Effects of Ginsenoside Rg₂ on Human Neuronal Nicotinic Acetylcholine Receptors

Materials and Methods

Oocyte Expression.

cDNAs of human neuronal nicotinic acetylcholine receptor subunits (α₇, α₃, α₄, β₂, and β₄) were inserted into the pSP64T vector (Krieg and Melton, 1984). Capped mRNA was synthesized in vitro using SP6 RNA polymerase. DefolliculatedXenopus oocytes were injected with 5 ng of total RNA in 50 nl of sterile water. The two subunits of heteromeric receptors were injected in an equimolar ratio. All measurements were made within 3 to 6 days after injection.

Electrophysiological Recordings.

Two-electrode voltage-clamp recordings were obtained as described previously (Stühmer, 1998) using a standard two-electrode voltage-clamp amplifier (Oocyte Clamp OC-725B; Warner Instrument, Hamden, CT). Oocytes were perfused (10–12 ml/min) with a modified frog Ringer's solution containing 82.5 mM NaCl, 2.5 mM KCl, 2.5 mM BaCl₂, 1 mM MgCl₂, and 5 mM HEPES, pH 7.4. With barium as the permeant divalent cation no calcium-activated chloride currents were detected, because no significant difference was found when BAPTA was injected into the oocytes. Holding potential was −80 mV unless otherwise indicated. Currents were filtered at 50 Hz with a low pass eight-pole Bessel filter (Frequency Devices, Haverhill, MA), sampled at 100 to 500 Hz and stored on hard disk for later analysis.

Low-resistance electrodes, filled with 2 M KCl, were used to get good voltage-clamp recordings, even with currents as large as 50 μA. The quality of the voltage clamp was always assessed by monitoring the output of the voltage electrode in an oscilloscope, and only the current records for which the voltage was completely flat were stored. Data acquisition and agonist application were controlled by a DigiData 1200 interface driven by pClamp 6.0 software (Axon Instruments, Union City, CA).

Rg₂ powder, obtained from Korea Ginseng and Tobacco Research Institute (Taejon, Korea) with a purity greater than 98%, was dissolved in dimethyl sulfoxide as a stock solution of 200 mM. Solutions containing ACh were freshly prepared every day directly in the recording solution. Control experiments showed that dimethyl sulfoxide alone, at the levels present in the solutions containing Rg₂ did not affect the ACh-induced currents.

Drug application was controlled by an electromechanical valve, by which the flow was fed under gravity through a 1-mm-diameter tube positioned very close to the oocyte. This system allowed a rapid application of the drugs, but in many cases, the impetus of the flow mechanically deformed the oocyte and gave rise to some irregularities in the current traces (Fig. 2). Solution exchange time, as assessed by changes in junction potentials of the voltage electrode in control current-clamp experiments, was in the range of 0.1 and 0.3 s (activation time measured as the time from 20 to 80% of the peak current in α₇ receptors was less than 75 ms). However, our system did not allow us to achieve a fast removal of the applied drugs, which was probably due to the large volume, approximately 1 ml, of the recording chamber. This prevented us from being able to measure reliably the closing kinetics of the currents.

We usually applied the drugs only for 2 s because this period of time was usually enough to reach the peak current in control conditions even for the slower heteromeric receptors (Figs. 1 and2). Also, for this time scale the desensitization of the heteromeric receptors, with the exception of α₃β₂, was very small or nonexistent, thus allowing to see more clearly the effect of Rg₂ on the kinetics of the currents.

Analysis.

Desensitization of currents for heteromeric receptors was quantified as the quotient of the current after 2 s of the pulse start to the peak current. Dose-response curves were fitted with the Hill equation using nonlinear fitting routines in Origin 5.0 (Microcal Software, Inc., Northampton, MA).

Averaged data are presented as means ± S.E.M. (number of cases). Statistical calculations were done with Statistica 5.0 (StatSoft, Tulsa, OK). Statistical significance is indicated by * (p < 0.05), ** (p < 0.01), or *** (p < 0.001).

