点击显示 收起
Departments of Ophthalmology and Visual Sciences and Biological Sciences, University of Illinois at Chicago, Chicago, Illinois
Abstract
Picrotoxin is a plant alkaloid that is often used to block the activity of neuronal GABA and glycine receptors. However, the mechanism by which picrotoxin inhibits these receptors is still in debate. In this study, we investigated the picrotoxin inhibition on perch- subunits expressed heterologously in Xenopus laevis oocytes, and on native GABAC receptors of perch bipolar cells. Both competitive and noncompetitive mechanisms were observed for picrotoxin inhibition of the GABAC receptor. In oocytes expressing the 1A subunit, terminating simultaneously the coapplication of GABA and picrotoxin induced a large rebound of membrane current. In addition, picrotoxin significantly accelerated the kinetics of GABA responses, particularly in the relaxation (offset) phase of GABA currents. Both current rebound and the large acceleration of GABA relaxation were unique to picrotoxin inhibition and were not observed with the competitive antagonist (1,2,5,6-tetrahydropyridin-4-yl)-methylphosphinic acid or the allosteric modulator zinc. The change in kinetics induced by picrotoxin was also observed on receptors formed by other GABA subunits, as well as on the GABAC receptors of retinal bipolar cells. Based on these observations, we proposed a model in which picrotoxin binds to the GABAC receptor in both channel open and closed states. Overall, this model provides a remarkably good approximation of the experimental findings we observed for picrotoxin inhibition of GABAC receptors. These results support an allosteric mechanism of picrotoxin inhibition of ligand-gated chloride channels.
GABA and glycine are the main inhibitory neurotransmitters in the nervous system. GABA activates three types of receptors (GABAA, GABAB, and GABAC) on neuronal membranes; GABAA and GABAC (or GABAA0r) receptors function as anion channels that are permeable to chloride, whereas GABAB receptors are metabotropic receptors that are linked through second messengers to potassium or calcium channels (Barnard et al., 1998; Costa, 1998; Qian and Ripps, 2001). Receptors activated by glycine also gate chloride channels, and thus both GABA and glycine receptors provide mechanisms for fast synaptic inhibition in neuronal tissues (Moss and Smart, 2001).
It has been recognized for decades that picrotoxin, a plant alkaloid, inhibits ligand-gated chloride channels (Bormann, 1988; Pribilla et al., 1992; Macdonald and Olsen, 1994). However, its binding site on these receptors has not been identified, and the inhibitory mechanism is still debated (Macdonald and Olsen, 1994; Luddens et al., 1995; Zhorov and Bregestovski, 2000). There are two hypotheses proposed to explain picrotoxin inhibition of GABA receptors. One hypothesis suggests that picrotoxin acts as a noncompetitive channel blocker for these receptors. This concept is supported by the fact that picrotoxin inhibits various anion-selective, ligand-gated channels independent of the receptor type. For example, besides GABA and glycine receptors, picrotoxin also effectively inhibits invertebrate chloride currents gated by glutamate (Cleland, 1996), acetylcholine (Yarowsky and Carpenter, 1978), and dopamine (Magoski and Bulloch, 1999). In addition, Inoue and Akaike (1988) noticed noncompetitive and apparent use-dependent qualities of picrotoxin inhibition in GABAA receptor-mediated responses on frog sensory neurons. Support for picrotoxin as a channel blocker is also provided by the finding that mutations in the intracellular half of the M2 segment (channel forming domain) of the subunit protein significantly affect the sensitivity of GABA receptors to picrotoxin (Gurley et al., 1995; Wang et al., 1995b; Zhang et al., 1995; Dibas et al., 2002). Furthermore, an anion-selective channel formed by a synthetic four-helix bundle protein composed of M2 segments from the glycine receptor can also be blocked by picrotoxin (Reddy et al., 1993).
However, there are other effects of picrotoxin that cannot be explained by a simple channel-block mechanism. Single channel activity of GABAA receptors recorded from rat sympathetic neurons suggests picrotoxin binds preferentially to agonist-bound forms of the receptor and stabilizes the channel in closed states (Newland and Cull-Candy, 1992). In addition, it has been demonstrated that there is a competitive component of picrotoxin inhibition of homomeric 1 glycine receptors (Lynch et al., 1995) and GABAC receptors (Qian and Dowling, 1994; Wang et al., 1995b; Zhang et al., 1995). These results suggest an allosteric mechanism for picrotoxin inhibition.
Although picrotoxin is widely used to inhibit neuronal GABA receptor activity, its effects on the kinetics of GABA-elicited responses have not been fully explored. No study, to our knowledge, has examined the effects of picrotoxin on relaxation of GABA-elicited whole-cell membrane currents. Here, we investigate the mechanism of picrotoxin inhibition using the GABAC receptor as a model system. Although the nomenclature is still in debate (Barnard et al., 1998), GABAC receptors referred to in this article are defined as those receptors formed at least in part by GABA subunits and the bicuculline- and benzodiazepine-insensitive GABA receptors such as those present on retinal neurons (Qian and Ripps, 2001). These GABAC receptors are sensitive to picrotoxin inhibition, and their current responses display very slow response kinetics, with offset time constants on the order of seconds (Qian and Ripps, 2001). This unique feature makes the GABAC receptor an ideal model system for investigating the action of picrotoxin on response kinetics. We found evidence that picrotoxin greatly accelerates the GABA response elicited from various types of GABAC receptors. This is in contrast to the competitive antagonist (1,2,5,6-tetrahydropyridin-4-yl)methylphosphinic acid (TPMPA), which has no effect, and the allosteric modulator zinc, which only slightly accelerates the GABA responses. Fitting the recorded responses to a mathematic model, our data indicate that picrotoxin could bind the receptor in both channel open and closed conformations. These results support an allosteric mechanism for picrotoxin inhibition of GABAC receptors.
