Introduction

Top Summary Introduction Procedures Results Discussion References

Neuronal nAChRs are ligand-gated ion channels present throughout the central and peripheral nervous systems. Multiple neuronal nAChR subunits (2-9 and 2-4) have been identified. These subunits combine in various compositions to form pentameric receptors with different functional and pharmacological characteristics (reviewed in McGehee and Role, 1995).

The 4 and 2 nAChR subunits are expressed abundantly in brain (Wada et al., 1989), and the most common nAChR in the brain is composed of these two subunits (Whiting et al., 1990; Flores et al., 1992). The subunit stoichiometry of 42 nAChRs expressed in Xenopus laevis oocytes has been deduced to be 4:2 = 2:3 (Anand et al., 1991; Cooper et al., 1991), and the subunits are assumed to be arranged around the central ion channel pore in the order (Anand et al., 1991). Patch-clamp recording from oocytes expressing pairs of  and  nAChR subunits revealed ACh-gated single channels with distinct conductances (Papke et al., 1989; Papke and Heinemann, 1991; Charnet et al., 1992; Kuryatov et al., 1997). The frequency of occurrence of the different single-channel conductance states depends on the : ratio of the mRNA mixture injected into the oocyte, indicating that receptors with different subunit stoichiometry are expressed (Papke et al., 1989).

The agonist binding sites of heteropentameric neuronal nAChRs generally are supposed to be located at the two subunit interfaces (reviewed in Bertrand and Changeux, 1995). Functional nAChRs with distinct subunit stoichiometries might differ with respect to the number and properties of agonist binding sites. We investigated the functional properties of the agonist binding sites in X. laevis oocytes expressing different levels of 4 and 2 nAChR subunits. The results on concentration-dependent effects of the agonist ACh and of d-TC, a competitive antagonist of 42 nAChRs (Bertrand et al., 1990), demonstrate four distinct subtypes of 42 nAChRs.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. X. laevis were bred and kept in the Hubrecht Laboratory (Utrecht, The Netherlands) or were bought from Nasco (Fort Atkinson, WI) and kept in our own laboratory. Neomycin was obtained from Sigma Chemical (St. Louis, MO). d-TC was from Fluka (Buchs, Switzerland). All other materials used came from sources identical to those described previously (Zwart and Vijverberg, 1997).

Receptor expression. Oocytes were prepared, injected, and incubated in Barth's solution containing 5 mg/liter of neomycin at 19° for 3-7 days as described previously (Zwart et al., 1995; Zwart and Vijverberg, 1997). Plasmids coding for 4 and 2 subunits of neuronal nAChRs were dissolved in distilled water at the approximately equal concentrations of 152 ± 15 and 136 ± 12 µg/ml (mean ± standard deviation of triplicate spectrophotometric determinations), respectively. Unless noted otherwise, mixtures of these solutions at 1:9, 1:1, and 9:1 ratios were coinjected at 18.4 nl/oocyte.

