Unorthodox Acetylcholine Binding Sites Formed by α5 and β3 Accessory Subunits in α4β2* Nicotinic Acetylcholine Receptors

Abstract

All nicotinic acetylcholine receptors (nAChRs) evolved from homomeric nAChRs in which all five subunits are involved in forming acetylcholine (ACh) binding sites at their interfaces. Heteromeric α4β2* nAChRs typically have two ACh binding sites at α4/β2 interfaces and a fifth accessory subunit surrounding the central cation channel. β2 accessory subunits do not form ACh binding sites, but α4 accessory subunits do at the α4/α4 interface in (α4β2)₂α4 nAChRs. α5 and β3 are closely related subunits that had been thought to act only as accessory subunits and not take part in forming ACh binding sites. The effect of agonists at various subunit interfaces was determined by blocking homologous sites at these interfaces using the thioreactive agent 2-((trimethylammonium)ethyl) methanethiosulfonate (MTSET). We found that α5/α4 and β3/α4 interfaces formed ACh binding sites in (α4β2)₂α5 and (α4β2)₂β3 nAChRs. The α4/α5 interface in (β2α4)₂α5 nAChRs also formed an ACh binding site. Blocking of these sites with MTSET reduced the maximal ACh evoked responses of these nAChRs by 30–50%. However, site-selective agonists NS9283 (for the α4/α4 site) and sazetidine-A (for the α4/β2 site) did not act on the ACh sites formed by the α5/α4 or β3/α4 interfaces. This suggests that unorthodox sites formed by α5 and β3 subunits have unique ligand selectivity. Agonists or antagonists for these unorthodox sites might be selective and effective drugs for modulating nAChR function to treat nicotine addiction and other disorders.

Introduction

Nicotinic acetylcholine receptors (nAChRs)² are ACh-gated ion channels formed from five homologous subunits organized like barrel staves around a central cation channel (1). There are homomeric nAChRs and heteromeric nAChRs. Homomeric α7 nAChRs have five α7 subunits with ACh binding sites between each of the five α7/α7 interfaces. All nAChR subunits exhibit basic homology throughout their sequences, indicating that all nAChRs evolved from a common ancestor (1, 2). Heteromeric nAChR agonist binding sites typically form at the interface between the primary (+) side of an α subunit, characterized by the presence of a C loop that closes over the site when the nAChR is activated by an agonist, and the complementary (−) side of a β2 or β4 subunit (3). Two such ACh binding sites assemble with an accessory subunit in a stoichiometry such as (α4β2)₂β2 (4, 5).

The (α4β2)₂α4 stoichiometry contains an unorthodox ACh binding site at the α4/α4 interface (6, 7). When this low affinity unorthodox site is bound by ACh or NS9283, an agonist specific for this site, nAChRs activate 3–4-fold more efficiently (8, 9). Blocking the α4/α4 site through alkylation of a cysteine introduced in the minus face of the α4 subunit blocked the activity of ACh and NS9283 (9). Histidine 142 (this is position 116 in the mature α4 peptide sequence) on the minus side of α4 is critical for the binding of NS9283 (10). Sazetidine-A is a high affinity agonist selective for α4/β2 sites that cannot bind to the α4/α4 interface because histidine 142 on the minus side of α4 prevents it from binding, whereas this amino acid is valine in β2 (11, 12). Because orthodox agonist sites are shared among nAChRs, subtype selectivity is more likely to be achieved by targeting unorthodox agonist sites.

It has been assumed that α5 and β3 function only as accessory subunits, and it is unknown whether they contribute to forming agonist binding sites (13–15). These subunits are closely related in sequence, but α5 has a cysteine pair at the tip of its C loop that defines it as an α subunit, and β3 does not. (α4β2)₂α5 and (α4β2)₂β3 nAChRs are found in the brain, as are (α6β2)(α4β2)β3 nAChRs (3, 16–20). α5 and β3 subunits have been implicated in nicotine addiction and are likely to be involved in many diseases associated with nAChRs (21–26). An α5 D398N and some β3 polymorphisms are associated with susceptibility to smoking (26, 27). The α5 polymorphism is also associated with susceptibility to lung cancer (26). Knock-out of α4, α6, or β2 subunits inhibits the rewarding properties and self-administration of nicotine (28), whereas knock-out of α5 inhibits the aversive effects of high nicotine concentrations and increases self-administration of nicotine (29). Thus, a drug that increases the function of α5* nAChRs might increase aversion to nicotine and promote cessation of smoking.

Unorthodox ACh binding sites are likely to be especially specific drug targets because α4/α4, α5/α4, and β3/α4 ACh binding sites have unique structural features compared with orthodox sites formed from combinations of α2, α3, α4, or α6 subunits with β2 or β4 or α7/α7 sites. Unorthodox sites are formed by incorporating α4 or α5 and perhaps other α and β3 subunits in pairs that do not include β2 or β4 subunits.

In this study, the effect of agonists at various subunit interfaces was determined through the selective blockage of specific subunit interfaces with the thioreactive agent, MTSET. Here we show that unorthodox ACh binding sites can form at the α5/α4 and β3/α4 interfaces in (α4β2)₂α5 and (α4β2)₂β3 nAChRs and at the α4/α5 in (β2α4)₂α5 nAChRs.

Results

Model Depicting nAChRs Formed by Concatamers

To study ACh homologue sites formed by α4 with α5 or β3 subunits, we constructed these subunit interfaces using concatamers with α4 and β2 subunits linked in different orders (Fig. 1). These illustrations are important for understanding the different concatamers used to form desired interfaces, such as the α5/α4 interface versus the α4/α5 interface. To represent the assembled orientation of linked subunits, nAChRs assembled from β2-α4 and a free subunit are noted as (α4β2)₂*. Correspondingly, (β2α4)₂* represents nAChRs assembled from a α4-β2 concatamer.

FIGURE 1.

Schematic diagram of (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs. A, this illustrates how the β2-α4 dimeric concatamer plus the free α5 or β3 subunit is thought to assemble into nAChRs. B, this illustrates how the α4-β2 dimeric concatamer plus the free α5 or β3 subunit is thought to assemble into nAChRs. Functional agonist sites, + and − sides of subunits, and (AGS)₆ linkers between the C terminus to the N terminus are illustrated. Existence of ACh sites at α5/α4, β2/α5, β3/α4, β2/β3, and α4/α5 will be explored in subsequent experiments.