Materials and Methods
All experimental procedures adhered to the Guidelines for the Care and Use of Laboratory Animals formulated by the National Institutes of Health and were approved by the Animal Care Committee of the University of Illinois at Chicago College of Medicine.
Subunit Expression and Current Recording from Xenopus laevis Oocytes. Gravid X. laevis females were purchased from Xenopus I (Dexter, MI) and housed in climate-controlled, light-cycled rooms at the University of Illinois at Chicago Biological Resources Laboratory. Ovarian lobes were removed under surgical anesthesia (0.02% 3-aminobenzoic acid ethyl ester; Tricaine) and incubated with constant agitation for 1 h in a calcium-free Ringer's solution containing collagenase (2 mg/ml) to remove the follicular layer. Stage V and VI oocytes were selected and kept at 16°C in a Ringer's solution containing 100 mM NaCl, 2 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, and 10 mM glucose, at pH 7.4. Plasmids containing perch- subunits that were linearized with BamHI, and capped mRNAs were synthesized with SP6 RNA polymerase using the mMessage mMachine (Ambion, Austin, TX) according to the manufacturer's instructions. Each oocyte was injected with 50 nl of mRNA (0.5 mg/ml), and after 2 to 5 days of expression, GABA-activated currents were recorded with the cell held at eC70 mV using a two-microelectrode voltage-clamp amplifier (GeneClamp 500; Axon Instruments Inc., Union City, CA). The recording electrodes were filled with 3 M KCl and had resistance about 1 M measured in Ringer's solution. A gravity flow system was used to deliver drugs to the preparation. Each superfusion channel was separately controlled by a solenoid manifold and operated under computer command. The recording chamber holding the oocyte had a volume of about 20 e蘬, and with a flow rate of 2.5 ml/min, the solution changing rate was 0.31 s based on measurements of the response to 25 mM potassium (Qian et al., 1997).
Patch-Clamp Recording from Isolated Retinal Bipolar Cells. Solitary bipolar cells isolated from the retina of hybrid bass (Morone chrysops crossed with Morone saxitilis) were used in this study. The cell isolation and recording procedure were the same as published previously (Qian and Ripps, 1999). In brief, animals were anesthetized with Tricaine (0.02%) and sacrificed by double pithing. The retina was isolated from the hemisectioned eyeball and incubated for 40 min in an enzyme solution made up of Leibovitz's L-15 culture medium (Invitrogen, Carlsbad, CA) containing 2 mg/ml papain (Calbiochem, San Diego, CA) and 1 mg/ml l-cysteine. Enzyme activity was arrested by washing the tissue five times in fresh L-15 medium. After trituration through a sterile pipette, aliquots of the supernatant containing dissociated cells were placed in 35-mm culture dishes containing 3 ml of the L-15 medium and stored in an incubator at 16°C. Bipolar cells were readily identified by their characteristic morphology (Qian and Dowling, 1995), and those with prominent cell bodies were selected for study.
Membrane currents were recorded using whole-cell voltage-clamp (Hamill et al., 1981). During the recording, cells were bathed in a Ringer's solution containing 145 mM NaCl, 2.5 mM KCl, 2.4 mM CaCl2, 1.5 mM MgCl2, 10 mM glucose, and 5 mM HEPES, pH 7.6, and the pipettes were filled with an intracellular solution containing 124 mM CsCl, 1 mM CaCl2, 11 mM EGTA, 2 mM MgCl2, and 10 mM HEPES, pH 7.4. Macroscopic currents were recorded with the cells clamped at eC60 mV. Because both GABAA and GABAC receptors are present on white perch bipolar cells, bicuculline was added to the bath solution to selectively block GABAA receptor activity. A Perfusion Fast-Step system (Warner Instrument, Hamden, CT) was used to deliver drugs onto the cell. The solution changing rate, measured by liquid junction potential, was 25 ms.
Data Analysis and Modeling. Data analysis was performed with pCLAMP software (Axon Instruments Inc.). The time constants () of the individual deactivation curves were calculated using the Clampfit program to fit the falling phase of the "off" response to a first-order exponential function. For the GABA-induced responses, the data were fit by the equation
(1)
where RX is the response induced by GABA at a concentration X, Rmax is the maximum GABA response, EC50 is the GABA concentration at which a half-maximum response is induced, and nH is the Hill coefficient. To study inhibition by the antagonists, the data were fit by the equation
(2)
where RX is the response in the presence of X concentration of antagonist, RC is the control response (i.e., the GABA response in the absence of the antagonist), IC50 is the antagonist concentration at which half of the GABA response is blocked, and nH is the Hill coefficient. The dose-response curves were fit to the Hill eq. 1 or 2 with the program Origin (OriginLab Corp., Northampton, MA). For each experiment, data were collected from five to eight cells and are presented as means ± 1 S.D.
