Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

NMDA receptors are involved in excitatory synaptic transmission and synaptic plasticity. Overactivation of these receptors can lead to neuronal cell death, and the receptors also play a role in seizure activity. Thus, NMDA receptors are potential targets for neuroprotective agents and anticonvulsants (Choi, 1988; Rogawski, 1992). NMDA receptors have a complex pharmacology and are targets for antagonists acting at the glutamate and glycine coagonist sites, at a large number of modulatory sites and at sites within the ion channel of the receptor (McBain and Mayer, 1994). The cDNAs encoding a number of subunits of NMDA receptors have been cloned, including the NR1 and NR2A, NR2B, NR2C and NR2D subunits (Hollmann and Heinemann, 1994; Moriyoshi et al., 1991). Native NMDA receptors are probably hetero-oligomers containing combinations of NR1 and NR2 subunits (Luo et al., 1997; Sheng et al., 1994). Recombinant NMDA receptors expressed from particular combinations of cloned subunits are a valuable experimental system with which to study the structure and function of the receptors and the site and mechanism of action of receptor antagonists.

A number of organic polycations, including the endogenous polyamines spermine and spermidine, are antagonists at native and recombinant NMDA receptors (Williams, 1997). When applied extracellularly, polyamines act as open-channel blockers at NMDA receptors and may also reduce currents through these channels by screening of surface charges (Araneda et al., 1993; Benveniste and Mayer, 1993; Rock and Macdonald, 1992). Spermine and spermidine are, however, very weak blockers of NMDA channels, being active at high micromolar to millimolar concentrations. Polyamine-conjugated spider and wasp toxins, containing an aromatic head group and polyamine tail, have been found to be potent glutamate receptor antagonists (Jackson and Usherwood, 1988). These toxins block invertebrate glutamate receptors, and in some cases, they also block mammalian NMDA and AMPA/kainate receptors. Such compounds are useful as tools to study the structure and function of glutamate receptors, although many of the toxins are unstable and are difficult to obtain because their syntheses are complex. We have recently found that N¹-dansyl derivatives of spermine and spermidine are potent NMDA channel blockers, being several hundred- to several thousand-fold more potent than the native polyamines (Chao et al., 1997). N¹-Dansyl-polyamines contain a hydrophobic substitution at one terminal amino group, but the remainder of the polyamine chain is intact. Because of the potent activities of dansyl-polyamines, we examined other polyamine derivatives with hydrophobic substitutions. Here, we describe the effects of benzyl-substituted polyamines. A number of di- and tri-benzyl polyamines were found to be potent NMDA receptor antagonists. Surprisingly, the most potent compounds were N,N,N-tri-benzyl triamines, such as the spermidine derivative TB-3-4, with a benzyl substitution on each of the three amino groups. Tri-benzyl-polyamines represent a new class of NMDA receptor antagonists that may be useful as tools to study NMDA channels and as lead compounds for novel neuroprotective or anticonvulsant agents.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

cDNA clones. The NR1 clone (Moriyoshi et al., 1991) and the NR1(N616Q) mutant were gifts from Dr. S. Nakanishi (Institute for Immunology, Kyoto University Faculty of Medicine, Kyoto, Japan). The NR1(N616R) mutant was a gift from Dr. R. J. Dingledine (Department of Pharmacology, Emory University, Atlanta, GA). The splice variant of NR1 used in these studies was NR1A (Moriyoshi et al., 1991; Sugihara et al., 1992). The NR2A and NR2B clones (Monyer et al., 1992) were gifts from Dr. P. H. Seeburg (Center for Molecular Biology, University of Heidelberg, Germany). The mouse NR2C and NR2D clones (3 and 4) (Ikeda et al., 1992; Kutsuwada et al., 1992) were gifts from Dr. M. Mishina (Department of Pharmacology, University of Tokyo, Japan). The NR1(N616G) mutant was prepared as described previously (Chao et al., 1997), and amino acids are numbered from the initiator methionine in NR1 and NR2 clones (Ishii et al., 1993; Moriyoshi et al., 1991). The GluR1 clone (Hollmann et al., 1989) was a gift from Drs. J. Boulter and S. F. Heinemann (The Salk Institute for Biological Studies, La Jolla, CA).

