SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  • 1
    The effects of neomycin on NMDA-evoked currents in isolated CA1 hippocampal pyramidal neurones were investigated and single channel activity was examined in outside-out patches taken from cultured hippocampal neurones. The effects of neomycin on two combinations of NMDA receptor subunits (NR1a-NR2A and NR1a-NR2B) expressed in human embryonic kidney (HEK293) cells were also studied.
  • 2
    Neomycin (0.01–1 mm) caused a potentiation of NMDA-activated currents in all neurones examined. No evidence of a voltage-dependent depression was observed in whole-cell recordings.
  • 3
    In outside-out patch recordings relatively low concentrations (30 and 100 μm) of neomycin caused a voltage-dependent reduction in single channel current amplitude as well as a large increase in the frequency of channel opening.
  • 4
    In saturating concentrations of glycine, neomycin enhanced NMDA-activated currents and this glycine-independent enhancement was confirmed using recombinant NR1a-NR2B receptors. Neomycin substantially increased the potency of glycine for the receptor by reducing the rate of dissociation of glycine from the receptor. Neomycin demonstrated a glycine-dependent enhancement of currents mediated by the NR1a-NR2A combination of subunits but a paradoxical depression was observed in saturating concentrations of glycine.
  • 5
    Neomycin increased the rate of deactivation of glutamate-activated currents consistent with neomycin causing a reduction in the affinity of the receptor for agonist.
  • 6
    These results indicate that neomycin has multiple and complex effects on NMDA receptors.

Paradoxically, Mg2+ (Paoletti et al. 1995; Wang & MacDonald, 1995) and Ca2+ (Gu & Huang, 1994) both enhance and depress NMDA-evoked responses in cultured and isolated neurones. They do so by increasing the apparent affinity of glycine for its co-agonist site (a glycine-dependent potentiation) and by enhancing responses in the presence of saturating concentrations of glycine (a glycine-independent potentiation). Polyamines such as spermine also show a similar spectrum of effects causing a voltage-dependent block as well as glycine-dependent and glycine-independent enhancement of NMDA-mediated responses (McBain & Mayer, 1994; Johnson, 1996; Williams, 1997). Some of the effects of divalent cations may be mediated through binding to the same sites as polyamines (Paoletti et al. 1995; Wang & MacDonald, 1995).

In cultured hippocampal neurones the effects of polyamines on NMDA-evoked currents were found to be highly variable ranging from a strong enhancement to a profound depression depending upon the particular cell examined (Benveniste & Mayer, 1993). This suggested that polyamines might have different effects depending upon which combinations of NMDA receptor subunits were expressed within a given cell. This, as well as considerable evidence from studies on recombinant NMDA receptors (Johnson, 1996; Williams, 1997), implies that there are multiple binding sites responsible for some or all of the different effects of polyamines (McBain & Mayer, 1994; Johnson, 1996; Williams, 1997).

At this juncture, neither divalent cations nor polyamines have proven to be selective for either blocking or enhancing NMDA-evoked currents. However, it has been reported that the aminoglycoside antibiotic neomycin may be more selective at the site(s) responsible for enhancement versus depression of NMDA receptor function whilst spermine appears to be much less selective (Pullan et al. 1992). If this were the case, neomycin might prove a useful agent for studying the mechanisms of the enhancement of NMDA channel function in the absence of a blockade. It might also provide a way of determining whether or not specific sites on the receptor mediate enhancement versus depression of NMDA-evoked responses. Therefore, the objective of the present study was to determine whether neomycin, spermine and Mg2+ act at the same site to enhance NMDA responses recorded from hippocampal neurones and to determine whether neomycin selectively enhances NMDA-activated currents.

Our previous understanding of the effects of polyamines on native NMDA receptors has come almost entirely from studies of cultured neurones where the effects of polyamines are highly variable (Benveniste & Mayer, 1993). It has also been the practice to lower the extracellular [Ca2+] in order to reduce a Ca2+-dependent inactivation of NMDA-evoked currents (Legendre & Westbrook, 1991; Lerma, 1992). Lowering the extracellular [Ca2+] activates a distinct non-selective cation current that may contaminate NMDA-evoked currents (Xiong et al. 1997). Therefore, we have controlled for the potential involvement of the non-selective cation current, as well as for the potential contribution of Ca2+ inactivation to NMDA receptor desensitization. Previous investigators have also controlled for the effects of lowered [Ca2+] (Benveniste & Mayer, 1993).

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements

Preparation of acutely isolated neurones

CA1 pyramidal neurones were acutely isolated from postnatal rat hippocampus according to the technique described previously (Wang & MacDonald, 1995). Briefly, Wistar rats (1–2 weeks old) were anaesthetized with halothane and killed by decapitation using a guillotine. The entire brain was rapidly removed and rinsed in cold extracellular solution (see below). The hippocampus was surgically isolated and then cut by hand with a razor blade into approximately 400–550 μm thick slices. Subsequently, the slices were incubated in oxygen-bubbled extracellular solution containing 0.5–0.8 mg ml−1 papain (derived from papaya latex, Sigma) at room temperature (20–22°C) for 30–35 min. The slices were then rinsed and kept in papain-free solution until cells were isolated. Prior to the isolation of individual neurones, a single slice was transferred to a plastic 35 mm tissue culture dish. Under a dissecting microscope, the CA1 region was separated from the slice and the remainder of the slice was discarded. The dish was placed in the recording chamber of an inverted phase-contrast microscope where CA1 pyramidal neurones were mechanically isolated by gently teasing the pyramidal cell layer apart using two fine glass probes. The isolated neurones were then allowed to sit for 10–15 min prior to recording. Only neurones which retained their pyramidal shape including a major primary and several secondary dendritic processes were used for recording. Some experiments were also performed upon primary cultured hippocampal neurones grown for approximately 2 weeks as described previously (MacDonald et al. 1989). All animals were killed in accordance with the guidelines of the Medical Research Council of Canada.

Recombinant receptors expressed in HEK293 cells

Recombinant NMDA receptor subunits (NR1a, NR2A and NR2B) were transiently transfected into human embryonic kidney (HEK293) cells (American Type Culture Collection). Constructs (cDNAs) for NMDA receptors were provided by Dr S. Nakanishi (Kyoto University, Kyoto, Japan) or Dr P. Seeburg (Max Planck Institute for Medical Research, Heidelberg, Germany) (cloned in pcDNA3). HEK293 cells were maintained in a mixture of minimal essential medium with Earle's salts and L-glutamine (Gibco BRL) with 10 % fetal bovine serum (Gibco) in a humidified atmosphere containing 5 % CO2. Combinations of NR1a-NR2A and NR1a-NR2B (1 : 3 ratio with a total of 3 μg cDNA per 35 mm culture dish) were transfected using the Perfect Lipid Method (Invitrogen). At 24 h prior to transfection, cells were plated to a density of about 106 cells per 35 mm dish. Following transfection the cells were cultured in the presence of 1 mm dl-2-amino-5-phosphono-valeric acid. Recordings were made 24–48 h after transfection.