Results

Rg₂ Affects Ionic Currents through Heteromeric Human Nicotinic Acetylcholine Receptors.

Application of 1 mM ACh to −80 mV voltage-clamped Xenopus oocytes expressing human neuronal nicotinic acetylcholine receptors elicits inward currents of different magnitude and kinetics depending on the particular receptor (Fig. 1). Homomeric α₇ receptors produce the fastest activating and inactivating currents. On the other hand, heteromeric α₃β₄, α₄β₄, α₄β₂, and α₃β₂ receptors produce slowly desensitizing currents. Although the rate in desensitization is variable, usually heteromeric α₃β₂ receptors produce faster desensitizing currents. In a smaller time scale it can be seen that the currents reach their peak values usually before 2 s, and that, for the heteromeric receptors, 2 s after the application of ACh the currents have decayed very little. Therefore, we decided to use pulses of ACh of 2 s because this was a convenient short duration where the peak of the current could still be reached and, except for α₃β₂, very little decay of the current had yet developed. On the other hand, recovery from desensitization was complete after 5 min of washing, and for that reason we always spaced by 5 min the application of agonist pulses.

Figure 2 shows currents elicited by 2-s pulses of 1 mM ACh. Coapplication of 100 μM Rg₂ did not produce any effect on the size or kinetics of the currents through α₇ receptors. In contrast, the effect of Rg₂ on the magnitude of the peak currents through heteromeric receptors was rather small, but the degree of desensitization was increased. This last effect was especially evident on the currents of receptors that desensitize very slowly in control conditions, i.e., in α₃β₄ and α₄β₂ receptors. The effect was completely reversible, and after the usual 5-min washing period the control current was completely recovered. Similar results were obtained with nicotine (data not shown). Thus, after these first results we decided to investigate the effect of Rg₂ only on heteromeric receptors.

Coapplication of Rg₂ Affects Mainly Desensitization of Receptors.

The statistics of the effect of coapplication of 100 μM Rg₂ on the peak current and desensitization proportion are shown in Fig. 2. Peak currents obtained in control conditions, i.e., 1 mM ACh and −80 mV membrane voltage, were measured 5 min before and 5 min after the coapplication of 1 mM ACh and 100 μM Rg₂, and the interpolated value of the peak control currents was used to normalize the peak current obtained in the presence of Rg₂ (Fig.3A). Coapplication of Rg₂ has no effect on the peak currents of α₇ receptors, and the effect on the peak currents of all heteromeric receptors, although statistically significant, is small, less than 30% reduction.

The desensitization of the heteromeric receptors, quantified as the quotient between the current level at the end of a 2-s pulse and the peak current,I_@2s/I_pk, is however strongly affected by the coapplication of Rg₂. In control conditions the currents of α₃β₄ and α₄β₂ hardly desensitize after 2 s (Fig. 2), andI_@2s/I_pkis close to 93 and 82%, respectively (Fig. 3B). However the coapplication of Rg₂ increases the desensitization, givingI_@2s/I_pkvalues close to 50 and 40%, respectively. α₄β₄ receptors usually gave less steady control currents during the pulse (Fig. 2), but statistically the desensitization at the end of the pulse was rather small, with anI_@2s/I_pkvalue close to 83% and the effect of Rg₂ on the desensitization of these receptors was very large, givingI_@2s/I_pkclose to 18%. Finally, the currents through α₃β₂ receptors are the ones that usually desensitize faster under control conditions, and are also strongly affected by the coapplication of Rg₂, givingI_@2s/I_pkclose to 52 and 13%, respectively.

Effect of Coapplication of Rg₂ on Desensitization Is Dose-Dependent.

Figure 4 shows dose-response curves for the effect of coapplication of Rg₂ on the desensitization of the different receptors. Data obtained at −80 or at −40 mV have been pooled. Data points have been fitted with the Hill equation with they_max fixed to the value of desensitization obtained in control conditions. Only α₃β₄ receptors have IC₅₀ greater than 100 μM, the maximum concentration of Rg₂ measured, whereas the other receptors have IC₅₀ values around 20 or 40 μM. The Hill coefficients are less than 1 for all receptors except for α₃β₂, the receptor that has the faster desensitization in control conditions. Insets in Fig. 4show representative records normalized to the peak value to show the kinetics of the currents with 1, 10, and 100 μM Rg₂ coapplication, where the dose dependence is more apparent.