Preparation of oocytes and voltage-clamp recording. The preparation and maintenance of oocytes were carried out as described previously (Williams et al., 1993). Capped cRNAs were prepared from linearized cDNA templates using mMessage mMachine kits (Ambion, Austin, TX). NR1 and NR2 subunits were injected in a ratio of 1:5 (0.25-4 ng of NR1 + 1.25-20 ng of NR2). Macroscopic currents were recorded with a two-electrode voltage-clamp using an OC-725 amplifier (Warner Instruments, Hamden, CT) or a GeneClamp 500 amplifier (Axon Instruments, Foster City, CA). Oocytes were continuously superfused (~5 ml/min) with a Mg⁺⁺-free saline solution (96 mM NaCl, 2 mM KCl, 1.8 mM BaCl₂, 10 mM HEPES, pH 7.5), which contained BaCl₂ rather than CaCl₂ to minimize Ca⁺⁺-activated Cl currents (Leonard and Kelso, 1990). In most experiments with NMDA receptors, oocytes were injected with K⁺-BAPTA (100 nl; 40 mM, pH 7.0) on the day of recording (Williams, 1993).

(1)

(2)

Synthesis of benzyl-polyamines. Benzyl-polyamines were prepared as the hydrochloride forms. The mono- and tri-benzyl derivatives 4-MB-3-3, 5-MB-4-4, TB-3-3, TB-3-4 and TB-4-4 (Niitsu and Samejima, 1986) and 4,9-DB-3-4-3 (Samejima et al., 1984) were synthesized as described previously.³ The other di-benzyl derivatives, DB-3, DB-7, DB-10, DB-3-3, DB-3-4, DB-4-4 and 1,12-DB-3-4-3, were synthesized using a method previously described for the synthesis of di-benzyl-putrescine but using an appropriate diamine (e.g., diaminoheptane for DB-7) or triamine (e.g., spermidine for DB-3-4) for synthesis of the corresponding di-benzyl-polyamine (Samejima et al., 1984).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Activities of benzyl-polyamines. A number of mono-, di- and tri-benzyl polyamines were studied. The effects of these compounds on responses to glutamate (10 µM; with 10 µM glycine) were measured at NR1/NR2A receptors in oocytes voltage-clamped at 70 mV (fig. 1). We initially studied mono- and di-benzyl polyamine derivatives. At concentrations of 1 to 10 µM, the di-benzyl-diamines DB-3, DB-7 and DB-10 inhibited responses to glutamate and were slightly more potent than di-benzyl-triamines (DB-3-3 and DB-3-4) of a similar chain length (fig. 1). The mono-benzyl triamines 4-MB-3-3 and 5-MB-4-4, which have a benzyl group attached to the central nitrogen, had activities similar to the terminal di-benzyl diamines (DB-7, DB-10) and triamines (DB-3-3 and DB-4-4) of equivalent chain length, suggesting that benzylation at either the terminal or central amino groups increases the potency of polyamines as NMDA antagonists. This conclusion was supported by the finding that the tri-benzyl triamines TB-3-3, TB-3-4 and TB-4-4 were more potent than their mono-benzyl analogs (4-MB-3-3 and 5-MB-4-4) or their di-benzyl analogs (DB-3-3, DB-3-4 and DB-4-4). Two spermine derivatives, 1,12-DB-3-4-3 and 4,9-DB-3-4-3, were also potent antagonists, with 4,9-DB-3-4-3 having activity similar to the spermidine derivative TB-3-4 (fig. 1).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Effects of polyamine analogs at NMDA receptors. The effects of spermidine, spermine and mono-, di- and tri-benzyl polyamines, each at 1 and 10 µM, on responses to glutamate (10 µM; with 10 µM glycine) were determined in oocytes expressing NR1/NR2A receptors and voltage-clamped at 70 mV. Data are expressed as a percentage of the control response to glutamate. Values are mean ± S.E.M. from four to seven oocytes for each compound. The structures of the analogs are also shown.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Subunit-specific block of NMDA receptors. A, The effects of various concentrations of TB-3-4 on responses to glutamate (glu, 10 µM; with 10 µM glycine) were determined in an oocyte expressing NR1/NR2A receptors and voltage-clamped at 70 mV. For illustration, traces have been normalized to the glutamate response at the beginning of the test pulse for each concentration of TB-3-4 because the response to glutamate showed a small (12%) rundown over the course of the experiment. Thus, calibration bars are shown next to the first and last traces. B, Concentration-inhibition curves for TB-3-4 were determined at NR1/NR2 receptors containing NR2A, NR2B, NR2C and NR2D subunits using protocols similar to those shown in A. Values are mean ± S.E.M. from four to seven oocytes for each subunit combination.