Fig. 1A illustrates how β2-α4 dimeric concatamer plus free α5 or β3 subunit is thought to assemble into nAChRs. Concatamers are generated by connecting the C-tail of one subunit to the N terminus of the next. The order of the subunits around the cation channel can be determined by the length of the linker between the subunits (30). When the 23-amino acid-long C-terminal tail of β2 subunits is linked by (AGS)₆, or a longer linker to the N terminus of α4, a binding site forms within the linked subunit pair to form an α4/β2 ACh binding site (30). This allows the free subunit to assemble as an accessory subunit in combination with two α4/β2 ACh binding sites.

Fig. 1B illustrates how α4-β2 dimeric concatamer plus free α5 subunit is thought to assemble into nAChRs. The linker of α4-β2 dimeric concatamer is much shorter than that of the β2-α4 dimeric concatamer. When the short 7-amino acid C-tail of α4, compared with the 23-amino acid-long C-tail of β2, is linked by (AGS)₆ to the N terminus of β2, the subunits are constrained to assemble so that α4/β2 ACh binding sites are formed between linked pairs of subunits, and α5 assembles with the plus rather than minus face of α4 to form an α4/α5 interface.

Immunoblots of Triton X-100 extracts from oocytes injected with concatamers confirmed the integrity of expressed proteins (Fig. 2). Concatamer β-6-α and α-6-β migrated at a molecular mass of ∼118 kDa. These data demonstrated that concatamer cRNAs were properly translated in the oocytes, and no degradation to unlinked subunits was seen.

FIGURE 2.

Western blots of wild type β-6-α and α-6-β concatamers co-expressed with α5 in oocytes. Expression of the concatamers β-6-α and α-6-β was detected by rat antisera to β2 or α4 followed by ¹²⁵I-labeled goat anti-rat IgG (2 nm). After washing, the blots were visualized by autoradiography. The Western blots showed that concatamers were not degraded to form any free subunits and that β-6-α and α-6-β migrated at a molecular mass of ∼118 kDa. On each lane the equivalent of half of an oocyte was loaded. These data also demonstrated that the concatamer cRNAs were properly translated in the oocytes, and no degradation to unlinked subunits (∼70 kDa for α4 and ∼57 kDa for β2) was seen.

Pharmacological Properties of (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs

Because the α4/α4 ACh site in (α4β2)₂α4 nAChRs changes the pharmacology of α4β2 nAChRs (6, 12), α5 and β3 subunits may alter agonist selectivity, potency, and efficacy of α4β2 nAChRs if they form agonist sites with α4. We used β2-α4 or α4-β2 concatamers with free subunits to express (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs in Xenopus oocytes. We first evaluated whether α5 and β3 subunits efficiently incorporated with dimeric concatamers.

If α5 or β3 failed to assemble with dimeric concatamers, the functional responses should represent activation of dimeric pentamers resulting from self-assembly of β2-α4 and α4-β2 (Fig. 3A) (30). These dimeric pentamers contain an α4/α4 interface (Fig. 3A). We used NS9283 to test for the presence of dimeric pentamers because this compound acts as a selective agonist for the α4/α4 ACh binding site (8, 9).

FIGURE 3.

Proof of assembly of free subunit with concatamers. A, schematic illustration of dipentameric nAChRs thought to be formed by the β2-α4 and α4-β2 dimeric concatamers alone. The dipentamer is composed of five pairs of linked subunits forming two pentamers and joined by one pair of linked subunits. The α4/α4 site in the dipentameric nAChRs is illustrated. B, NS9283 increased activation of dipentameric nAChRs but not nAChRs expressed from concatamer and free subunit (β2-α4 and α4-β2 only versus β2-α4 + α5, β2-α4 + β3, and α4-β2 + α5). This suggests that sufficient β3 and α5 were used to ensure assembly of all concatamers into monopentamers.

NS9283 increased the responses evoked by ACh almost 6-fold when dimeric concatamers alone were expressed in Xenopus oocytes (Fig. 3B). However, when the dimeric concatamers were expressed with free α5 or β3, NS9283 did not potentiate the responses (Fig. 3B). The absence of potentiation indicates that the cRNA ratios used to obtain (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 only formed monopentamers.

Then we characterized the activation of (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs by ACh, nicotine, and various partial agonists. (α4β2)₂α4 nAChRs exhibited biphasic concentration/response curves with a high sensitivity component caused by the two α4/β2 sites (EC₅₀ = ∼0.7 μm ACh) and a 3–4-fold larger component caused by the third low sensitivity α4/α4 site (EC₅₀ = ∼80 μm ACh) (7, 12, 31, 32). Assaying (α4β2)₂α5 nAChRs using closely spaced ACh concentrations resolved a biphasic concentration/response curve (Fig. 4A). This suggests the presence of high sensitivity (EC₅₀ = 1.38 ± 0.33 μm) ACh binding sites at the two α4/β2 interfaces and a lower sensitivity (EC₅₀ = 22.6 ± 5.8 μm) ACh binding site at the α5/α4 interface (Table 1). (β2α4)₂α5 nAChRs exhibited a monophasic concentration/response curve with an EC₅₀ of 1.25 ± 0.12 μm ACh.

FIGURE 4.

ACh and Sazetidine-A concentration/response curves for (α4β2)₂α5 and (α4β2)₂β3 nAChRs. A, a biphasic ACh concentration/response curve was observed for (α4β2)₂α5 nAChRs. A high sensitivity EC₅₀ = 1.38 ± 0.33 μm and a low sensitivity EC₅₀ = 22.60 ± 5.75 μm were determined. Sazetidine-A is a partial agonist with I_max = 36.3 ± 1.01%. B, a biphasic ACh curve for (α4β2)₂β3 nAChRs was not detected. A single high sensitivity EC₅₀ = 1.56 ± 0.14 μm was determined. Sazetidine-A is a partial agonist with I_max = 33.2 ± 4.9%. I_max is the maximum current peak at the saturating concentration.

TABLE 1.

Pharmacological profiles of (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs for ACh, nicotine, and several partial agonists

The values for the EC₅₀, Hill coefficients, and their relative maximum efficacy to ACh are listed. ND represents values not determined. n ≥ 3 for all concentration/response curves. HS and LS are the high sensitivity and low sensitivity components of the biphasic concentration/response curve, respectively.

