Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

N-Methyl-D-aspartate (NMDA) receptor activation has been linked to many important cellular processes in the brain, including neuronal development, synaptic plasticity and excitotoxic cell death (Michaelis, 1998). It is therefore not surprising that the function of the NMDA receptor itself is tightly regulated by a variety of endogenous factors such as magnesium, protons, polyamines, and zinc (McBain and Mayer, 1994; Dingledine et al., 1999). Redox-active agents can also modify NMDA receptor activity, with reducing agents potentiating the response to NMDA, and oxidizing agents reversing the actions of reductants and sometimes attenuating the response to the agonist (Aizenman et al., 1989; Tang and Aizenman, 1993a). A physiological role for this modulatory site was suggested following the discovery of a number of redox-active substances of endogenous origin, which could alter the redox state of the receptor in dissociated mammalian neurons (Aizenman, 1994; Sinor et al., 1997). In fact, alteration of the redox site by endogenously derived factors has been associated with certain forms of long-term potentiation and long-term depression in hippocampal slices (Gozlan et al., 1995). The redox-active substances responsible for these effects in more intact brain tissue remain to be identified.

In recombinant receptors, two separate redox sites on the NMDA receptor have been proposed to exist, each on separate subunits. Using site-directed mutagenesis, Sullivan et al. (1994) determined that two cysteines on the NMDA receptor subunit (NR) 1, Cys744 and Cys798, were largely responsible for redox modulation of receptors composed of NR1/NR2B, NR1/NR2C, or NR1/NR2D subunit combinations. As the sensitivity of NR1/NR2A receptors to thiol agents was only slightly diminished when these two cysteine residues were mutated, an additional redox site, unique to NR1/NR2A receptors, was suggested. Using chimeric constructs, Köhr et al. (1994) concluded that the amino terminus of the NR2A subunit contained such site. However, to date, the amino acid residues constituting the redox site on NR2A have not been identified. Mutating various cysteines located in the amino terminus of this subunit, either individually (Sullivan et al., 1994; Köhr et al., 1994), or in tandem (Choi et al., 1997), does not totally abolish the redox sensitivity of NR2A-containing receptors. Although the existence of a separate redox site on NR2A has been questioned (Paoletti et al., 1997), the recent discovery of an NR2A-selective reducing agent strongly argues for the presence of such a site (Arden et al., 1998).