View this table: [in this window] [in a new window]   TABLE 1 Effects of TB-3-4 at wild-type and mutant NMDA receptors The effects of TB-3-4 were measured at NR1/NR2 receptors containing wild-type and mutant NR1 subunits in oocytes voltage-clamped at 70 mV. Values for the IC50, mean (S.E.M., +S.E.M.), and Hill slope (nH; mean ± S.E.M.) were determined from concentration-inhibition curves with five or six concentrations of TB-3-4 at each receptor type.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Effects of polyamine analogs at NMDA and AMPA receptors. A and B, Effects of TB-3-4, 4,9-DB-3-4-3 and N1-DnsSpm were determined at NR1/NR2A receptors (A) activated by glutamate (glu, 10 µM; with 10 µM glycine) and at GluR1 receptors (B) activated by 100 µM kainate (KA) in oocytes voltage-clamped at 70 mV. All horizontal scale bars in A and B are 30 sec. Recovery from block by 10 µM 4,9-DB-3-4-3 at NR1/NR2A receptors and by N1-DnsSpm at GluR1 receptors was extremely slow and sometimes incomplete. The breaks in the traces correspond to 13 min (NR1/NR2A) and 8 min (GluR1). C, Responses to glutamate or kainate measured in the presence of polyamines are expressed as a percentage of the control response at each receptor type. Values are mean ± S.E.M. from four to seven oocytes.

