Dual Action of n-Butanol on Neuronal Nicotinic α4β2 Acetylcholine Receptors

Materials and Methods

Cell Preparations.

Human α4β2 AChR subunit combination was stably expressed in the HEK293 cell line. Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 2 mMl-glutamine, 100 U of penicillin, 100 μg of streptomycin (Invitrogen, Carlsbad, CA), 6% iron-supplemented calf serum (Sigma-Aldrich, St. Louis, MO), and 100 μg/ml G418 (Mediatech, Herndon, VA). Cells were kept at 34.7°C in air + CO₂ (93 + 7%, by volume). For patch-clamp experiments, cells were plated on glass coverslips coated with poly-l-lysine and cultured for 1 to 5 days.

Electrophysiological Recording.

Whole-cell currents were recorded with an Axopatch 200 patch-clamp amplifier (Axon Instruments, Inc., Foster City, CA) at room temperature (20–25°C). Recorded currents were directly digitized at 1 to 10 kHz via a Digidata 1200 ADC/DAC interfaced with a microcomputer under the control of the ClampEx module of the pClamp6 software package (Axon Instruments, Inc.). The holding potential was −50 mV. The external solution contained 150 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 1 mM MgCl₂, 10 mM glucose, 5.5 mM HEPES acid, and 4.5 mM Na-HEPES, pH adjusted to 7.3 with HCl, and osmolarity 320 mOs. The internal solution contained 140 mM K-gluconate, 2 mM MgCl₂, 1 mM CaCl₂, 11 mM EGTA, 10 mM HEPES acid, 2 mM Mg²⁺ ATP, and 0.2 mM Na⁺ GTP, pH adjusted to 7.3 with KOH, and osmolarity 300 mOs. The patch-clamp pipettes were pulled from Clark patch glass capillaries (PG120T-10, 1.2-mm o.d., 0.93-mm i.d., 10-cm length; Warner Instrument, Hamden, CT) in two stages on a vertical pipette puller (PP-830; Narishige, Tokyo, Japan) and lightly fire-polished to a final resistance of 1.5 to 2 MΩ when filled with internal solution. Recording was started about 5 to 10 min after the rupture of the cell membrane to allow for the internal solution to reach the equilibration with intracellular milieu.

Drug Application.

All drugs were applied to the cell by a modified computer-operated U-tube perfusion system (Marszalec and Narahashi, 1993). The speed of solution exchange was measured by the junction potential change using the patch pipette placed at a distance from the opening of the U-tube or near the U-tube opening. The rise time (10–90%) of junction potential change was found to be 50 ms when the patch pipette was placed 200 μm from the opening of the U-tube and was reduced to less than 10 ms when the patch pipette was placed near the opening of the U-tube. When the recording cell was placed at 200 μm from the opening of the U-tube, the solution exchange on the cell surface was, however, found to complete within around 200 ms by measuring the rate of changes in ACh-induced current in response to changes in sodium ion concentration (Liu and Dilger, 1991; Mori et al., 2001). In a few experiments, the recording cell was brought near the opening of the U-tube. The rise time of the ACh response was less than 10 ms. In the present experiments, short-pulse drug exposure time was 250 ms unless otherwise stated, and long-pulse experiments were performed with 1 to 10 s pulses. In all these cases, the intervals between pulses were 2 min to avoid current reduction due to agonist-induced desensitization. Control currents (30 μM ACh alone) were usually checked before and after each application of a test drug to ensure the stability of control responses. n-Butanol was purchased from Aldrich Chemical Co. (Milwaukee, WI). DHβE and mecamylamine were purchased from Sigma/RBI (Natick, MA).

Data Analysis.

Recorded currents were initially analyzed by the Clamp-Fit module of the pClamp6 to assess whole-cell current amplitudes and decay kinetics. Data were then exported from pClamp6 to a Microsoft Excel program (Microsoft Office 2000) for statistical analysis. The concentration-response data were subsequently fitted to a single or double Hill logistic equations and compiled for graphical analysis in SigmaPlot 5.0 (SPSS Science, Inc., Chicago, IL). Data were expressed as mean ± S.E.M. unless otherwise mentioned.