Single-channel studies have revealed that reducing agents potentiate NMDA-induced responses of NR1/NR2A receptors by increasing both the open dwell-time of the channel and increasing the frequency of channel opening relative to the oxidized state (Brimecombe et al., 1997). A sulfhydryl agent-induced alteration in NR1/NR2B- or NR1/NR2C-mediated channel activity, however, is manifested only as a change in opening frequency (Brimecombe et al., 1997). The present study attempted to determine whether the two proposed redox sites on NR1 and on NR2A could act independently of each other in affecting channel function. Because only NR1/NR2A receptors display an increase in open dwell-time in the presence of reductants, we hypothesized that the putative NR2A redox site alone could affect this channel property. Furthermore, because redox agents influence the frequency of channel opening in all receptor configurations, and all receptors contain the NR1 subunit, we further hypothesized that cysteines 744 and 798 of NR1 were critical for this type of channel alteration. We have therefore determined the extent of redox modulation of receptors composed of NR2A, NR2B, or NR2C when coexpressed with an NR1 subunit in which Cys744 and Cys798 have been mutated.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Transfection Protocol. Chinese hamster ovary (CHO)-K1 cells were grown in Ham's F-12 nutrient medium with 10% fetal bovine serum and 1 mM glutamine and passaged (<30 times) at a 1:10 dilution when 80% confluent, approximately every 2 days. The cDNAs for NR1a (hereafter referred to simply as NR1; a gift from Dr. S. Nakanishi, Kyoto University, Kyoto, Japan), NR2A, and NR2C (gifts from Dr. P. Seeburg, Max Planck Institute for Medical Research, Heidelberg, Germany) were previously subcloned into a mammalian expression vector (Boeckman and Aizenman, 1994, 1996). NR1(C744A, C798A) was a kind gift from Dr. S. Traynelis, Emory University, Atlanta, GA. The mouse clones 1 (NR2A) and 2 (NR2B) were previously subcloned into pRK7 (Meguro et al., 1992; Gallagher et al., 1996) and were generous gifts from Dr. D. Lynch, University of Pennsylvania, Philadelphia, PA, and Dr. M. Mishina, University of Tokyo, Tokyo, Japan. NR2A and 1 were used interchangeably. For this study we used 2 exclusively rather than NR2B (Boeckman and Aizenman, 1996), because functional expression of receptors with this latter vector was unexplainably lost. Although we noted an increase in the mean open dwell-time of NR1/1 and NR1/2 receptors when compared with those previously reported for NR1/NR2A and NR1/NR2B channels (Brimecombe et al., 1997) all other receptor properties, including redox sensitivity, were virtually identical between the rat and mouse subunits. The vector for a positive transfection marker, green fluorescent protein (pCI/GFP; a gift from Dr. M. Chalfie, Columbia University, New York, NY), was also generated previously (Brimecombe et al., 1997). CHO cells were seeded at 2.5 × 10⁵ cells/well in 6-well plates approximately 24 h before transfection with 1.3 µg of total DNA and 5 µl of lipofectAMINE (Gibco-BRL) in 1 ml of serum-free CHO media per 35-mm dish. The ratio of pCI/GFP to total other DNA was 1:4.3, and the ratio of NR1 to NR2 was always 1:3 (Cik et al., 1993). After a 4- to 5-h incubation at 37°C, cells were refed with CHO medium containing 300 µM ketamine to prevent cell death which ensues following functional receptor expression (Boeckman and Aizenman, 1996). Cells were used for recording 40 to 50 h after transfection.