	(α4β2)2α5	(α4β2)2β3	(β2α4)2α5
ACh
Biphasic
EC50,HS (μm)	1.38 ± 0.33	ND	ND
EC50,LS (μm)	22.6 ± 5.8
Monophasic
EC50 (μm)	1.85 ± 0.23	1.39 ± 0.33	1.24 ± 0.12
Efficacy (%)	99.0 ± 2.1	95.3 ± 2.1	94.4 ± 2.5
nHill	0.918 ± 0.074	0.951 ± 0.093	0.910 ± 0.049
Nicotine
EC50 (μm)	0.951 ± 0.338	0.602 ± 0.110	0.487 ± 0.127
Efficacy (%)	35.4 ± 3.0	23.8 ± 1.2	59.1 ± 4.8
nHill	0.910 ± 0.154	1.32 ± 0.26	0.883 ± 0.060
Varenicline
EC50 (μm)	0.139 ± 0.044	0.111 ± 0.043	0.058 ± 0.026
Efficacy (%)	11.5 ± 0.7	11.6 ± 0.8	14.4 ± 1.23
nHill	0.856 ± 0.189	0.560 ± 0.091	0.538 ± 0.045
Sazetidine-A
EC50 (nM)	8.68 ± 0.92	7.08 ± 3.49	1.31 ± 0.30
Efficacy (%)	36.3 ± 1.0	33.2 ± 4.9	45.7 ± 2.6
nHill	1.20 ± 0.12	1.29 ± 0.45	1.17 ± 0.26
Cytisine
EC50 (μm)	0.501 ± 0.085	2.52 ± 0.77	0.444 ± 0.090
Efficacy (%)	8.02 ± 0.44	11.0 ± 0.7	8.51 ± 0.34
nHill	0.973 ± 0.164	0.562 ± 0.057	0.852 ± 0.119

A biphasic concentration/response curve was not resolved for (α4β2)₂β3 nAChRs. A single EC₅₀ (1.56 ± 0.14 μm) for ACh was determined (Fig. 4B and Table 1). This indicates that either β3/α4 sites have the same sensitivity as α4/β2 or the functional effect of β3/α4 sites is not detectable.

The potencies and efficacies of nicotine and various partial agonists were also determined (Table 1). Pharmacological profiles of the (α4β2)₂α5 and (α4β2)₂β3 nAChRs were very similar. However, (α4β2)₂α5 nAChRs were 5-fold more sensitive to cytisine (EC₅₀ = 0.501 ± 0.095 μm) than (α4β2)₂β3 nAChRs (EC₅₀ = 2.25 ± 0.77 μm). Potencies and efficacies of nicotine, sazetidine-A, and varenicline were similar for both subtypes.

Sazetidine-A acts as a partial agonist for both (α4β2)₂α5 and (α4β2)₂β3 nAChRs (Fig. 4) despite acting as a full agonist for the (α4β2)₂β2 subtype (11). This is because sazetidine-A is only able to bind to two of the three sites, at the α4/β2 interfaces, but not the third unorthodox site at the α5/α4 and β3/α4 interfaces in which there is a histidine 142 at the minus face of α4 to prevent binding of sazetidine-A (11, 12).

The pharmacological properties of (β2α4)₂α5 nAChRs were very similar to those of (α4β2)₂α5 nAChRs; however, (β2α4)₂α5 nAChRs were almost 3-fold more sensitive to varenicline and almost 2-fold more sensitive and efficacious to nicotine than (α4β2)₂α5 nAChRs (Table 1). In addition, nicotine was ∼2-fold more efficacious for (α4β2)₂α5 nAChRs compared with (β2α4)₂α5 nAChRs. The difference in varenicline and nicotine sensitivities was within the range of error, but the difference in nicotine efficacies was not. Such pharmacological differences may result from the different assembly of the two subtypes. (α4β2)₂α5 nAChRs contain two α4/β2 sites and one α5/α4 site, whereas (β2α4)₂α5 nAChRs contain only one α4/β2 site and one α4/α5 site.

Cysteine Mutants Used to Study Specific Interfaces Formed by β3 and α5 Subunits

Because sazetidine-A is a partial agonist for (α4β2)₂α5, (α4β2)₂β3, and (β2α4)₂α5 nAChRs, it is likely that sazetidine-A only binds to the α4/β2 interfaces, but ACh binds to both the α4/β2 and additional sites formed by β3 or α5 subunits (Fig. 1). To test this hypothesis and evaluate which interfaces forms ACh sites, we mutated a cysteine at the minus site of various subunit interfaces (Figs. 5 and 6 and Table 2). Alkylation of this cysteine residue will block the ACh homologue sites in the corresponding subunit interfaces (9, 33). If this interface forms an agonist site, MTSET will block the activation of ACh. No change of response is expected if MTSET reacts with the cysteine at interfaces that do not bind an agonist (Figs. 5 and 6) (9, 33).

FIGURE 5.

Schematic illustration of alkylation of (α4β2)₂α5 nAChRs. nAChRs were expressed from wild type or mutant β2-α4 dimeric concatamers co-expressed with wild type or mutant α5 subunits. The α4^T126C mutation was introduced to provide a cysteine at the minus face of the α5/α4 interface, and the α5^T139C mutation was introduced to provide a cysteine at the minus face of the β2/α5 interface, where alkylation of the cysteine by MTSET allowed for the selective blockage of the desired interface. The free cysteine at position 2 of α5 was mutated to a cysteine-null pseudo wild type α5^C2S.

FIGURE 6.

Schematic illustration of alkylation of (α4β2)₂β3 nAChRs. nAChRs were expressed from wild type or mutant β2-α4 dimeric concatamers co-expressed with wild type or mutant β3 subunits. The α4^T126C mutation was introduced to provide a cysteine at the minus face of the β3/α4 interface, and the β3^T123C mutation was introduced to provide a cysteine at the minus face of the β2/β3 interface, where alkylation of the cysteine allowed for the selective blockage of the desired interface by MTSET.

TABLE 2.

Mutants used for MTSET alkylation experiments

Locations of mutations are underlined. Binding site mutations are in loop E on the minus side of the site. The C2S mutation is near the N terminus.

Mutant	Location of mutations
α4T126C	VQWTPPA
α5T139C	VTWTPPA
β2L121C	IFWLPPA
β3T123C	VVWTPPA
α5C2S	AGRCGLA

The α5/α4 and β3/α4 interfaces can be selectively blocked by mutating threonine 126 on the minus face of the α4 subunit to a cysteine residue followed by alkylation of the cysteine with MTSET. The β2/α5 and β2/β3 sites were investigated by mutating the threonine 139 residue on the minus face of the α5 subunit and mutating the threonine 123 residue on the minus face of the β3 subunit. The α5 subunit contains a free cysteine at position 2. To prevent nonspecific modification or any potential disulfide formations between the single free cysteine at position 2, we mutated the cysteine to a serine (33). The mutations did not significantly change the level of expression. Maximum responses evoked by ACh were similar among WT and mutants: 1.62 ± 0.21 μA for (α4β2)₂α5, 1.78 ± 0.14 μA for (α4^T126Cβ2)₂α5, 1.21 ± 0.56 μA for (α4β2)₂α5^T139C, 1.11 ± 0.29 μA for (α4β2)₂β3, 1.70 ± 0.37 μA for (α4^T126Cβ2)₂β3, and 1.58 ± 0.45 μA for (α4β2)₂β2^T123C nAChRs.