Electrophysiology and data analysis. Oocytes were placed in a silicon rubber tube (diameter, 3 mm), penetrated by two microelectrodes, and voltage clamped at 80 mV. ACh and d-TC, dissolved in external solution, were applied by perfusion of the tube at a rate of ~20 ml/min. The external saline contained 115 mM NaCl, 2.5 mM KCl, 1 mM CaCl₂, and 10 mM HEPES, pH 7.2, with NaOH. Voltage-clamp equipment, experimental protocols, and data acquisition were exactly as described previously (Zwart and Vijverberg, 1997). All amplitudes of ACh-induced ion currents were normalized to the amplitude of alternately evoked control responses to a near-maximum effective concentration of ACh to adjust for small variations in response amplitude over time. Data are expressed as mean ± standard deviation of number of oocytes. ANOVA was performed using Microsoft Excel. Results were compared using a two-tailed Student's t test. Concentration-effect curves were fitted, using Jandel SigmaPlot software, to data obtained in separate experiments, and mean ± standard deviation values of estimated parameters were calculated for three oocytes. Curves drawn in the figures were generated using the mean values of the estimated parameters. Agonist data were fitted according to the equations i/i_max = E_max/[1 + (EC₅₀/[ACh])^n_H] and i/i_max = E_maxa/(1 + EC_50a/[ACh]) + E_maxb/(1 + EC_50b/[ACh]) for one- and two-component activation curves, respectively. Antagonist data were fitted according to the equations: i/i_max = 1/[1 + ([dTC]/IC₅₀)^n_H] and i/i_max = E_m/[1 + ([dTC]/IC_50a)^n_Ha) + (1  E_m)/[1 + ([dTC]/IC_50b)^n_Hb] for one- and two-component inhibition curves, respectively. In these equations, [ACh] and [dTC] are the concentrations of ACh and d-TC, respectively; n_H is the Hill coefficient; and i/i_max is the normalized current amplitude. Goodness-of-fit was judged by analysis of the residuals (run test).

	Results

Top Summary Introduction Procedures Results Discussion References

Effects of 4:2 ratio on ion current amplitude. Superfusion of voltage-clamped oocytes, expressing 42 nAChRs after the injection of 4 and 2 subunit cDNAs in the : ratios 1:9, 1:1, and 9:1, with 300 µM ACh results in ligand-gated ion currents. The effect of subunit ratio on response amplitude was investigated in the same batch of oocytes because expression levels vary between batches. The amplitude of the inward current induced by the near-maximal effective concentration of 300 µM ACh at the holding potential of 80 mV depends on the 4:2 subunit cDNA ratio injected into the oocytes (Fig. 1). The largest inward currents are observed in oocytes injected with 4 and 2 subunit cDNAs in the 1:1 ratio. At both the 1:9 and 9:1 : ratios, amplitudes of inward currents evoked by 300 µM ACh are reduced significantly (t test; p < 0.001). Some oocytes (9 of 54) that did not respond to ACh at all, most likely because the cDNA was not properly injected into the nucleus, were excluded from the analysis. The results show that, after injection of the 4 and 2 subunit cDNAs in extreme ratios, while maintaining the total cDNA injected approximately constant, inward current amplitudes are reduced. This indicates that after injection of cDNAs in the : ratios of 1:9 and 9:1, response amplitude is limited by the availability of  and  subunits, respectively. After injection of the same total amounts of either cDNA encoding the 4 subunit only or cDNA encoding the 2 subunit only, oocytes did not respond to 300 µM ACh (results not shown). This demonstrates that 4 and 2 subunits do not form functional homopentameric nAChRs. The absence of ACh-induced inward current also shows that 4 and 2 do not form functional heteromeric nAChRs with native nAChR subunits that possibly are present in X. laevis oocytes (Buller and White, 1990).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Peak inward current amplitudes evoked by superfusion with 300 µM ACh in X. laevis oocytes expressing 42 nAChRs. Oocytes from a single batch were injected with approximately equal amounts of cDNAs encoding 4 and 2 nAChR subunits in the ratios indicated. Bars, mean ± standard deviation values of peak amplitudes from the number of oocytes indicated (between brackets). Peak amplitudes obtained for the 1:9 and 9:1 : ratios are significantly reduced compared with those obtained for the 1:1 : ratio (t test; p < 0.001).

