Pentylenetetrazole-Induced Inhibition of Recombinant γ-Aminobutyric Acid Type A (GABA_A) Receptors: Mechanism and Site of Action

Materials and Methods

Cloned GABA_A Receptors.

Human embryonic kidney cell lines (HEK293) stably expressing varying configurations of recombinant GABA_A receptors were studied in the present investigation. Cells expressing rat α1β2γ2, α3β2γ2, α6β2γ2, α1β2, and β2γ2 receptors (short isoform of the γ2 subunit in all cases) were generously supplied by Pharmacia-Upjohn (Kalamazoo, MI). A detailed description of the preparation of HEK293 cells stably expressing GABA_A receptors has been published previously (Hamilton et al., 1993). Cells stably expressing human α1β2γ2 receptors were also studied. A complete description of preparation of this cell line has been published previously (Hawkinson et al., 1996).

Electrophysiology.

Whole-cell and outside-out patch recordings were made at room temperature (22–25°C). Except during acquisition of current-voltage relationships, cells were voltage-clamped at −60 mV. Patch pipettes of borosilicate glass (1B150F; World Precision Instruments, Inc., Sarasota, FL) were pulled (Flaming/Brown, P-87/PC; Sutter Instrument Co., Novato, CA) to a tip resistance of 1 to 2.5 MΩ for whole-cell recordings and 8 to 12 MΩ for outside-out single-channel recordings. The pipette solution contained 140 mM CsCl, 10 mM EGTA, 10 mM HEPES, 4 mM Mg-ATP, pH 7.2. Coverslips containing cultured cells were placed in a small chamber (∼1.5 ml) on the stage of an inverted light microscope (Olympus IMT-2; Olympus, Tokyo, Japan) and superfused continuously (5–8 ml/min) with the following external solution containing 125 mM NaCl, 5.5 mM KCl, 0.8 mM MgCl₂, 3.0 mM CaCl₂, 20 mM HEPES, 10 mM d-glucose, pH 7.3. GABA-induced Cl⁻ currents from the whole-cell or outside-out configuration of the patch clamp technique were obtained using an Axoclamp 200A amplifier equipped with a CV-4 headstage (Axon Instruments, Foster City, CA). For whole-cell recording, GABA-induced Cl⁻ currents were low-pass filtered at 5 kHz, monitored on an oscilloscope and a chart recorder (Gould TA240; Gould, Cleveland, OH), and stored on a computer (pClamp 6.0, Axon Instruments) for subsequent analysis. Series resistance compensation (60–80%) was applied at the amplifier. To monitor the possibility that access resistance changed over time or during different experimental conditions, we measured and stored on our digital oscilloscope the current response to a 5 mV voltage pulse at the initiation of each recording. This stored trace was continually referenced throughout the recording. If a change in access resistance was observed during the recording period, the patch was aborted, and the data were not included in the analysis. For single-channel recordings, the currents were filtered with a low-pass Bessel filter (80 dB/decade) at a cut-off frequency of 1 to 2 kHz and simultaneously recorded on a video cassette recording system (Sony SLV-420; Sony, Tokyo, Japan) via a digital data recorder (VR-10B CRC; Instrutech Corp., Great Neck, NY).

Experimental Protocol.

For all single-channel recordings and whole-cell recordings in HEK293 cells, GABA with or without PTZ was prepared in the extracellular solution and then applied from independent reservoirs by gravity flow for 10 to 20 s to cells or membrane patches using a Y-shaped tube positioned within 100 μm of the cells or the membrane patch. With this system, the 10 to 90% rise time of the junction potential at the open tip is 12 to 51 ms. We did not attempt to study receptors lacking the β subunit, because this subunit may be required for plasma membrane targeting and functional expression (Connolly et al., 1996). Receptors were typically activated with roughly the EC₃₀ GABA concentration; this concentration was chosen because minimal desensitization, which may confound interpretation of results, was elicited. Once a control GABA response was determined, the effect of PTZ on the response was examined. Recovery from PTZ-induced inhibition was readily obtained, thus full inhibition curves for PTZ could generally be obtained from studying one cell. GABA applications were separated by at least 2-min intervals to ensure both adequate washout of GABA from the bath and recovery of receptors from desensitization, if present. Typically, to monitor GABA response, a GABA pulse was applied before, during, and after incubation.