The ACh dose-response curve was fitted by a single Hill equation or by the sum of two Hill equations:where y is the normalized peak currents, andx is the agonist concentration. EC_50Hand EC_50L are the half-effective concentrations for the high- and low-affinity receptors, respectively.n_H1 andn_H2 are the Hill coefficients of the high- and low-affinity receptors, respectively, andp₁ is the percentage of receptors in the high-affinity state.

Simulation.

The kinetic simulation was carried out with C⁺⁺ programs for numerical solution of differential equations based on a kinetic scheme (Scheme I in Fig. 9). Butanol is assumed to have two kinds of action on nnAChRs, as a partial agonist and as a channel blocker.

Results

Both ACh and n-Butanol Activate nnAChR Currents.

The currents of the α4β2 AChRs expressed in HEK cells rose rapidly, reaching a peak in less than 100 ms at low ACh concentrations, and its rising phase became faster with increasing ACh concentrations. At 3 mM, the time to peak is less than 10 ms (Figs.1A and 2A). During application of 10-s pulse of 3 mM ACh, the ACh-induced currents decayed with two exponential phases (Fig. 1A), reflecting receptor desensitization. The time constants (τ) of fast and slow desensitization were 350 ± 30 ms (τ_fast) and 2270 ± 70 ms (τ_slow), respectively (n = 7). About 22% of the current was associated with the fast component, whereas the rest was associated with the slow one.

When the dose-response relationship was fitted to a single Hill equation (Fig. 1C, dotted line), an EC₅₀ value of 68 μM and an n_H of 0.7 ± 0.1 were obtained (n = 7–28). This EC₅₀ value is much higher than that measured on the α-bungarotoxin-insensitive nnAChRs in rat cortical neurons (EC₅₀ of 2.7 μM; Aistrup et al., 1999). The high EC₅₀ value and a smalln_H of the ACh dose-response relationship for the expressed α4β2 nnAChRs has also been reported previously (Chavez-Noriega et al., 1997; Zuo et al., 2001; Buisson and Bertrand, 2001).

One possibility of the low n_H for the ACh dose-response curve is due to the existence of two receptor pools with different affinities for ACh, as suggested by Buisson and Bertrand (2001). To examine this possibility, additional experiments covering a wider range of ACh concentrations were performed and the dose-response relationship was fitted with the sum of two Hill equations. In this case, 20% (p₁) of the receptors have a high affinity for ACh, with an EC_50H value of 1.0 ± 2.5 μM and an n_H1 of 0.8 ± 0.7, whereas the remaining 80% receptors have a low affinity, with an EC_50L value of 63 ± 14 μM and an n_H2 value of 1.3 ± 0.2 (n = 7–28) (Fig. 1C, solid line). The large standard deviations for the EC_50L andn_H2 for the high-affinity receptor are probably due to the fact that it constitutes a minor component of the total responses. These results resemble those of Buisson and Bertrand (2001) in that high-affinity receptors have an EC_50H value of 1.60 μM and ann_H1 of 0.92, whereas the low-affinity receptors have an EC_50L value of 68 μM and ann_H2 of 1.60.

To test an alternative hypothesis that the small Hill coefficient is due to rapid receptor desensitization, an improvement in our perfusion system was made by bringing a smaller cell to the site close to the opening of the U-tube so that the solution exchange was greatly speeded up to less than 10 ms. Figure 2 depicts examples of current traces and the dose-response relationship. Even at low ACh concentrations, the ACh-induced currents reached the peak in less than 20 ms. The decay time constant estimated from the current induced by 3 mM ACh was 310 ± 40 ms (n = 5). No decay time constant faster than 100 ms was detected. When the data were fitted to the single Hill equation, the EC₅₀ andn_H were estimated to be 38 ± 9 μM and 0.8 ± 0.2, respectively (Fig. 2B, dotted line). Thus, it is rather unlikely that the small n_His due to the complication from the receptor desensitization because the peak of ACh current was measured at less than 20 ms and because the receptor underwent desensitization with a time constant of more than 300 ms. However, when one assumed 80% of the data coming from the low-affinity receptor with 20% from the high-affinity receptor, an EC₅₀ and an n_Hof 60 ± 13 μM and 1.3 ± 0.3, respectively, were obtained for the low-affinity receptor (Fig. 2B, solid line). The 20% high-affinity receptor was based on the analysis of Fig. 1C. To simplify the simulation, we simulated the effect of butanol only on the low-affinity receptor, because the low-affinity receptor represents the majority of the receptors in our cell lines, and also because most of the experiments were performed at high concentrations of ACh (>10 μM).