Effects of 4:2 ratio on the agonist concentration-effect curve. Agonist concentration-effect curves were obtained from oocytes expressing 42 nAChRs after the injection of subunit cDNAs in the : ratios of 1:9, 1:1, and 9:1. Examples of ACh-induced ion currents are shown in Fig. 2A for the different subunit cDNA ratios tested.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Concentration-dependent effects of ACh on 42 nAChRs expressed in X. laevis oocytes after injection of cDNAs in the : ratios indicated. a, Inward currents evoked by 10-sec periods of superfusion with external saline containing ACh at the concentrations indicated for three oocytes at the 1:1 (), 9:1 (), and 1:9 () : ratios. Calibration bars, 4.5 µA (1:1) and 350 nA (9:1 and 1:9). b, Dependence of inward current amplitude, normalized to the amplitude of control responses induced by 1 mM ACh, on the concentration of ACh. Values are mean ± standard deviation obtained from three oocytes. The data obtained for the 1:1 and 9:1 : ratios were fitted adequately by one-component concentration-effect curves (drawn lines). Data obtained at concentrations of >300 µM ACh, which induced ion channel block, are not included in the fit. The data obtained for the 1:9 : ratio could not be fitted by a one-component concentration-effect curve. The two-component curve drawn was fitted to all data obtained at concentrations of 300 µM ACh with slope factors set to 1. The concentration-effect curve describing the component more sensitive to ACh, fitted separately to the data obtained at 30 µM ACh, is also drawn. The estimated parameters of the fitted curves are presented in Table 1.

View this table: [in this window] [in a new window]   TABLE 1 Effects of ACh on 42 nAChRs expressed in Xenopus oocytes injected with cDNAs encoding  and  subunits in the ratios indicated For each experimental condition, three concentration-effect curves, each obtained from a different oocyte, were fitted, and the values represent mean ± standard deviation of the estimated parameters.

Effects of 4:2 ratio on the antagonist concentration-effect curve. Concentration-effect curves for the competitive nAChR antagonist d-TC also were obtained from oocytes expressing 42 nAChRs after the injection of subunit cDNAs in the : ratios of 1:9, 1:1, and 9:1. Examples of ion currents evoked by superfusion of 300 µM ACh in the absence and the presence of various concentrations of d-TC are shown in Fig. 3A for the different subunit cDNA ratios tested. d-TC was superfused 4 min before an ACh-induced current was recorded in the presence of d-TC. For three oocytes of each : subunit cDNA ratio tested, peak inward current amplitudes were normalized to those of control inward currents evoked by 300 µM ACh in the same oocyte. The normalized data are plotted against d-TC concentration in Fig. 3A. ANOVA of the inhibitory effects of d-TC showed that the data for the different subunit cDNA ratios all are statistically different at the p < 0.001 level. For the 9:1 and 1:1 : subunit cDNA ratios, data could be fitted by a one-component concentration-effect curve. However, for the 1:9 ratio, the concentration dependence of inhibition by d-TC seemed to be biphasic, and fit was significantly improved by fitting a sum of two concentration-effect curves to the data (Fig. 3A). Estimated values of the parameters of the fitted concentration-effect curves are presented in Table 2. The inhibitory effects of d-TC on 42 nAChRs at the 1:1 : ratio also were investigated at a holding potential of 40 mV (result not shown). The IC₅₀ value for d-TC of 0.91 ± 0.48 µM (three oocytes) at 40 mV could not be distinguished from that obtained at the holding potential of 80 mV. Thus, it seems that voltage-dependent ion channel block, observed previously for muscle-type nAChRs at high concentrations of d-TC (Colquhoun et al., 1979), does not strongly contribute to the inhibitory effects of d-TC observed at concentrations of 10 µM.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent effects of d-TC on 42 nAChRs expressed in X. laevis oocytes after injection of cDNAs in the : ratios indicated. a, Top, inward currents evoked by 10-sec periods of superfusion with external saline containing 300 µM ACh in the presence of d-TC at the concentrations indicated for three oocytes at the 1:1 (), 9:1 (), and 1:9 () : ratios. Bottom, dependence of inward current amplitude, normalized to the amplitude of control responses induced by 300 µM ACh, on the concentration of d-TC. Values are mean ± standard deviation obtained from three oocytes. Drawn lines, mean of one-component agonist concentration-effect curves fitted to the data obtained for the 1:1 and 9:1 : ratios and of two-component concentration-effect curves fitted to the data obtained for the 1:9 : ratio to account for a second component of inhibition. The estimated parameters of the fitted curves are presented in Table 2. Calibrations bars, 2 µA (1:1), 500 nA (9:1), and 100 nA (1:9). b,Inhibition of inward currents evoked by 10-sec periods of superfusion with external saline containing 10 µM ACh in the presence of d-TC at the concentrations indicated. The low concentration of ACh selectively activates the more-sensitive 42 nAChRs expressed at the 1:9 : ratio (see Fig. 2B). Calibration bar, 50 nA. Bottom, concentration-dependent inhibition of inward current amplitude by d-TC (), which is fitted by a monophasic inhibition curve (drawn line). The biphasic inhibition curve obtained for the inhibition of 300 µM ACh-induced inward current is drawn for comparison. Dashed inhibition curves (A and B), results of fitting a model assuming two sites for the inhibition by d-TC contributing differentially to the curves obtained at the different : ratios (see Discussion).