Preapplication of Rg₂ Decreases Peak Currents in a Dose-Dependent Manner.

The effect of coapplication of 100 μM Rg₂ with 1 mM ACh on the peak current, although significant, was small for all the receptors, thus the dose-response curves, measured with concentrations from 0.1 to 100 μM Rg₂ show little concentration dependence (Fig.5, open circles, dotted lines). However, when Rg₂ was preapplied for 5 min before the pulse of agonist, a significant dose-dependent effect was observed, (Fig. 5, solid circles, solid lines). The dependence is not very steep, only α₄β₄ receptors have an absolute value of the n of Hill greater than 1. Preapplications of 5 min 100 μM Rg₂ produced a reduction of the peak current of more than 85% in α₄β₄, α₃β₂, and α₄β₂ receptors, but only a reduction of about 40% in α₃β₄ receptors. Insets in Fig. 5 show representative current traces for 5-min preapplication of Rg₂ at concentrations of 1, 10, and 100 μM, where the dose dependence is more clearly seen. Preapplication of 10 μM Rg₂ also produced a small decrease in the peak current of α₇ receptors of approximately 10% (data not shown), compared with reductions of 15, 30, 50, and 60% for α₃β₄, α₄β₄, α₃β₂, and α₄β₂, respectively.

Effect of Rg₂ on ACh Dose-Response Curves of α₄β₂ and α₃β₄Receptors.

To further investigate the mechanism by which Rg₂ inhibits nAChRs we have chosen to pursue our study in more detail with α₄β₂ and α₃β₄ receptors, the most abundant in the central and peripheral nervous system, respectively.

We obtained ACh dose-response curves, measured at the peak of the currents, with and without coapplication of 10 μM Rg₂ for α₄β₂ and α₃β₄ receptors, as shown in Fig. 6, top. In these experiments, a smaller concentration of 10 μM Rg₂ was chosen instead of 100 μM because, as indicated in Fig. 5, for α₄β₂ and α₃β₄ receptors the same reduction of the peak current is obtained with either 10 or 100 μM Rg₂ coapplication. Dose-response data were fitted with the Hill equation with an inverse error weight-fitting routine. In both receptors, Rg₂ coapplication produced a reduction of the maximum response of a little more than 10%, without affecting much neither the EC₅₀ nor then_H.

For α₄β₂ receptors data points obtained at low concentrations of ACh suggest the presence of a high-affinity site (Covernton and Connolly, 2000), although not enough data were obtained to perform a two-site fit.

Effect of Rg₂ on Desensitization at Different ACh Concentrations.

The effect of coapplication of Rg₂ on the desensitization of the currents was also dependent on the ACh concentration used to elicit the current as shown in Fig. 6, bottom. Although in control conditions no dependence of desensitization on the ACh concentration was seen, in the presence of 10 μM Rg₂ the level of desensitization was almost constant for low concentrations of ACh, but then started to increase with the ACh concentration. Insets in Fig. 6show representative currents for α₄β₂ and α₃β₄ receptors, normalized at their peak values to compare the kinetics, at three ACh concentrations, 100, 300, and 1000 μM with 10 μM Rg₂ coapplication, where the concentration dependence of the desensitization is apparent.

Voltage Dependence of Rg₂ Effect.

To further investigate the mechanism of action of Rg₂ we studied the voltage dependence of its effect both on the peak currents and on the desensitization. Figure 7, top, shows peak current-voltage relationships for α₄β₂ and α₃β₄ receptors obtained with 1 mM ACh, with or without 100 μM Rg₂coapplication. The effect of Rg₂ coapplication is small, although the relative reduction of the peak current seems to be larger at more depolarized potentials. See also insets in Fig. 7 where representative records, obtained at −100 and +40 mV are shown for α₄β₂ and α₃β₄ receptors both in the absence and in the presence of Rg₂coapplication.