Mathematical modeling of the action of picrotoxin on GABA-evoked responses was performed with Matlab (MathWorks Inc., Natick, MA) by digitally solving differential equations derived for each step in the model shown in Fig. 5 (Colquhoun and Hawkes, 1981). An automated search program was developed under Matlab to achieve the minimal square error of means between calculated and recorded responses. Each parameter in the model was varied independently with 2 times original value in parameter space searched in each iteration. The loop continued until all the parameters converged. Three different sets of seed (initial parameter values) were used for each fit, and when the optimized parameters meet within a margin of 5%, the average was used as an estimate of the best rate. Responses from six oocytes were used for the modeling study. For each cell, GABA activation parameters (k1, k2, , and ) were determined from GABA-elicited response in Ringer's solution. Parameters related to picrotoxin inhibition were calculated from traces elicited by coapplication of GABA with 10, 30, and 100 e picrotoxin. Each model was judged by comparing the residual error between the recorded response and that predicted by the model and by a comparison of other aspects of picrotoxin inhibition observed on oocyte and those predicted by the model [i.e., GABA dose-response curves in the presence of picrotoxin, and IC50 and Hill coefficient values of picrotoxin inhibition (Tables 3 and 4)].
Results
Picrotoxin Inhibition of GABA-Mediated Responses. The homomeric receptors formed by GABA subunits are highly sensitive to picrotoxin inhibition. Figure 1A illustrates the inhibitory effects of picrotoxin on GABA-activated currents recorded from an oocyte expressing the perch-1A subunit. In these experiments, GABA (10 e) was coapplied with picrotoxin at concentrations indicated by the numbers above each trace. Note the extremely slow kinetics that characterize the GABA-evoked responses, especially the long time course for the off responses after the termination of GABA application (Qian et al., 1998). In the absence of picrotoxin (trace zero), superfusion of 10 e GABA-elicited an inward current of about 1.2 e藺 in this cell, and it returned slowly toward baseline after the termination of a 20-s GABA pulse. The magnitude of the GABA current response was progressively reduced by coapplication of increasing concentrations of picrotoxin. Interestingly, when the two drugs were simultaneously withdrawn, membrane current increased transiently (Fig. 1A, arrow). The magnitude of this rebound current increased with increasing concentrations of picrotoxin, but after reaching its peak, the membrane currents gradually returned to baseline with a similar time course as the GABA-elicited currents.
The current rebound that followed termination of the coapplied drugs is unique for picrotoxin. As shown in Fig. 1, B and C, both the GABAC receptor competitive antagonist TPMPA (Ragozzino et al., 1996) and the allosteric modulator zinc (Chang et al., 1995; Wang et al., 1995a), when coapplied with GABA, inhibited GABA-elicited membrane currents in a concentration-dependent manner. However, even at high concentration (100 e), simultaneous withdrawal of GABA with these antagonists did not produce a rebound in the membrane current. Instead, the membrane current decayed back to baseline and followed a similar time course as the GABA-elicited response in Ringer's solution. Similar results were observed on five oocytes with TPMPA and four cells for zinc inhibition.
The sensitivity to picrotoxin on 1A receptors was dependent on GABA concentration. Figure 2A shows picrotoxin inhibition data obtained with two concentrations of GABA; responses were normalized to the maximum amplitudes elicited by GABA alone. When tested with 1 e GABA, the IC50 value of picrotoxin inhibition was about 0.59 e. However, with a 10-fold higher concentration of GABA, the picrotoxin inhibition curve was shifted to the right and the IC50 value increased to 25 e. The dependence of picrotoxin sensitivity on GABA concentration suggests an apparent competitive inhibitory mechanism on the GABAC receptor. Similar behavior was observed for TPMPA and zinc inhibition. As shown in Fig. 2B, the IC50 values for TPMPA inhibition on the perch-1A receptor also varied with GABA concentration. When tested with 1 e GABA, the IC50 value for TPMPA was 2.1 e, and the value increased to 32.6 e when tested with 10 e GABA. Figure 2C illustrates zinc inhibition curves tested with high and low GABA concentrations. The IC50 value for zinc increases from 6.4 e in the presence of 1 e GABA to 125.6 e with 10 e GABA.
Although the sensitivity of the three antagonists exhibited a similar dependence on GABA concentration, each compound had different effects on the GABA dose-response relations. Figure 2D shows the GABA dose-response curves for the perch-1A receptor measured in the presence of various picrotoxin concentrations. Two effects are observed: 1) the GABA dose-response curves were shifted to the right along the abscissa by picrotoxin in a concentration-dependent manner; and 2) there was a reduction in the maximum response elicited by GABA in the presence of picrotoxin. These results suggest that both competitive and noncompetitive mechanisms exist for picrotoxin inhibition of GABAC receptors (Qian et al., 1998).
On the other hand, TPMPA proved to be a competitive antagonist of GABAC receptors. Only a rightward shift of the GABA dose-response curve was observed in the presence of TPMPA (Fig. 2E), and no significant reduction of the maximum GABA-elicited current was observed. It has been reported that zinc acts as an allosteric modulator of the receptor formed by human 1 subunits and that zinc inhibition contains both competitive and noncompetitive mechanisms (Chang et al., 1995; Wang et al., 1995a). We observed a similar effect on perch-1A receptors; i.e., zinc reduced both the sensitivity and maximum response to GABA (Fig. 2F).