The concentration of MTSET used for alkylation experiments was 0.5 mm for α5 containing nAChRs because higher concentrations resulted in a significant decrease in the ACh response of wild type nAChRs. The concentration of MTSET used for alkylation experiments for β3 containing nAChRs was 2 mm, the concentration at which the wild type response was not affected after alkylation. With 5 mm MTSET, blockage (∼30%) was similar to that with 2 mm.

MTSET Blocks Activation from the α5/α4 Site

The β2-α4^T126C dimeric concatamer was expressed with α5 to allow blockage of the α5/α4 interface (Fig. 5). (α4^T126Cβ2)₂α5 nAChRs have intact agonist sites at the α4/β2 interfaces, a blockable site at the α5/α4 interface, and a blockable site at the nonfunctional β2/α4 interface (Fig. 5). Alkylation of the cysteine mutation at the β2/α4 interface does not block activation (9). Therefore, any decrease in response after MTSET treatment of (α4^T126Cβ2)₂α5 nAChRs can be attributed to specific block of the α5/α4 interface.

Wild type nAChRs were tested before and after MTSET treatment as a negative control for the mutant nAChRs (Fig. 7A). Incubation of oocytes with MTSET for 5 min reduced activation by ACh for (α4^T126Cβ2)₂α5 nAChRs by 45–50% of the original ACh response (Figs. 6C and 7B). Although response evoked by 30 μm ACh was blocked ∼50% (p = 0.0016), MTSET did not significantly affect responses of wild type or mutant (α4β2)₂α5 nAChRs to a saturating (100 nm) concentration of sazetidine-A (Fig. 7B). This confirms the idea that sazetidine-A can bind to and activate α4/β2 sites but not α5/α4 sites, which have a histidine 142 residue on the minus side of α4 that prevents binding of sazetidine-A (10–12).

FIGURE 7.

Effect of MTSET alkylation on activation of (α4β2)₂α5 nAChRs. A, concentration/response curves for wild type (α4β2)₂α5 nAChRs before and after alkylation with no significant block in response. B, MTSET blocked activation by ACh (30 μm), but not activation by sazetidine-A (100 nm) of mutant (α4^T126Cβ2)₂α5 nAChRs. MTSET had no effect on the negative control, wild type (α4β2)₂α5 nAChRs. A Student t test was used to compare responses before and after treatment of MTSET (0.5 mm). **, p < 0.01. C–F, ACh concentration/response curves before and after alkylation of mutant (α4β2)₂α5 nAChRs. MTSET blocked responses of (α4^T126Cβ2)₂α5, (α4^T126Cβ2)₂α5^C2S, and (α4^T126Cβ2)₂α5^{C2S, T139C} nAChRs but not of (α4β2)₂α5^{C2S, T139C} nAChRs. There is no significant difference in the block of response when compared with dimeric concatamer plus wild type α5 and the α5^C2S.

When α5^C2S was co-expressed with the β2-α4^T126C dimeric concatamer, there was no additional blockage by MTSET compared with using wild type α5, indicating that the block observed in the wild type α5 was solely due to blockage of activation by ACh at the α5/α4 interface (Fig. 7D).

Unlike the α5/α4 interface, the β2/α5 interface did not form a functional ACh binding site. When α5^{C2S, T139C} was co-expressed with the β2-α4 dimeric concatamer, the response after alkylation was similar to the wild type. In addition, when the β2-α4^T126C dimeric concatamer was co-expressed with α5^{C2S, T139C}, there was no significant increase in the blockage as seen previously with (α4^T126Cβ2)₂α5 nAChRs (Fig. 7E).

This suggests that (α4β2)₂α5 nAChRs contain three functional ACh binding sites: two within the dimeric concatamers at the two α4/β2 interfaces and one at the α5/α4 interface. All the values for the ACh concentration/response curves before and after MTSET alkylation for α5 containing nAChRs are presented in Table 3.

TABLE 3.

Summary of potencies and efficacies of ACh activating wild type and mutant (α4β2)₂α5 nAChRs before and after MTSET treatment

The nAChRs are expressed from concatamer β2-α4 and free α5 subunit in Xenopus oocytes. Efficacy indicates the maximum efficacy relative to ACh before MTSET treatment. n indicates the number of oocytes tested.

	β2-α4 + α5	β2-α4T126C + α5	β2-α4T126C + α5C2S	β2-α4 + α5C2S,T139C	β2-α4T126C + α5C2S,T139C
No MTSET
EC50 (μm)	1.85 ± 0.23	2.77 ± 0.38	2.22 ± 0.30	1.81 ± 0.16	1.31 ± 0.42
Efficacy (%)	99.0 ± 2.1	96.0 ± 2.4	100.4 ± 2.5	102.2 ± 1.8	103.0 ± 5.8
nHill	0.918 ± 0.074	0.824 ± 0.076	0.911 ± 0.094	1.08 ± 0.16	0.821 ± 0.186
n	7	7	3	4	4
After MTSET
EC50 (μm)	2.36 ± 0.22	2.97 ± 0.33	1.33 ± 0.46	1.60 ± 0.18	1.33 ± 0.31
Efficacy (%)	92.9 ± 1.2	56.6 ± 1.3	61.4 ± 2.3	91.2 ± 2.3	58.5 ± 1.7
nHill	0.893 ± 0.039	1.05 ± 0.10	0.898 ± 0.238	1.28 ± 0.16	1.11 ± 0.13
n	7	7	3	6	10

MTSET Blocks Activation from the β3/α4 Site

β2-α4^T126C dimeric concatamer was expressed with the free β3 subunit to allow blockage of the β3/α4 interface (Fig. 6). Similarly to (α4^T126Cβ2)₂α5 nAChRs, these nAChRs have intact agonist sites at the α4/β2 interfaces, a blockable site at the accessory β3/α4 interface, and a blockable site at the nonfunctional β2/α4 interface (Fig. 6).