View this table: [in this window] [in a new window]   TABLE 2 Inhibitory effects of d-TC on 42 nAChRs expressed in Xenopus oocytes injected with cDNAs encoding  and  subunits in the ratios indicated Responses were evoked by superfusion with 300 µM ACh. For each experimental condition, three concentration-effect curves, each obtained from a different oocyte, were fitted, and the values represent mean ± standard deviation of the estimated parameters.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Ion currents with very low sensitivity to d-TC in oocytes injected with 4 and 2 cDNAs at the 1:9 : ratio and the inhibitory effect of a high concentration of d-TC. Oocytes were superfused continuously with 30 µM d-TC to inhibit the d-TC-sensitive component of ACh-induced ion current (see Fig. 3A). a, Inward currents evoked with 100 µM ACh before and 4 min after raising the d-TC concentration to 330 µM at the holding potential of 60 mV. b, Inward currents from the same oocyte demonstrating the inhibitory effect of d-TC at the holding potential of 100 mV. c, Inhibitory effect of coapplication of 100 µM ACh and 330 µM d-TC illustrating a reduced peak inward current amplitude and accelerated inward current decay. Horizontal calibration bar, 10 sec. Vertical calibration bar, 100 nA (A and C) and 200 nA (B).

Effects of nAChR expression level on agonist and antagonist sensitivities. Because it cannot be excluded that nAChR expression level by itself affects agonist and antagonist sensitivity, experiments were performed on oocytes injected with diluted 1:1 : cDNA solution at constant injection volume. In oocytes injected with a 10-fold diluted cDNA solution, superfusion with 1 mM ACh did not evoke detectable (>5 nA) inward currents (10 oocytes). This indicates that nAChR expression is greatly reduced, consistent with previous results on the expression level of chick 42 nAChRs in oocytes (Bertrand et al., 1991). After the injection with a 4-fold diluted cDNA solution, ACh-induced ion currents were evoked, which were smaller than those obtained before but had a similar shape (Fig. 5A). The inward current decay at the end of ACh application amounted to 48 ± 15% (five oocytes) of the peak current. This cannot be distinguished from the degree of decay at the higher expression level (t test; p = 0.24) but differs from the degree of inward current decay obtained for the 1:9 : ratio (p < 0.01). Agonist and antagonist sensitivities were determined from six oocytes with a maximal peak inward current amplitude of 398 ± 82 nA (i.e., in the same range as with the 1:9 and 9:1 : ratios). The results obtained with ACh at concentrations of 1 mM from oocytes with high and low level nAChR expression at the 1:1 : ratio are not identical (two-way ANOVA; F_1,28 = 7.556, p = 0.01). Although the fitted concentration-effect curve of ACh (Fig. 5A, Table 1) shows a small shift toward higher concentrations, the EC₅₀ and n_H values cannot be distinguished from those obtained for the higher expression level (t test; p > 0.08). The concentration-effect curve of d-TC at the low expression level (Fig. 5B, Table 2) cannot be distinguished statistically from that obtained with the high expression level at the 1:1 : ratio (two-way ANOVA; F_1,20 = 2.290, p = 0.15). However, this curve differs from the inhibition curves obtained for the 9:1 and 1:9 : ratios (two-way ANOVA; p < 0.001). Further statistical analysis showed that the expression level does not influence the IC₅₀ value of d-TC (t-test; p = 0.33) and that regardless of expression level, distinct IC₅₀ values are obtained for the different subunit ratios (t test; p values < 0.01). From the combined results, it is concluded that reducing the nAChR expression level causes only a small decrease in ACh sensitivity, which is opposite the 30-fold increase in sensitivity observed at the 1:9 : ratio.