Solutions, drugs and the perfusion systems

The extracellular solution contained (mm): NaCl, 140; CaCl2, 1.3 or 2, or 0.2; KCl, 5.0; Hepes, 25; glucose, 33. The pH was adjusted to 7.4 with NaOH and the osmolarity of the solution ranged between 320–335 mosmol l−1. During the recording, various concentrations of glycine (Sigma), NMDA (Sigma), neomycin (Sigma), spermine (Sigma), MgCl2 (Aldrich Chemical Co.) and CaCl2 (BDH) were added to give the concentrations indicated in the text and figure legends. In all experiments the NMDA-containing solutions contained the same concentration of Ca2+ as found in the control extracellular solution, unless otherwise stated, in order to prevent contamination of NMDA-evoked currents with a non-selective cation current (Xiong et al. 1997). No responses to glycine alone (leqslant R: less-than-or-eq, slant 10 μm) were detected in any of the neurones tested demonstrating that few if any glycine receptors were expressed in these neurones.

Whole-cell patch electrodes were filled (unless otherwise specified) with (mm): CsF, 140; CsOH, 35; Hepes, 10; MgCl2, 2; CaCl2, 1; EGTA, 11; tetraethylammonium (TEA), 2; Na2ATP or MgATP, 4. The pH was adjusted to 7.3 using CsOH or NaOH and the osmolarity was 300 mosmol l−1. In some experiments EGTA was replaced with 11 or 20 mm BAPTA with or without Ca2+ added to the solution.

A multibarrelled perfusion system was employed to achieve a rapid exchange of solutions. Briefly, the perfusion barrels consisted of three square capillary glass tubes, each of which was connected via capillary tubes to a separate reservoir of perfusion fluid. Barrels were mounted on a solenoid-driven mini-manipulator (SF 77 Perfusion Fast Step, Warner Instruments Corp.), which was controlled by the acquisition program (pCLAMP 6; Axon Instruments Inc.). Isolated CA1 pyramidal neurones were first patched and then lifted into the outflow of the control barrel. In order to determine the rapidity of the change of the solution, the block of the NMDA currents by Mg2+ (1 mm) was examined. The maximum block of whole-cell NMDA-evoked currents by Mg2+ occurred within 12 to 45 ms (n= 4).

Patch clamp recordings

Patch clamp recordings were made in the whole-cell or in the outside-out patch configurations. All recordings were performed at room temperature. Recording electrodes with resistances of 3–5 MΩ were constructed from thin-walled borosilicate glass (1.5 mm diameter; WPI) using a two-stage puller (PP83; Narishige). Using an Axopatch-1B or Axopatch 200 amplifier (Axon Instruments Inc.), whole-cell recording data were digitized, filtered (2 kHz) and acquired on-line using the pCLAMP 6 program. Population data were expressed as means ±s.e.m. and the ANOVA test (two-way) or Student's paired t test was employed when appropriate to examine the statistical significance of the differences between groups of data. For outside-out membrane patch recordings an Axopatch-1B amplifier was employed and single channel events (filtered at 2 kHz) were first collected on videotape with a digital data recorder (VR-10; Instrutech Corp.) and later played back and acquired using the pCLAMP 6 program.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements

Neomycin-induced enhancement of NMDA-evoked currents in isolated CA1 hippocampal neurones

Neomycin (0.01–1.0 mm), in the continuous presence of 1.3 mm extracellular Ca2+, exclusively enhanced NMDA-evoked currents recorded in the presence of near-saturating concentrations of glycine (3 μm; see Fig. 8) in all isolated neurones examined (holding potential, −60 mV; n= 310; Fig. 1A). During relatively short periods of application both the initial peak of the NMDA-evoked current (Ip) and the quasi-steady-state current (Iss) were enhanced by the co-application of neomycin (Fig. 1A). These observations are consistent with the possibility that neomycin acts as a selective activator at the polyamine site mediating glycine-independent potentiation of NMDA receptor function. In a proportion of the neurones (about 11 %), application of 1 mm neomycin also evoked small outward currents. An example of such an outward current is shown in Fig. 1B (arrow). We attribute this outward current to the rapid blockade of a residual non-selective cation current that could still be detected in the presence of 1.3 mm Ca2+ (Xiong et al. 1997). In subsequent experiments only cells demonstrating little or no outward current responses to 1 mm neomycin (1.3 mm extracellular Ca2+) and leak currents of less than 100 pA (−60 mV) were studied.

image

Figure 8. Neomycin increases the apparent affinity of the NMDA receptor for glycine

A, the concentration of glycine was varied (0.01–10 μm) in the presence of a constant concentration of NMDA (100 μm; −60 mV). These recordings included experiments in which the background contamination of glycine was probably greater than for those experiments shown in Fig. 7A and B. The applications were then repeated in the presence of 100 μm neomycin. Neomycin potentiated these currents and, as shown in B, shifted the concentration-response curves for glycine to the left (Ip, control: EC50= 117 ± 32 μm, nH= 0.96 ± 0.08; Ip plus 100 μm neomycin: EC50= 48 ± 9 μm, nH= 0.88 ± 0.07). Normalized peak currents were obtained from 7 cells in the absence (Control) and the presence of neomycin. Each current was normalized to that induced by the application of 10 μm glycine. C, example traces showing the decay of currents following cessation of the application of glycine (300 nm) evoked in the absence (Control) or presence (+ Neo) of neomycin (100 μm). Data points were fitted using a first-order exponential equation. The value of the time constant of the offset of glycine was increased in the presence of neomycin (control τoff= 288 ms; neomycin τoff= 861 ms). In contrast, the time constant of onset (τon) was slightly decreased in the presence of neomycin (not shown). D, plots of 1/τonversus the concentration of glycine in the presence (n= 6) or absence (Control; n= 8) of neomycin. These relationships were near-linear and the slope of these curves was used to obtain an estimate of the forward rate of the glycine response (without neomycin: k+= 1.30 × 10−7 M−1 s−1; with neomycin: k+= 1.49 × 10−7 M−1 s−1). These values were not significantly different from each other. E, plots of 1/τoffversus the concentration of glycine in the absence (Control) and presence of neomycin. The y-intercepts of these curves yield estimates of the reverse rate of the glycine response (without neomycin: k-= 2.66 s−1; with neomycin: k-= 1.36 s−1). Neomycin significantly reduced the apparent rate of dissociation of glycine (P < 0.001, two-way ANOVA).

Download figure to PowerPoint

image

Figure 1. Enhancement by neomycin of NMDA-induced currents in acutely isolated hippocampal CA1 pyramidal neurones in the presence of a near-saturating concentration of glycine (3 μm)

A, application of neomycin (1 mm) together with, or prior to and during, application of NMDA (100 μm) enhanced NMDA-evoked currents in an isolated neurone. All three responses are shown superimposed in the recording on the right. B, in some of the cells examined, application of neomycin induced small outward currents (indicated by the arrow; dashed line indicates zero current level) which we attribute to a depression of the non-selective cation current still detectable in these neurones in an extracellular concentration of Ca2+ of 1.3 mm. Superimposed currents are also shown on the right. C, in the presence of a near-saturating concentration of glycine (3 μm), prolonged application of NMDA (100 μm) could induce substantial glycine-independent desensitization. Co-application of neomycin in this example greatly enhanced the amplitude of the initial peak (Ip) of the response to NMDA but the amplitude of the steady-state current (Iss) was much less affected. Thus, in this particular neurone the extent of desensitization was increased (Iss/Ip was decreased). D, initial portion of the two currents illustrated in C, shown normalized and superimposed, showing that neomycin (0.5 mm) paradoxically reduced the rate of desensitization.

