α-Conotoxin AuIB Isomers Exhibit Distinct Inhibitory Mechanisms and Differential Sensitivity to Stoichiometry of α3β4 Nicotinic Acetylcholine Receptors*

Non-native disulfide isomers of α-conotoxins are generally inactive although some unexpectedly demonstrate comparable or enhanced bioactivity. The actions of “globular” and “ribbon” isomers of α-conotoxin AuIB have been characterized on α3β4 nicotinic acetylcholine receptors (nAChRs) heterologously expressed in Xenopus oocytes. Using two-electrode voltage clamp recording, we showed that the inhibitory efficacy of the ribbon isomer of AuIB is limited to ∼50%. The maximal inhibition was stoichiometry-dependent because altering α3:β4 RNA injection ratios either increased AuIB(ribbon) efficacy (10α:1β) or completely abolished blockade (1α:10β). In contrast, inhibition by AuIB(globular) was independent of injection ratios. ACh-evoked current amplitude was largest for 1:10 injected oocytes and smallest for the 10:1 ratio. ACh concentration-response curves revealed high (HS, 1:10) and low (LS, 10:1) sensitivity α3β4 nAChRs with corresponding EC50 values of 22.6 and 176.9 μm, respectively. Increasing the agonist concentration antagonized the inhibition of LS α3β4 nAChRs by AuIB(ribbon), whereas inhibition of HS and LS α3β4 nAChRs by AuIB(globular) was unaffected. Inhibition of LS and HS α3β4 nAChRs by AuIB(globular) was insurmountable and independent of membrane potential. Molecular docking simulation suggested that AuIB(globular) is likely to bind to both α3β4 nAChR stoichiometries outside of the ACh-binding pocket, whereas AuIB(ribbon) binds to the classical agonist-binding site of the LS α3β4 nAChR only. In conclusion, the two isomers of AuIB differ in their inhibitory mechanisms such that AuIB(ribbon) inhibits only LS α3β4 nAChRs competitively, whereas AuIB(globular) inhibits α3β4 nAChRs irrespective of receptor stoichiometry, primarily by a non-competitive mechanism.

Conotoxins are short disulfide-rich bioactive peptides that have been originally isolated from venoms of carnivorous mollusk cone snails, belonging to the genus Conus. ␣-Conotoxins are among the largest class of conotoxins found in the venom of most cone snail species (1). This class of conotoxins targets various subtypes of nicotinic acetylcholine receptors (nAChRs) 2 and is distinguished by four cysteines arranged in a CC-C-C pattern.
␣-Conotoxins have attracted considerable attention as some of them, such as Vc1.1 and RgIA, have been shown to possess analgesic activity in rodent behavioral models of neuropathic pain (2,3). Interestingly, AuIB has recently been shown to be analgesic in vivo despite the fact that it acts on the ␣3␤4 nAChR subtype different from the ␣9␣10 nAChR targeted by Vc1.1 and RgIA. 3 Vc1. 1 and RgIA have been shown to suppress N-type Ca 2ϩ channel currents in dorsal root ganglion (DRG) neurons of neonatal and adult rats and wild type and ␣9 knock-out mice via activation of GABA B G protein-coupled receptors (2). Similarly, AuIB inhibits N-type Ca 2ϩ channels in rat DRG neurons analogous to Vc1.1 and RgIA and its effect can be blocked with selective GABA B receptor antagonists. 4 GABA B -mediated inhibition of N-type Ca 2ϩ channels is proposed as an analgesic mechanism for ␣-conotoxins Vc1.1, RgIA, and AuIB (2), which can reconcile obvious differences in their nAChR subtype selectivity.
␣-Conotoxin AuIB has been characterized on oocyte-expressed nAChRs and shown to be selective primarily for the ␣3␤4 nAChR subtype (3). The ␣3␤4 nAChR subtype is a predominant subtype in autonomic ganglia, adrenal medulla, and in subpopulations of central nervous system neurons such as medial habenula and dorsal medulla (4). AuIB is a 15-residue conotoxin with an unusual 4/6 intercysteine spacing. Native AuIB peptide found in the venom, like the vast majority of other ␣-conotoxins, has 2-8, 3-15 (Cys 1 -Cys 3 , Cys 2 -Cys 4 ) cystine globular connectivity (5). When AuIB is synthesized chemically, a disulfide bond isomer having 2-15, 3-8 (Cys 1 -Cys 4 , Cys 2 -Cys 3 ) connectivity is co-produced as a by-product, which is the ribbon isoform of AuIB (Fig. 1A) (5). Surprisingly, the ribbon isomer of AuIB has been shown previously to be ϳ10fold more potent at nAChRs of rat parasympathetic ganglion neurons compared with the globular (native) peptide isoform (5). However, when AuIB(ribbon) was probed on rat nAChRs heterologously expressed in Xenopus oocytes it was reported to be less active than the globular isomer (6). The difference in * This work was supported by a National Health & Medical Research Council activity of the two AuIB isomers on native versus recombinant nAChRs remains to be elucidated. Taken together, there is an incomplete understanding of diversity of mechanisms of action of ␣-conotoxin AuIB and its isoforms. Here, we further explore the inhibitory mechanisms of globular and ribbon isomers of AuIB on rat ␣3␤4 nAChRs expressed in Xenopus oocytes.