Picrotoxin Accelerates Relaxation of GABA-Elicited Responses. The presence of picrotoxin significantly altered the kinetics of the GABA-induced responses. In particular, the rate at which the GABA-elicited responses returned to baseline was greatly accelerated by picrotoxin. Figure 3A shows typical current traces elicited from perch-1A receptors by 10 e GABA (trace 0) or by coapplication of 10 e GABA with varying concentrations of picrotoxin (0.5eC10 e). Termination of GABA in the presence of picrotoxin caused a rapid return of membrane current; the rate of return increased with rising concentrations of picrotoxin. To quantitate this effect, we measured the time constants for the current decay (summarized in Fig. 3B). As the picrotoxin concentration increased, the time constant of the GABA offset current progressively decreased. Pairwise comparisons of the time constants in Ringer's solution and in the various concentrations of picrotoxin were significantly different.
The large acceleration of GABA relaxation is unique to picrotoxin. After a similar protocol, we tested the effects of TPMPA and zinc on the time course of GABA offset responses. As shown in Fig. 3C, in the presence of 30 e TPMPA, the GABA offset response followed a similar time course as that observed in Ringer's solution. The time constant of the GABA offset response in the presence of TPMPA was 0.98 ± 0.03 (n = 4) of that measured in Ringer. The allosteric modulator zinc accelerated the GABA relaxation only slightly (Fig. 3D). In the presence of 30 e zinc, the time constant of the GABA offset response was 0.76 ± 0.08 (n = 4) of the value measured in Ringer's solution.
Effects of Picrotoxin on Other GABAC Receptors. Because both the kinetics and the amplitude of the inhibitory current contribute significantly to the shape of the neuronal signal conveyed across the synapse, we next explored whether a similar acceleration of GABA responses by picrotoxin could also be observed on other GABAC receptors. Specifically, we determined the effect of picrotoxin on the time constant of the off response for the receptors formed by various perch- subunits that we had previously cloned from a perch retinal cDNA library (Qian et al., 1998). A similar but abbreviated protocol was used in this study. For the receptors formed by GABA subunits, the kinetics of the GABA offset response was largely independent of the GABA concentration (Qian et al., 1998). In these experiments, a GABA concentration eliciting a submaximum response was used, and a picrotoxin concentration was chosen to inhibit 50% of these GABA-elicited responses. As shown in Fig. 4A, picrotoxin reduced the time constants of the offset currents for each of the perch GABAC receptors (plotted on a logarithmic scale). Among them, picrotoxin produced the largest (5-fold) reduction in the time constant of the GABA-offset response on perch-1A receptors; for the other subunits, the decay rates were accelerated between 1.4- and 2.1-fold (Fig. 4A).
Having shown that picrotoxin could significantly accelerate the GABA responses of GABAC receptors expressed in oocytes, we next tested whether the native GABAC receptors present on retinal neurons showed a similar behavior. Figure 4B shows an example of a GABAC receptor-mediated response recorded in Ringer's solution from an individual bipolar cell isolated from the hybrid bass retina; the cell was incubated in 200 e bicuculline to block the activity of the endogenous GABAA receptors. As shown previously (Qian and Ripps, 1999), coapplication of 1 e picrotoxin with 30 e GABA reduced by about half the GABA-elicited current. In addition, the GABA offset response was accelerated in the presence of picrotoxin. This is shown more clearly in Fig. 4C, where each response is normalized to its initial amplitude and plotted on an expanded time scale. Superimposed on the current traces are single exponential decay functions (thin continuous lines). Note that they provide good approximations for the decay phase of the GABA-elicited responses. The time constants of the GABA offset responses measured in Ringer's solution, or in picrotoxin, are presented in the bar graphs to the right of Fig. 4A. The reduction in the time constants from 1.6 s in Ringer to 0.65 s in picrotoxin reflects the approximately 2.4-fold increase in the decay rate of the GABA responses in the presence of picrotoxin. The large difference in the response kinetics of native GABAC receptors on bipolar cells compared with those from homomeric GABA receptors expressed in X. laevis oocytes might reflect the incorporation of the GABAA receptor 2 subunits in the native GABAC receptors, as we have reported previously (Qian and Ripps, 1999; Qian and Pan, 2002).
Schematic Model for Picrotoxin Inhibition. To account for the data observed with picrotoxin inhibition, we adopted a model of picrotoxin inhibition that incorporates its effect on both agonist binding and channel gating of the GABAC receptor (Smart and Constanti, 1986). We modeled the GABA-elicited response and picrotoxin inhibition in two sequential steps. In the first step, a simple binding and gating process was used to model GABA-activated responses (Fig. 5A). Asterisks () denote the receptor with its channel in the open conformation. We considered two models, three binding steps and one gating step (Fig. 5B), and five binding steps and three gating steps (Fig. 5C). Amin and Weiss (1996) have shown that, for the human 1 receptor, binding of three GABA molecules is required to open the channel. A scheme similar to Fig. 5B has previously been successfully used to describe the GABA-elicited currents of the human GABA 1 receptor (Chang and Weiss, 1999). On the other hand, homomeric GABA receptors are thought to be pentamers, and five GABA binding sites are available on the receptor (Amin and Weiss, 1996; Qian and Ripps, 2001), and the scheme of Fig. 5C might better describe the GABA-elicited responses. We tested both models with GABA-elicited responses obtained from six oocytes. Although both models fit the current traces equally well, the model in Fig. 5C provides a much better description of picrotoxin inhibition (see below). Figure 6A illustrates an example of fitting the recorded GABA-elicited responses (thick gray lines) to the model C (thin lines). The average parameters derived from numerical model fitting of GABA-elicited responses from six oocytes are listed in Table 1.