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Agonist and antagonist sensitivities of 42 nAChRs expressed in oocytes injected with 4-fold diluted 1:1 : cDNA solution. a, Inward currents evoked by 10-sec periods of superfusion with external saline containing ACh at the concentrations indicated (calibration bar, 200 nA). Bottom, mean ± standard deviation of peak inward current amplitudes normalized to that of 1 mM ACh-evoked inward currents from three oocytes. The data were fitted by a one-component concentration-effect curve (drawn line). For comparison, the concentration-effect curve obtained for 42 nAChRs at the higher expression level with the 1:1 : ratio (Fig. 2A) is drawn (dashed line). Data obtained at 1 mM ACh, which induced ion channel block, are not included in the fit. b, Inhibition of 300 µM ACh-induced inward current by d-TC at the concentrations indicated (calibration bar, 500 nA). Bottom, mean ± standard deviation (three oocytes) of peak inward current amplitudes normalized to that of 300 µM ACh-evoked control responses. The two data points without error bars, from one oocyte. The data were fitted by a one-component concentration-effect curve (drawn line). The estimated parameters of the curves fitted to the data in A and B are presented in Tables 1 and 2, respectively.

	Discussion

Top Summary Introduction Procedures Results Discussion References

Differences in agonist and antagonist sensitivities of receptors expressed in oocytes after injection of cDNAs coding for the 4 and 2 subunits at 1:9, 1:1, and 9:1 ratios demonstrate heterogeneity of 42 nAChRs. The results show that combinations of 4 and 2 subunits form four pharmacologically distinct subtypes of heteromeric nAChRs. Two subtypes of nAChRs, which are distinguished for the 1:1 and 9:1 : ratios, have similar low sensitivities to ACh (EC₅₀  60 µM) but distinct sensitivities to d-TC (IC₅₀ = 0.20 and 0.53 µM, respectively). Two additional, distinct subtypes of nAChRs are formed at the 1:9 : ratio: one subtype with markedly enhanced sensitivity to ACh (EC₅₀ = 1.8 µM) and decreased sensitivity to d-TC (IC₅₀ = 2.04 µM), and a second subtype with low sensitivity to ACh and very low sensitivity to d-TC (IC₅₀ = 163 µM). Heterogeneity of neuronal nAChRs has been suggested before from the observation that the occurrence of distinct single-channel conductances in oocytes expressing neuronal-type nAChRs depends on : ratio (Papke et al., 1989).

The effects of varying : ratio on the properties of nAChRs were investigated in oocytes injected with approximately equal total amounts of cDNA (2.5-2.8 ng/oocyte). The reduced response amplitudes at the 1:9 and 9:1 ratios indicate that the availability of  and  subunits limits functional receptor expression at these cDNA ratios (Fig. 1). A comparison of the properties of nAChRs expressed at a reduced level for the 1:1 : ratio (Fig. 5) with those of nAChRs expressed at the 1:9 and 9:1 ratios (Figs. 2, 3) demonstrates that the 30-fold difference in sensitivity to ACh, the shift in sensitivity to d-TC, and the differences in the degree of inward current decay are due to changes in subunit ratio and not to differences in expression level.