In the absence of ACh, application of long pulses (2–5 s) ofn-butanol generated inward currents in a dose-dependent manner at a holding potential of −50 mV (Fig.3A). The currents induced byn-butanol were small compared with ACh-induced currents. For example, the currents induced by 300 mM butanol ranged from 40 to 80% of the currents induced by 30 μM ACh. Unlike ACh-induced currents, the butanol-induced currents decayed with a single exponential time course (Fig. 3A). For currents induced by 300 mM butanol, the time constant of decay was around 2 s. A logistic equation with three parameters was used to fit the butanol data, giving an EC₅₀ value of 230 ± 90 mM, ann_H of 1.8 ± 0.4, and the saturation currents estimated to be 1.6 ± 0.5 times that of the currents induced by 300 mM butanol (n = 6) (Fig. 3B). The butanol-induced current had a much slower onset (200–300 ms) compared with the ACh-induced current, suggesting either a lower binding rate of butanol to nnAChRs or a slower open rate of butanol-bound receptors. Furthermore, no current was evoked by butanol in the cells that failed to respond to ACh. These results suggest thatn-butanol is capable of activating nnAChRs albeit with less potency and less efficacy than ACh.

To further test the hypothesis that butanol-induced current was due to the activation of nnAChRs, we tested whether specific AChR inhibitors could block the currents. DHβE has been widely used as a nonselective and competitive nAChR antagonist. It has a nanomolar affinity for α4β2 nnAChRs (Dwoskin and Crooks, 2001). The previous study of human α4β2 nnAChRs showed that the inhibitory action of mecamylamine is voltage-dependent and noncompetitive, suggesting that it acts as an open channel blocker (Papke et al., 2001). Both 1 μM DHβE (ACh receptor blocker) and 50 μM mecamylamine (ACh channel blocker) blocked the currents induced by either 300 mM butanol or 30 μM ACh when preperfused in the bath for 2 min and coapplied with the agonist for 2 or 5 s (Fig. 4). Both DHβE and mecamylamine blocked the peak ACh- and butanol-induced currents almost to the same extent (Table 1).

The blocking kinetics in the presence of DHβE and mecamylamine were significantly different between the ACh-opened (Fig. 4, A and B) and the butanol-opened channels (Fig. 4, C and D). In the presence of 1 μM DHβE, the butanol-induced current rose more slowly than the butanol control current (Fig. 4C), whereas the ACh-induced current was simply scaled down without changing kinetics (Fig. 4A). When the block was measured at the end of agonist pulse, ACh-induced currents were more sensitive than butanol-induced currents to both blockers (Table1). DHβE at 1 μM reduced butanol-induced currents at the end of agonist pulse to 33 ± 1% (n = 3) of control, while reducing ACh-induced currents to 12 ± 3% (n = 3) of control. Similarly, 50 μM mecamylamine reduced butanol-induced currents to 29 ± 2% (n = 3) of control compared with 1.0 ± 0.4% (n = 5) for ACh-induced currents (Table 1). The slow rise in the butanol-induced current in the presence of DHβE is consistent with the notion that butanol at high concentrations might compete against DHβE for binding to the receptors. Thus, the DHβE-bound receptors become unblocked as more butanol molecules competitively bind to the receptors to open the channel. Without doing more detailed Schild analysis of competition between butanol and DHβE, one could not rule out an alternative explanation, namely, a different conformational change of the receptor occurs in the presence of butanol and DHβE. Mecamylamine blocked the ACh-induced current in a time-dependent manner (Fig. 4B), leading to a near complete block at the end of 2 s pulse (Table 1). The time-dependent enhancement of block of mecamylamine was not seen with butanol-induced current (Fig. 4D).

Thus, the butanol-induced current seems to be generated by the activation of AChRs, and butanol acts as a partial agonist for nnAChRs. However, the butanol-opened channels seem to be less sensitive to DHβE and mecamylamine than the ACh-opened channels. The reduced sensitivity to DHβE is probably due to the displacement of DHβE from the binding site by high concentrations of butanol, whereas the reduced sensitivity to mecamylamine block may represent a less sensitivity of the butanol-activated channel to open channel blockers.

Modulation of ACh Dose-Response Curve by Butanol.