Chemicals.

GABA and PTZ were obtained from Sigma (St. Louis, MO). αIMGBL was synthesized as described previously (Canney et al., 1991). GABA and PTZ stocks were made in double-distilled H₂O. αIMGBL was made in dimethyl sulfoxide and diluted in saline so that the final dimethyl sulfoxide concentration (v/v) was <0.2%.

Data Analysis.

For whole-cell recording, all data were recorded on a chart recorder and stored on a computer for subsequent off-line analysis (pClamp 6.0). GABA concentration-response profiles were constructed from whole-cell recordings and fitted to the following equation: I/I_max= [GABA]ⁿ/([GABA]ⁿ + EC₅₀ⁿ), where I andI_max represent the normalized GABA-induced current at a given concentration and the maximum current induced by a saturating [GABA], EC₅₀ is the half-maximal effective GABA concentration, and n is the Hill coefficient. PTZ inhibition-response relationship was fitted with the equation I/I_max = [PTZ]ⁿ/([PTZ]ⁿ + IC₅₀ⁿ), where I is Cl⁻ current amplitude normalized to control, IC₅₀ is the half-blocking concentration, andn is the Hill coefficient. A minimum of three (typically five to eight) individual experiments were conducted for each paradigm.

Single-channel currents were replayed from tape and digitized at 53 kHz sampling frequency. Single-channel current amplitudes and durations were determined by computer using Fetchan and pStat (pClamp 6.0 software). Channel openings and closings were detected with the 50% threshold crossing method. Openings briefer than 1 ms (≈ the system dead time) were not detectable. Only patches demonstrating infrequent multiple openings (no more than two simultaneous openings apparent) were used for kinetic analysis. To reduce errors due to multichannel patches, we excluded overlaps of these infrequent simultaneous openings in the analysis of closed dwell-time data. The presence of multiple openings would decrease the apparent duration of longer close components but would have no effect on the open state properties. Duration histograms were fitted by a maximum likelihood method. The number of exponential functions required to fit distribution was increased until additional components did not significantly improve the fit. Open channel probability (P_o) was calculated as P_o = time in open channel state/(total time × number of channels in patch).

All data were presented as means ± S.E.M. Student's ttest (paired or unpaired), or one-way ANOVA test was used to determine statistical significance (p < 0.05).

Results

Evaluation of GABA_A Receptor Subunit-Dependent Inhibition by PTZ.

Other ligands that are presumed to bind at the picrotoxin site of the GABA_A receptor display some subunit-dependent effects (Bell-Horner et al., 2000). We thus evaluated whether changes in GABA_A receptor subunit isoforms influence the ability of PTZ to block the GABA_A receptor. Figure1 illustrates PTZ inhibition in several configurations of the receptors. In all receptors tested, PTZ inhibited GABA-mediated current in a concentration-dependent manner. PTZ had similar effects on peak and steady-state currents and did not enhance current decay rate (Fig. 1A). This contrasts with the block induced by picrotoxin and the picrotoxin-site ligands TBPS and U-93631, which have minimal effects on peak amplitude but significantly enhance rate of current decay (Yakushiji et al., 1987; Dillon et al., 1993, 1995a). The PTZ-induced block reached steady state during the first drug application, as repeated applications of PTZ plus GABA did not result in further current inhibition. Table 1shows IC₅₀ and Hill coefficients for PTZ inhibition in the receptors evaluated. PTZ displayed comparable efficacy and affinity at the different receptor configurations tested (Fig. 1B; Table 1). PTZ IC₅₀ values were near 1 mM (range = 0.6–2.2 mM) for all subunit configurations tested, including those with varying α subunit isoforms and those lacking an α or γ subunit. Moreover, the IC₅₀ values were comparable in rat and human receptors.