The properties of ligand-gated ion channels depend on subunit composition. Combining 4 and 2 subunits into a pentameric nAChR protein theoretically can yield eight distinct subtypes of the 42 nAChR. The two homopentameric 4 and 2 nAChRs are not expressed or are not functional. Six heteropentameric subunit assemblies remain: four : subunit stoichiometries (1:4, 2:3, 3:2, and 4:1) and two alternative subunit arrangements for the 2:3 and the 3:2 stoichiometries (Fig. 6). Of these six, at least four have been demonstrated to form pharmacologically distinct, functional nAChRs. This implies that functional 42 nAChRs are formed with other than the 2:3 : subunit stoichiometry.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Schematic representation of the six possible arrangements of  and  subunits within heteropentameric nAChRs. Note the two alternative subunit arrangements for the 2:3 and 3:2 : subunit stoichiometries.

Functional heteropentameric neuronal nAChRs presumably contain two binding sites for ACh located at the interfaces (Luetje and Patrick, 1991; Bertrand and Changeux, 1995). However, only two of all possible subunit assemblies contain two interfaces (Fig. 6). The conclusion that four distinct types of 42 nAChRs are formed implies that agonist binding at one or more of the , , or subunit interfaces may be involved in ion channel activation or that binding of a single agonist molecule is sufficient to activate 42 nAChRs. The latter would be consistent with the results because the slopes of the concentration-effect curves of ACh all are close to 1 (Table 1). Shallow slopes of the concentration-effect curve of ACh on 42 nAChRs generally are reported: in oocyte studies, 0.81 for rat 42 (Stafford et al., 1994), 1.5 for chick 42 (Bertrand et al., 1990), and 1.02 for human 42 (Chavez-Noriega et al., 1997), and in human embryonic kidney 293 cells, 1.2 for human 42 (Buisson and Bertrand, 1998). In 7 nAChRs, interfaces contribute to ligand binding. This homopentameric nAChR seems to contain five binding sites for the competitive antagonist methyllycaconitine (Palma et al., 1996). The 7 subunit contains the principal component (loops A-C) as well as the highly conserved residue Trp54, which constitutes part of a complementary component of the proposed agonist binding site. These two components of the agonist binding site in 7 are thought to be equivalent to those contained by the muscle-type 1 and the  and  nAChR subunits, respectively. The 4 subunit contains the principal component of the agonist binding site as well as a tryptophan residue homologous to Trp54 in 7. The 2 subunit contains the complementary component of the agonist binding site as well as loops A and B of the principal component (Corringer et al., 1995). The homology among 4, 2, and 7 suggests that , , or subunit interfaces also might form functional agonist or antagonist binding sites. The shift in agonist sensitivity seems to be specific for the 42 nAChRs because it is not observed when 32 and 44 nAChRs are expressed at 1:1 and 1:9 : cDNA ratios (Luetje and Patrick, 1991). This indicates that neither nor subunit interfaces are directly involved in the change in sensitivity to ACh.