Voltage-dependence and permeation through NMDA channels. TB-3-4 did not act as a competitive antagonist at the glutamate or glycine sites because inhibition by TB-3-4 (0.3 µM) was not overcome by increasing the concentrations of glutamate and glycine over the range of 0.3 to 10 µM (data not shown). Benzyl-polyamines including TB-3-4 may act as NMDA channel blockers, similar to spermine itself and to polyamine analogs such as N¹-DnsSpm. Consistent with this idea, the block by TB-3-4 was strongly voltage dependent, being more pronounced at hyperpolarized membrane potentials (fig. 4).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Voltage-dependent block by TB-3-4. A, I-V curves were constructed by voltage ramps (150 to +30 mV over 6 sec; inset) at NR1/NR2B and NR1/NR2D receptors activated by glutamate (10 µM; with 10 µM glycine) in the absence and presence of 1 µM TB-3-4. Leak currents have been subtracted. B, Currents measured in the presence of TB-3-4 (Iglu+TB-3-4) are expressed as a fraction of the control current (Iglu) for each oocyte. The solid lines are fits to the Woodhull model (text equation 2), which was fit from +30 to 100 mV (NR1/NR2B) and from +30 to 115 mV (NR1/NR2D) and extrapolated to 150 mV. Data around the reversal potential were masked during the fitting procedure.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Permeation of TB-3-4 through mutant NMDA channels. A, I-V curves were constructed by voltage-ramps from 185 to +40 mV (see inset in NR1/NR2A) at cells expressing NR1/NR2A, NR1(N616Q)/NR2A and NR1(N616G)/NR2A receptors. Responses to 10 µM glutamate (with 10 µM glycine) were measured in the absence (control) and presence of 0.3 µM TB-3-4. Leak currents have been subtracted. B, Glutamate currents measured in the presence of 0.3 µM or 3 µM TB-3-4 (Iglu + TB-3-4) are expressed as a fraction of the control glutamate current (Iglu) at 100 and 170 mV. Data were obtained from voltage ramps similar to those shown in A and are mean ± S.E.M. from five to nine oocytes for each subunit combination.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we found that N-benzyl polyamines are potent NMDA receptor antagonists. The rationale for studying benzyl-polyamines was based on the finding that linear polyamines with a hydrophobic substitution at one end of the polyamine chain, such as the polyamine-conjugated spider toxins or the spermine derivative N¹-DnsSpm, are potent NMDA channel blockers. In these molecules, the linear polyamine chain remains intact, and we have proposed that the polyamine tail of N¹-DnsSpm may enter deep into the ion channel pore with the head group of the molecule interacting with more peripheral residues in the mouth of the channel (Chao et al., 1997; Kashiwagi et al., 1997). Furthermore, bis(ethyl)penta-amines, which have hydrophobic substitutions on the terminal amino groups, are potent NMDA channel blockers (Igarashi and Williams, 1995). We hypothesized that benzyl substitutions at the terminal amino groups of polyamines may increase their potencies as NMDA channel blockers. We also wanted to determine the effects of benzyl-substitutions at the central imino groups of polyamines because hydrophobic substitutions at these positions have not been studied previously. The N-benzyl substitutions increased the potencies of polyamines, and the most potent benzyl derivatives were N⁴,N⁹-di-benzyl-spermine (4,9-DB-3-4-3), in which the benzyl groups are attached to the central imino groups, and tri-benzyl-triamines such as N¹,N⁴,N⁸-tri-benzyl-spermidine (TB-3-4), which has benzyl substitutions at the terminal amino and central imino groups.

Many of the polyamine-conjugated spider toxins block AMPA or kainate receptors in addition to NMDA receptors, and none of the known toxins are highly selective for NMDA channels. Similarly, N¹-DnsSpm is a potent voltage-dependent blocker of GluR1 AMPA channels⁴ as well as NMDA channels (Chao et al., 1997). Notably, TB-3-4, at a concentration 50-fold higher than its IC₅₀ value at NR1/NR2A receptors, was almost inactive at GluR1 channels. Thus, the pharmacology of TB-3-4 is markedly different from that of N¹-DnsSpm, with TB-3-4 being a selective NMDA antagonist. Compounds such as TB-3-4 and N¹-DnsSpm are useful new tools to study the structural properties of glutamate receptor channels and to study differences in channel structure between different classes of receptors.

There are a number of possibilities that may account for the different profiles seen with simple N-substituted polyamines such as TB-3-4 and N¹-DnsSpm. The potency of TB-3-4 (IC₅₀ = 0.2 µM) is similar to that of N¹-DnsSpm (IC₅₀ = 0.3 µM) and is several thousand-fold greater than that of spermine or spermidine (IC₅₀ = 500-1500 µM) (Chao et al., 1997). The increased potency of N¹-DnsSpm compared with spermine is due in large part to an increase in voltage-dependence rather than to an increase in the affinity of binding, with the value of z being 2.6 for N¹-DnsSpm and 1.1 for spermine, and the K_d(0) value being 800 µM for N¹-DnsSpm and 7400 µM for spermine (Chao et al., 1997). The K_d(0) value for TB-3-4 was 5 µM, suggesting that the affinity of the binding site for TB-3-4 is considerably higher than that for spermine or N¹-DnsSpm and that the increase in the potency of TB-3-4 compared with spermine and spermidine is due largely to an increase in the affinity of binding with only a small increase in the voltage-dependence of block. Indeed, the value of z for block by TB-3-4 (z = 1.41) is much less than that of N¹-DnsSpm, and the calculated value of , the average depth of the membrane electric field sensed by the blocker, is much smaller for TB-3-4 ( = 0.47) than for N¹-DnsSpm ( = 0.87). This suggests that the two polyamine analogs may bind to different sites within the channel pore of NMDA receptors. However, there are a number of limitations to these interpretations and to the use of the Woodhull model to characterize block of N-substituted polyamines. For example, it is not known whether only one molecule of the polyamine enters and binds to the channel or whether two or more molecules can simultaneously enter the channel. Similarly, it is not known whether all three of the charged amino groups of TB-3-4 and N¹-DnsSpm enter the membrane electric field. If only one molecule of each polyamine enters the channel at a time, then TB-3-4 and N¹-DnsSpm may have separate binding sites. However, if two molecules of N¹-DnsSpm but only one of TB-3-4 can enter the channel, this may account for the observation that the value of z for N¹-DnsSpm is about twice that of the z value for TB-3-4.

