![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
NEUROPHARMACOLOGY
Center for Neuroscience and Aging, The Burnham Institute, La Jolla, California (H.-S.V.C., S.A.L.); and Departments of Cardiology (H.-S.V.C.) and Neurosciences (S.A.L.), University of California-San Diego, La Jolla, California
Received for publication
February 25, 2005
Accepted
May 11, 2005.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsDiscussionReferences |
|---|
). Excessive activation of NMDARs by excitatory amino acids such as glutamate, however, contributes to neuronal damage in many neurological disorders (Lipton and Rosenberg, 1994; Choi, 1998). To avoid side effects, blockade of excessive NMDAR activity should be achieved without interfering with normal function. High-affinity NMDAR antagonists are toxic to neurons and produce neurobehavioral side effects (Rogawski and Wenk, 2003). In contrast, memantine is a low-affinity, uncompetitive open-channel blocker of the NMDAR and has excellent safety and efficacy profiles for treatment of Alzheimer's disease (Reisberg et al., 2003; Tariot et al., 2004). The molecular mechanism underlying the clinical tolerability of memantine remains contentious (Rogawski and Wenk, 2003; Lipton and Chen, 2004). Identification of the precise location(s) of memantine interaction on NMDARs would provide insight into the drug's clinical tolerability and aid in the search for better memantine-like agents.
We previously reported that memantine, at low micromolar concentrations, blocked NMDA-evoked responses via a bimolecular reaction and interacted with extracellular Mg2+ in NMDA-gated channels (Chen et al., 1992; Chen and Lipton, 1997). Because of its interaction with external Mg2+ and mutational analysis by others (Kashiwagi et al., 2002), the specific site of memantine action is assumed to be near the external Mg2+ blocking site at the selectivity filter region of the NMDAR-associated channel (Sakurada et al., 1993; Danysz and Parsons, 2003), which is formed by asparagine (N) residues at the "N-site" of NR1 and "N+1 site" of NR2 subunits (Dingledine et al., 1999). Compared with physiological block by external Mg2+, a common explanation for the safety and effectiveness of memantine is that memantine is a "better magnesium", manifesting a somewhat slower unblocking rate, moderate voltage dependence, and slightly higher affinity (Danysz and Parsons, 2003). However, Mg2+ interacts differentially with the N-site residues of NR1 and NR2 subunits when applied from the intracellular versus extracellular surface (Wollmuth et al., 1998a,b). The N-site asparagine of the NR1 subunit represent the dominant blocking site for intracellular Mg2+, whereas the N and N + 1 site asparagines of the NR2A subunit form the critical blocking site for extracellular Mg2+. Heretofore, the molecular interaction of memantine with these critical residues at the selectivity filter region has not been clearly demonstrated.
Additionally, using mutation/substitution analysis, Kashiwagi and colleagues (2002) showed that several residues in NR1 and NR2B affect antagonism by 1 µM memantine (on NR1: W608/W611 in the M2, residues near A645 in M3, and A653/V656/L657 in the post-M3 segments; on NR2B: W607L in the M2 segment). These residues, however, are located far apart, precluding the possibility that they all contribute to the binding site of 1 µM memantine. Here we demonstrate the mechanism by which these residues located in the channel selectivity filter and along the permeation pathway affect memantine action.
Several studies have also reported a second binding site for memantine in NMDA-gated channels (Antonov and Johnson, 1996; Bresink et al., 1996; Blanpied et al., 1997; Sobolevsky et al., 1998). This second site was reported to have a much lower affinity, minimal voltage dependence, and a noncompetitive mechanism of block. There were, however, several differences among these studies, including the estimated IC50, the precise degree of voltage dependence, and the location of this blocking site in the NMDA-gated channel. Additionally, memantine, unlike Mg2+ and MK-801, displays minimal NR2 subtype-specific difference for its blocking action among the various combinations of heteromeric NMDAR channels (Bresink et al., 1996). The influence of the eight splice variants of the NR1 subunit (Sugihara et al., 1992) on memantine blockade of NMDAR-associated channels remains unknown.
|
). At the channel selectivity filter, however, a common assumption has been that all of these blockers act like extracellular Mg2+. In contrast, our new results indicate that the predominant site for specific memantine action is at the blocking site for intracellular Mg2+ and not extracellular Mg2+. The second, very low-affinity site is located at a shallow position in the extracellular vestibule that overlaps the site of action of nonspecific blockers. The N and N+1 site asparagines of the NR2A subunit stabilize memantine binding electrostatically. The pharmacological implications of our findings are discussed with regard to development of improved NMDAR antagonists. |
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsDiscussionReferences |
|---|
) and NR2A (Monyer et al., 1992) were used as the template for all mutations. We numbered the amino acid residues beginning with the initiator methionine in the NR1 and NR2 clones (Moriyoshi et al., 1991). Thus, NR1(N616) and NR2A(N614) are the N-site equivalents in the M2 region. The N to R mutant NR1(N616R) was obtained from Dr. R. Dingledine's laboratory (Emory University, Atlanta, GA). The remaining mutant NMDAR subunits were all generated in our laboratory using the polymerase chain reaction-based QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Multiple amino acid mutants on the same subunit were generated sequentially. All point mutations were verified by DNA sequencing of the mutated region (400 to 900 base pairs around the mutation site), and the whole clone was verified by restriction mapping. cRNA was transcribed in vitro using T3 or T7 mMESSAGE mMACHINE transcription kits (Ambion, Austin, TX). The diagram in Fig. 1A illustrates the relative positions of selected residues for mutations in the NR1 and NR2A subunits. Representative recordings of blockade by magnesium or various concentrations of amantadine and memantine of NMDA-evoked currents from wild-type or mutant NR1/NR2A channels are shown in Fig. 1B.