The results show that the inhibitory effects of d-TC on 42 nAChRs are not due to ion channel block. Low concentrations of d-TC have been shown to inhibit neuronal nAChRs by a competitive interaction before (Bertrand et al., 1990), but the exact nature of the inhibitory effect of the very high concentrations of d-TC on nAChRs at the 1:9 : ratio remains to be investigated. However, the four distinct IC₅₀ values for d-TC do not necessarily imply that 42 nAChRs contain four distinct binding sites for d-TC. Apart from the component with very low sensitivity, the effects of d-TC consist of monophasic inhibition curves with shallow slopes, which gradually shift to higher concentrations with decreasing : ratio (Fig. 3; Table 2). This indicates that high and low affinity sites differentially contribute to the inhibitory effect of d-TC and cause shift of the IC₅₀. Three two-component inhibition curves of d-TC were simultaneously fitted to the three sets of data obtained for the 9:1, 1:1, and 1:9 (10 µM ACh) : ratios. Hill coefficients were set to 1, and the IC₅₀ values of d-TC and the relative proportions of the two effects of d-TC were estimated. This yielded IC₅₀ values of 0.16 and 4.3 µM d-TC, regardless of : ratio. The estimated proportions of the more- and less-sensitive components changed with : ratio. The size of the more sensitive component decreased from 88% at the 9:1 : ratio to 62% and further to 26% at the 1:1 and 1:9 : ratios, respectively. The curves fitted according to the two-component inhibition model (Fig. 3) demonstrate that for 42 nAChRs, at least two sites with an estimated 27-fold difference in sensitivity to d-TC are required to account for the inhibitory effects observed. This is in the same order of magnitude as the 15- and 27-fold differences in sensitivity of 11 and 11 muscle-type nAChRs expressed in fibroblasts and human embryonic kidney 293 cells to d-TC and dimethyl-d-TC, respectively (Blount and Merlie, 1989; Sine, 1993). The shift in the inhibition curve of d-TC can be accounted for by assuming a minimum of two distinct sites for d-TC. The differential contribution of these sites to the inhibitory effect at different : ratios suggests that they are located on distinct subtypes of 42 nAChRs. The possibility that the high and low affinity components reflect effects of d-TC on a single population of receptors is highly unlikely. If the two sites would be present on the same nAChR, only the inhibition caused by the binding of d-TC to the high affinity site would have been detected. Some receptors seem to lack both sites because they are affected by d-TC at very high concentrations only.

The results with d-TC indicate that two subtypes of 42 nAChRs may be distinguished in oocytes injected with cDNAs at the 1:1 : ratio. In oocytes injected with cRNAs or cDNAs encoding chick 4 and 2 subunits at a 1:1 ratio, nAChRs seem to have a 2:3 : stoichiometry (Anand et al., 1991; Cooper et al., 1991). The current and the previous results are compatible when two subunit arrangements, which are possible for the 2:3 : stoichiometry (Fig. 6), are formed and have distinct sensitivities to d-TC. The diverging results on the number of conductance levels of single ion channels formed by 42 nAChRs at the 1:1 : ratio provide no conclusive evidence on 42 nAChR heterogeneity (Papke et al., 1989; Cooper et al., 1991; Charnet et al., 1992; Kuryatov et al., 1997). Single-channel properties of 42 nAChRs at other than the 1:1 : ratio have not been reported. Whether the arrangement of subunits within a receptor affects single-channel properties remains unknown. The current results demonstrate that changing the relative abundance of subunits expressed by changing the cDNA ratio leads to alternative subunit assemblies with sensitivities to ACh and d-TC distinct from those obtained for nAChRs at the 1:1 : ratio.

The ratios of subunit cDNAs used in the current study for the expression of distinct subtypes of 42 nAChRs in oocytes are in the same order of magnitude as those found in brain. The ratio of 4:2 mRNA levels in various regions of rat brain ranges between 1:1 and 1:15 (Liu et al., 1996). For -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-type glutamate receptors, it has been shown that the number of glutamic acid type 2 receptor subunits incorporated depends on the relative abundance of the glutamic acid type 2 receptor subunit expressed in oocytes as well as in rat hippocampal interneurons. The resulting differences in subunit stoichiometry of native and heterologously expressed -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors are associated with changes in channel function (Washburn et al., 1997). Although it was assumed previously that nAChRs have a fixed subunit stoichiometry, the current results show that 42 nAChR subunit stoichiometry also can be changed by altering the relative abundance of  and  subunits. Thus, changing the subunit stoichiometry of receptors by selective up- or down-regulation of the expression level of specific subunits seems to be a more general mechanism to tune the properties of ligand-gated ion channels. The current observation that agonist and antagonist sensitivities depend on subunit stoichiometry provides a rational basis for the existence of heteromeric neuronal nAChRs.