Figure 7, bottom, shows the voltage dependence of desensitization for α₄β₂ and α₃β₄ receptors both in control conditions and in the presence of 100 μM Rg₂ coapplication. For α₄β₂ receptors the linear fit to the data points indicates that there is no significant voltage dependence, neither in control nor with Rg₂ coapplication conditions, because in both cases the slope of the fitted straight lines is not statistically different from zero. For α₃β₄ receptors, however, the voltage dependence of desensitization can be fitted with a straight line with a very statistically significant nonzero slope, both in control (p < 0.05) and with Rg₂ coapplication (p < 0.001) conditions. The desensitization increases at more depolarized potentials and this trend is increased by the presence of Rg₂.

Discussion

In this work we have studied the effect of Rg₂, a major active component of ginseng, on the function of human nAChR expressed in Xenopus oocytes.

The function of the receptors was assessed by measuring the ionic current elicited by ACh pulses in voltage-clamp conditions. The general characteristics of the human nAChR currents recorded inXenopus oocytes, i.e., the size of the peak current and the kinetics of activation and desensitization measured in control conditions, were similar to those reported by Chavez-Noriega et al. (1997). With the short pulse used in this work, our current readings usually show complex kinetics. This is probably due to a mixture of perfusion artifacts and an irregular distribution of receptors on the oocyte surface, which result in nonsynchronized activation.

Coapplication of 100 μM Rg₂ produced a small reduction of peak current in all the heteromeric receptors investigated but not in homomeric α₇ receptors. A similar differential blockade by ω-conotoxin MVIIC, ω-conotoxin GVIA, and diltiazem has been reported for rat α₃β₄ and α₇ neuronal nicotinic receptors (Herrero et al., 1999).

The effect of Rg₂ on the peak currents was small and not very dose-dependent when coapplied; however, the main effect of Rg₂ coapplication was to increase the rate of decay of the currents of heteromeric receptors. This result could indicate an increase of the desensitization rate or could simply reflect the progression of the channel block by Rg₂.

Desensitization is a general property of ligand-gated ion channels and it has been known that the subunit composition of heteromeric nAChRs determines the degree of desensitization. This has also been shown to be the case for human nAChRs (Gerzanich et al., 1998). The factors by which desensitization is modulated include receptor phosphorylation, calcium, and noncompetitive blockers (Ochoa et al., 1989), and although the molecular mechanism of desensitization is not known it has been shown recently that some regions of the N-terminal domain of the β₂ subunit determine the rate of desensitization (Bohler et al., 2001). The modulatory effect of Rg₂ on desensitization was dose-dependent, and could not be easily related to the presence of the β₂ subunit. Thus, although IC₅₀ was smaller andn_H was larger for α₃β₂ compared with α₃β₄, the reverse was true for α₄β₂ compared with α₄β₄.

A lack of dose dependence for the reduction of the peak currents together with a clear dose dependence of the value of desensitization could suggest that the mechanism of action of Rg₂could be similar to a slow open channel blocker, so that the current would have time to fully activate before Rg₂progressively blocks the channel. However as discussed below, Rg₂ did not behave as a typical open channel blocker (Buisson and Bertrand, 1998).

Contrasting with the lack of dose dependence of the effect of Rg₂ on the peak currents when coapplied, the reduction of the peak currents was clearly dose-dependent when preincubated for 5 min, reaching 90% with 100 μM, the largest concentration applied. These results suggest that Rg₂ is acting on resting state of the receptor, either stabilizing it or promoting desensitization from it. The receptors with the β₂ subunit had smaller IC₅₀ values than the corresponding receptors with the β₄ subunits. However, the differences are small, and taking into account the statistical errors of the fit, those values overlap considerably.

On the other hand, the receptors with the β₂subunit also had more shallow inhibition curves with Hill coefficients less than 1. In our case, this cannot be due to heterogeneity in the receptor population, thus suggesting that more than one molecule of Rg₂ binds to those receptors with negative cooperativity.