Electrophysiological Recordings of Recombinant Channels from Oocytes. Preparation of Xenopus oocytes and injection of cRNA were performed as previously described (Choi et al., 2000). Whole-cell recordings were performed as previously described (Choi et al., 2000) at a holding potential of –60 mV unless otherwise indicated. The recording solution was comprised of Barium Ringer's solution containing 1 mM BaCl2, 95 mM NaCl, 2 mM KCl, 5 mM HEPES, pH adjusted to 7.5 with NaOH. Glycine (10 µM) and 200 µM NMDA were used to elicit NMDA-activated currents. Data were analyzed with MacLab v3.6 and Axograph v3.5.5 software (Axon Instruments, Union City, CA). Solutions were applied to oocytes with an array of pipettes similar to the "sewer pipe system" as described in Choi et al. (2000). All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated.
Protocols for Studying the Voltage Dependence of Memantine Block. As we previously described, intracellular monovalent permeant ions affect the apparent on rate of memantine blockade, but this effect is minimal at holding potentials more negative than –60 mV (Chen and Lipton, 1997). Additionally, the effect of extracellular permeant ions on the on and off rates of memantine-like compounds is predominantly voltage-independent (Antonov et al., 1998). At holding potentials more negative than –60 mV, the voltage-dependent ionic interactions would be minimized so that the true voltage dependence of memantine blockade can be adequately assessed. Therefore, steady-state blockade by memantine at –60, –80, –100, and –120 mV was monitored to analyze the degree of voltage dependence of the various mutations.
Unlike other investigators, we avoided the voltage-ramp method to study the voltage-dependent effect of various mutations because, at voltages more positive than –60 mV, memantine binding shifts toward the shallow site and also manifests more complicated ionic interactions with intracellular permeant ions (Antonov et al., 1998). Therefore, 
(the electrical distance in Woodhull's model), calculated by the voltage ramp method, represents the result of mixed-binding at the deep and shallow sites, as well as ionic interactions. Because of the difference in apparent IC50 values between uncompetitive and noncompetitive blocking sites of memantine, 1 µM memantine was used to assess voltage dependence to minimize any binding at the superficial, noncompetitive site. If we found that the affinity of memantine decreased with a given point mutation, we used an increased memantine concentration of 10 µM for the determination of voltage dependence. At these two concentrations and at holding potentials 
–60 mV, memantine is a very specific open-channel blocker with 1:1 binding stoichiometry (Chen and Lipton, 1997). This type of action, displaying minimal voltage dependence due to ionic interactions of permeant ions, approximates Woodhull's model (Woodhull, 1973). Each data point in the voltage-dependent analysis represents the mean ± S.E.M. of the responses obtained from three to seven oocytes.
Chemical Modification by 2-Aminoethyl-methanethiosulphonate (MTSEA). The substituted cysteine accessibility method (SCAM) (Karlin and Akabas, 1998; Beck et al., 1999) was used to determine the location of residues that could affect memantine binding. For this purpose, we added the small sulfhydryl-modifying agent MTSEA (Toronto Research Chemicals, Inc., Toronto, ON, Canada), which reacts with free cysteine thiols to determine their accessibility after reaction with memantine. It should be noted, however, that a single endogenous cysteine residue, NR2A(C399), can be modified by MTSEA and would result in a decrease in NMDA-gated current that could obfuscate our results (Choi et al., 2000). To avoid this pitfall and to test the accessibility of cysteine-substituted mutants around the channel mouth, an NR2A(C399A) mutation was used in conjunction with other mutants to form heteromeric channels. For example, the NR2A(N614C) mutation was constructed with NR2A(C399A) as the template to generate the final mutant NR2A(C399A/N614C).
The effect of MTSEA in the presence or absence of 200 µM memantine was assessed 15 s after the onset of memantine washout. This time point was chosen to allow 
90% recovery from memantine block while minimizing any significant recovery from MTSEA modulation. While assessing the protection from SCAM by memantine (Protection-SCAM, Figs. 3 and 4), MTSEA was applied 10 s after memantine administration and was subsequently washed out 15 s before memantine. To analyze the "on rate" of MTSEA modification with a pulsatile protocol in the presence of an open-channel blocker like memantine, the washout rate of the open-channel blocker would have to be sufficiently fast to allow repeat application of MTSEA. Thus, we did not use a pulsatile protocol to monitor the on rate of modification of exposed cysteines by MTSEA because the washout of millimolar concentrations of memantine is too slow, on the order of 6 to 8 min, in the oocyte recording system (Sobolevsky et al., 2002). However, this problem of slow washout will affect kinetic analyses but not steady-state analyses. In fact, a slow washout rate actually ensures the accuracy of Protection-SCAM analysis at steady state because memantine remains at the blocking site for a longer period. Therefore, to allow sufficient recovery from block by millimolar concentrations of memantine, the effect of MTSEA was measured 40 s after initiating washout of 2 mM memantine. For millimolar concentrations of hexamethonium, a multicharged molecule, complete recovery from blockade occurs 30 s after washout. Protection from MTSEA modulation by hexamethonium was therefore assessed at 15 and 30 s after washout.