Some NMDA channel blockers can permeate the channel if the driving force for the blocker is made sufficiently large. Permeation of native or wild-type recombinant NMDA channels by low concentrations of Mg⁺⁺, by spermine and by linear polyamine analogs such as BE4444 has been reported (Benveniste and Mayer, 1993; Igarashi and Williams, 1995; Mayer and Westbrook, 1987). Asparagine residues in NR1 and NR2 subunits, including N616 in NR1, control cation permeability and appear to form the narrowest constriction of the ion channel pore (Wollmuth et al., 1996). An asparagine-to-glycine (N-to-G) mutation at N616 increases the apparent size of this constriction from 0.55 to 0.75 nm. Paradoxically, an asparagine-to glutamine (N-to-Q) mutation (the side chain of Q is bulkier than of N) also produces a small increase in pore size, possibly because the bulky side chain of the Q residue does not pack well and disrupts channel structure (Wollmuth et al., 1996). N¹-DnsSpm does not permeate wild-type NMDA channels or channels containing an NR1(N616Q) mutation but can easily permeate channels with the NR1(N616G) mutant (Chao et al., 1997). In this study, we found that TB-3-4 can apparently permeate some mutant NMDA channels at extreme negative membrane potentials.

Block by TB-3-4 showed little or no recovery at extreme negative potentials, suggesting that TB-3-4 does not easily permeate wild-type NMDA channels. Some recovery from block was seen at receptors containing the N-to-Q mutation and, in particular, at receptors containing the N-to-G mutation at NR1(N616). The diameter of the largest portion of TB-3-4 was estimated to be 0.6 to 0.65 nm, which is larger than the narrow constriction in wild-type channels (0.55 nm; Wollmuth et al., 1996; Zarei and Dani, 1995) but smaller than the diameter of the naphthalene ring of N¹-DnsSpm (0.8-0.85 nm; Chao et al., 1997). This is consistent with the idea that TB-3-4 can permeate wild-type channels only poorly but can more easily permeate channels with the NR1(N616Q) or NR1(N616G) mutations. The smaller size of TB-3-4 compared with N¹-DnsSpm could also account for the observation that TB-3-4 can permeate mutant channels more easily than N¹-DnsSpm, an effect that is manifest as a more pronounced recovery from block at extreme negative potentials with TB-3-4 than with N¹-DnsSpm (Chao et al., 1997). The effects of NR1(N616) mutations on block and permeation of TB-3-4 appear to be volume specific, influencing permeation of TB-3-4 with only modest effects on the potency of block. However, an N-to-R mutation, which introduces a positive charge at position N616, drastically reduced block by TB-3-4, and it is possible that NR1(N616) normally contributes to part of a binding site for TB-3-4 or that the introduction of a positive arginine residue at this position causes electrostatic repulsion of one or more of the amino groups in TB-3-4.

In conclusion, benzyl-polyamines such as TB-3-4 are a novel class of NMDA channel blocker that show selectivity for NMDA over AMPA channels. Benzyl-polyamines represent new tools to study GluR channels, and these compounds may bind to sites within the channel that are different from the binding sites for other polyamine derivatives, such as the dansyl-polyamines.