Having fixed the parameters for GABA activation, we next examined picrotoxin inhibition responses. In this model (Fig. 5D), which is based on the GABA activation scheme shown in Fig. 5C, picrotoxin interacts with the ligand-bound GABA-receptor complex in both channel closed and channel open states (Smart and Constanti, 1986). Three pairs of parameter were used: k3 and k4 describe the GABA binding ability of picrotoxin-bound receptor, k5 and k6 indicate picrotoxin affinity to ligand-bound channel closed receptor, and k7 and k8 illustrate binding of picrotoxin to channel open state (Table 2). A few other models were also tested in this study. 1) Model based on three-step binding and one-step gating for GABA activation (Fig. 5B) gave a less satisfactory fit than the model in Fig. 5D. The ratio of residual error derived from these two models was 2.57 ± 0.75 (n = 6). Furthermore, the model in Fig. 5D provided a better prediction for the Hill coefficient of picrotoxin inhibition (Tables 3 and 4). 2) Incorporating steps for picrotoxin binding to the ligand-unbound receptor provided little improvement in reducing the error (about 25%) between predicted and recorded responses. In addition, the rate constants for this step exhibited a large divergence. These results indicate that picrotoxin binding to the ligand-unbound receptor is not necessary for modeling picrotoxin inhibition on GABA receptors. 3) When an additional step was introduced for picrotoxin-bound channel open state, the model yielded a poor fit to the recorded responses. The residual errors produced by this model were more than 5 times larger than those generated by model shown in Fig. 5D. 4) We also attempted to further reduce the number of free parameters in the model by considering the affinity of picrotoxin for the channel open and closed states of the receptor to be equal (i.e., eliminating parameter k7 and k8). However, such change significantly increased the residual error by 2.7 ± 0.5 times (n = 6). Because large conformational changes are required for channel gating, it is reasonable to speculate that the receptor in channel closed and open conformations exhibit different affinities for picrotoxin. For these reasons, we retained k7 and k8 in our model.
Figure 6B shows an example of fit our model (thin lines) to responses elicited by coapplication of GABA and 10, 30, and 100 e picrotoxin (thick gray lines). The model faithfully predicts rebound in membrane current after the termination of GABA and picrotoxin (Fig. 6B, arrow), and it reproduces the enhancement of the rebound current with increasing picrotoxin concentrations. The average parameters derived from numerical model fitting of responses elicited by coapplication of GABA and picrotoxin from six oocytes are listed in Table 2.
The model also faithfully reproduced the acceleration of GABA relaxation in the presence of picrotoxin. As shown in Fig. 6C, the model predicted responses in the presence of picrotoxin are very similar to the oocyte recordings shown in Fig. 3A. To further test the accuracy of our model for picrotoxin inhibition of GABAC receptors, we compared the predicted responses and the actual records from oocytes under low GABA concentrations. The predicted current responses to 1 e GABA exhibit both slow off and slow onset responses (Fig. 6D). Coapplication of picrotoxin reduced the amplitude of GABA-elicited current. Interestingly, however, no significant current rebound was predicted under these conditions when GABA and picrotoxin were terminated (Fig. 6D), in contrast to those observed when a high (10 e) GABA concentration was used (Fig. 6B). The actual recorded current responses to 1 e GABA on an oocyte expressing perch-1A subunit is shown in Fig. 6E. Although there is some deviation in relative amplitude in the presence of picrotoxin, the traces derived from the model reproduce both the slow onset of GABA response and the lack of rebound GABA and picrotoxin application. Overall, the model provided a good approximation for both GABA-elicited responses and picrotoxin inhibition.
The simple model shown in Fig. 5D also provided a good estimate for other aspects of picrotoxin inhibition on the perch-1A receptor. Table 3 compares the GABA dose-response properties (EC50 and Hill coefficient, nH, and maximum GABA responses in the presence of various concentration of picrotoxin) recorded from perch-1A receptor on oocytes and those predicted by the model; and Table 4 lists the properties of picrotoxin inhibition curve (IC50 and nH) measured from oocytes versus those predicted by the model. Overall, the model provides a fairly good approximation for the responses recorded in Ringer's solution and picrotoxin inhibition when tested with 10 e GABA, although some deviation occurred when low GABA and high picrotoxin concentrations were used.