|
|
). Each point on the dose-response curve represents the mean ± S.E.M. of the response obtained from 3 to 12 oocytes. The data were weighted with the reciprocal of the variance and fitted by SigmaPlot software using the Marquardt-Levenberg algorithm with an empirical Hill equation:
 | (1) |
The voltage dependence of inhibition by memantine was determined at various holding potentials by measuring the residual current after reaching a steady state of blockade with 1 or 10 µM memantine. The equilibrium dissociation constant (Ki) was calculated as previously described (Chen et al., 1992) using the following equation:
 | (2) |
), and EC50 is the apparent equilibrium dissociation constant for NMDA, which is assumed to be 50 µM. The Ki calculated from the degree of blockade was then fitted with the Woodhull equation (Woodhull, 1973) to estimate the fraction of the transmembrane field sensed by memantine (MEM):
 | (3) |
, the fraction of the transmembrane potential field sensed by extracellular application of MEM; VH, the holding potential; z, the charge of MEM at neutral pH which is assumed to be +1 (Schneider et al., 1984); and F, R, and T have their usual meanings. Although the Woodhull equation is a simplified way to look at the voltage dependence of a channel blocker, it does allow us to compare the influence of the various N-site mutants on the voltage dependence of memantine blockade.
Interaction of Memantine with the Intracellular Magnesium Blocking Site at the Narrow Constriction of the Selectivity Filter Region. The specific memantine blocking site has been assumed to be near the narrow constriction region of the NMDAR-associated ion channel due to its interaction with extracellular Mg2+ as well as evidence from mutational analyses (Chen et al., 1992; Chen and Lipton, 1997; Kashiwagi et al., 2002). However, low micromolar concentrations of memantine have also been shown to interact with additional amino acid residues in the M2, M3, and post-M3 segments of the NR1 subunit (Kashiwagi et al., 2002). Amino acid substitutions for residues that lie in the permeation pathway could influence memantine action by direct, indirect, or allosteric mechanisms of interaction. To determine more precisely the location of memantine action, we first investigated the effect of N-site mutations. Figure 1 shows the diagram of mutations and representative tracings as described under Materials and Methods. Q (glutamine) or R (arginine) substitution of the N-site asparagine in either NR1 or NR2A subunits decreases the degree of block by both extracellular and intracellular Mg2+ on recombinant NMDAR channels (Wollmuth et al., 1998a,b).
|
). Additionally, the positively charged R substitution in either NR1 or NR2A reduced antagonism by memantine more than the equivalent Q substitution of the N-site asparagine.
In addition to the change in affinity of memantine antagonism caused by the various N-site mutations, Fig. 2B demonstrates the change in voltage dependence that these mutations engendered. The assessments of memantine blockade were made at a number of steady-state holding potentials (see Materials and Methods for details). The electrical distance from the extracellular surface of the membrane () for each mutant was then calculated as described under Materials and Methods. The N to Q substitution in the NR2A subunit moved the calculated by 25% toward the extracellular side of the membrane's electrical field [from 0.84 for NR1/2A to 0.57 for NR1/NR2A(N614Q), or from 0.97 for NR1(N616Q)/NR2A to 0.73 for NR1(N616Q)/NR2A(N614Q)]. The N to Q mutation in the NR1 subunit, however, shifted the calculated 14% toward the intracellular side of the membrane [from 0.84 for NR1/NR2A to 0.97 for NR1(N616Q)/NR2A, or from 0.57 for NR1/NR2A(N614Q) to 0.73 for NR1(N616Q)/NR2A(N614Q)]. These two opposite effects on electrical distance brought about by N to Q substitutions on NR1 versus NR2A subunits were additive if both subunits were mutated (i.e., if the N-site of NR1 and NR2A is each mutated to Q). Therefore, in comparison with NR1/NR2A wild-type receptors (with = 0.84), the NR1(N616Q)/NR2A(N614Q) mutant (with = 0.73) shifted toward the extracellular side of the membrane by 11% (= 25–14%). The effect on voltage dependence of the N to R mutation in the NR1 subunit was not tested because of the requirement for much higher concentrations of memantine to achieve substantial blockade. At such high concentrations, memantine is not specific for the NMDAR open-channel blocking site (Chen et al., 1992, and see later section for Fig. 2D).
The asparagine residue at the N+1 position of the NR2 subunit (N615) forms part of the channel selectivity filter (Kuner et al., 1996; Wollmuth et al., 1996). However, N615Q mutation of the NR2A subunit decreased the IC50 of memantine blockade by only 6-fold (IC50 = 6.3 ± 1.8 µM), while dramatically decreasing the voltage dependence of block ( = 0.40, Fig. 2B). The NR1/NR2A(N615R) combination resulted in no expression of functional heteromeric NMDA channels and therefore could not be tested further. Combining the affinity and voltage-dependent effects of mutations at the N and N+1 sites on the NR2A subunit, we found that these two residues contribute only slightly to the specific binding site for memantine (as evidenced by their small effect on apparent affinity) but exert a predominant influence on the electrostatic interaction of memantine binding (as seen by their effect on voltage dependence). Interestingly, these two NR2A residues have been shown to affect block by intracellular Mg2+ in a similar manner (Wollmuth et al., 1998b). In contrast, the dramatic effect of mutation of the N-site asparagine of the NR1 subunit on memantine affinity leads us to conclude that this residue (also representing the intracellular Mg2+ blocking site) is the predominant memantine blocking site.