Patch-Clamp Recordings and Analysis. The extracellular solution contained: 150 mM NaCl, 2.8 mM KCl, 1.0 mM CaCl₂, 10 mM HEPES, and 0.01 mM glycine 0.01, pH adjusted to 7.2 with 0.3 N NaOH. The pipette solution contained: 14 mM CsF, 10 mM EGTA, 1.0 mM CaCl₂, and 10 mM HEPES, pH adjusted to 7.2 with CsOH. Whole-cell measurements were performed at 60 mV with 2 M electrodes; methods for acquisition and analysis of whole-cell data have been described previously (Tang and Aizenman, 1993a). NMDA (30 µM), dithiothreitol (DTT; 4 mM) and 5,5'-dithio-bis-(2-nitrobenzoic acid) (DTNB; 0.5 mM) were applied to the cells by a multibarrel fast perfusion system (Warner Instruments, Hamden, CT). Single-channel recordings were performed at 60 mV with the same recording solutions on outside-out patches using 10 to 15 M electrodes. Drugs were applied to the patch by complete bath exchange (Brimecombe et al., 1997).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Mutating Cys744 and Cys798 on the NR1 subunit does not alter the single-channel redox properties of NR2A-containing receptors. A, representative 10 µM NMDA-induced single channel events from an outside-out membrane patch excised from a CHO cell transfected with NR1(C744A, C798A)/NR2A in the continuous presence of either 1 mM DTT or 0.1 mM DTNB. B, open dwell-time histograms for the events from the patch represented in A. The exponential functions fitting both distributions denote a change in time constants between the reduced (7.5 ms) and oxidized (5.4 and 6.3 ms) conditions. C, frequency of NMDA channel opening during the DTT and DTNB treatments for the same patch (hence the lack of error bars). The average change in open channel frequency between the two redox treatments (FDTT/FDTNB) for the five patches obtained from cells transfected with this subunit combination are shown in Table 1.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Single-Channel Redox Properties of NR1-Mutated NMDA Receptors. NMDA (10 µM)-activated channels were first recorded from patches excised from cells transfected with the NR1(C744A, C798A)/NR2A subunit combination (n = 5; Fig. 1A). The single-channel amplitudes of the unitary currents mediated by these channels were not altered by redox treatments (3.4 ± 0.1 pA in DTT, 3.5 ± 0.1 pA in DTNB; p > .05, paired t test). As seen in wild-type NR1/NR2A receptors (Brimecombe et al., 1997), the open dwell-time of the events increased significantly in the presence of DTT when compared with DTNB (5.0 ± 0.5 ms in DTT, 3.3 ± 0.6 ms in DTNB; p < .01, paired t test; Fig. 1B and Table 1). The mean change in open dwell-time, which is the difference in the open time between reduced and oxidized conditions for each patch, averaged across all patches [ open time (OT) in Brimecombe et al., 1997] was 1.9 ± 0.2 ms for this receptor configuration. Unexpectedly, DTT increased the frequency of channel opening 5.0- ± 1.6-fold (n-fold = F_DTT/F_DTNB, where F is opening frequency), when compared with the oxidized state (Fig. 1C and Table 1). These values are very similar to those previously obtained with wild-type NR1/NR2A channels (OT = 1.6 ms, F_DTT/F_DTNB = 5.7; Brimecombe et al., 1997). In fact, there are no significant differences in OT or F_DTT/F_DTNB between wild-type and mutant receptors (p > .05, unpaired t tests).