Download figure to PowerPoint

In the present study the degree of desensitization (Iss/Ip) with a fixed concentration of NMDA (i.e. 100 μm) varied substantially between individual neurones and depended in part upon the duration of the application (compare Fig. 1A with C). Some of this variability probably reflects the different degrees of glycine-sensitive desensitization. We therefore waited for approximately 10 min following formation of the patch in order to permit stabilization of this process (Sather et al. 1991). In cultured neurones it has previously been reported that spermine decreases glycine-dependent desensitization (increases Iss/Ip) and reduces the rate of onset of desensitization (Lerma, 1992; Benveniste & Mayer, 1993). In contrast, however, in the present study, neomycin and spermine enhanced glycine-independent desensitization (Fig. 1C and see Fig. 3) in isolated neurones by enhancing the peak of the current relative to the steady state (decreasing Iss/Ip). In spite of this enhancement of desensitization, the time constant of desensitization was actually increased during the co-application of neomycin (desensitization rate slowed; Fig. 1D) and it simply took longer to reach the steady state in the presence than in the absence of neomycin. We then considered whether Ca2+-dependent inactivation may have contributed to the amplitude of the steady-state currents in isolated neurones.

image

Figure 3. Concentration-dependent enhancement of NMDA-evoked currents by neomycin in a near-saturating concentration of glycine (3 μm) and depression of these currents by spermine at relatively low and high concentrations

A illustrates the enhancement observed with increasing concentrations of neomycin (−60 mV). Peak current (Ip) was enhanced to a greater degree than steady-state current (Iss) (P < 0.001, two-way ANOVA). B, the relative enhancement of NMDA-activated current ((Ineomycin/Icontrol) - 1) plotted against the concentration of neomycin. Concentration-response relationships were fitted using a logistic equation. C, spermine also induced a concentration-dependent enhancement of NMDA-evoked current. D, concentration-response relationships showing, however, that in some cells a small depression, particularly of Iss, was observed following application of low concentrations of spermine (10 or 50 μm), and that currents were also consistently depressed at much higher concentrations (i.e. 3 mm). As a consequence, the enhancement of Ipversus Iss was significantly different at the highest concentration of spermine (P < 0.05). The potency of spermine (Ip: EC50= 181 ± 14 μm; Iss: EC50= 162 ± 17 μm; n= 7) to enhance NMDA-activated currents was slightly lower than that observed for neomycin (Ip: EC50= 138 ± 11 μm, nH= 1.29 ± 0.15, n= 8; Iss: 142 ± 17 μm, nH= 1.42 ± 0.24, n= 8). E and F, concentration-response relationships for neomycin in a different group of neurones held at +40 mV in order to test for the presence of a voltage-dependent block of the currents by neomycin. Neomycin caused a similar enhancement of NMDA-evoked currents at this holding potential (Ip: EC50= 192 ± 28 μm, Iss: EC50= 207 ± 47 μm; n= 6).

Download figure to PowerPoint

Enhancement of desensitization and possible contribution of Ca2+-dependent inactivation

The method most often employed to reduce Ca2+-dependent inactivation is to simultaneously reduce the extracellular [Ca2+] during application of NMDA. Any prolonged exposure to a low extracellular [Ca2+] is thought to lead to a large leakage current indicative of cell damage. Indeed, reducing the [Ca2+] from 1.3 to 0.2 mm evoked slowly developing inward currents in almost all of the acutely isolated neurones we examined (Fig. 2A). However, this inward current was mediated by the activation of a specific non-selective cation channel that we recently described (Xiong et al. 1997). Furthermore, this non-selective cation current was also sensitive to neomycin and other polyamines (Fig. 2A). Therefore, it was impractical to simultaneously lower [Ca2+] and apply NMDA as both NMDA-evoked and non-selective cation currents would be superimposed (Xiong et al. 1997).

image

Figure 2. Calcium-dependent inactivation and the non-selective cation current in isolated neurones

A shows the non-selective cation current activated in an isolated neurone by rapid reduction of the extracellular Ca2+ (Ca2+o) concentration from 1.3 to 0.2 mm (- Neo). In the presence of neomycin (+ Neo, 10 μm) this current was substantially depressed. Hence, we did not change the Ca2+ concentration during application of NMDA. Although the absolute steady-state value of NMDA currents may have been influenced by Ca2+-dependent inactivation we found that manoeuvres designed to minimize the influx of Ca2+ into the neurones failed to prevent the strong glycine-independent desensitization observed in these neurones. B, example of a recording of an NMDA (100 μm)-evoked current recorded in the presence of 100 μm neomycin with [Ca2+] maintained at 1.3 mm. C, when [Ca2+] was simultaneously reduced to 0.2 mm during the application of NMDA there was a proportionately greater increase in the size of the steady-state current versus the peak (Iss/Ip increased from 0.12 to 0.15). D, however, when [Ca2+] was reduced prior to the application of NMDA there was no change in the degree of desensitization. This is also illustrated in E where the normalized currents in B and D are shown superimposed. F and G, changes in the extracellular concentration of Ca2+ ranging from 0.2 to 5 mm had no effect on the degree of desensitization of NMDA-evoked currents in isolated neurones. In G (right panel), superimposed, normalized currents are also illustrated. All recordings in this figure were done at a holding potential of −60 mV.

Download figure to PowerPoint

In order to determine whether or not Ca2+-dependent inactivation was depressing steady-state NMDA-evoked currents we recorded under conditions that would minimize the non-selective cation current. We accomplished this (Fig. 2) by recording responses to NMDA in the continuous presence of neomycin (100 μm) as it is also a potent antagonist of the non-selective cation channel (IC50= 4 μm). Note that despite its high potency neomycin cannot suppress the non-selective cation current (Xiong et al. 1997). We used neomycin in this manner because there are no blockers of the non-selective cation current that do not also block NMDA channels (e.g. gadolinium potently blocks non-selective cation and NMDA channels).

Simultaneous reduction of extracellular [Ca2+] and application of NMDA appeared to reduce the degree of desensitization (Fig. 2B and C). However, when [Ca2+] was reduced prior to the application of NMDA a residual (not blocked by neomycin) inward non-selective cation current was observed (Fig. 2D) and this current accounted entirely for the apparent reduction in the degree of desensitization (Fig. 2E). Furthermore, we also examined similar NMDA responses over a wider range of extracellular Ca2+ concentrations and found no correlation between [Ca2+] and the degree of desensitization in isolated neurones (Fig. 2F and G). The degree of desensitization was also similar when cells were held at positive potentials or when the patch pipettes were filled with high concentrations (20 mm) of the more rapid Ca2+ buffer BAPTA. Therefore, we conclude that Ca2+ inactivation played little if any role in determining the amplitude of the steady-state currents we observed in isolated neurones and, unlike cultured neurones (Legendre et al. 1993), EGTA within the patch pipette appeared to be sufficient to prevent Ca2+-dependent inactivation.