We also tested the effect on memantine blockade of mutating residues around the N-site asparagine in the M2 region of the NR1 subunit. In agreement with a recent study characterizing many such residues (Kashiwagi et al., 2002), we found that S604A, S617N, and G618D mutations in the M2 domain of the NR1 subunit, which affect blockade by external Mg2+ or phencyclidine (Yamakura et al., 1993), exert minimal effects on the specific memantine channel-blocking site. W611L mutation of the NR1 subunit decreased memantine affinity slightly as well as voltage dependence ( = 0.63). Mutation of the A645 residue to serine in the M3 domain of the NR1 subunit, which affects MK-801 affinity, decreased the IC50 of memantine by 4-fold without changing its voltage dependence ( = 0.8). The tryptophan residue (W606L) at the –8 position from the N-site of the NR2A subunit, unlike its effect on NR1/NR2B receptors (Kashiwagi et al., 2002), decreased memantine blockade by only 2-fold (IC50 = 1.5 ± 0.9 µM) and slightly increased the voltage dependence ( = 0.88).
Use of Protection-SCAM to Locate the Memantine Binding Sites. To locate precisely the memantine binding sites on the NMDARs, we present in Fig. 3 a modified method for protecting exposed cysteine residues with memantine from reaction with methanethiosulfonate (MTS) agents (Protection-SCAM, see also Sobolevsky et al., 2002). The NR1(N616C) mutation decreased the IC50 of memantine to 40 µM (Fig. 2C) due to the critical role of the N-site on the NR1 subunit for specific memantine binding. In contrast, the NR2A(N614C) N-site mutation manifested a smaller effect on the potency of memantine (IC50 = 1.4 ± 0.6 µM) and decreased voltage dependence of block ( = 0.45). This result is consistent with the finding above that the N-site asparagine of NR2A affects memantine action primarily via electrostatic interaction. Since the N-site asparagines of the NR1 and NR2 subunits are located close together (Amar et al., 2001), we used the NR2A(N614C) mutant, which manifests less effect on affinity than the NR1 mutant, to determine whether memantine can prevent accessibility of MTS agents to the region controlling the channel selectivity filter. C399A was included in this mutant [NR2A(C399A/N614C)] to avoid reaction of MTS with endogenous C399 (see Materials and Methods). The NR2A(C399A) mutation did not affect the affinity or voltage dependence of memantine action.
A relatively high concentration of memantine (200 µM) was used in this experiment to ensure blockade of 100% of the specific memantine binding sites (on the other hand, 30% of the nonspecific sites were blocked by this concentration, Fig. 2A). We found that memantine prevented most of the effect of 20 µM MTSEA on the NR2A N-site mutant. To show this, we measured residual NMDA-evoked current after the addition of MTSEA in the presence and absence of memantine (Fig. 3A). To compare the effect of memantine on the various single cysteine-substituted mutants in the channel mouth, we measured the parameter of relative degree of modulation (block or potentiation) by MTSEA in the presence or absence of memantine (designated the "relative degree of modulation" or RDM; Fig. 3). MTSEA reacted with NR2A(N614C) and blocked the current by 79.5 ± 1.6%. Using the RDM, we found that memantine prevented 86.8 ± 3.1% of this MTSEA modulation. This result confirms that the selectivity filter region is the major site for specific memantine action (Chen and Lipton, 1997; Kashiwagi et al., 2002; Sobolevsky et al., 2002). Note that measurement of the slowing of the on rate of MTSEA action by memantine was not used in this study because of the relatively slow washout of memantine in the oocyte expression system (for details, see Materials and Methods).
Location of a Second Site of Memantine Action. As mentioned earlier, there appears to be a second, very low affinity site for memantine action, and amino acid residues in the M3 and post-M3 segments of the NR1 subunit can affect this blocking action of memantine. In agreement with the results of Beck et al. (1999), we found that MTSEA decreased the current of NR1(L651C)/NR2A(C399A), NR1(F654C)/NR2A(C399A), and NR1(R659C)/NR2A(C399A) channels, yet potentiated the current of NR1(A653C)/NR2A(C399A) channels. To locate the second, very low-affinity site of memantine action, the RDM was calculated for cysteine-substituted mutants in the channel mouth using higher concentrations of memantine. N to R substitution in NR2A changed the of memantine block to 0.5, a value similar to that of nonspecific polar organic ions such as Tris ( = 0.51) and N-methyl-D-glucamine ( = 0.51) (Villarroel et al., 1995). The value of for NR1(L651C), NR1(A653C), and NR1(F654C) mutations after exposure to MTS reagents was 
0.3, 0.0, and 0.0, respectively (Sobolevsky et al., 2002). We thus hypothesized that these residues must be located at or near the outer edge of the second, nonspecific site of memantine binding at the channel vestibule.
From the dose-response curve for memantine inhibition of NR1(N616R)/NR2A (Fig. 2A), we knew that 200 µM memantine blocked 30% of the nonspecific sites, whereas 2 mM blocked 90%. These two concentrations were therefore chosen to more precisely locate the second site of memantine action using Protection-SCAM. We found that 200 µM memantine prevented 50 to 60% of the modulation by MTSEA at L651C, A653C, and F654C (Fig. 4B, left panel). However, increasing the concentration of memantine to 2 mM protected L651C from MTSEA modulation even better. Nonetheless, we observed that there was less protection by 2 mM memantine as the location of the cysteine moved in the extracellular direction (residues A653C and F654C), suggesting that the second site of memantine binding was located at or very near L651 and internal to F654C (Fig. 4B, right panel).