Kinetic Parameters for PTZ Interaction with GABA_AReceptors.

The above results demonstrated that the IC₅₀ for PTZ inhibition of GABA-gated current (millimolar range) is approximately 1000-fold higher than that reported for PTX (Krishek et al., 1996; Bell-Horner et al., 2000), which has in general been defined in the low-micromolar range. The rapid onset and recovery from PTZ inhibition, compared with PTX (Dillon et al., 1995a), suggested that this difference in functional affinity is at least in part due to changes in dissociation rate constants (k₋₁). To quantify the kinetic interactions of PTZ with the GABA-bound receptor, we used the following one-site model:where R* is the GABA-bound receptor, R*PTZ is the PTZ-bound nonconducting receptor, and k₊₁ andk₋₁ are the drug association and dissociation rates, respectively. Interaction of the receptor with PTZ will proceed to equilibrium with an exponential time constant, τ, equal to 1/(k₊₁[PTZ] +k₋₁). Thus, based on this one-site model, a plot of the 1/τ versus PTZ concentration should yield data that can be fitted to a linear function. The slope of the line will equal the k₊₁, and the y-intercept will equal the k₋₁ for PTZ interacting with GABA-bound receptors. This model has been used previously to define the kinetic parameters for other picrotoxin-site ligands (Dillon et al., 1993, 1995a). To define these parameters for PTZ, human α1β2γ2 receptors were incubated with GABA at 5 μM; this concentration activated a stable current that displayed negligible desensitization. PTZ at varying concentrations (0.1–3 mM) was then coapplied with 5 μM GABA for 10 s. As shown in Fig.2A, the PTZ-induced current decay was concentration-dependent. The reduction of initial current induced by PTZ was nicely fitted with a monoexponential function, yielding a τ for inhibition at each concentration tested. By plotting the data as described above, we obtained an association rate (k₊₁) of 1.14 × 10³ M⁻¹s⁻¹, a dissociation rate (k₋₁) of 0.476 s⁻¹, and a dissociation constant (K_d =k₋₁/k₊₁) of 0.418 mM for PTZ in the presence of GABA (Fig. 2B). Thek₋₁ for PTZ can also be estimated from the recovery of the GABA current upon removal of PTZ. Thus, the τ for relaxation back to the control GABA current following PTZ inhibition was determined. This τ was calculated to be 3.47 ± 0.16 s, the inverse of which yielded ak₋₁ of 0.3 ± 0.01 s⁻¹. This estimate ofk₋₁ is comparable with that determined using the one-site model, thus validating the use of the model to define these kinetic parameters.

Effect of PTZ on GABA Current Voltage Relationship.

Because inhibition of ionic currents by many channel blockers is influenced by transmembrane voltage, we assessed the potential voltage dependence of PTZ-induced block of GABA_A receptors. In human α1β2γ2 receptors, current activated by 10 μM GABA was recorded over a range of holding potentials in the presence or absence of 1 mM PTZ. GABA induced Cl⁻ current was outward-rectifying. PTZ did not significantly change the reversal potential (4.1 ± 1.0 mV in control and 4.4 ± 1.0 mV in the presence of 1 mM PTZ, p > 0.05, paired ttest, n = 4) or the rectification profile. In addition, the magnitude of PTZ-induced inhibition was similar when tested at −60, −30, +30, and +60 mV (one-way ANOVA, p > 0.05,n = 4). Although not definitive, a lack of voltage dependence argues against a classical open channel blocking mechanism for PTZ.

Effect of PTZ on the GABA Concentration-Response Curve.