Wild type nAChRs were tested before and after MTSET treatment to demonstrate that there was no block in response and to provide a comparison with the mutant nAChRs (Fig. 8A). Incubation of oocytes with MTSET for 5 min reduced activation by ACh for (α4^T126Cβ2)₂β3 nAChRs by ∼25–30% of the original ACh response (p = 0.0061) (Fig. 8, B and C). However, there was no block of response to a saturating (100 nm) concentration of sazetidine-A in this mutant (Fig. 8B). This confirms the idea that sazetidine-A can bind to and activate the α4/β2 sites but not β3/α4 sites.

FIGURE 8.

Effect of MTSET alkylation on activation of (α4β2)₂β3 nAChRs. A, concentration/response curves of wild type (α4β2)₂β3 nAChRs before and after 2 mm MTSET treatment. There was no significant block in response. B, MTSET (2 mm) blocked maximal responses of ACh (30 μm) but not sazetidine-A (100 nm) for (α4^T126Cβ2)₂β3 nAChRs. Student t test was used to compare responses before and after treatment of MTSET. **, p < 0.01. C and D, ACh concentration/response curves before and after alkylation of mutant (α4β2)₂β3 nAChRs. MTSET blocked responses of (α4^T126Cβ2)₂β3 nAChRs but not of (α4β2)₂β3^T123C nAChRs.

In addition to the β3/α4 interface, we also investigated whether the β2/β3 interface formed a functional ACh binding site by mutating threonine 123 at the minus face of β3 to a cysteine. Unlike the β3/α4 interface, the β2/β3 interface did not form an ACh binding site. When β3^T123C was co-expressed with the β2-α4 dimeric concatamer, the response after alkylation was similar to the wild type subtype (Fig. 8D).

This suggests that (α4β2)₂β3 nAChRs contain three functional ACh binding sites: two within the dimeric concatamers at the two α4/β2 interfaces and one at the β3/α4 interface. The potencies and efficacies of ACh before and after MTSET alkylation for β3 containing nAChRs are presented in Table 4.

TABLE 4.

Summary of potencies and efficacies of ACh activating wild type and mutated (α4β2)₂β3 nAChRs before and after MTSET treatment

The nAChRs are expressed from concatamer β2-α4 and free α5 subunit in Xenopus oocytes. Efficacy indicates the maximum efficacy relative to ACh before MTSET treatment. n indicates the number of oocytes tested.

	β2-α4 + β3	β2-α4T126C + β3	β2-α4 + β3T123C
No MTSET
EC50 (μm)	1.56 ± 0.14	2.00 ± 0.20	2.11 ± 0.30
Efficacy (%)	101.2 ± 1.9	102.5 ± 2.2	101.5 ± 2.7
nHill	1.10 ± 0.09	0.890 ± 0.064	0.993 ± 0.103
n	8	9	5
After MTSET
EC50 (μm)	1.48 ± 0.14	1.03 ± 0.17	1.84 ± 0.31
Efficacy (%)	109.1 ± 2.3	78.6 ± 2.5	104.9 ± 3.4
nHill	1.27 ± 0.13	0.980 ± 0.131	1.07 ± 0.16
n	5	6	5

MTSET Blocks Activation from the α4/α5 Site

We next investigated whether an ACh binding site existed at the α4/α5 interface that formed when co-expressing a constrained α4-β2 dimeric concatamer with α5. To characterize any block of ACh activation from the α4/α5 site, the α5^{C2S, T139C} subunit was co-expressed with α4-β2 dimeric concatamer (Fig. 9A). The cysteine-null α5^C2S subunit was used as the negative control (Fig. 9A).

FIGURE 9.

Single-site activation of nAChRs. A, schematic illustration of nAChRs expressed from concatamers and free subunit. MTSET does not react with wild type cysteine-null pseudo wild type (β2α4)₂α5^C2S or (α4β2α4β2)β2 nAChRs. MTSET blocks one or two ACh sites and leaves a single ACh site in (α4^L121Cβ2α4β2)β2, (β2α4)₂α5^C2S,T139C, and (α4β2^L121C)₂α4 nAChRs. B, effect of MTSET on activation of nAChRs by ACh (300 μm). Almost full response was seen for cysteine-null pseudo wild type (β2α4)₂α5^C2S or wild type (α4β2α4β2)β2 nAChRs, partial response (45.6 ± 5.3 and 39.1 ± 5.1%) was seen for (β2α4)₂α5^{C2S, T139C}and (α4β2α4β2^L121C)α4 nAChRs, and little response was seen after alkylation of (α4β2^L121C)₂α4 nAChRs.

MTSET blocks activation from the α4/α5 site, reducing the response to 45.80 ± 0.05% after alkylation (Fig. 9B). These nAChRs are thought to have one conventional agonist site at the α4/β2 interface and one unorthodox site at the α4/α5 interface that can be blocked by MTSET. Therefore any decrease in response seen after MTSET treatment of (β2α4)₂α5 nAChRs can be attributed to specific block of the α4/α5 interface. MTSET only blocked ∼54% of maximal response of (β2α4)₂α5 nAChRs. This remaining response could result from activation from the single conventional ACh site in the α4/β2 interface or inefficient blockage of the α4/α5 site.

To study whether nAChRs can be activated by a single-site, the β2^L121C-α4-β2-α4 tetrameric concatamer was co-expressed with β2. Free β2 assembles efficiently with β2-α4-β2-α4 tetrameric concatamer to form monopentamers, ensuring the block seen was attributed to specific block of the α4/β2 site (9). When one of the two α4/β2 sites is alkylated, there is still 39.1 ± 5.1% activation (Fig. 9). This supports the idea that one site can activate the channel, which is consistent with previous reports (34–36).

To check efficiency of blocking ACh activation under these conditions, the corresponding amino acid in the β2 subunit of the β2-α4 dimeric concatamer, leucine 121, was mutated to cysteine, as previously described (9). When both of the α4/β2 sites were blocked with MTSET, blockage of response was essentially complete (Fig. 9).

This indicates that one α4/β2 was able to sufficiently activate the channel by itself if another functional binding site was deactivated, but when both of the α4/β2 sites were deactivated, the channel was not able to open with only the α4/α4 binding site. Therefore, the remaining response seen in MTSET-treated (β2α4)₂α5 nAChRs is likely from activation from the single conventional ACh site in the α4/β2 interface.

Presence of Pentameric α5β2 Assemblies without Functional ACh Binding Sites

Because MTSET blocks activation from α5/α4 and α4/α5 sites, α5 can assemble like a conventional α4 or β2 subunit to form an ACh binding site. We showed that the β2/α5 site did not form an ACh site (Figs. 5 and 7E). To investigate whether α5/β2 forms an ACh binding site, we injected oocytes with free β2 and α5. Any response detected should result from activation of α5/β2 site.