For the post-M3 segment, we know from previous work that V656A and L657A mutations of the NR1 subunit affect the potency of memantine (Kashiwagi et al., 2002). The L657C mutation was minimally modified by MTSEA in our study. Therefore, the more externally located R659C mutation of NR1 was used to assess memantine protection from SCAM. However, we found that 2 mM memantine offered no protection of the R659C residue from reaction with MTSEA (Fig. 4B, right panel). As a technical note for these experiments, the residues for cysteine substitution were chosen because of their spatial proximity to the second memantine binding site, which influenced MTSEA modulation but did not significantly alter memantine affinity per se (data not shown).
Significance and Nonspecific Nature of the Second Site of Memantine Action. Hexamethonium (HME) is an acetylcholine receptor blocker that can nonspecifically block the NMDAR channel at millimolar concentrations at –100 mV (Villarroel et al., 1995). We found that 2 and 20 mM HME, at a holding potential of –100 mV, blocked NMDA-gated current by 50 and 90%, respectively. We then used the Protection-SCAM technique described above to localize the site of nonspecific HME binding in the NMDAR channel. NR1(L651C) modulation was slightly affected by 2 mM HME, whereas 20 mM drastically reduced L651C modulation. In contrast, 2 mM HME did not affect modulation of A653C or F654C, and 20 mM had only a modest effect (30–40% protection), suggesting that A653 and F654 are located at the outer edge of the HME binding site (Fig. 5). Additionally, the residue R659C in the post-M3 segment of the NR1 subunit was not protected from MTSEA modulation by 20 mM HME. Thus, the nonspecific site of HME binding is virtually the same as the nonspecific site of memantine action.
|
|
300 and 700 µM, respectively. Since at hundreds of micromolar, both memantine and amantadine exhibit effects on other receptors (Chen et al., 1992), these results suggest that this block is relatively nonspecific and that dose-response curves for memantine and amantadine approach similar values after losing their interaction with the specific binding site in the channel. Furthermore, we suggest that the ratio of affinities between the specific and nonspecific blocking sites for amantadine (20-fold) is too small compared with memantine (300-fold) to allow clinical application of amantadine as a specific and clinically tolerated NMDA antagonist (Fig. 2D). We speculate that the method of comparing affinities between sites demonstrated in Fig. 2D will prove useful in searching for additional clinically tolerated, "low-affinity" open-channel blockers (see Discussion).
Lack of Effect of Variant NR1 Isoforms on Memantine Potency. The experiments described above were all performed with the NR1–1a splice variant, which has two C-terminal inserts and no N-terminal insert. We therefore coexpressed NR1–1b (containing all N- and C-terminal inserts), NR1–4a (lacking all N- and C-terminal inserts), or NR1–4b (containing only the N-terminal insert) with NR2A to form heteromeric channels in the oocyte expression system. None of the NR1 splice variants differentially affected the affinity of memantine antagonism. Neither was the voltage dependence of memantine blockade differentially influenced (Table 1).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsDiscussionReferences |
|---|
). Additionally, a second, more superficial site for memantine antagonism that manifests a much lower affinity is located on NR1 at the level of residue L651. This second site appears to be nonspecific and overlaps the site occupied by nonspecific pore blockers, such as hexamethonium. We show that a large difference in the affinity between these two sites is crucial for maintaining the selectivity of a low-affinity NMDAR open-channel blocker like memantine. Finally, we find amino acid residues in the more external post-M3 segment are not directly involved in memantine binding.
Many possible factors have been suggested for memantine's clinical tolerability, including moderate-to-low affinity, fast unblocking kinetics, and moderate voltage dependence in the NMDAR-associated channel (Rogawski and Wenk, 2003). These macroscopic explanations suggest that memantine acts like a better magnesium based on comparison with the kinetic properties of block by "extracellular" Mg2+ at the channel selectivity filter (Danysz and Parsons, 2003). In the present study, however, we show that this comparison may be misleading. Instead, memantine interacts with the intracellular Mg2+ blocking site, which is located slightly deeper than the extracellular Mg2+ blocking site (Kuner et al., 1996). This finding clearly indicates that memantine should not be simply considered a "better external Mg2+ blocker".
Our experiments on the voltage dependence of memantine blockade revealed further details of the molecular interaction of memantine with amino acid residues in the selectivity filter region. N-to-Q substitution at the N-site of the NR2A subunit shifted the calculated 25% toward the extracellular side of the membrane electric field. In contrast, the N-to-Q mutation at the N-site of the NR1 subunit [NR1(N616Q)] shifted the calculated approximately 14% toward the "intracellular" side. These two opposing effects on electrical distance were additive if both subunits were mutated. One possible explanation for this opposing effect on the calculated is that the carbonyl oxygen in the N-site of the NR2A subunit points toward the extracellular side of the membrane, whereas the carbonyl oxygen in the N-site of the NR1 subunit points intracellularly.
Furthermore, the NR1(N616Q) mutation affected the voltage dependence of memantine blockade less than half that of the equivalent substitution on the NR2A subunit. A simple model (Fig. 6) suggests that the carbonyl oxygen of the N-site on the NR1 subunit is located relatively horizontal to the long axis of the channel pore, although the carbonyl oxygen side chain of the NR2A N-site manifests a more vertical orientation. We also found that the N-to-Q mutation at the N+1 site of the NR2A subunit decreased the calculated more than the N-site substitution, suggesting a steric effect in addition to a simple electrostatic interaction and compatible with a critical role of the N+1 site in forming the selectivity filter. The differential orientation of carbonyl oxygens in NR1 and NR2A subunits at the channel selectivity filter may explain why other open-channel blockers differ in their pharmacological properties from memantine if they predominantly interact with the extracellular Mg2+ site. To support this concept, we modeled the M2-to-M3 segment of the NR1 subunit as an inverted K+ channel based on the latter's crystal structure. The resulting model showed that the orientation of the carbonyl oxygen of the N-site of the NR1 subunit is very close to that proposed here. Due to a staggered configuration relative to the NR1 subunit (Wollmuth and Sobolevsky, 2004), the M2-to-M3 segments of the NR2A subunit cannot be simply modeled in this manner (see legend Fig. 6).