Picrotoxin-induced block of GABA receptors has generally been defined as noncompetitive (Akaike et al., 1985; Yakushiji et al., 1987; Yoon et al., 1993) or mixed, having both noncompetitive and competitive components (Smart and Constanti, 1986; Krishek et al., 1996). The mechanism by which PTZ inhibits GABA-induced currents was studied by determining the concentration-response relationship of GABA in the absence and presence of PTZ (1, 5, and 20 mM) in human α1β2γ2 receptors. Figure 3B shows that PTZ significantly increased the GABA EC₅₀ in a concentration-dependent manner (a 7-fold increase in GABA EC₅₀ in the presence of 20 mM PTZ,p < 0.05, one-way ANOVA), whereas the Hill coefficient was not significantly altered (p > 0.05, one-way ANOVA). Although there was a modest tendency for maximal GABA current to be reduced by PTZ, the difference was not significant (p > 0.05, paired t test, n= 4–5 for each PTZ concentration group). Similar results were obtained when experiments were conducted using rat α1β2γ2 receptors (not shown). Thus, whereas picrotoxin is a mixed antagonist, the antagonism of GABA-gated current by PTZ is predominantly competitive in nature.

Effect of Intracellular PTZ.

It has been reported in theXenopus oocyte expression system that PTZ is able to cross the plasma membrane of cells and possibly acts at an intracellular site(s) (Bloms-Funke et al., 1996). To test this, we examined effects of extracellularly applied PTZ in recombinant human α1β2γ2 receptors when a near saturating PTZ concentration (20 mM) had already been added to the pipette solution and thus pre-equilibrated intracellularly. Because HEK cells are known to be coupled, these experiments were conducted on isolated single cells to ensure that PTZ was fully equilibrated in the cell under study. In the presence of intracellular PTZ, extracellular coapplication of 1 and 20 mM PTZ resulted in 45.5 ± 2.5% and 89.7 ± 5.2% inhibition of the current amplitude induced by 10 μM GABA (Fig.4A; n = 6). This degree of inhibition was not different from that recorded in cells that were not exposed to intracellular PTZ (50 ± 2.7% inhibition by 1 mM PTZ and 91.4 ±1.5% by 20 mM PTZ, n = 5,p > 0.05, unpaired t test, Fig. 4B). In addition, the effects of extracellular application of PTZ were highly reversible and repeatable in the cells preloaded with PTZ. These data only conclusively demonstrate that the presence of PTZ inside the cell does not prohibit block by extracellular application; they do not rule out the possibility of an additional intracellular site of action of PTZ. To further assess the possibility that PTZ may have an intracellular site of action, we conducted the following experiments. GABA-activated currents were recorded from isolated single cells with or without intracellular 20 mM PTZ introduced into the cell via the patch pipette. The presence of 20 mM intracellular PTZ did not significantly influence the current amplitude induced by 10 μM GABA, as the current amplitude was 1539 ± 317 pA in PTZ-preloaded cells (n = 18) and 1476 ± 346 pA in control cells (n = 16, p > 0.05, unpairedt test). These results do not support the hypothesis that PTZ has an additional intracellular site of action. Moreover, the data also suggest that PTZ can only access its binding site to GABA_A receptors via an extracellular pathway.

Effect of PTZ on GABA-Activated Single-Channel Currents.

To explore at the single-channel level the effect of PTZ on GABA_A receptor gating kinetics, we recorded single-channel currents in the absence or presence of PTZ (1 mM, close to its IC₅₀ in the whole-cell recordings) using excised outside-out patches. Because the PTZ action among various subunit compositions or species was indistinguishable at the whole-cell level, we studied kinetics of single-channel activity in human α1β2γ2 receptors only.