We first confirmed that α5 and β2 assembles into pentameric nAChRs using sucrose gradient sedimentation (Fig. 10A). These sucrose gradient fractions were immunoisolated by mAb210 (against α5 subunit) and detected by ¹²⁵I-mAb 295 (against β2 subunit) to study assembly of nAChRs containing both β2 and α5 subunits. β2 and α5 monomeric subunits assembled to form aggregates that sedimented at nearly the size of 9.5S Torpedo nAChR monomers and 13S dimers. Aggregates at the size of 13S dimers of Torpedo nAChRs suggest the presence of dipentamers (Fig. 10B). Dipentamers might form through disulfide linking of cysteines near the N terminus of α5. Torpedo nAChR dipentamers are linked by disulfide bonds between cysteines at the C terminus of δ subunits (37).

FIGURE 10.

Sucrose gradient sedimentation and surface expression of oocytes injected with free β2 + free α5. A, sucrose gradient sedimentation of Triton X-100 extract from oocytes injected with β2 and α5. Components containing α5 subunits of the 63 sucrose fractions were immunoisolated using mAb 210-coated wells. α5β2 nAChRs in these wells were then quantified using I¹²⁵-mAb 295 (against β2 subunit). An internal standard in the gradient was Torpedo nAChR that was resolved into 9.5S monomers and 13S dimers. Arrows indicate the locations of the monomer and dimer peaks of Torpedo nAChRs. The gradient revealed that the wild type β2 and α5 free subunits associated to form aggregates that sedimented at nearly the size of 9.5S Torpedo nAChR monomers and 13S dimers of Torpedo nAChRs. B, surface expression of nAChRs in oocytes injected with β2 and α5 subunits. Radioactively labeled ¹²⁵I-mAb 295 to β2 subunit was used to determine surface expression of nAChRs, and uninjected oocytes were used to determine nonspecific binding.

We next studied whether these pentameric and dipentameric α5β2 nAChRs reached the surface and responded to ACh activation. Live Xenopus oocytes were labeled with 1.13 ± 0.15 fmol of ¹²⁵I-mAb 295, ∼4-fold higher than nonspecific labeling (Fig. 10B). Although α5β2 nAChRs expressed on the oocyte surface, no functional response was detected when various concentrations of ACh were tested. In addition, there were no high affinity epibatidine binding sites detected when the extracts were incubated with 1 nm ³H-labeled epibatidine for 2 h at room temperature on mAb 295-coated Immulon wells. This suggests that the α5β2 oligomers most likely form both β2/α5 and α5/β2 interfaces, neither activated by ACh.

Discussion

Discovery of an α4/α4 interface that formed an unorthodox ACh binding site that greatly potentiated the effect of the two orthodox α4/β2 ACh binding sites and is a target for the site-selective agonist drug NS9283 altered our understanding of how nAChRs work (6, 9). Here we show that α5 and β3 subunits, which were previously not thought to form ACh binding sites, can form α5/α4, β3/α4, and α4/α5 ACh binding sites. The α5/α4 ACh binding site increases the response from (α4β2)₂α5 nAChRs almost 2-fold. α5 subunit greatly increases the Ca²⁺ permeability of (α4β2)₂α5 nAChRs (38). Increased response and high Ca²⁺ permeability caused by the presence of α5 makes the α5/α4 ACh binding site on this nAChR subtype an appealing drug target. Unorthodox sites may be useful for targeting drugs that can be exceptionally specific for important nAChR subtypes. The unique binding sites at the α5/α4 and β3/α4 interfaces are sites at which ACh site-selective agonists might act, much as NS9283 is able to act as an ACh site-selective agonist at the unique α4/α4 interfaces of (α4β2)₂α4 nAChRs to potentiate their function (8, 9).

It has been previously assumed that α5 and β3 subunits cannot participate in forming ACh binding sites but do play a role in increasing Ca²⁺ permeability, sensitivity to activation, expression, and further up-regulation by nicotine for various α4β2* nAChRs (13–15). In (α4β2)₂α4 nAChRs, the α4 subunit occupies the accessory position and forms a low affinity ACh site at the α4/α4 interface. We investigated whether the α5 and β3 subunits act similar to the α4 subunit in α4β2 nAChRs to form an ACh binding site at the α5/α4 and β3/α4 interfaces of (α4β2)₂α5 and (α4β2)₂β3 nAChRs.

First, we showed that the α5 subunit can form a functional ACh binding site at the α5/α4 interface in (α4β2)₂α5 nAChRs where its primary (+) face contributes to forming the subunit interface and also a site at the α4/α5 interface in (β2α4)₂α5 nAChRs where its complementary (−) face contributes to forming the subunit interface. Second, we showed that α5 can only form ACh binding sites with α4 but not β2. This is surprising because α2, α3, α4, and α6 readily form ACh binding sites with β2. However, this is consistent with the idea that unorthodox ACh binding sites have novel features that make them highly specific drug targets. The α5/α4 binding site contributes almost half of the total ACh response of (α4β2)₂α5 nAChRs. Lastly, we showed that, similar to α5, β3 subunits are able to form a functional ACh binding site at β3/α4 interfaces in (α4β2)₂β3 nAChRs. The accessory β3 subunit, in (α4β2)₂β3 nAChRs, was able to form a functional ACh binding site at the β3/α4 interface, contributing ∼30% of the ACh response.

A previous study concluded that a functional α5/α4 binding site was not present in mouse (α4β2)₂α5 nAChRs (39). In their studies, when the aromatic box residues in α5 were mutated to inhibit the interaction with the quaternary amine of ACh in α5, they observed no change in EC₅₀ and concluded that there is no ACh binding site present at the α5/α4 interface. However, these mutations did cause a large decrease in the amplitude of response, which is consistent with our findings that the response is potentiated by a third ACh site present at the α5/α4 interface.

It has previously been reported that α4/α5 may form a functional ACh binding site (40). Our results are consistent with this observation.

Our data suggest that neither α5/β2 nor β2/α5 can form ACh binding sites. It is known that α3/β2 forms ACh binding sites but β2/α3 does not (41). However, β2/α3 can be activated by the cholinergic agonist morantel, allowing morantel to potentiate the ACh response of (α3β2)₂β2 AChRs (41). This suggests that with a suitable cholinergic agonist, β2/α5 or α5/β2 might also be activated.