In general, the concentrations at which Rg₂affects these human neuronal nicotinic receptors was in the low micromolar range. These concentrations are similar to the EC₅₀ value (around 30 μM) for the increase of Ca²⁺ activated Cl⁻currents in Xenopus oocytes by Rg₂(Choi et al., 2001). Also, the concentrations of Rg₂ that reduce ACh-evoked Ca²⁺ and Na⁺ influx in bovine chromaffin cells are in the same range (Tachikawa et al., 1995). Other ginsenosides, such as Rf for example, also affect Ca²⁺ channels with an IC₅₀of 40 μM approximately (Nah et al., 1995). Similarly, 1 mg/ml American ginseng extract containing a mixture of ginsenosides at a concentration between 1 and 50 μM also blocks brain Na⁺ channels.

We studied the mechanism by which Rg₂ inhibits nAChRs, measuring the effect of 10 μM Rg₂ on different concentrations of ACh for α₄β₂ and α₃β₄ receptors. Dose-response curves obtained in control conditions were similar to those reported previously (Chavez-Noriega et al., 1997) but, in control conditions the desensitization did not depend much on the agonist concentration. In both receptors, a negligible shift in the ACh dose-response curves was produced, the inhibition of the current was not relieved at high ACh concentrations, and the increase produced by Rg₂ on desensitization was larger at high ACh concentrations. All these results suggest a noncompetitive mechanism with no subtype selectivity. The effect of the noncompetitive antagonists mecamylamine and d-tubocurarine, on agonist-stimulated intracellular calcium concentration elevations has also been shown to be nonselective between α₂β₄, α₃β₄, and α₄β₄ receptors. In contrast, the effect of the competitive antagonist dihydro-β-erythroidine was highest with α₃β₄ receptors expressed in human embryonic kidney 293 cells (Stauderman et al., 1998). Other drugs have an action mechanism dependent on the receptor type (Chiodini et al., 2001).

Finally, we investigated the possibility that Rg₂could act as an open channel blocker by measuring its effect on current-voltage relationships. Current-voltage relationships show a strong inward rectification for both receptors either in control conditions or in the presence of 100 μM Rg₂. It has been shown that ginseng extracts and ginsenoside Rb1 block Na⁺ channels in a voltage-dependent manner (Liu et al., 2001). In contrast, Rg₂ reduction of peak currents of α₄β₂, and α₃β₄ receptors does not show a substantial voltage dependence, thus suggesting that Rg₂ does not act as an open channel blocker. On the other hand, we found that the voltage dependence of the desensitization, and its modification by Rg₂coapplication, were different for α₄β₂ and α₃β₄ receptors. Thus, although α₄β₂ did not show any appreciable voltage dependence of desensitization neither in control conditions nor in the presence of 100 μM Rg₂, the currents of α₃β₄ receptors desensitized faster at more depolarized potentials and this dependence was increased by the presence of Rg₂. This is different from the behavior of rat α₃β₄ receptors stably transfected in human embryonic kidney cells (Zhang et al., 1999) where the time constant of desensitization of nicotine-activated current was concentration-dependent, but not voltage-dependent. Unfortunately, we were unable to wash out the drugs fast enough to be able to resolve tail currents that would have provided information about the recovery from a possible open channel block.

In this work, we have used the term desensitization to indicate the reduction of the current during steady application of ACh, either alone or together with Rg₂. No mechanistic implications were intended. It is possible that during coapplication of Rg₂ the “intrinsic” mechanism of the desensitization of the receptors is proceeding normally, i.e., is not being affected by Rg₂. However, the kinetics of onset of block by Rg₂ produces an increase in the decay rate of the current that is seen as an “apparent” increase in desensitization rate. As discussed above, our data do not allow us to rule out this last possibility.

In conclusion, our results show that ginsenoside Rg₂ has several effects on human heteromeric nAChR, principally altering its desensitization characteristics, while leaving unaffected homomeric α₇ receptors. These differential inhibitory effects of one of the main components of ginseng could be one of the mechanisms of its many pharmacological and therapeutic properties.