Multiple effects of polyamines: comparison of neomycin and spermine effects on NMDA-evoked currents in high concentrations of glycine

We then compared the concentration-response relationship for application of neomycin with that for spermine on NMDA-evoked currents recorded in a relatively high concentration of glycine (3 μm). These relationships were determined by applying various concentrations of polyamine during co-application of a fixed concentration of NMDA (100 μm) whilst the membrane potential was set at −60 mV. Neomycin enhanced both Ip (EC50= 138 ± 11 μm, Hill coefficient (nH) = 1.29 ± 0.15, n= 8; 94 ± 11 % enhancement at 3 mm, −60 mV) and Iss (EC50= 142 ± 17 μm, nH= 1.42 ± 0.24, n= 8; 63 ± 14 % enhancement) (Fig. 3A and B) although the enhancement of the peak current was significantly greater (P < 0.001, two-way ANOVA).

Spermine, as in the case of cultured spinal cord neurones (Lerma, 1992), also enhanced NMDA-evoked currents although application of both the lowest and the highest concentrations of spermine was associated with some evidence of a depression of steady-state NMDA-activated currents (Fig. 3D). Peak currents were also enhanced to a greater degree than steady-state currents (Fig. 3C and D). Under these recording conditions (high glycine, −60 mV), spermine demonstrated a somewhat lower potency and efficacy for the enhancement of NMDA-evoked currents than did neomycin (spermine, Ip: EC50= 181 ± 14 μm, nH= 1.5 ± 0.2, n= 7; 66 ± 10 % enhancement at 3 mm, −60 mV; Iss: EC50= 162 ± 17 μm, nH= 1.7 ± 0.3, n= 7; 30 ± 9 % enhancement). These calculated values are only rough estimates as a spermine-induced depression was observed at this holding potential. Both neomycin and spermine increased the time constants of desensitization (i.e. slowed the rate of desensitization; τcontrol= 609 ± 80 ms; τneomycin= 909 ± 82 ms (1 mm); n= 8; P < 0.05; Student's t test).

Neomycin did not overtly block NMDA-evoked currents at −60 mV but we nevertheless repeated the concentration- response relationship at a holding potential of +40 mV in order to minimize any potential voltage-dependent block which might have been present at −60 mV (Fig. 3E and F). Depolarization to positive membrane potentials was difficult to maintain due to the increase in outward or leak currents in isolated neurones. However, in six stable recordings at +40 mV there was little evidence that neomycin caused a proportionately greater enhancement of NMDA-activated outward currents (Ip: EC50= 192 ± 28 μm, nH= 1.35 ± 0.22; Iss: EC50= 207 ± 47 μm, nH= 2.17 ± 0.24; n= 6).

Presence of a voltage-dependent block

Unlike neomycin, there was clear evidence that spermine was blocking NMDA-evoked currents at −60 mV. Spermine also causes a substantial and strongly voltage-dependent block of NMDA-activated currents in cultured neurones (Benveniste & Mayer, 1993). Therefore we compared the effects of 1 mm neomycin to those evoked by the application of the same concentration of spermine over a range of different holding potentials. A fixed concentration of NMDA (100 μm) was applied at holding potentials ranging from −100 to +60 mV (Fig. 4). The change in the apparent steady-state currents was measured at each of the various holding potentials. Neomycin enhanced NMDA-evoked currents at each of the holding potentials tested (Fig. 4A and B). The current-voltage relationships were near-linear (Fig. 4B) and the percentage enhancement induced by neomycin was not obviously voltage dependent (Fig. 4B, inset; n= 6). In contrast, spermine clearly caused a voltage-dependent depression of the NMDA-evoked currents (Fig. 4C and D).

image

Figure 4. Enhancement induced by a relatively high concentration (1 mm) of neomycin demonstrated little dependence upon holding potential whilst the same concentration of spermine revealed a voltage-dependent depression of NMDA-activated currents

A, representative traces showing superimposed whole-cell currents evoked by application of NMDA (100 μm, 3 μm glycine) and 1 mm neomycin at holding potentials ranging from −100 to +60 mV in an acutely isolated CA1 pyramidal neurone. B, current-voltage relationships of steady-state NMDA-evoked currents recorded from a second isolated neurone, in the absence (Control) and presence of neomycin. The relative enhancement of current by neomycin was about the same regardless of the holding potential. This is further illustrated in the inset which shows the normalized values for 6 different neurones as a function of holding potential. C, representative traces from another isolated neurone under the same recording conditions as A except that spermine (1 mm) was applied. At negative holding potentials spermine clearly depressed Iss. Cessation of the spermine application was also associated with a prominent ‘rebound’ current (see inset). D, current-voltage relationships for this neurone demonstrating the presence of a voltage-dependent depression of the currents at negative membrane potential values.

Download figure to PowerPoint

At negative membrane potential values (i.e. −100 mV) the superimposed application of spermine was associated with a depression followed by a characteristic rebound of inward current following cessation of the application (Fig. 4C, inset). Such a rebound current is often interpreted to represent blocked channels that must re-enter a conducting state prior to closing but in this case it was more probably a consequence of the slower recovery kinetics of the potentiation versus the depressant effects of spermine (Benveniste & Mayer, 1993).

Although neomycin did not obviously depress NMDA-evoked currents, we anticipated that if the concentration of neomycin was increased further a block might become apparent. Indeed, when the concentration of neomycin was increased to 5 mm the current-voltage relationship demonstrated a clear but small outward rectification at negative holding potential values (not shown). Inspection of individual responses also demonstrated the presence of a very small inward rebound current at negative potentials following the termination of application of this concentration of neomycin. Nevertheless, neomycin enhanced NMDA-evoked currents at all holding potentials examined.

Do polyamines and Mg2+ interact at the same site?

The experiments described so far were performed in the absence of added Mg2+ in order to avoid the potent voltage-dependent block of NMDA currents by this divalent cation. However, Mg2+ may interact at one or more of the same sites as polyamines and protons (Paoletti et al. 1995). We therefore examined the potential interaction of Mg2+ with the glycine-independent enhancement of NMDA-evoked currents by neomycin.

The holding potential was maintained at a positive value (+50 mV) and neomycin concentration-response relationships were examined. Several fixed concentrations of Mg2+ were co-applied with NMDA and neomycin (Fig. 5A). At concentrations of neomycin below 300 μm the effects of Mg2+ were clearly additive with those of neomycin; however, at concentrations above this value Mg2+ depressed the enhancement induced by neomycin (Fig. 5B) suggesting that Mg2+ acts at the same site as neomycin. Similar results were observed with an interaction between neomycin and spermine. At a fixed concentration of neomycin (100 μm) spermine potentiations were additive below 1 mm but were occluded at this concentration (Fig. 5C).

image

Figure 5. Neomycin, Mg2+ and spermine probably interact at the same site to induce a glycine-independent enhancement of NMDA-evoked currents

A, example traces showing the interaction between neomycin and Mg2+ on currents activated by NMDA (100 μm) in the presence of a near-saturating concentration of glycine (3 μm). The neurone was held at +50 mV. The left panel shows the responses to application of neomycin (10–1000 μm) superimposed upon that to NMDA itself. The right panel shows the same responses repeated in the presence of Mg2+ (3 mm). Note that Mg2+ further enhanced the responses at low concentrations of neomycin but depressed those at higher concentrations. B, effects of three concentrations of Mg2+ (0 (Control), 3 and 8 mm) on the proportionate enhancement of steady-state NMDA-activated current induced by various concentrations of neomycin (0 Mg2+: nH= 1.11 ± 0.12; 3 mm Mg2+: nH= 0.57 ± 0.14; 8 mm Mg2+: nH= 0.29 ± 0.06). Mg2+ clearly was additive at low concentrations of neomycin and depressed the enhancement seen at high concentrations. C illustrates another type of interaction, namely the actions of a fixed concentration of neomycin (100 μm) imposed upon various concentrations of spermine (0 neomycin (Control), nH= 0.83 ± 0.18; 100 μm neomycin, nH= 0.40 ± 0.12). At high concentrations of spermine the additive effects of neomycin were lost.