In conclusion, unorthodox ACh binding sites can form at the α5/α4 and β3/α4 interfaces in (α4β2)₂α5 and (α4β2)₂β3 nAChRs and at the α4/α5 in (β2α4)₂α5 nAChRs. Looking forward, these unique interfaces can be targets for unique drugs, specifically site-selective agonists to bind to them for potential treatment of diseases or nicotine addiction. These unorthodox sites provide reason to believe that unorthodox sites might also exist at β3/α6 interfaces. This site would provide a specific target on the complex (α6β2)(α4β2)β3 nAChR subtype that is important in regulating dopamine release and potentially important in smoking cessation therapy and other applications associated with learning and motor control (15, 16). The synthesis of specific drugs that can selectively bind to these unorthodox sites might play a significant role in the treatment of nicotine addiction through activation of (α4β2)₂α5 (29) or antagonism of (α6β2)(α4β2)β3 nAChRs (18).

Experimental Procedures

Chemicals

MTSET was purchased from Toronto Research Chemicals Inc. (North York, Canada). NS9283 was synthesized as described (42). 25 and 10 mm stocks of NS9283 were prepared in DMSO. MTSET solutions were freshly prepared from solid daily and kept on ice until used. Dilutions of all drugs were prepared daily in ND96 testing buffer before use. All other chemicals were purchased from Sigma-Aldrich unless otherwise noted.

Preparation of cDNAs and cRNAs, Mutagenesis

Human β2 and α4 cDNAs were cloned as described (43, 44). cDNA for human α5 was provided by Dr. Francesco Clementi (University of Milan, Milan, Italy), and human β3 was provided by Christopher Grantham (Janssen Research Foundation, Beerse, Belgium) (15). Syntheses of concatamers, β2-(AGS)₆-α4 (abbreviated β2-α4), α4-(AGS)₆-β2, and β2-(QAP)_n-α4-(QAP)_n-β2 (abbreviated β2-α4-β2) (abbreviated α4-β2) were described (30, 45).

Point mutations to various subunits were introduced using the PfuUltra high fidelity DNA polymerase (Agilent, Santa Clara, CA) or the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla CA), following the manufacturer's instructions. Two separate point mutations were introduced in α5. A threonine to cysteine mutation at position 139 was introduced to allow alkylation of ACh site at minus face of α5 subunit. To prevent nonspecific modification or any potential disulfide formations between the single free cysteine at position 2 of α5, this cysteine was mutated to serine. One point mutation was introduced in β3. The threonine present at position 123 in the minus side of β3 was mutated to a cysteine. Mutated dimeric concatamers, β2^L121C-(AGS)₆-α4 (abbreviated β2^L121C-α4) and β2-(AGS)₆-α4^T126C (abbreviated β2-α4^T126C), were prepared as described (9). All oligonucleotides and the amino acids mutated are listed in Table 2. The mature amino acid sequences were used to number nAChR subunits.

To prepare the β2^L121C-α4-β2-α4 tetrameric concatamer, the β2^L121C-α4 dimeric concatamer and the β2-α4-β2 trimeric concatamer were cut by BsiWI restriction enzyme unique to β2 subunit. The 3390-bp BsiWI fragment from β2-α4-β2 trimeric concatamer was inserted into β2^L121C-α4 concatamer using T4 DNA ligase (New England Biolabs). The correct clone was chosen from the one that contained two fragments, 3400 and 7100 bp, after digestion by BstEII restriction enzyme (unique for α4).

All mutagenesis and ligation products were transformed into XL-10 Gold Ultracompetent cells (Stratagene) using the protocol provided with the kit. After performing a miniprep using the FastPlasmid mini kit (5Prime, Hilden, Germany), the right clone was chosen through sequencing of the miniprep DNA. High quality DNA was purified using Qiagen plasmid midiprep kit (Qiagen). The cDNAs were linearized using restriction enzymes (AseI for β2-α4, α4, α4-β2, and β2-α4-β2-α4; PvuII for β2; and EcoRI for α5 and β3). Linearized cDNA was used as a template to prepare cRNAs using mMessage mMachine kits (Ambion, Austin, TX). SP6 RNA polymerase was used for transcribing cDNAs of α4, α5, β2, β3, β2-α4, and β2-α4-β2-α4 in pSP64 vectors. The concentrations of synthesized cDNAs and cRNAs were measured by a spectrophotometer, and their quality was determined by agarose gel electrophoresis.

Xenopus Oocyte Preparation

Oocytes were harvested from Xenopus laevis frogs in accordance to the approved institutional animal care and use committee protocol as described (9, 46). The oocytes were washed three times with OR2 (85.5 mm NaCl, 2 mm KCl, 1 mm MgCL₂, and 5 mm HEPES, pH 7.5) solution and three times with 50% Leibovitz-15 (L-15) medium (Invitrogen), 10 mm HEPES, pH 7.5, 10 units/ml penicillin, 10 μg/ml streptomycin. The oocytes were then placed in Leibovitz-15 medium for a couple hours to allow for recovery before injection.

cRNA Microinjection

Subunit cRNAs were mixed in different ratios to produce the desired constructs. Proper ratios were determined by performing expression tests with ACh only and ACh coappplied with 10 μm NS9283. If potentiation was seen with NS9283, this indicated the presence of dipentamers and the ratio was adjusted until no potentiation was seen. To obtain (α4β2)₂α5 nAChRs, oocytes were injected with a total of 5–10 ng of cRNA, two parts of β2-α4 to one part of free α5. To obtain (α4β2)₂β3 nAChRs, the oocytes were injected with a total of 90 ng of cRNA, one part of β2-α4 to two parts of free β3. To obtain (β2α4)₂α5 nAChRs, the oocytes were injected with 150 ng of cRNA, six parts of α4-β2 to one part of free α5. To obtain (α4β2α4β2)β2 nAChRs, the oocytes were injected with 30 ng of cRNA, two parts of β2-α4-β2-α4 and one part of β2. For surface labeling assays, 60 ng of free β2 and free α5 was injected at a one to one ratio. These cRNA ratios are indicated in terms of weight. Because cysteine mutations in this study did not change protein expression, the cRNA ratios of mutants used were the same as wild type. After microinjection, the oocytes were incubated in medium made up of L-15 with 50 μg/ml gentamycin and switched out to L-15 without gentamycin before electrophysiological recording. Function was assayed 3–7 days after injection.