Picrotoxinin Mediates Inhibition on Perch-1A Receptor. Picrotoxin is an equal molar mixture of picrotin and picrotoxinin. Whereas picrotoxinin is the only active component inhibiting GABAA responses, picrotin is equally effective when blocking certain types of glycine receptors (Lynch et al., 1995; Steinbach et al., 2000). Here, we have shown that picrotoxin inhibition on perch-1A receptor contains both competitive and noncompetitive components and picrotoxin is likely to bind the ligand-bound receptor in both channel open and closed states. It is possible, therefore, that each component of picrotoxin inhibition is mediated by separate compounds present in picrotoxin (i.e., picrotin and picrotoxinin). To test this hypothesis, we evaluated the effectiveness of picrotin and picrotoxinin in blocking GABA-elicited currents on oocytes expressing the perch-1A subunit under high and low GABA concentrations (Fig. 7). With low concentration of GABA (1 e), coapplication of 0.5 e picrotoxin inhibits about half of GABA-elicited current (Fig. 7A). The same amount of inhibition was observed with picrotoxinin at an equal concentration, whereas picrotin did not significantly reduce GABA-elicited current. Picrotin (20 e) was also ineffective in blocking response elicited with high concentration of GABA (5 e) (Fig. 7B). Again, only picrotoxinin showed significant inhibition of the GABA currents. The averaged values from five oocytes for reduction of the GABA-elicited currents observed in the presence of picrotoxin, picrotoxinin, and picrotin are shown in the bar graph in Fig. 7C. It is clear that the picrotoxin inhibition observed for the perch-1A receptor is solely mediated by picrotoxinin.
Discussion
The results of the present study demonstrate several unique features for picrotoxin inhibition on receptors formed by GABA subunits. A large response rebound was observed when coapplication of GABA and picrotoxin was terminated simultaneously (Fig. 1). This phenomenon is not observed with either the competitive antagonist TPMPA or the allosteric modulator zinc. In addition, among the three antagonists tested, only picrotoxin exhibited a significant capacity to alter the kinetics of the GABA-elicited responses from receptors formed by GABA subunits (Fig. 3). This finding was also true for the native GABAC receptors present on retinal bipolar cells (Fig. 4) and was observed on human 1 receptor (Goutman and Calvio, 2004). Specifically, picrotoxin reduced the time constant of the GABA-offset decay currents in a concentration-dependent manner. Accelerating the decay phase of the GABA response on postsynaptic neurons would reduce the amount of negative charge flowing through the GABA-gated channels and lead to a reduction in GABA-mediated inhibition. Therefore, the effects of picrotoxin on the kinetics of GABA-elicited responses could act synergistically with its effect on response amplitude to shape the GABA signal in neurons.
Picrotoxin inhibition of GABA receptor activity exhibits both competitive and noncompetitive mechanisms. The significant rebound in membrane currents after termination of picrotoxin and GABA coapplication is consistent with a noncompetitive (channel blocker) mechanism of picrotoxin action. The slow kinetics of GABA relaxation observed for GABAC receptors make this phenomenon more prominent (Qian et al., 1998; Qian and Ripps, 2001). A similar phenomenon was observed for the GABAC receptors present on catfish cone-driven horizontal cells (Dong and Werblin, 1996). On the other hand, the dependence of the IC50 value of picrotoxin inhibition on GABA concentration (Fig. 2A) and the increase in the EC50 of the GABA dose-response curve in the presence of picrotoxin (Fig. 2D) are indicative of a competitive mechanism for picrotoxin inhibition (Woodward et al., 1992).
To account for the properties of picrotoxin inhibition on GABAC receptors, we adopted a simple model that incorporated the binding of picrotoxin to both the open and closed states of the receptor. The parameters derived from fitting the model to the recorded data yielded values for GABA binding and gating that were similar to those determined for the human 1 receptor (Chang and Weiss, 1999). Modeling picrotoxin binding to ligand-bound receptors in both the channel open (GnR) and closed states (GnR) was necessary to provide a reasonable fit. On the other hand, the binding of picrotoxin to ligand-unbound receptors was not required. The prominent rebound after terminating the coapplication of GABA and picrotoxin (Fig. 1A) is one of the signatures of an open channel blocker. If picrotoxin acted purely as a channel blocker, coapplication of GABA with picrotoxin would significantly retard the off response (Drapeau and Legendre, 2001). However, the off response after terminating the coapplication of GABA and picrotoxin followed a similar time course as that elicited by GABA alone (Fig. 1A). In addition, modeling of picrotoxin as a pure channel blocker of the GABAC receptor failed to predict other properties of picrotoxin inhibition, especially the competitive component.
Using the values listed in Tables 1 and 2, the model in Fig. 5D provides a good estimate of the various aspects of picrotoxin inhibition of perch-1A receptors (Tables 3 and 4; Fig. 6). For example, the sensitivity (IC50 value) for picrotoxin inhibition varied with GABA concentration (Fig. 2A; Table 4), and the presence of picrotoxin shifted the GABA dose-response curve to the right and reduced the maximum response (Fig. 2D; Table 3). These results indicate that the model provides a good estimate of both the competitive and noncompetitive components for picrotoxin inhibition on GABAC receptors. Furthermore, the model predicted both the slow onset GABA response and the lack of rebound when coapplication of 1 e GABA and picrotoxin was terminated (Fig. 6, D and E), even though some deviation occurred with respect to the response amplitude in the presence of picrotoxin. Recording single channel GABAC receptor activity would provide an alternative way to test our model. However, the small single channel conductance gated by the GABAC receptor makes such a task technically difficult (Chang and Weiss, 1999).