Download figure to PowerPoint

Effects of neomycin on agonist potency

One of the other effects of spermine is to cause a small decrease in the apparent affinity of recombinant NR1a-NR2B receptors for glutamate and NMDA (Williams, 1994). We therefore examined whether or not neomycin could alter agonist potency. NMDA concentration-response relationships were generated in the presence of near-saturating concentrations of glycine (3 μm) but with or without 100 μm neomycin (Fig. 6A and B). The peak and steady-state currents in response to applications of NMDA were measured. However, we were unable to detect any change in the potency using this approach (Fig. 6B), probably because NMDA is a relatively low-affinity agonist for its receptor and because its rate of dissociation from the receptor is rapid compared with that for a high-affinity agonist such as glutamate (Sather et al. 1992).

image

Figure 6. Neomycin may decrease the apparent affinity of the NMDA receptor for its agonist

A, example traces showing currents evoked by various concentrations of NMDA (10–1000 μm) in the absence or presence of neomycin (100 μm). Currents were recorded in the presence of 3 μm glycine in an isolated neurone held at −60 mV. B, no change in the EC50 values for either the peak (Ip) or steady-state (Iss) currents was observed following application of neomycin (see text). In each case, currents were normalized to the response to 1 mm NMDA (n= 11). C, top: 2 example traces showing the currents activated by application of 300 μm glutamate (−60 mV in the presence of 1 μm glycine) before (- Neo) and following (+ Neo) application of neomycin (300 μm). Bottom: the deactivation phase of the response is illustrated using superimposed and normalized currents. Single exponential fits to the traces demonstrated that the time constant of decay (τoff) of the glutamate-activated current was reduced by neomycin. D, summary data from 9 neurones. Neomycin reduced the time constant of deactivation from 381 ± 32 to 268 ± 22 ms (* P < 0.50), consistent with a possible reduction in the affinity of the receptor for glutamate.

Download figure to PowerPoint

Therefore, we examined the time course of deactivation of glutamate-evoked currents in the presence and absence of neomycin. Neomycin consistently decreased the time constants of deactivation of these currents (Fig. 6C and D). This result is consistent with a decrease in the affinity of the receptor for glutamate but it could also reflect changes in the rates of transition to and from various states of the receptor (e.g. desensitized states).

Neomycin enhances the potency of glycine for the receptor

The next aspect considered was the effect of neomycin on the potency of glycine for the receptor. For example, Mg2+ causes a greater enhancement of NMDA-evoked currents when the concentration of glycine is low than at saturating concentrations (Wang & MacDonald, 1995). Therefore, we compared neomycin concentration-response relationships in a low concentration of added glycine (300 nm) versus a saturating concentration (10 μm; Fig. 7). As anticipated, the degree of enhancement by neomycin was far greater at the low (Fig. 7A and C) than at the high (Fig. 7B and D) concentration of glycine. A similar effect was seen with cells held at +40 mV (not shown). Furthermore, in 300 nm glycine the enhancement of Ip was proportionately much greater than that of Iss (P < 0.001, two-way ANOVA), although Ip was also significantly more enhanced than Iss in the presence of 10 μm glycine (P < 0.001, two-way ANOVA). The EC50 values for Ip and Iss were similar (Fig. 7). The potency of neomycin was also greater at the low concentration of glycine (Fig. 7) than at the saturating concentration (300 nm glycine, Ip: EC50= 53 ± 13 μm, Iss: EC50= 61 ± 11 μm, n= 7; 10 μm glycine, Ip: EC50= 106 ± 32 μm; Iss: EC50= 183 ± 29 μm, n= 7; P < 0.05, Student's t test) indicating that neomycin is apparently more potent in inducing a glycine-dependent enhancement. When these experiments were repeated with a high concentration of BAPTA in the patch pipette, in order to avoid any possible Ca2+-dependent inactivation, the selective effect of neomycin on the peak versus steady-state currents was not altered (not shown).

image

Figure 7. The enhancement by neomycin, and its effects on desensitization, are strongly dependent upon the concentration of glycine

A, in the presence of 300 nm added glycine, NMDA (100 μm)-mediated currents were substantially enhanced by neomycin (+ Neo, 100 μm). Peak currents were clearly potentiated more than steady-state currents. This isolated cell was held at −60 mV. B, in the same cell, in the presence of a saturating concentration of glycine (10 μm), neomycin also enhanced the current (glycine-independent enhancement) but the proportionate increase in the peak and steady-state currents was not as great as that seen at the lower concentration of glycine. C, relative enhancement ((Ineomycin/Icontrol) - 1) of Ip and Iss plotted against the concentration of neomycin (300 nm glycine always present). The enhancement of the peak current was significantly greater than that of the steady-state current (P < 0.001, two-way ANOVA). The EC50 values for the enhancement of the peak and steady-state currents were 53 ± 13 and 61 ± 11 μm, respectively (n= 7) (Ip: nH= 0.92 ± 0.18; Iss: nH= 1.04 ± 0.27). D, in contrast, at a saturating concentration of glycine (10 μm) the enhancement of Ip and Iss was smaller (Ip: EC50= 106 ± 32 μm, nH= 1.27 ± 0.29; Iss: EC50= 183 ± 29 μm, nH= 1.73 ± 0.33; n= 7). Nevertheless, the proportionate increase in Ip was still greater than that of Iss (P < 0.001, two-way ANOVA).

Download figure to PowerPoint

We then examined concentration-response relationships for glycine itself. Increasing concentrations of glycine enhanced both the peak and steady-state NMDA-evoked currents (Fig. 8A). However, the steady-state current saturated at lower concentrations of glycine than did the peak current. Co-application of neomycin potentiated the effects of glycine and increased the potency of glycine for the receptor by a factor of about 2.4 (Ip without neomycin: EC50= 117 ± 32 μm, n= 7; with neomycin: EC50= 48 ± 9 μm, n= 7; Fig. 8B; P < 0.05, Student's t test). Neomycin also consistently decreased the rate of recovery from the responses to glycine (increased the time constant for the offset of glycine, see Fig. 8). This observation is consistent with neomycin increasing the potency for glycine largely by decreasing its rate of dissociation from its binding site (Fig. 8E).