In our previous study (Zuo et al., 2001), butanol was thought to have a direct inhibitory action and its IC₅₀ value was estimated to be 6.8 mM by extrapolation of the relationship between the IC₅₀ and carbon number of longer chainn-alcohols. Because the estimated IC₅₀value of butanol is far lower than its EC₅₀ value of 230 mM, one could evaluate its direct inhibitory action on ACh dose-response relationship. Figure 5A shows the effects of a low concentration (10 mM) of butanol on the currents induced by various concentrations of ACh. Butanol inhibited currents induced by 10 to 3000 μM ACh about 30%. The EC₅₀ value of ACh dose-response curve was slightly and insignificantly (P > 0.5) shifted in the direction of higher ACh concentrations by 10 mM butanol coapplication, without change in the Hill coefficient (EC₅₀ of 66 ± 25 μM; n_H of 0.6 ± 0.1 without butanol versus EC₅₀ of 96 ± 45 μM; n_H, 0.6 ± 0.1 with 10 mM butanol; Fig. 6A). At lower concentrations, butanol clearly suppressed the ACh-induced current, but the suppression was less than that predicted by IC₅₀ value of 6.8 mM. These results suggest that butanol is not a simple channel blocker.

The effect of n-butanol on the ACh dose-response relationship was also evaluated in the presence of high butanol concentration at which it could act as a partial agonist. Butanol at 300 mM modulated ACh-induced currents in a complex manner (Fig. 5B). Currents induced by low concentrations of ACh (3 and 10 μM) were potentiated by butanol, whereas those induced by higher concentrations of ACh (30–3000 μM) were inhibited, with the maximum inhibition occurring at 100 μM ACh (Fig. 6A). Again, this V-shape (Fig. 6A) ACh-dose-response relationship in the presence of 300 mM butanol is not consistent with a simple partial agonist model, which predicts a monophasic ACh dose-response curve with butanol raising the foot at low ACh concentrations without altering the maximum response.

Interaction between Butanol and ACh.

Because butanol acts as a partial agonist only at higher concentrations and as a channel blocker at lower concentrations, one would expect that the interaction between butanol and ACh on the α4β2 receptor depends on the concentrations of both butanol and ACh. Thus, the dose-response relationship for butanol action was examined in the presence of 30 μM and 1 mM ACh.

The inhibition of ACh-induced currents by low-to-medium butanol concentrations up to 100 mM was slightly dependent on the ACh concentration (Figs. 7 and8A). However, the effect of butanol at 300 mM was greatly dependent on ACh concentration: either no change or further inhibition was observed. Butanol exhibited a biphasic inhibitory dose-response curve at low ACh concentrations (30 and 100 μM ACh, around ACh EC₅₀) (Figs. 7A and 8A). At 1 mM ACh, which activates most receptors, butanol exhibited a monophasic dose-response relationship for inhibiting the ACh-induced current (Figs. 7B and 8A), because butanol not only competed with ACh for the binding site but also acted as a blocker. Thus, the inhibition by low-to-medium butanol concentrations up to 100 mM was slightly dependent on the ACh concentration (Figs. 7 and 8A). At 300 mM butanol one would not expect to see much current because butanol could effectively abolish the current by competing with ACh for binding to the receptor and by blocking the open channel. Because a substantial current actually remains, butanol must be less effective in blocking the butanol-opened channel than the ACh-opened channel, a result consistent with the weak blocking action of mecamylamine (Fig. 4D). At higher concentrations of ACh, a monophasic inhibition is expected to be observed because both the partial agonist action and the channel blocking action would contribute to the butanol inhibitory action.

Experimentally, the coapplication of 30 μM ACh and 300 mM butanol produces highly variable results compared with the control current produced by 30 μM ACh. The examples shown in Figs. 7 and 10 represent extremes of the inhibitory and potentiating responses. This is due to the fact that butanol is a weak agonist with an EC₅₀ value around 230 mM and with a large S.E.M. of 90 mM.

Simulation of Butanol Action.

