitory effects of Zn²⁺ on NMDA responses recorded at −60 mV in oocytes expressing wild-type NR1a–NR2A receptors or mutant NR1a–NR2A(N595K) receptors. Shown are superimposed individual responses to a 20 sec pulse of glutamate and glycine (100 μm each) recorded in the presence of a Zn²⁺ chelator (“0”Zn²⁺ concentration; 1 μm TPEN for wild-type receptors and 2 μm DTPA for mutant receptors) or in the presence of 1 or 100 μm added Zn²⁺. Both types of receptors are strongly inhibited by 1 μm Zn²⁺, but the presence of 100 μm Zn²⁺ fails to produce an additional inhibition on the mutant receptors. B, Leak-substractedI–V curves from Xenopus oocytes expressing the mutant NR1a–NR2A(N595K) receptors in the absence (“0” Zn²⁺; 2 μmDTPA) or in the presence of added Zn²⁺ (1 and 100 μm). The traces obtained in 1 and 100 μm Zn²⁺ are superimposed almost perfectly over the whole voltage range, indicating that the voltage-dependent inhibition by Zn²⁺ has been suppressed by the mutation.

Fig. 8.

The NR2A subunit-specific potentiation produced by reducing agents is probably the consequence of Zn²⁺chelation. Shown are glutamate-evoked currents recorded in an HEK 293 cell expressing NR1a–NR2A receptors. The cell was exposed either to a control solution or to TPEN (1 μm) and DTE (3 mm) applied separately or simultaneously. The holding potential was −50 mV. A, Each trace is an individual response to a 2 sec pulse of glutamate (100 μm) applied on a background of glycine (100 μm). The potentiations produced by DTE andTPEN are not additive. B, Glutamate off-relaxations are shown on an expanded time scale and after normalization to the steady-state response amplitude. The decay of the current was fit by a single exponential function (solid lines superimposed on the current traces), with a time constant τ_off indicated in the figure. TPEN andDTE applied separately or simultaneously induce similar accelerations of the off response.