The model also provides a good representation of the acceleration of GABA relaxation in the presence of picrotoxin. As shown in Fig. 6C, the traces predicted by the model nicely illustrate that the time course of the GABA off responses were significantly faster in the presence of picrotoxin, similar to those observed for GABAC receptors (Fig. 3). For the simple kinetic model shown in Fig. 5A, the off response time constant () exhibits the following relation: 1/ = k2 x (/( + )) (Bean, 1990). Thus, binding of picrotoxin to both GnR and GnR contributes to the acceleration of the GABA off response. In the presence of 1 e picrotoxin, there was approximately 1.83-fold change in the off rate for GnR, whereas the off rate for GnR was only 1.07 times that of the rate in Ringer. Therefore, the 2.5-fold reduction in the time constant of GABA off responses induced by 1 e picrotoxin on perch-1A receptor is mainly contributed by the enhancement of the off rate for GnR (Fig. 3B).
Our model suggested that the affinity for picrotoxin (0.15 e for GnR, and 37 e for GnR) is altered by receptor channel gating. This difference in picrotoxin affinity nicely explains the behavior of picrotoxin inhibition on GABAC receptors. Because GABAC receptors share a high degree of homology with GABAA and glycine receptors, the mechanisms of picrotoxin inhibition revealed in this study might therefore be extended to these receptors. Picrotoxin inhibition of GABAA receptors is mainly noncompetitive (Korpi et al., 2002), whereas competitive mechanisms dominate for glycine receptors (Lynch et al., 1995). These diverse actions of picrotoxin can easily be adopted in the model by altering the affinity of picrotoxin for the receptor in the open and closed states. For example, high picrotoxin affinity for channel open states of GABAA receptors will generate predominately noncompetitive inhibition, whereas high picrotoxin affinity for closed channel states of glycine receptors will exhibit mostly competitive inhibition. Therefore, the allosteric mechanism revealed in this study for picrotoxin inhibition of the GABAC receptor may be extended to the actions of picrotoxin on other ligand-gated chloride channels.
Acknowledgements
We thank Drs. Harris Ripps and Robert Paul Malchow and David Ramsey for helpful comments on this manuscript. Dr. Paul Brown (Purdue University, West Lafayette, IN) provided hybrid bass for this study. Wen Wang provided excellent technical assistance.
doi:10.1124/mol.104.003996.
References
Amin J and Weiss DS (1996) Insights into the activation mechanism of rho1 GABA receptors obtained by coexpression of wild type and activation-impaired subunits. Proc R Soc Lond Ser B Biol Sci 263: 273eC282.
Barnard EA, Skolnick P, Olsen RW, Mohler H, Sieghart W, Biggio G, Braestrup C, Bateson AN, and Langer SZ (1998) International Union of Pharmacology. XV. Subtypes of gamma-aminobutyric acidA receptors: classification on the basis of subunit structure and receptor function. Pharmacol Rev 50: 291eC313.
Bean BP (1990) ATP-activated channels in rat and bullfrog sensory neurons: concentration dependence and kinetics. J Neurosci 10: 1eC10.
Bormann J (1988) Electrophysiology of GABAA and GABAB receptor subtypes. Trends Neurosci 11: 112eC116.
Chang Y, Amin J, and Weiss DS (1995) Zinc is a mixed antagonist of homomeric rho 1 gamma-aminobutyric acid-activated channels. Mol Pharmacol 47: 595eC602.
Chang Y and Weiss DS (1999) Channel opening locks agonist onto the GABAC receptor. Nat Neurosci 2: 219eC225.
Cleland TA (1996) Inhibitory glutamate receptor channels. Mol Neurobiol 13: 97eC136.
Colquhoun D and Hawkes AG (1981) On the stochastic properties of single ion channels. Proc R Soc Lond B Biol Sci 211: 205eC235.
Costa E (1998) From GABAA receptor diversity emerges a unified vision of GABAergic inhibition. Annu Rev Pharmacol Toxicol 38: 321eC350.
Dibas MI, Gonzales EB, Das P, Bell-Horner CL, and Dillon GH (2002) Identification of a novel residue within the second transmembrane domain that confers usefacilitated block by picrotoxin in glycine 1 receptors. J Biol Chem 277: 9112eC9117.
Dong CJ and Werblin FS (1996) Use-dependent and use-independent blocking actions of picrotoxin and zinc at the GABAC receptor in retinal horizontal cells. Vision Res 36: 3997eC4005.
Drapeau P and Legendre P (2001) Neuromuscular transmission on the rebound. Recept Channels 7: 491eC496.
Goutman JD and Calvo DJ (2004) Studies on the mechanisms of action of picrotoxin, quercetin and pregnanolone at the GABA 1 receptor. Br J Pharmacol 141: 717eC727.
Gurley D, Amin J, Ross PC, Weiss DS, and White G (1995) Point mutations in the M2 region of the , , or subunit of the GABAA channel that abolish block by picrotoxin. Recept Channels 3: 13eC20.
Hamill OP, Marty A, Neher E, Sakmann B, and Sigworth FJ (1981) Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflueg Arch Eur J Physiol 391: 85eC100.
Inoue M and Akaike N (1988) Blockade of -aminobutyric acid-gated chloride current in frog sensory neurons by picrotoxin. Neurosci Res 5: 380eC394.