We previously performed a kinetic simulation for the dual action of n-alcohols on nnAChRs (Zuo et al., 2001). The original scheme based on that study was expanded to Scheme I (Fig. 9) to include the assumption that butanol acts as a partial agonist on nnAChRs. In addition, we assumed that the ACh receptor-channel can be opened by two ACh molecules or two butanol molecules, but not by one ACh molecule and one butanol molecule. In addition, butanol is able to block the two ACh-opened channel but not the butanol-activated channel. Despite some indication of ACh “self-block” at higher concentrations (3 mM; Figs. 1A and 10, F and G), it was not included in the kinetic simulation. This is because most of butanol and ACh interaction studies were performed at ACh concentrations less than 3 mM. To simplify the simulation, we focused on the effect of butanol on the low-affinity receptor, which constitutes at least 80% of the total current we measured.

Because data on single-channel recording and kinetic analysis of α4β2 nnAChRs are limited, most of the parameters used in our simulation (Fig. 9) were chosen based on previous publications on muscle nAChRs (Colquhoun and Sakmann, 1985; Franke et al., 1991, 1993). We started with the parameters from Colquhoun and Sakmann (1985):k_on1 = 10⁸M⁻¹ s⁻¹,k_off1 = 8,000 s⁻¹, α = 700 s⁻¹, and β = 30,000 s⁻¹. Their rate constants were modified based on our experimental results and for best fitting to our data. The modification of parameters was performed in several steps. First, the rate constants for ACh binding and unbinding rates used in our simulation were as follows: ak_on1 of 10⁸M⁻¹ s⁻¹ and ak_off1 of 6,000 s⁻¹. These values gave aK_d of 60 μM for ACh binding, very close to the observed EC_50L. Second, for butanol binding, the slower onset of butanol-induced currents compared with ACh-induced currents (200–300 ms versus less than 20 ms) can be explained by either the slow binding of butanol (k_on), or the slow opening (β) of the butanol-bound channel. In our model, butanol binds to the receptor much more slowly than ACh, whereas the opening rates are assumed to be the same in both cases. The rate constantsk_on2 (400 M⁻¹s⁻¹) and k_off2(100 s⁻¹) were chosen to generate a butanol activation dose-response relationship similar to experimental observation (K_d = 250 mM for each binding step). Both rate constants for butanol were smaller than those of ACh, because butanol induced currents with a much slower onset. Third, Two-step desensitization process, RA₂^D1 and RA₂^D2, was used because the ACh-induced current decayed with a biexponential time course (Fig. 1A). The desensitization rates δ₁ and δ₃ were chosen as 3 and 0.4 s⁻¹, respectively, with the resensitization rates (γ) 1/10 of desensitization rates. Fourth, desensitization occurred much more slowly for butanol-activated currents than for ACh-activated currents. The decay for the current induced by 300 mM butanol is well fitted with one exponential time course, with τ around 2 s. Thus, δ₂ is estimated to be around 0.5 s⁻¹, with γ₂1/10 of δ₂. Fifth, single-channel study of the muscle receptor showed a very high open probability [P_open = β/(α+β)] of about 0.98 (Colquhoun and Sakmann, 1985). Previous data suggest that the nnAChRs have a much lower open probability, although no maximumP_open value is available (Weaver and Chiappinelli, 1996). Because no such data are available, we chose 0.67 for P_open of ACh-opened channels, as in our previous simulations (Zuo et al., 2001). The same opening and closing rate constants are assigned to butanol-opened channels to simplify the simulation.

The currents were simulated from Scheme I (Fig. 9) at various concentrations of ACh. The simulated current traces are illustrated in Fig. 1B and the peak amplitude normalized to that evoked by 3 mM ACh is plotted against ACh concentration (Fig. 1C). The EC₅₀ and n_Hvalues were estimated to be 61 μM and 1.4, respectively. The simulated butanol dose-response relationship exhibited an EC₅₀ value of 180 mM and ann_H of 1.7 (Fig. 3C). The EC₅₀ values for both ACh and butanol are close to those of the experimental results.

Butanol effects on ACh dose-response curve were simulated with the incorporation of the blocking action of butanol on the ACh-activated currents. The following blocking parameters were used: the rate constant for butanol block of the two ACh-opened channel (b₁), 1.5 × 10⁴M⁻¹ · s⁻¹; and the butanol unbinding rate constant (ub₁), 100 s⁻¹. These values give us theK_d around 6.7 mM, similar to the extrapolated IC₅₀ value for butanol block (Zuo et al., 2001). Based on weak blocking action of the open channel blocker mecamylamine on the butanol-activated currents, we assumed that the butanol-opened channels are not sensitive to butanol block at all. The simulated results are plotted in Fig. 6B, which captures many features of the observed ACh dose-response curves with both 10 and 300 mM butanol.