Furthermore, the reduced voltage dependence of memantine action after N-to-Q mutation at the N or N+1 sites of the NR2A subunit suggests that the externally pointing carbonyl oxygens provide significant electrostatic stabilization of memantine when it is bound to the N-site of the NR1 subunit. Conceivably, this electrostatic stabilization would affect many drug properties such as slowing the unblocking rate, changing the voltage dependence, and enhancing trapping of memantine. This complex effect may explain why prior studies, focused on low-affinity, slow kinetics of block, voltage dependence, and partial trapping, have not identified a single factor for the tolerability of memantine (Parsons et al., 1995). The true molecular binding site(s) of other open-channel blockers and their interaction with NMDARs should be further investigated to find a safer and more effective antagonist.
Other studies have shown that amino acid residues in addition to the N-site asparagine in the M2 and M3 segments of NR1 and NR2 subunits affect Mg2+ and memantine blockade of NMDARs (Dingledine et al., 1999; Kashiwagi et al., 2002). In the present study, we performed assays of both affinity and voltage dependence to assess the effects of these other amino acid substitutions on memantine blockade. We found that residues located near the selectivity filter region in either the M2 or M3 segment of NR1(W611 and A645) or NR2A(W606) slightly affected the affinity or voltage dependence of memantine blockade. This result is consistent with the notion that these residues exert an allosteric effect on memantine action.
We also precisely located a second, superficial binding site for memantine near the level of residue L651 in the M3 segment of the NR1 subunit. Amino acid residues located externally to NR1(F654) were not protected by memantine even at very high (millimolar) concentrations in the Protection-SCAM assay and are therefore not directly involved in memantine binding. In contrast to our results, A653T in the M3 segment and V656A, L657A in the post-M3 segment of NR1, were reported to decrease antagonism by 1 µM memantine (Kashiwagi et al., 2002). However, the A653C and L657C residues manifest differential modification rates by MTSEA in the open versus closed conformation (Sobolevsky et al., 2002). Thus, these externally located residues are more likely to be involved in a conformational change during channel gating. Their mutations, therefore, may alter the structure or the open-closed equilibrium of the permeation pathway, resulting in a change in affinity of various open-channel blockers.
Alternatively, the effect of the NR1(A653T) mutation might suggest that binding of memantine to this superficial site is required prior to a rapid transition into the deeper selectivity filter site (Antonov et al., 1998). At low micromolar concentrations and hyperpolarized potentials, this latter scenario would have minimal significance after considering the very low affinity of memantine (>300 µM at –60 mV) at this superficial site, which would result in such an ephemeral dwell time that it would preclude any physical meaning. However, this transition may become more significant for very low affinity open-channel blockers like amantadine and at more depolarized potentials, such as pathological depolarization by toxic levels of glutamate.
Additionally, we established a simple method using NR1(N616R) versus the wild-type NR1 subunit to quickly evaluate the relative affinity between the two sites of memantine action (as in Fig. 2D). For so called low-affinity NMDA open-channel blockers, the apparent affinity, apparent unblocking rate, and degree of trapping of each open-channel blocker will represent "mixed" properties of both sites if the relative affinities are not too far apart. Therefore, despite prior reports, none of these properties at one site alone can explain the variable clinical tolerability of low-affinity NMDAR antagonists. In the case of memantine, however, the affinities of the two sites are sufficiently distinct that the pharmacological properties of the specific site may account for its lack of side effects (Lipton and Chen, 2004).
Another important therapeutic implication of our findings concerns uncompetitive antagonism. Uncompetitive, unlike competitive or noncompetitive, antagonists can block excessive activation of NMDARs while sparing normal neurotransmission (Lipton and Chen, 2004). Memantine blocks NMDARs via an uncompetitive mechanism at low micromolar concentrations, yet possesses a noncompetitive component (via partial trapping in the channel) at higher concentrations (Blanpied et al., 1997; Chen and Lipton, 1997). Lipophilic leak of memantine from its blocking site cannot explain this noncompetitive component (Mealing et al., 2001; Bolshakov et al., 2003). Instead, here we identify a second site of memantine binding of very low affinity at the channel vestibule. Occupancy by memantine of this shallow site may allow dissociation of the drug in either the open or closed conformation, resulting in a form of noncompetitive antagonism. This superficial site of memantine action is nonspecific and may also explain the noncompetitive component of many other very low-affinity, open-channel blockers. Importantly, we describe a method to rapidly evaluate the relative affinity of drugs between the superficial and deep sites of the NMDAR. This relative affinity may be used to predict the "noncompetitive" versus "uncompetitive" component of action of an NMDAR channel blocker and therefore will be an important tool in the search for future therapeutic agents in this class.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
ABBREVIATIONS: NMDAR, N-methyl-D-aspartate receptor; MK-801, dizocilpine; MTSEA, 2-aminoethyl-methanethiosulphonate; SCAM, substituted cysteine accessibility method; MEM, memantine; MTS, methanethiosulfonate; RDM, relative degree of modulation; HME, hexamethonium.