Proton-dependent potentiation

Neomycin and spermine enhanced NMDA-activated currents in isolated neurones in the presence of saturating concentrations of glycine indicating the presence of a glycine-independent mechanism of potentiation. The potentiation by spermine is probably mediated through a relief of the proton block of NMDA receptors and at low pH values the potentiation is proportionately greater than at high values (Traynelis et al. 1995). This suggests that the potentiation by neomycin should also be dependent upon the degree of proton block. We therefore examined the concentration-response relationships for the effects of neomycin at two additional values of pH (6.8 and 8.0). The enhancement by neomycin was dependent upon extracellular pH (Fig. 9A–C) with the greatest potentiation observed in higher concentrations of protons. The potency of neomycin was also dependent upon pH (Ip, pH 6.8: EC50= 167 ± 19 μm, nH= 1.29; pH 8.0: EC50= 235 ± 24 μm, nH= 1.32; n= 6; P < 0.001, two-way ANOVA). One possibility is that neomycin may alter the sensitivity of the receptors for protons, as previously suggested (Kashiwagi et al. 1996; Gallagher et al. 1997). We therefore examined the effects of pH upon NMDA-activated currents in the absence or presence of neomycin (Fig. 9D–F). Neomycin at a concentration of 1 mm shifted the pH versus peak response relationship to the left by 0.8 pH units demonstrating that this polyamine substantially modifies the sensitivity of the channels to protons.

image

Figure 9. Neomycin-induced enhancement of NMDA-evoked currents depends upon extracellular pH and neomycin shifted the sensitivity to pH

A-C, concentration-response relationships for the effect of neomycin on responses to NMDA (100 μm, 3 μm glycine) were recorded from isolated neurones at three different pH values (6.8, 7.4, 8.0). A, sample recordings of responses at a pH of 6.8. B, similar currents recorded at pH 8.0. All cells were held at +40 mV. C, concentration-response relationships for neomycin for a series of cells examined at each pH value. The degree of potentiation by neomycin was dependent upon the concentration of protons. D and E, the response to an application of NMDA (100 μm, 10 μm glycine) was repeated for a series of different values of pH. Example recordings from a single neurone are shown in the absence (D) and presence (E) of 1 mm neomycin. F, similar results from 5 neurones plotted as the peak current normalized to the maximal peak current versus pH. In the presence of neomycin the EC50 value was shifted by 0.8 pH units to the left (control: EC50= 7.0 ± 0.02 pH units; with neomycin: EC50= 6.2 ± 0.05 pH units; P < 0.001, two-way ANOVA).

Download figure to PowerPoint

Observations of neomycin effects on single channel activity

The major actions of neomycin were to enhance NMDA-evoked currents (glycine dependent and glycine independent) with only a weak voltage-dependent blockade discernible at particularly high concentrations of neomycin (e.g. 5 mm). Relative to spermine and Mg2+, the apparent affinity of neomycin for the blocking site appeared to be almost negligible. This suggests that neomycin might indeed be considered a selective activator of the polyamine site(s) that mediates the enhancing actions on NMDA channels.

We examined this question further by looking at the effects of neomycin on NMDA channel activity recorded from outside-out patches taken from cultured hippocampal neurones, as we have found that it is difficult to make such patches from acutely isolated neurones. We therefore confirmed that relatively low concentrations of neomycin (30 or 100 μm) enhanced currents in cultured hippocampal neurones and that the enhancement was not voltage dependent in whole-cell recordings (not shown). In outside-out patch experiments, channel activity was assessed in low concentrations of glycine (300 nm) and patches were selected for relatively infrequent channel events. Neomycin was included at concentrations near or below the EC50 values calculated from whole-cell recordings.

Neomycin at a concentration as low as 30 μm had the dramatic effect of increasing the frequency of channel opening (Fig. 10A) and revealed multiple openings indicative of the presence of more than a single channel in each patch. However, along with this increase in channel open frequency, the amplitude of individual channel currents was reduced. In four selected patches, which initially demonstrated a low frequency of activity and which were held at −60 mV, neomycin (100 μm) decreased single channel currents by about 30 % (control, 2.12 ± 0.08 pA; neomycin, 1.48 ± 0.17 pA; Fig. 10). Individual openings appeared to ‘flicker’ between an open and a non-conducting state suggestive of a rapid open channel block (Fig. 10A). It was not possible to resolve individual open times under the conditions of the present recordings and no further analysis was attempted. Depolarization of the patches to +40 mV reduced the apparent ‘flickering’ of the channels and greatly diminished the reduction in mean channel amplitude. Thus, these recordings clearly demonstrated the presence of a voltage-dependent blockade of NMDA channel activity with concentrations of neomycin at and below the EC50 for enhancement. At positive potentials we also detected a small reduction in single channel currents in the absence of any apparent channel ‘flickering’ (example from one patch; control, 2.04 pA; 100 μm neomycin, 1.80 pA; +40 mV).

image

Figure 10. Neomycin reduces single channel conductance and enhances the probability of opening of NMDA channels recorded in outside-out patches taken from cultured hippocampal neurones

A, representative traces of channel activity (−60 mV patch potential; 5 μm NMDA) are shown in the absence and presence of two concentrations of neomycin (30 and 100 μm). Even though the frequency of channel activity was increased in the presence of neomycin, channels also appeared to ‘flicker’ between open and non-conducting states. B shows the same patch held at +40 mV, and demonstrates that the ‘flickering’ was diminished. C, amplitude histograms for channel activity in the absence (Control) or presence of 100 μm neomycin, with the patch held at −60 mV. Note the increase in the number of peaks in the presence of neomycin. The probability of channel opening (Po) was also increased by neomycin (without neomycin (Control): Po= 0.08 ± 0.03; with neomycin: Po= 0.28 ± 0.05; n= 4), but the value of single channel current was reduced (control: 2.12 ± 0.08 pA; with neomycin: 1.48 ± 0.17 pA; P < 0.05; n= 4). D, similar plots to those in C, with the same patch held at +40 mV. The peaks demarcate the multiple channels present in the patch.

Download figure to PowerPoint

Effects of neomycin on recombinant NMDA receptors

Polyamines both enhance and block responses of recombinant NMDA receptors expressed in Xenopus oocytes (Williams, 1997). NMDA channels demonstrate glycine-dependent and/or glycine-independent forms of potentiation in response to spermine depending upon the identity of the co-expressed NR2 subunit. NR1a-NR2B channels are reported to possess a glycine-independent component whilst those formed from the NR1a-NR2A combination do not (Williams, 1994). We therefore examined whether or not a similar relationship could be observed for neomycin. A concentration of neomycin (1 mm) which caused a near-maximal enhancement of NMDA-evoked currents in isolated hippocampal neurones was employed.

At this concentration, neomycin enhanced responses mediated by both combinations of subunits. It should be noted that the NR1a-NR2B subunit combination has a much higher affinity for glycine (0.2 μm) than does the NR1a-NR2A channel (2.1 μm) (Kutsuwada et al. 1992) and we did not examine responses of the NR1a-NR2B combination to concentrations of glycine below its EC50 value.