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Single channel redox properties of the cysteine-mutated NR1 subunit coexpressed with either NR2B or NR2C. A, open dwell-time histograms of NMDA-activated events obtained from a representative NR1(C744A, C798A)/NR2B-containing patch (left). The open dwell-time constants did not differ between the reduced and oxidized conditions (4.1 and 3.9 ms, respectively). In contrast, the frequency of channel opening was decreased in the presence of the oxidant when compared with the open channel frequency in the presence of the reductant (right). The average values from the seven patches obtained from cells transfected with this receptor combination are shown in Table 1. B, NR1(C744A, C798A)/NR2C-containing receptors were insensitive to redox agents. There was no change in the open dwell-time between the reduced and oxidized conditions (1.7 and 1.8 ms, respectively; left). Furthermore, there was no change in the frequency of channel opening between the two redox states (right). The average values from the five patches obtained from cells expressing NR1(C744A, C798A)/NR2C are summarized in Table 1.

Whole-Cell Measurements Suggest Presence of an Additional Redox Site on NR2B. NMDA (30 µM)-induced whole-cell responses were obtained from cells transfected with NR1 or NR1(C744A, C798A) together with either NR2A, NR2B, or NR2C. Current amplitude measurements were obtained after a 3-min incubation in 4 mM DTT and after a 1-min incubation with the oxidizing agent DTNB (0.5 mM). Mutating cysteines 744 and 798 on the NR1 subunit did not alter the extent of redox modulation of NR1/NR2A receptors (Fig. 3). The I_DTT/I_DTNB peak current ratio was 3.6 ± 0.3 for wild-type NR1/NR2A receptors and 4.0 ± 0.3 for the NR1(C744A, C798A)/NR2A subunit configuration (Table 2). These current ratios were not significantly different from each other (p > .05, unpaired t test), and, as such, this finding indicates that cysteines 744 and 798 are not required for the redox modulation of NR1/NR2A receptors, similar to what the single-channel data revealed.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Redox modulation of whole-cell currents mediated by NR2A-containing receptors. A, representative whole-cell responses recorded from a CHO cell expressing wild-type NR1/NR2A receptors during application of 30 µM NMDA. Currents were measured during control conditions, following a 3-min incubation with 4 mM DTT and after a 1-min incubation with 0.5 mM DTNB. B, similar measurements were obtained from another CHO cell transfected with NR1(C744A, C798A)/NR2A subunits. Responses were obtained from a total seven cells transfected with the wild-type NR1, and from five cells transfected with the cysteine mutated subunit. Peak current amplitudes from traces such as these ones were used to obtained the values shown in Table 2.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Residual redox sensitivity of NR1(C744A, C798A)/NR2B receptors. A, representative whole-cell responses recorded from a CHO cell expressing NR1/NR2B receptors during application of 30 µM NMDA. Currents were measured during control conditions, following a 3-min incubation with 4 mM DTT and after a 1-min treatment with 0.5 mM DTNB. Similar measurements were obtained in six additional cells. B, responses obtained from a CHO cell transfected with NR1(C744A, C798A)/NR2B subunits. Note that although the DTT treatment does not substantially enhance the amplitude of the control current, DTNB depresses the response in a DTT-reversible fashion. Similar recordings were obtained in total of five cells transfected with this subunit configuration. Data are summarized in Table 2.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The experiments presented here were aimed at determining the functional effects of thiol modification of two putative redox sites on the NMDA receptor. We have observed that Cys744 and Cys798 in the NR1 subunit are critical for redox modulation of NR2C-containing receptors, and that these residues contribute to some, but not all, of the redox sensitivity of the NR1/NR2B subunit combination. These findings are mostly in agreement with the results of Sullivan et al. (1994), although the residual redox sensitivity of NR1(C744A, C798B)/NR2B receptors had not been reported previously and is suggestive of yet an additional redox site on NR2B. The most surprising result obtained in the present study, however, is that the redox site on NR1 appears not to contribute to the actions of DTT and DTNB on NR1/NR2A receptor function, at least within the parameters examined here.

Evidence suggesting a redox sensitive in the extracellular amino terminus of NR2A was obtained by Köhr et al. (1994) using chimeric subunits. In that study, it was shown that the redox properties of NR1/NR2A receptors were substantially different from those observed in all other receptor configurations tested (NR1/NR2B, NR1/NR2C, NR1/NR2D). First, although the effects of DTT on whole-cell currents mediated by all other subunits combinations were relatively slow to develop (minutes), NR1/NR2A responses were enhanced very quickly (seconds) after exposure to the reducing agent. This rapid effect of DTT on NR1/NR2A was followed by a subsequent slow rise in current amplitude with a time course reminiscent of the effects of this substance on the other subunit configurations. Second, the actions of the reductant on NR1/NR2B, NR1/NR2C, and NR1/NR2D receptors were very long-lasting and only fully reversible following treatment with DTNB. In contrast, the rapid current-enhancing actions of DTT on NR1/NR2A receptors were also quick to reverse, leaving behind a DTNB-sensitive, more persistent component, which, again, was very similar to what was observed with the other receptors. The rapidly reversibly DTT effect on NR1/NR2A was abolished in receptors composed of a NR2A chimeric construct in which the amino terminus of the latter subunit was substituted for the analogous segment of NR2C (NR2[CA]). This suggested to the authors that an NR2A redox-sensitive site was localized to this region. Third, all receptor configurations, except for NR1/NR2A, could be permanently potentiated (that is, be rendered insensitive to subsequent DTNB oxidation) when treatment with an alkylating agent immediately followed the DTT exposure. Alkylation has also been produced in native neuronal receptors (Tang and Aizenman, 1993a), and it has been a useful tool for establishing redox modulation as the mechanism of action of endogenous redox-active substances (Aizenman et al., 1992; Tang and Aizenman, 1993b). Whether alkylation can be produced in an NR1/NR2[CA] receptor remains to be established.