The characteristics of single GABA-activated channels in human recombinant α1β2γ2 receptors are shown in Figs. 5 and 6 and Table 2. It should be noted that events briefer than 1 ms were beyond the resolution of the data as collected. Thus, we may have underestimated the number of openings and closings per patch. However, these very brief events contribute less than 10% of the total open duration and less than 1% of the total closed duration in recombinant GABA_A receptors (Fisher et al., 2000); thus, they probably have minimal impact with regard to PTZ inhibition. No spontaneous single-channel activity was observed in HEK cells expressing human α1β2γ2 receptors, in the absence of GABA (a1 and a2 in Fig. 5A). Application to an excised patch of 5 μM GABA for 20 s elicited bursts of channel openings displaying a mean amplitude of 1.49 pA at a holding potential of −40 mV (∼37 pS for a Cl⁻ reversal potential of 0 mV). Subconductance openings contributed a small proportion of total membrane current. Thus, formal analyses were performed on only the large conductance state. Histograms for 5 μM GABA-induced channel openings were best fitted with two-exponential functions, indicating the presence of two open states with mean durations of 2.6 and 10.7 ms (Fig. 6A). Channel open frequency during control conditions was 15.9 ± 6.7 s⁻¹. Closed dwell-time distributions for single-channel currents were binned logarithmically and fitted to a logarithmic scale (Fig. 6B). Exponential fitting of these data indicated three closed states with mean durations of 3.4, 31.8, and 239.1 ms.

Coapplication of PTZ (1 mM) with GABA significantly altered several single-channel characteristics of human α1β2γ2 receptors. PTZ markedly and reversibly decreased the frequency of GABA-gated channel openings by 49% at a holding potential of −40 mV (Fig. 5A). PTZ had no effect on the mean current amplitude (Fig. 5B) or duration of long or short open states (Fig. 6A; Table 2). A shift in the relative contribution of longer open states was observed in the presence of PTZ. The predominant effects of PTZ were on channel closed states. The durations of the long and intermediate closed states were significantly increased by PTZ (Fig. 6B; Table 2). In addition, the relative contribution of the longest closed state was increased, whereas the weight of the short and intermediate closed states was correspondingly decreased (Table 2). These changes led to a significant reduction inP_o in the presence of PTZ. Effects of PTZ on single-channel characteristics of human recombinant α1β2γ2 GABA_A receptors are summarized in Table 2.

Effect of αIMGBL on PTZ Inhibition.

The γ-butyrolactone derivative αIMGBL is an antagonist of the action of picrotoxin (Holland et al., 1990; Yoon et al., 1993). We used αIMGBL as a probe to determine whether functional domains for PTZ and picrotoxin overlap. αIMGBL (5 mM) was chosen because it efficiently blocked picrotoxin inhibition of GABA-activated current in rat α1β2γ2 receptors (Fig. 7, A and C). αIMGBL also significantly reduced the inhibition of GABA-activated current produced by 1 mM and 20 mM PTZ (p < 0.05, paired ttest, n = 5–7). Application of αIMGBL alone (n = 2) had no effect on the whole-cell recording, and it did not affect GABA-activated current (Fig. 7B; n = 6). Protection of PTZ inhibition with the picrotoxin antagonist αIMGBL supports the suggestion that PTZ and picrotoxin share a related functional domain.

Discussion

Based largely on radioligand binding studies, PTZ is believed to interact with the picrotoxin site of the GABA_Areceptor (Ramamjaneyulu and Ticku, 1984; Squires et al., 1984). Remarkably few studies, however, have evaluated directly the actions of PTZ on GABA_A receptors and its mechanism and molecular site of action. Thus, in an attempt to more precisely define these parameters, we used an array of experimental manipulations to assess quantitatively the mechanism and site of action of PTZ.

Several lines of evidence obtained in the present report support the contention that the site of action of PTZ is similar to the picrotoxin site of the receptor. First, we demonstrate that the γ-butyrolactone compound αIMGBL, shown previously to antagonize the actions of picrotoxin (Holland et al., 1990), can also block the inhibitory actions of PTZ. Both findings are consistent with a common site of action of the two convulsant agents; second, single-channel recordings from human α1β2γ2 GABA_A receptors demonstrated that PTZ inhibition of GABA_Areceptors is due to a decrease in the frequency of channel opening. PTZ did not affect mean-channel open duration or single-channel conductance. These effects of PTZ at the single-channel level are comparable with those elicited by picrotoxin (Newland and Cull-Candy, 1992) and the picrotoxin-site ligands dieldrin (Ikeda et al., 1998) and U-93631 (Dillon et al., 1995b).