A similar simulation was also performed for the butanol inhibitory dose-response relationship (Fig. 8B). A biphasic inhibitory dose-response curve was obtained using 30 μM ACh with a maximal inhibition at 100 mM butanol, whereas a monophasic inhibitory relationship was observed at 1 mM ACh. Thus, simulation yields the butanol dose-response relationships similar to those of experimental results.

Butanol as Channel Blocker.

n-Butanol also acted as a channel blocker for nnAChRs, in a way similar to other long-chain alcohols (Murrell and Haydon, 1991). Although butanol was a weak agonist with an EC₅₀ value around 200 mM, it was much more potent as a channel blocker of the ACh-opened channels. The blocking rate constant of 1.5 × 10⁴M⁻¹ · s⁻¹ for the two ACh-opened channel, combined with the unbinding rate constant of 100 s⁻¹ gave rise to aK_d of 6.7 mM. This difference between the agonist EC₅₀ value and the blockingK_d was reflected in currents induced by coapplication of ACh and butanol (Fig.10). Butanol at 10 mM showed hardly any agonist action (Fig. 10A), but caused a significant inhibition of currents induced by 30 and 3000 μM ACh (Fig. 10, D and G). Butanol at 300 mM had a significant agonist action (Fig. 10B), potentiated the current evoked by 30 μM ACh (Fig. 10E), and inhibited the current evoked by 3000 μM ACh (Fig. 10H).

Discussion

Multiple Action of n-Butanol.

The present study showed that n-butanol exerted multiple actions on the α4β2 nnAChRs stably expressed in HEK293 cells. In the absence of ACh, n-butanol generated a small current, and in the presence of ACh, it either potentiated or inhibited ACh-induced currents, depending on the concentrations of ACh and butanol. Most of the features of these multiple actions could be simulated by a model based on the hypothesis that n-butanol acts both as a partial agonist to induce currents and as an open-channel blocker (Fig.9).

Contributions of Different Actions of Butanol to the Biphasic Dose-Response Relationship.

To understand the contributions of two major actions of butanol as a partial agonist and as a channel blocker to the overall action of butanol on nnAChRs, we analyzed the biphasic nature of the ACh dose-response curve at 300 mM butanol in detail. At low concentrations of ACh, 3 μM for example, we observed over 3-fold potentiation by coapplication of 300 mM butanol (Fig. 6A). This is mainly due to the agonist action of butanol at such a high concentration to open the ACh channel. Based on our simulation (Fig.11), 300 mM butanol itself can open about 40% of the total AChR channels, similar to what 60 μM ACh does. Thus, with 3 μM ACh and 300 mM butanol, over 99.8% of the open channels are occupied by two butanol molecules. However, because the affinity of ACh for the receptor is more than 4000 times higher than that of butanol (ACh K_d value of 60 μM versus butanol EC₅₀ value of 250 mM), butanol's contribution as a partial agonist decreases dramatically as the ACh concentration increases, because more and more receptors are bound by ACh instead of butanol. The most significant change in the composition of the overall open channels occurs between the ACh concentrations of 30 and 100 μM: the fraction of the two ACh-open channel increases from 10% of total opened channels at 30 μM ACh to more than 99% at 100 μM ACh, and at the same time, the two butanol-opened channel drops from 90% of total open channels to less than 1%. This trend continues until the concentrations are increased to 3000 μM ACh, when almost all the open channels are occupied by two ACh molecules, whereas butanol hardly opens any channel by itself (<0.01%).

Figure 11 depicts the results of simulation of the fraction of various receptor states at the peak current in the presence of various concentrations of ACh (Fig. 11A) and butanol (Fig. 11B). Complex dose-response curves are obtained showing an apparent potentiating action of butanol at low ACh concentrations and a biphasic inhibitory action at moderate-to-high ACh concentrations (Fig. 11A, open squares) as seen experimentally (Fig. 6A). Figure 11A illustrates that the initial level (∼35%) occurring at 1 μM ACh is mainly due to the activation of the receptors by butanol. Because butanol blocks the ACh-opened channels only, the blocking action takes place only when there are many ACh-opened channels. This is shown by the fact that the percentage of butanol block of the total active receptors is 1% at 1 μM ACh compared with 82% at 30 μM ACh. Because the blocking action of butanol takes place at a rate much lower than that of ACh to open the channel, as the concentration of ACh increases, the current peaks sooner and butanol blocks less. Thus, only 48% block occurs at 3000 μM ACh compared with 82% block at 30 μM ACh. Thus, the combination of both the partial agonist action and the kinetics of channel blocking action generates a V-shaped ACh dose-response curve at 300 mM butanol (Fig. 11A, squares).