Address correspondence to: Dr. Huei-Sheng Vincent Chen, The Burnham Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037. E-mail: hsv_chen{at}burnham.org
|
References |
|---|
|
TopAbstractMaterials and MethodsDiscussionReferences |
|---|
Amar M, Perin-Dureau F, and Neyton J (2001) High-affinity Zn block in recombinant N-methyl-D-aspartate receptors with cysteine substitutions at the Q/R/N site. Biophys J 81: 107–116.[Medline]
Antonov SM, Gmiro VE, and Johnson JW (1998) Binding sites for permeant ions in the channel of NMDA receptors and their effects on channel block. Nat Neurosci 1: 451–461.[CrossRef][Medline]
Antonov SM and Johnson JW (1996) Voltage-dependent interaction of open-channel blocking molecules with gating of NMDA receptors in rat cortical neurons. J Physiol 493: 425–445.
Beck C, Wollmuth LP, Seeburg PH, Sakmann B, and Kuner T (1999) NMDAR channel segments forming the extracellular vestibule inferred from the accessibility of substituted cysteines. Neuron 22: 559–570.[CrossRef][Medline]
Blanpied TA, Boeckman FA, Aizenman E, and Johnson JW (1997) Trapping channel block of NMDA-activated responses by amantadine and memantine. J Neurophysiol 77: 309–323.
Bolshakov KV, Gmiro VE, Tikhonov DB, and Magazanik LG (2003) Determinants of trapping block of N-methyl-D-aspartate receptor channels. J Neurochem 87: 56–65.[CrossRef][Medline]
Bresink I, Benke TA, Collett VJ, Seal AJ, Parsons CG, Henley JM, and Collingridge GL (1996) Effects of memantine on recombinant rat NMDA receptors expressed in HEK 293 cells. Br J Pharmacol 119: 195–204.[Medline]
Chen HS and Lipton SA (1997) Mechanism of memantine block of NMDA-activated channels in rat retinal ganglion cells: uncompetitive antagonism. J Physiol 499: 27–46.
Chen HS, Pellegrini JW, Aggarwal SK, Lei SZ, Warach S, Jensen FE, and Lipton SA (1992) Open-channel block of N-methyl-D-aspartate (NMDA) responses by memantine: therapeutic advantage against NMDA receptor-mediated neurotoxicity. J Neurosci 12: 4427–4436.[Abstract]
Choi DW (1998) Antagonizing excitotoxicity: a therapeutic strategy for stroke? Mt Sinai J Med 65: 133–138.[Medline]
Choi YB, Tenneti L, Le DA, Ortiz J, Bai G, Chen HS, and Lipton SA (2000) Molecular basis of NMDA receptor-coupled ion channel modulation by S-nitrosylation. Nat Neurosci 3: 15–21.[CrossRef][Medline]
Danysz W and Parsons CG (2003) The NMDA receptor antagonist memantine as a symptomatological and neuroprotective treatment for Alzheimer's disease: preclinical evidence. Int J Geriatr Psychiatry 18 (Suppl 1): S23–S32.[CrossRef][Medline]
Dingledine R, Borges K, Bowie D, and Traynelis SF (1999) The glutamate receptor ion channels. Pharmacol Rev 51: 7–61.
Karlin A and Akabas MH (1998) Substituted-cysteine accessibility method. Methods Enzymol 293: 123–145.[CrossRef][Medline]
Kashiwagi K, Masuko T, Nguyen CD, Kuno T, Tanaka I, Igarashi K, and Williams K (2002) Channel blockers acting at N-methyl-D-aspartate receptors: differential effects of mutations in the vestibule and ion channel pore. Mol Pharmacol 61: 533–545.
Kuner T, Wollmuth LP, Karlin A, Seeburg PH, and Sakmann B (1996) Structure of the NMDA receptor channel M2 segment inferred from the accessibility of substituted cysteines. Neuron 17: 343–352.[CrossRef][Medline]
Lipton SA and Chen HS (2004) Paradigm shift in neuroprotective drug development: clinically tolerated NMDA receptor inhibition by memantine. Cell Death Diff 11: 18–20.[CrossRef][Medline]
Lipton SA and Rosenberg PA (1994) Mechanisms of disease: excitatory amino acids as a final common pathway for neurologic disorders. N Engl J Med 330: 613–622.
Mealing GA, Lanthorn TH, Small DL, Murray RJ, Mattes KC, Comas TM, and Morley P (2001) Structural modifications to an N-methyl-D-aspartate receptor antagonist result in large differences in trapping block. J Pharmacol Exp Ther 297: 906–914.
Monyer H, Sprengel R, Schoepfer R, Herb A, Higuchi M, Lomeli H, Burnashev N, Sakmann B, and Seeburg PH (1992) Heteromeric NMDA receptors: molecular and functional distinction of subtypes. Science (Wash DC) 256: 1217–1221.
Moriyoshi K, Masu M, Ishii T, Shigemoto R, Mizuno N, and Nakanishi S (1991) Molecular cloning and characterization of the rat NMDA receptor. Nature (Lond) 354: 31–37.[CrossRef][Medline]
Nakanishi S (1992) Molecular diversity of glutamate receptors and implications for brain function. Science (Wash DC) 258: 597–603.
Parsons CG, Quack G, Bresink I, Baran L, Przegalinski E, Kostowski W, Krzascik P, Hartmann S, and Danysz W (1995) Comparison of the potency, kinetics and voltage-dependency of a series of uncompetitive NMDA receptor antagonists in vitro with anticonvulsive and motor impairment activity in vivo. Neuropharmacology 34: 1239–1258.[CrossRef][Medline]
Reisberg B, Doody R, Stoffler A, Schmitt F, Ferris S, and Mobius HJ; Memantine Study Group (2003) Memantine in moderate-to-severe Alzheimer's disease. N Engl J Med 348: 1333–1341.