Surface Labeling

Radioactively labeled ¹²⁵I-mAb 295 to β2 was used to determine surface expression of the nAChRs (47). Surface labeling was performed on the same day as electrophysiological recording. Each set of oocytes, injected with different constructs, was placed in separate Eppendorf tubes with a total volume of 500 μl of L-15 medium containing 3% bovine serum albumin with 5 nm ¹²⁵I-mAb 295 at room temperature for a minimum of 3 h (47). The oocytes were washed three times to remove unbound ¹²⁵I-mAb 295 to β2 or until there was no change in activity measured by a Geiger counter. The oocytes were put into individual tubes, and bound ¹²⁵I was measured using a γ-counter. Uninjected oocytes were used as the control for nonspecific binding.

Sucrose Gradient Sedimentation

A group of 60 oocytes was homogenized in 1 ml of buffer A (50 mm Na₂HPO₄, pH 7.5, 50 mm NaCl, 5 mm EDTA, 5 mm EGTA, 5 mm benzamidine, 15 mm iodoacetamide, and 2 mm phenylmethylsulfonyl fluoride) and then pelleted by centrifuging at 13,400 rpm for 15 min at 4 °C. The pellets were resuspended by pipetting, and membrane proteins were solubilized in 150 μl of buffer A containing 2% Triton X-100 for 1 h at room temperature on a rotator. Debris was removed by centrifugation at 13,400 rpm for 15 min at 4 °C. Aliquots (150 μl) of the lysates, mixed with 1 μl of 2 mg/ml purified Torpedo californica electric organ nAChR, were loaded onto 11.4 ml sucrose gradients (linear 5–20% sucrose (w/w) in 10 mm sodium phosphate buffer, pH 7.5, that contained 100 mm NaCl, 1 mm NaN₃, 5 mm EDTA, 5 mm EGTA, and 0.5% Triton X-100). Gradients were centrifuged for 16 h at 40,000 rpm in a SW-41 rotor (Beckman Coulter, Fullerton, CA) at 4 °C. Fractions were collected at 15 drops/well from the bottom of the tubes and used for additional analysis. Then 50 μl of each fraction were transferred to mAb 210-coated wells to isolate α5-containing nAChRs and incubated with 5 nm ¹²⁵I-mAb 295 to detect β2-containing nAChRs. Another 20 μl of each fraction were transferred to mAb 210-coated wells incubated with 1 nm ¹²⁵I-α-bungarotoxin overnight at 4 °C to isolate and detect α1-containing T. californica nAChRs, which are used as molecular mass standards. Afterward, the wells were washed three times with PBS and 0.5% Triton X-100, and bound ¹²⁵I-mAb 295 and ¹²⁵I-α-bungarotoxin were determined by γ-counting.

Solid Phase Radioimmunoassay

Total epibatidine binding was determined using an increasing amount of extract loaded with 1 nm [³H]epibatidine (PerkinElmer Life Sciences) in a total volume of 100 μl in PBS buffer containing 0.5% Triton X-100 and 10 mm NaN₃ (47). Binding took place on a horizontal rotator for a minimum of 2 h at room temperature or overnight at 4 °C. After incubation, the wells were washed three times with 0.5% Triton X-100 in PBS before elution with 30 μl of 0.1 m NaOH solution. Bound radioactivity was determined using a 1450 Trilux Microbeta liquid scintillation counter (PerkinElmer Life Sciences) with OptiPhase scintillation fluid as described (47). Nonspecific binding was determined using uninjected oocytes.

Electrophysiology

Functions of nAChRs were measured by two-electrode voltage clamp using either a manual two-microelectrode voltage clamp amplifier setup (oocyte clamp OC-725; Warner Instrument, Hamden, CT) or OpusXpress 6000A (Molecular Devices, Union City, CA), an integrated system that provides automated impalement, voltage clamp, and simultaneous drug delivery to eight oocytes in parallel (48) (49). Voltage was held at −50 mV.

In the agonist concentration/response experiments, the oocytes first received two control applications of 300 μm ACh followed by applications of increasing concentrations of agonists. The peak amplitudes were normalized to the average of the maximum ACh response evoked by initial two controls.

Alkylation experiments were performed with the sulfhydryl agent, MTSET. The oocytes were placed in eight separate chambers and received 0.5 mm MTSET, for α5 containing nAChRs, or 2 mm MTSET, for β3 containing nAChRs, for 60 s at a rate of 0.9 ml/min, and then incubated in the MTSET for 5 min with no flow of ND96 buffer to retain the reagent in the bath. MTSET concentration was determined by finding the highest concentration that did not block the wild type nAChRs. After incubation, flow continued, and oocytes were washed for 287 s. Immediately after the washing steps, increasing concentrations of ACh were applied. The oocytes were discarded after MTSET treatment because of chemical modification. In the MTSET experiments, the peak amplitudes were normalized to the maximum ACh response prior to MTSET treatment.

In NS9283 experiments, 10 μm NS9283 was co-applied with 3 μm ACh and normalized to two controls of the same ACh concentration in which 0.1% DMSO was used as the vehicle control, instead of the 10 μm NS9283. The mean and standard error were both calculated from normalized responses in all experiments. All concentration/response values were expressed ± standard error.

The Hill equation was fit to the concentration/response relationship using a nonlinear least squares curve fit method (Kaleidagraph; Abelbeck/Synergy, Reading, PA): I(x) = I_max [x^nH/(x^nH + EC₅₀^nH)], where I(x) is the peak current measured at the agonist concentration x, I_max is the maximum current peak at the saturating concentration, EC₅₀ is the agonist concentration required to achieve half of the maximum response, and n_H is the Hill coefficient. All statistical analyses were performed using a Student's t test (two-tailed; *, p < 0.05; **, p < 0.01; and ***, p < 0.001).

Western Blots

Proteins from oocytes expressing linked subunits were solubilized in 2% Triton X-100, resolved into subunits by SDS-PAGE, and then transferred to Immun-Blot polyvinylidene difluoride membrane (0.2 μm; Bio-Rad). The blots were probed with rat antiserum to α4 (diluted 1:1000) or β2 (diluted 1:500) followed by ¹²⁵I-labeled goat anti-rat IgG (2 nm). After washing, blots were visualized by autoradiography.

Author Contributions

A. J., A. K., and J. L. designed the experiments. A. J., J. L., and J. W. wrote the manuscript. A. J. performed electrophysiological, epibatidine, and sucrose gradient sedimentation assays and analyzed data. A. J. and J. W. prepared mutant cDNAs. AK prepared concatamers and performed western blots. J. W. provided preliminary data. T. M. K. provided NS9283. All authors revised and approved the final version of the manuscript.