Additional characteristics of block are common to both PTZ and PTX. For instance, block by both PTX (Yakushiji et al., 1987; Newland and Cull-Candy, 1992; Yoon et al., 1993) and PTZ (present results) is voltage-independent. Moreover, we show that PTZ can only gain access to its binding site extracellularly. This is evident by the finding that intracellular PTZ, at a near saturating concentration, did not alter GABA-activated current. Our results conflict with those of Bloms-Funke et al. (1996), who found that intracellular PTZ could inhibit GABA_A receptors expressed in Xenopusoocytes. The discrepancy is likely due to the differences in experimental preparations. Intracellular picrotoxin was reported to block GABA currents in one study (Akaike et al., 1985) but had no effect in a separate study (Cull-Candy et al., 1987).

The fact that a number of characteristics of block by PTZ and PTX are similar does not indicate that the two drugs share a common site of action. Studies have shown that amino acids at the 2′ and 6′ positions in the cytoplasmic end of TM2 are involved in picrotoxin blockade of the GABA_A receptor (ffrench-Constant et al., 1993; Gurley et al., 1995; Xu et al., 1995). Gurley et al. (1995)demonstrated that picrotoxin sensitivity was abolished when the 6′ threonine of the β2 subunit was mutated to phenylalanine (T6′F). We have recently shown that this mutation also abolishes the inhibitory actions of PTZ (Dibas and Dillon, 2000). The dependence of both PTZ and PTX on the nature of this 6′ TM2 residue for GABA_A receptor block provides a physical basis for the similarity of action of the two drugs and supports the suggestion that they interact at the same or overlapping domains.

Although several lines of evidence support a common site of action of PTZ and picrotoxin, there are several important distinguishing features between the two compounds, some of which shed additional light on their respective functional domains. The affinity of PTZ is roughly 1000-fold lower than that of PTX and other presumed PTX-site ligands (Dillon et al., 1993, 1995a; Krishek et al., 1996). Our kinetic analysis illustrates that this is predominantly due to differences in the dissociation rate (k₋₁). In the present experiments, we defined the PTZk₋₁ in GABA-bound receptors at 0.476 s⁻¹ and k₊₁ at 1.14 × 10³ M⁻¹s⁻¹. The same parameters for PTX were shown to be 0.0058 s⁻¹ and 1.3 × 10⁴ M⁻¹s⁻¹, respectively (Dillon et al., 1995a). The equilibrium dissociation constants (K_d) are thus 418 μM for PTZ and 443 nM for PTX. By definition, the formation of drug-receptor complexes at the equilibrium concentration of a drug is equivalent to the dissociation of drug-receptor complexes. Although the association rate (k₊₁) for PTZ is about 10-fold lower on a molar basis than that of PTX, at an equilibrium concentration thek₊₁ is equivalent to thek₋₁ (0.476 s⁻¹for PTZ versus 0.0058 s⁻¹ for PTX). Thus, the much greater effect on initial peak current observed with PTZ inhibition, compared with PTX inhibition (Yakushiji et al., 1987; Yoon et al., 1993; Dillon et al., 1995), can be fully explained by the differences in kinetic constants described here.