Korpi ER, Grunder G, and Luddens H (2002) Drug interactions at GABAA receptors. Prog Neurobiol 67: 113eC159.
Luddens H, Korpi ER, and Seeburg PH (1995) GABAA/benzodiazepine receptor heterogeneity: neurophysiological implications. Neuropharmacology 34: 245eC254.
Lynch JW, Rajendra S, Barry PH, and Schofield PR (1995) Mutations affecting the glycine receptor agonist transduction mechanism convert the competitive antagonist, picrotoxin, into an allosteric potentiator. J Biol Chem 270: 13799eC13806.
Macdonald RL and Olsen RW (1994) GABAA receptor channels. Annu Rev Neurosci 17: 569eC602.
Magoski NS and Bulloch AG (1999) Dopamine activates two different receptors to produce variability in sign at an identified synapse. J Neurophysiol 81: 1330eC1340.
Moss SJ and Smart TG (2001) Constructing inhibitory synapses. Nat Rev Neurosci 2: 240eC250.
Newland CF and Cull-Candy SG (1992) On the mechanism of action of picrotoxin on GABA receptor channels in dissociated sympathetic neurones of the rat. J Physiol (Lond) 447: 191eC213.
Pribilla I, Takagi T, Langosch D, Bormann J, and Betz H (1992) The atypical M2 segment of the subunit confers picrotoxinin resistance to inhibitory glycine receptor channels. EMBO (Eur Mol Biol Organ) J 11: 4305eC4311.
Qian H and Dowling JE (1995) GABAA and GABAC receptors on hybrid bass retinal bipolar cells. J Neurophysiol 74: 1920eC1928.
Qian H, Dowling JE, and Ripps H (1998) Molecular and pharmacological properties of GABA- subunits from white perch retina. J Neurobiol 37: 305eC320.
Qian H and Dowling JE (1994) Pharmacology of novel GABA receptors found on rod horizontal cells of the white perch retina. J Neurosci 14: 4299eC4307.
Qian H, Hyatt G, Schanzer A, Hazra R, Hackam AS, Cutting GR, and Dowling JE (1997) A comparison of GABAC and rho subunit receptors from the white perch retina. Vis Neurosci 14: 843eC851.
Qian H and Pan Y (2002) Co-assembly of GABA subunits with the GABAA receptor 2 subunit cloned from white perch retina. Brain Res Mol Brain Res 103: 62eC70.
Qian H and Ripps H (1999) Response kinetics and pharmacological properties of heteromeric receptors formed by coassembly of GABA - and 2-subunits. Proc R Soc Lond B Biol Sci 266: 2419eC2425.
Qian H and Ripps H (2001) The GABAC receptors of retinal neurons, in Progress in Brain Research (Kolb H, Ripps H, and Wu S eds) pp 295eC308, Elsevier Science B.V., Amsterdam, The Netherlands.
Ragozzino D, Woodward RM, Murata Y, Eusebi F, Overman LE, and Miledi R (1996) Design and in vitro pharmacology of a selective -aminobutyric acidC receptor antagonist. Mol Pharmacol 50: 1024eC1030.
Reddy GL, Iwamoto T, Tomich JM, and Montal M (1993) Synthetic peptides and four-helix bundle proteins as model systems for the pore-forming structure of channel proteins. II. Transmembrane segment M2 of the brain glycine receptor is a plausible candidate for the pore-lining structure. J Biol Chem 268: 14608eC14615.
Smart TG and Constanti A (1986) Studies on the mechanism of action of picrotoxinin and other convulsants at the crustacean muscle GABA receptor. Proc R Soc Lond Ser B Biol Sci 227: 191eC216.
Steinbach JH, Bracamontes J, Yu L, Zhang P, and Covey DF (2000) Subunit-specific action of an anticonvulsant thiobutyrolactone on recombinant glycine receptors involves a residue in the M2 membrane-spanning region. Mol Pharmacol 58: 11eC17.
Wang TL, Hackam A, Guggino WB, and Cutting GR (1995a) A single histidine residue is essential for zinc inhibition of GABA rho 1 receptors. J Neurosci 15: 7684eC7691.
Wang TL, Hackam AS, Guggino WB, and Cutting GR (1995b) A single amino acid in -aminobutyric acid 1 receptors affects competitive and noncompetitive components of picrotoxin inhibition. Proc Natl Acad Sci USA 92: 11751eC11755.
Woodward RM, Polenzani L, and Miledi R (1992) Characterization of bicuculline/baclofen-insensitive -aminobutyric acid receptors expressed in Xenopus oocytes. I. Effects of CleC channel inhibitors. Mol Pharmacol 42: 165eC173.
Yarowsky J and Carpenter DO (1978) A comparison of similar ionic responses to -aminobutyric acid and acetylcholine. J Neurophysiol 41: 531eC541.
Zhang D, Pan ZH, Zhang X, Brideau AD, and Lipton SA (1995) Cloning of a -aminobutyric acid type C receptor subunit in rat retina with a methionine residue critical for picrotoxinin channel block. Proc Natl Acad Sci USA 92: 11756eC11760.
Zhorov BS and Bregestovski PD (2000) Chloride channels of glycine and GABA receptors with blockers: Monte Carlo minimization and structure-activity relationships. Biophys J 78: 1786eC1803.