Paoletti et al. (1997) recently suggested that some of the effect of DTT on NR1/NR2A-mediated currents are due to zinc chelation by the reductant itself, with the ensuing removal of a high-affinity tonic receptor block by background contamination levels of this metal. These investigators demonstrated that rapid potentiating effects of DTT were indistinguishable from those produced by more conventional metal chelators. Hence, the authors concluded that the amino terminus of NR2A likely contained a Zn²⁺-sensitive site, rather than a true redox-sensitive site. Yet, in spite of these findings, there are two lines of evidence that still favor the existence of a unique redox site on NR2A. One, mutation of Cys744 and Cys798 in NR1, dramatically reduces the affinity of NR1/NR2A receptors for zinc (Zheng et al., 1998; Traynelis et al., 1998). In fact, Zheng et al. (1998) observed virtually no potentiation of NR1(C744A, C798A)/NR2A responses with concentrations of a metal chelator that were very effective in enhancing NR1/NR2A-mediated currents. As is evident in the present report, the effects of DTT on wild-type and mutant NR2A-containing receptors are essentially indistinguishable. Second, recent experiments conducted in our laboratory (Arden et al., 1998) demonstrated that cyanide can selectively potentiate NMDA-induced currents mediated by NR1/NR2A receptors. The cyanide-induced current enhancement occurred slowly (minutes) and was readily reversible by DTNB. Mutation of Cys744 of NR1 alone, sufficient to abolish DTT potentiation in NR1/NR2B receptors (Sullivan et al., 1994), and decrease high-affinity zinc block in NR1/NR2A (Traynelis et al., 1998), did not block the actions of cyanide or DTT on NR2A-containing receptors. Interestingly, the mitochondrial poison depressed NR1/NR2B-mediated currents, reinforcing the notion that the redox sensitivity of NR1/NR2A is indeed unique. Finally, cyanide was able to potentiate NR1/NR2A responses in the presence of a metal chelator. It is noteworthy that in our laboratory we have not reliably seen potentiation by zinc chelators in wild-type NR1/NR2A receptors (Arden et al., 1998). This likely indicates low background levels of contaminating metals in our solutions, or that endogenous tyrosine kinase Src activity is high in CHO cells, as NMDA receptor phosphorylation by this enzyme reduces tonic zinc inhibition in NR1/NR2A receptor channels (Zheng et al., 1998). The fact that NR1/NR2A receptors cannot be alkylated (Köhr et al., 1994) is further suggestive that the redox site on NR1 does not mediate the effects of thiol substances on this subunit combination.

The single-channel experiments in the present investigations were initially designed to test the hypothesis that the redox-sensitive site on NR1 would influence the frequency of channel opening and that the NR2A redox site would modify the open dwell-time of the channel. Namely, we were interested in investigating whether the two separate redox sites acted independently of each other. However, our data strongly suggest that Cys744 and Cys798 on the NR1 subunit are not necessary for the redox modulation of NR1/NR2A receptors. What other residues may be involved in the thiol sensitivity of this receptor configuration? In a preliminary report (Choi et al., 1997), the DTT sensitivity of NR1/NR2A receptors in which five cysteines were mutated simultaneously (Cys744 and Cys798 on NR1 and Cys87, Cys231, and Cys320 in the amino terminus of NR2A) was evaluated. Even after these mutations, NMDA-activated currents were still potentiated by DTT, implying that additional cysteines are involved in redox modulation of this receptor. Second, logical guesses are the cysteines in NR2A corresponding to Cys744 and Cys798 in the NR1 subunit, which would provide a "redundant" redox site in NR1/NR2A receptors. By analogy, the homologous cysteines in NR2B may impart the residual redox sensitivity to NR1(C744A, C798A)/NR2B receptors. Although these cysteines are located on all the NR2 subunits, including NR2C, there may be structural limitations in an NR1/NR2C receptor, which do not allow for these two cysteines to be accessible to redox-active substances. In fact, there are no instances where the NR2A or NR2B subunits contain homologous cysteines that are not present in NR2C in regions that face the extracellular milieu (Monyer et al., 1992). Clearly, additional work is required to resolve these issues. Determining the manner in which individual redox sites alter channel function will likely aid in elucidating how the various domains of NMDA receptor subunits modify channel gating.