Glycine (0.3–10 μm) substantially enhanced the NMDA-evoked currents of the NR1a-NR2A combination (Fig. 11A and B). However, as the concentration of glycine was increased the enhancement of the currents by neomycin was proportionately less demonstrating the presence of a glycine-dependent potentiation. Surprisingly, at a concentration of glycine of 10 μm the enhancement was replaced by a prominent depression of the amplitude of NMDA-evoked currents (Fig. 11A and B). Note that the absolute value of current in the presence of neomycin plus 10 μm glycine was actually smaller than that recorded in the lower concentrations of glycine suggesting that neomycin had directly inhibited the effects of high concentrations of glycine. Neither the potentiation nor the depression induced by neomycin were strongly dependent upon holding potential (Fig. 11B).

image

Figure 11. Neomycin causes a glycine-dependent enhancement of NMDA-evoked currents in HEK293 cells expressing NR1a-NR2A subunits as well as a glycine-independent enhancement of the NR1a-NR2B combination

A, various concentrations of glycine were applied to a neurone expressing the NR1a-NR2A receptor. In low concentrations of glycine, application of neomycin (1 mm) increased NMDA (100 μm)-evoked currents by a factor of about three. With increasing concentrations of glycine, the proportionate enhancement was reduced. In near-saturating concentrations, neomycin inhibited these currents. B, the voltage dependence of the effect of neomycin was examined in a low (1 μm) and high (10 μm) concentration of glycine. Both the enhancement in low glycine and the depression in high concentrations of glycine were similar at hyperpolarized and depolarized holding potential values (n= 6 for each concentration of glycine; −60 to +60 mV). C, similar recordings in a cell expressing the NR1a-NR2B subunit combination. Note that the lowest concentration of glycine used (0.3 μm) is above the EC50 value for the co-activation of this receptor (0.2 μm, see text). Neomycin (1 mm) enhanced NMDA-evoked currents to about the same absolute value regardless of the concentration of glycine. D, little voltage dependence was observed for the effects of neomycin on this subunit combination (n= 6; −60 to +40 mV; [glycine], 10 μm). E, relative enhancement (Ineomycin/Icontrol) for responses of cells expressing the NR1a-NR2A combination (n= 4–6). The enhancement by neomycin (1 mm) decreased and was replaced by depression at near-saturating concentrations of glycine. F, same plot as in E for responses of the NR1a-NR2B subunit combination (n= 4–5). The enhancement did not depend on the glycine concentration.

Download figure to PowerPoint

In the case of the NR1a-NR2B subunit combination a glycine-independent enhancement by neomycin was observed (Fig. 11C and D). No glycine sensitivity of the potentiation was observed at positive holding potentials, neither was a prominent voltage dependence observed with this concentration of neomycin (Fig. 11D).

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements

NMDA-activated currents demonstrate: (1) several types of desensitization, (2) Ca2+-dependent inactivation, and (3) ‘run-down’ during prolonged whole-cell recordings (McBain & Mayer, 1994). This can make it difficult to determine the mechanisms of action of polyamines and cations. For example, they may enhance NMDA-evoked currents by glycine-dependent and glycine-independent mechanisms whilst inhibiting by both voltage-dependent and voltage-independent mechanisms.

Desensitization and Ca2+ inactivation of NMDA-activated currents in isolated neurones

We found that in isolated CA1 pyramidal neurones NMDA-evoked currents primarily demonstrated glycine-insensitive desensitization (time constants of desensitization leqslant R: less-than-or-eq, slant 1 s). This desensitization was not a consequence of Ca2+ inactivation of NMDA channels because over a wide range of extracellular Ca2+ concentrations and Ca2+ buffering conditions we could detect no change in the decay of NMDA-activated currents. There was an apparent reduction in desensitization if extracellular [Ca2+] was simultaneously reduced during the application of NMDA but it was clear that this effect was a consequence of the activation of a contaminating Ca2+-dependent non-selective cation current (Xiong et al. 1997).

Reduction in single channel conductance

Calcium, Mg2+ (Ascher, 1988; Ascher & Nowak, 1988; Paoletti et al. 1995) and polyamines (Rock & Macdonald, 1992) reduce the single channel conductance of NMDA channels, probably by shielding membrane surface charge (Paoletti et al. 1995). At physiological pH neomycin possesses five charged primary and secondary amine moieties suggesting that it could bind to various anionic sites on many different proteins. In the presence of neomycin we also detected a reduction of single channel currents even at positive values of membrane potential suggesting that this polyamine causes a similar shielding of surface charge.

Possible voltage-dependent block by neomycin

Neomycin (leqslant R: less-than-or-eq, slant 1 mm) enhanced NMDA-evoked whole-cell currents in isolated hippocampal neurones without demonstrating any signs of an underlying blockade. The degree of enhancement was not dependent upon holding potential also suggesting that there was no underlying voltage-dependent block by this polyamine. Paradoxically, single channel recordings from outside-out patches demonstrated that neomycin probably caused a voltage-dependent and rapid open block of the channel. The discrepancy between single channel and whole-cell effects can be reconciled if neomycin were to have caused a transient block of the open channel without altering the total time that the channel remained open (Nowak et al. 1984). Each sojourn to the open state might have simply been interrupted by transient blockages that subsequently delayed channel closure. For example, the protein kinase inhibitor H7 (1-(5-isoquinolynesulfonyl)-2-methylpiperazine) reduces the mean open time of the NMDA channel but at the same time increases the length of each channel burst (Amador & Dani, 1991). Consequently, H7 does not cause a voltage-dependent block of whole-cell NMDA-evoked currents even though a blockade is observed in single channel recordings. Thus, we have grossly underestimated the blocking potency of neomycin from measurements of whole-cell currents.

Spermine also enhanced NMDA-evoked currents in isolated rat CA1 pyramidal neurones at both negative and positive holding potentials. Nevertheless, at negative holding potentials depression (voltage-dependent block), as well as potentiation, was observed. Spermine appeared to show less voltage-dependent block than previously reported for cultured rat hippocampal neurones (Benveniste & Mayer, 1993) but a similar degree of block by spermine was observed on cultured spinal cord neurones (Lerma, 1992).

Glycine-independent enhancement

Concentration-response curves for glycine demonstrated maximal enhancement of NMDA-evoked currents with concentrations of glycine ranging from 3 to 10 μm (Fig. 8). At these concentrations of glycine, neomycin caused a glycine-independent enhancement of NMDA-evoked currents (see Fig. 7). The role of glycine-independent enhancement would appear to be relatively minor in isolated CA1 hippocampal neurones particularly in the presence of physiological concentrations of extracellular Mg2+ (see below).

In response to applications of spermine, NMDA receptors composed of the NR1a-NR2B combination of subunits give rise to currents that demonstrate a glycine-independent enhancement whilst those from the NR1a-NR2A combination do not (Johnson, 1996; Williams, 1997). In contrast, currents from both NR1a-NR2B and NR1a- NR2A combinations of subunits demonstrate glycine-dependent enhancement by spermine. This provides the strongest evidence that both glycine-dependent and glycine-independent potentiation by polyamines are mediated via separate sites on the receptor (Johnson, 1996; Williams, 1997). In HEK293 cells expressing the NR1a-NR2B combination of subunits neomycin enhanced NMDA-evoked currents by about 3-fold. The enhancement was similar regardless of the concentration of glycine, again demonstrating that neomycin caused a glycine-independent enhancement of NMDA receptors.