Rogawski MA and Wenk GL (2003) The neuropharmacological basis for the use of memantine in the treatment of Alzheimer's disease. CNS Drug Rev 9: 275–308.[Medline]
Sakurada K, Masu M, and Nakanishi S (1993) Alteration of Ca2+ permeability and sensitivity to Mg2+ and channel blockers by a single amino acid substitution in the N-methyl-D-aspartate receptor. J Biol Chem 268: 410–415.
Schneider E, Fischer PA, Clemens R, Balzereit F, Fünfgeld EW, and Haase HJ (1984) Effects of oral memantine administration on Parkinson symptoms. Results of a placebo-controlled multicenter study. Dtsch Med Wochenschr 109: 987–990.[Medline]
Sobolevsky AI, Beck C, and Wollmuth LP (2002) Molecular rearrangements of the extracellular vestibule in NMDAR channels during gating. Neuron 33: 75–85.[CrossRef][Medline]
Sobolevsky AI, Koshelev SG, and Khodorov BI (1998) Interaction of memantine and amantadine with agonist-unbound NMDA-receptor channels in acutely isolated rat hippocampal neurons. J Physiol 512: 47–60.
Sugihara H, Moriyoshi K, Ishii T, Masu M, and Nakanishi S (1992) Structures and properties of seven isoforms of the NMDA receptor generated by alternative splicing. Biochem Biophys Res Commun 185: 826–832.[CrossRef][Medline]
Sullivan JM, Traynelis SF, Chen HS, Escobar W, Heinemann SF, and Lipton SA (1994) Identification of two cysteine residues that are required for redox modulation of the NMDA subtype of glutamate receptor. Neuron 13: 929–936.[CrossRef][Medline]
Tariot PN, Farlow MR, Grossberg GT, Graham SM, McDonald S, Gergel I; Memantine Study Group (2004) Memantine treatment in patients with moderate to severe Alzheimer disease already receiving donepezil: a randomized controlled trial. JAm Med Assoc 291: 317–324.
Villarroel A, Burnashev N, and Sakmann B (1995) Dimensions of the narrow portion of a recombinant NMDA receptor channel. Biophys J 68: 866–875.[Medline]
Wollmuth LP, Kuner T, and Sakmann B (1998a) Adjacent asparagines in the NR2-subunit of the NMDA receptor channel control the voltage-dependent block by extracellular Mg2+. J Physiol 506: 13–32.
Wollmuth LP, Kuner T, and Sakmann B (1998b) Intracellular Mg2+ interacts with structural determinants of the narrow constriction contributed by the NR1-subunit in the NMDA receptor channel. J Physiol 506: 33–52.
Wollmuth LP, Kuner T, Seeburg PH, and Sakmann B (1996) Differential contribution of the NR1- and NR2A-subunits to the selectivity filter of recombinant NMDA receptor channels. J Physiol 491: 779–797.
Wollmuth LP and Sobolevsky AI (2004) Structure and gating of the glutamate receptor ion channel. Trends Neurosci 27: 321–328.[CrossRef][Medline]
Woodhull AM (1973) Ionic blockage of sodium channels in nerve. J Gen Physiol 61: 687–708.
Yamakura T, Mori H, Masaki H, Shimoji K, and Mishina M (1993) Different sensitivities of NMDA receptor channel subtypes to non-competitive antagonists. Neuroreport 4: 687–690.[Medline]
This article has been cited by other articles:
|
|
 |
|
  S. E. Kotermanski, J. T. Wood, and J. W. Johnson Memantine binding to a superficial site on NMDA receptors contributes to partial trapping J. Physiol., October 1, 2009; 587(19): 4589 - 4604. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Nakanishi, S. Tu, Y. Shin, J. Cui, T. Kurokawa, D. Zhang, H.-S. V. Chen, G. Tong, and S. A. Lipton Neuroprotection by the NR3A Subunit of the NMDA Receptor J. Neurosci., April 22, 2009; 29(16): 5260 - 5265. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. E. Kotermanski and J. W. Johnson Mg2+ Imparts NMDA Receptor Subtype Selectivity to the Alzheimer's Drug Memantine J. Neurosci., March 4, 2009; 29(9): 2774 - 2779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Blanke and A. M. J. VanDongen The NR1 M3 Domain Mediates Allosteric Coupling in the N-Methyl-D-aspartate Receptor Mol. Pharmacol., August 1, 2008; 74(2): 454 - 465. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Wrighton, E. J. Baker, P. E. Chen, and D. J. A. Wyllie Mg2+ and memantine block of rat recombinant NMDA receptors containing chimeric NR2A/2D subunits expressed in Xenopus laevis oocytes J. Physiol., January 1, 2008; 586(1): 211 - 225. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Villmann and C.-M. Becker On the Hypes and Falls in Neuroprotection: Targeting the NMDA Receptor Neuroscientist, December 1, 2007; 13(6): 594 - 615. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Jin, H. Sugiyama, M. Takigawa, D. Katagiri, H. Tomitori, K. Nishimura, N. Kaur, O. Phanstiel IV, M. Kitajima, H. Takayama, et al. Comparative Studies of Anthraquinone- and Anthracene-Tetraamines as Blockers of N-Methyl-D-aspartate Receptors J. Pharmacol. Exp. Ther., January 1, 2007; 320(1): 47 - 55. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