A somewhat unexpected but notable finding was that PTZ antagonized GABA-activated current exclusively via competitive inhibition. This contrasts with PTX-induced block, which displays noncompetitive (Akaike et al., 1985; Yakushiji et al., 1987; Yoon et al., 1993) and competitive components (Smart and Constanti, 1986; Krishek et al., 1996) of block. The PTZ-induced block is unlikely to be a true competition at the GABA binding site, as PTZ does not inhibit binding of [³H]muscimol to GABA_Areceptors (Ticku and Maksay, 1983). The competitive component of picrotoxin block may be due to its ability to allosterically stabilize the receptor in an inactivated state, instead of true competition at the agonist binding site (Smart and Constanti, 1986, and below). A similar mechanism is likely to account for PTZ-induced block. Other examples of allosterically mediated “competitive” antagonism have been demonstrated (Bertrand et al., 1992; Lynch et al., 1995). Indeed, picrotoxin itself blocks human glycine α1 receptors through allosteric competitive inhibition (Lynch et al., 1995); an arginine residue at the extracellular region of TM2 was shown to be involved in this antagonism. With regard to PTX and PTZ block of the GABA_A receptor, the 6′ threonine of TM2 may be involved in their ability to allosterically stabilize the receptor in a nonconducting state.

A difference between PTZ and PTX inhibition that provides considerable information about their relative functional domains is revealed by our assessment of subunit-dependent effects of PTZ. In general, picrotoxin has been considered to display relatively nonspecific interactions with various configurations of GABA_A receptors (Newland and Cull-Candy, 1992; Krishek et al., 1996). However, recent work has shown that GABA_A receptors lacking an α subunit are significantly (10- to 20-fold) more sensitive to PTX than those composed of αβγ subunits (Bell-Horner et al., 2000). This is also true for the PTX-site ligand U-93631 (Bell-Horner et al., 2000) and the insecticide dieldrin (C. L. Bell-Horner and G. H. Dillon, in preparation). Interestingly, our present results demonstrate that PTZ inhibition is not enhanced in receptors lacking an α subunit. The enhanced sensitivity to PTX, U-93631, and dieldrin is presumably due to the relative abundance of alanines at the 2′ position in TM2 (Bell-Horner et al., 2000). This amino acid position is equivalent to that shown by ffrench-Constant et al. (1993) to confer resistance to dieldrin in mutant Drosophila GABA receptors that have serine instead of alanine in this position. A recent model (Zhorov and Bregestovski, 2000) hypothesizes on the interaction of PTX in the Cl⁻ channel. According to the model, PTX enters deep into the channel, and its hydrophobic domain interacts with residues at the TM2 2′ position, whereas its electronegative domain hydrogen bonds with threonine residues at the 6′ position. Considering that both the 2′ and 6′ positions have been implicated in block by PTX, U-93631, and dieldrin, this model is plausible for the actions of these antagonists. As noted, the 6′ residue is also involved in PTZ inhibition (Dibas and Dillon, 2000). However, the current results would suggest the postulated hydrophobic interactions at the 2′ position are of minor significance for the actions of PTZ. The model by Zhorov and Bregestovski was formulated to describe noncompetitive blockade. Considering that the 6′ position is involved in blockade for both PTX and PTZ (which displays only competitive antagonism), it is possible that this site is responsible for the allosteric competitive blockade for both compounds. Two possibilities could mediate the noncompetitive PTX blockade. It could be mediated at a distinct, undefined site; two sites of action for PTX have been proposed (Yoon et al., 1993). Alternatively, PTX interaction at the 2′ position may induce a structural change at the gate of the channel (Xu et al., 1995), which results in noncompetitive antagonism. The lack of effect of the 2′ position on PTZ-induced block is consistent with this view. Additional studies are necessary to more fully delineate the domains of the two ligands.

In summary, our results support the contention that the central nervous system convulsant pentylenetetrazole competitively antagonizes the GABA_A receptor, likely through an allosteric interaction in the Cl⁻ channel. A number of characteristics of PTZ-induced blockade are similar to that induced by picrotoxin, and block by both ligands is affected by the 6′ position of TM2. However, the different characteristics of PTZ inhibition that we have described, compared with PTX inhibition, indicate the domains of interaction cannot be identical. Our results further underscore the complexity of the convulsive site.