Glycine-independent potentiation: interactions of neomycin, spermine, Mg2+ and protons

In isolated CA1 neurones the Mg2+-induced potentiation of NMDA-evoked currents was additive with that caused by low concentrations of neomycin. However, Mg2+ depressed the enhancement observed at concentrations of neomycin above 1 mm. This suggests that Mg2+ and neomycin may interact at the same binding site to produce glycine-independent potentiation. We also observed similar interactions between neomycin and spermine suggesting that Mg2+, spermine and neomycin probably interact at the same site or sites to cause glycine-independent enhancement. Consistent with this conclusion, Mg2+ and spermine also cause a less than additive glycine-independent enhancement of NMDA-evoked currents for the NR1a-NR2B combination of subunits (Paoletti et al. 1995). Furthermore, Mg2+, neomycin and spermine probably act at the same site to enhance the binding of TCP ([1(2-theinyl)-cyclohexyl]-piperidine) to the NMDA receptor (Reynolds, 1990; Sacaan & Johnson, 1990; Pullan et al. 1992).

Spermine and Mg2+ induce a glycine-independent enhancement, at least in part by relieving a tonic inhibition of the receptor by protons (Paoletti et al. 1995; Traynelis et al. 1995). We also found that the glycine-independent potentiation by neomycin was strongly dependent upon extracellular pH. For example, we found that neomycin caused a large shift in the pH sensitivity of the NMDA receptor and this result is consistent with the enhancement by neomycin being mediated by a relief of the tonic block by protons. This evidence provides further support for a common site of interaction for neomycin, Mg2+ and protons.

Glycine dependence of neomycin

Neomycin (100 μm) more than doubled the potency of glycine for the NMDA receptors (peak currents) of isolated neurones. Neomycin (100 μm) enhanced the potency for glycine by reducing its rate of offset (decreased the ‘off’ rate: 2.66 to 1.36 s−1) suggesting that it slowed the dissociation of glycine from the receptor. The apparent affinity for glycine changed from about 200 to 90 nm in the presence of neomycin. Similar changes in the ‘off’ rate have been detected with a concentration of 3 mm Mg2+ in isolated neurones (Wang & MacDonald, 1995); however, substantially larger shifts in the potency for glycine are seen in high concentrations of spermine (1 mm) in cultured neurones (3.5- to 5.4-fold; Benveniste & Mayer, 1993).

As anticipated, neomycin also caused a glycine-dependent enhancement of NMDA-evoked currents in HEK293 cells expressing the NR1a-NR2A subunit combination. The proportionate enhancement by neomycin decreased as the concentration of glycine was increased to saturating concentrations and no enhancement was observed in saturating concentrations of glycine. This observation is consistent with an enhanced potency of glycine and with a lack of glycine-independent potentiation. Perhaps somewhat surprising was our observation that the absolute amplitude of the currents recorded during the application of neomycin was inversely related to the concentration of glycine. The neomycin-induced enhancement seen with low concentrations of glycine was replaced by a depression in high concentrations of glycine (Fig. 11A and B). We have also observed a similar glycine-dependent enhancement and voltage-independent depression of NMDA-evoked currents by high concentrations of the neurotrophin BDNF (brain derived neurotrophic factor) and by dynorphin (Jarvis et al. 1997).

Polyamines do not bind to the glycine site of the NMDA receptor (Ransom & Stec, 1988) suggesting that neomycin is unlikely to have directly antagonized the binding of glycine. Another possible explanation is that the binding of neomycin reduced the affinity of the receptor for NMDA. In this respect, we have provided evidence that the potency for glutamate may be reduced by neomycin, and polyamines have previously been shown to decrease [3H]glutamate binding to the NMDA receptor (Pullan & Powel, 1991).

Effects of neomycin on desensitization

Spermine reduces the rate of glycine-insensitive desensitization of NMDA-evoked currents (increased the time constants of desensitization) and diminishes the extent of desensitization (increased the ratio of steady-state to peak currents) in recordings from cultured hippocampal neurones (Benveniste & Mayer, 1993). Lerma (1992) made a similar observation in recordings from cultured spinal cord neurones in the presence of 1 μm glycine. Furthermore, in ‘well-dialysed’ neurones spermine decreases the rate of onset of desensitization and reduces the extent of desensitization (Lerma, 1992). Although we also observed a slowing of desensitization in subsaturating concentrations of glycine we found that the extent of desensitization of NMDA-activated currents was, in contrast, accentuated by applications of polyamines to isolated neurones. This discrepancy may be accounted for simply on the basis of a residual glycine-sensitive component of desensitization.

The potency of neomycin determined from concentration- response relationships in the presence of a low concentration of glycine (0.3 μm, Ip: EC50= 53 μm; Iss: EC50= 61 μm; P < 0.05) was about twice that in the presence of a high concentration of glycine (10 μm, Ip: EC50= 106 μm; Iss: EC50= 183 μm). The difference in the potency of glycine in the presence of 100 μm neomycin was also about 2-fold, calculated from the ratio of the ‘on’ and ‘off’ rates of glycine. This evidence suggests that neomycin might interact at two distinct sites to produce a glycine-dependent versus glycine-independent enhancement.

Summary of the effects of neomycin

We have demonstrated that neomycin is not a selective activator for the enhancement of NMDA-evoked responses. Quite the contrary, we have identified at least five potential effects of neomycin on the NMDA responses in isolated CA1 pyramidal neurones: (1) it causes a voltage-dependent block of NMDA-evoked currents, (2) it causes both a glycine-dependent, and (3) a glycine-independent enhancement of these currents, (4) it may reduce the affinity of the receptor for its endogenous agonist, glutamate, and (5) it may also inhibit NMDA-evoked currents by shielding surface charge. However, neomycin is more efficacious than spermine at enhancing NMDA-evoked currents simply because it is much less effective than spermine at causing a voltage-dependent block of whole-cell currents. We attribute the apparent selectivity of neomycin versus spermine to the different characteristics of the kinetics of the blockade of single NMDA channels by these polyamines.

Effects of polyamines and Mg2+ on synaptic transmission

Endogenous polyamines can be released from depolarized brain slices (Harman & Shaw, 1981; Fage et al. 1992) and might therefore potentially alter NMDA receptor-mediated transmission. However, whether or not these polyamines would enhance, depress or have no effect on synaptic transmission would depend entirely upon the relative concentrations of glycine and divalent cations present in the synaptic cleft. It is possible that the relatively high concentrations of endogenous Mg2+ and glycine found in the synaptic cleft would act to protect NMDA receptors from transient modulation by released polycationic compounds.

Acknowledgements

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements

This work was supported by a grant from the Medical Research Council of Canada. Z.-G. X. is a fellow of the Medical Research Council of Canada and W.-Y. L. a fellow of the National Centres of Excellence and a fellow of the Heart & Stroke Foundation of Ontario. We thank Ms E. Czerwinska for preparation of NMDA receptor cDNAs. We thank Ms E. Czerwinska and L. Brandes for preparation of hippocampal tissue cultures.