The simulation of the V-shape dose-response curve for butanol inhibition at 30 μM ACh (Fig. 8B) could be similarly accounted for as illustrated in Fig. 11B (squares). At low concentrations of butanol (1–30 mM), more than 99% of the total open channels bind two ACh molecules and are very sensitive to butanol blocking action. However, as the butanol concentration increases, butanol starts competing with 30 μM ACh to open the channel. For example, 100 and 300 mM butanol contribute 11 and 90%, respectively, to the total channel conductance in the presence of 30 μM ACh and butanol. Thus, a biphasic inhibitory dose-response relationship was obtained at 30 μM (Fig. 8).

When the α4β2 receptors were activated by 1 mM ACh, butanol exerted a monophasic inhibition, yielding an apparent IC₅₀ value around 100 mM. A similar dose-response relationship was simulated by the kinetic model in which butanol acted as a partial agonist and as an open channel blocker. In the simulation, however, the IC₅₀ value for open channel block was assumed to be 6.7 mM, a value similar as the extrapolated IC₅₀ value from the previous study of the chain-length dependence of long-chain alcohols to inhibit the whole-cell current of α4β2 receptors (Zuo et al., 2001).

The obvious discrepancy between the measured IC₅₀value of 100 mM and the extrapolated one of 6.7 mM is due to two factors, both of which would reduce IC₅₀ value measured at the whole-cell level. One factor is the butanol's dual action because the channel blocking action could not be cleanly separated from its partial agonist action at the whole-cell level. Another factor contributing to the lesser potency of block is the kinetic effect of channel gating on the equilibrium block. Given the blocking and unblocking rate constants used in the simulation, butanol should reach an equilibrium block at the peak of ACh current. However, in the constant presence of ACh during coapplication of ACh and butanol, the rate constants governing channel opening and closing would reduce the butanol block. This gating effect on the open channel block was recognized in open channel block of other channels as well (Coronado and Miller, 1979; French and Shoukimas, 1981).

Comparison of multiple actions with previous studies.

Multiple actions have been reported previously for the n-octanol action on the GABA_A receptor (Kurata et al., 1999); the actions of d-tubocurarine (Steinbach and Chen, 1995), metocurine, and atracurium (Fletcher and Steinbach, 1996) on the fetal nAChRs; and the actions of d-tubocurarine on nnAChRs containing β4 subunit (Cachelin and Rust, 1994). However, in most cases, all the observations were satisfactorily explained by a model in which the drug acted as a weak agonist to open the channel either by itself or by cobinding with another agonist. The latter example was used in the model to explain the dual action of atropine (Zwart and Vijverberg, 1997) and d-tubocurarine (Cachelin and Rust, 1994) on nnAChRs because the receptors occupied by two atropine or tubocurarine molecules do not conduct current. In our case, butanol acts as a partial agonist at high concentrations and as channel blocker at low concentrations on the α4β2 receptor.

Inasmuch as channel opening and block are such distinct actions, the multiple effects of the agents suggest multiple sites of action on nnAChRs. Our previous simulation based on a model in which long-chain alcohols both block the open ACh receptor-channel and interfere with ACh binding explains satisfactorily the alcohol's inhibitory action (Zuo et al., 2001). It was also suggested that the inhibitory action of long-chain alcohols is due to both alcohol block of open channels and alcohol interference with ACh binding (decrease ink_on). One possible mechanism ofk_on reduction is that long-chain alcohols can compete with ACh for the ACh binding site. Because butanol has a similar size to ACh, it is capable of opening the channel after binding to ACh binding site. This competition between long-chain alcohols and ACh for binding to the agonist sites would manifest as a slowing in k_on, the rate constant for ACh to bind the agonist site in the previous simulation (Zuo et al., 2001).