Manipulations of the ␣3:␤4 subunit ratios show that ␣3␤4 nAChRs expressed in oocytes are present in different subunit stoichiometries and inhibition by AuIB(ribbon), but not AuIB-(globular), is limited to one of the receptor stoichiometries. Inhibition by AuIB(ribbon) is consistent with a competitive antagonism, whereas AuIB(globular) inhibits ␣3␤4 nAChRs via a non-competitive mechanism.
Circular Dichroism (CD) Spectroscopy-CD spectroscopy was performed on a Jasco J-810 spectropolarimeter. Spectra were recorded at room temperature under nitrogen atmosphere. Peptides were dissolved in 20 mM phosphate buffer, containing 30% trifluoroethanol at pH 7. The peptide concentration was determined by quantitative RP-HPLC. The peptides were transferred into a 0.01-cm path length demountable cell and data were recorded over 5 scans, from 260 to 185 nm at 10 nm/min, with a resolution of 1 nm and a response time of 0.25 s. CD data in ellipticity was converted to mean residue ellipticity ([]R) using the equation where is the ellipticity in millidegrees, C is the peptide molar concentration (M), l is the cell path length (cm), and Np is the number of peptide residues. CD was used to confirm the globular and ribbon AuIB structure. AuIB(globular) showed CD spectra with ␣-helical content as opposed to that of AuIB(ribbon), which has less secondary structure (Fig. 1C).
RNA Preparation-Plasmid DNAs encoding rat ␣3 and ␤4 nAChR subunits were obtained from J. Patrick (Baylor College of Medicine, Houston, TX). After multiplication plasmid DNA was linearized with appropriate restriction enzymes and cRNA was synthesized in vitro using a SP6 in vitro transcription kit (mMessage mMachine; Ambion, Foster City, CA). RNA for ␣3 and ␤4 subunits was synthesized in parallel on the same day using identical procedures to maximize consistency between subunits in concentration and purity. RNA concentration was controlled spectrophotometrically for each new aliquot of RNA and before injections. Total amount of RNA injected per oocyte was ϳ5 ng in ϳ50 nl/volume. Stage V to VI oocytes obtained from Xenopus laevis were subsequently incubated for 2-8 days at 18°C before electrophysiological recordings as described previously (6).
Electrophysiological Recordings and Data Analysis-Oocytes were transferred to the recording chamber (ϳ50 l volume) and perfused at 3-5 ml/min with ND96 solution (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM HEPES, pH 7.5) by a gravity fed perfusion system. nAChR-mediated currents were evoked by pipetting 100 l of ACh-containing solutions into the bath with the perfusion stopped. Oocytes were preincubated with the peptide for ϳ5 min prior to ACh application and subsequently ACh was co-applied together with the peptide. Two-electrode voltage clamp recordings from Xenopus oocytes were made at room temperature (21-23°C) using a GeneClamp 500B amplifier (Molecular Devices Corp., Sunnyvale, CA) at a holding potential of Ϫ80 mV. ACh-induced currents were recorded and their peak amplitudes were measured with pClamp 9 software (Molecular Devices Corp.). AChevoked currents recorded following exposure to peptides were normalized to an average of 2-6 control ACh-induced currents. Data from several experiments were pooled and each data point represents the average of 3-8 cells Ϯ S.E. Estimates of toxins potencies (IC 50 , n H ) were obtained by fitting data points to the equation: % response ϭ 100/{1 ϩ ([toxin]/IC 50 ) nH } with SigmaPlot8 (Systat Software Inc., San Jose, CA).
Molecular Modeling and Docking-The molecular model of rat ␣3 subunit was generated using ␣1 nAChR subunit (Protein Data Bank (PDB) code 2QC1 (8)). The rat ␤4 subunit homology model was generated using the ␤ chain of Torpedo acetylcholine receptor (PDB ID 2BG9 (9)), as several attempts by using high resolution AChBP crystal structures were not satisfactory. The open and closed C-loop conformations of ␣3 subunit were resembled using co-crystal structures of AChBP in complex with different molecular sizes of ligands (Ac-AChBP-TxIA(A10L) (PDB code 2UZ6 (10)), Ac-AChBP-methyllycaconitine (PDB code 2BYR (11)), Ac-AChBP-nicotine (PDB code 1UW6 (12)), and Ls-AChBP-HEPES (PDB code 1I9B (13)). Templates and interfaces were assembled using Aplysia californica acetylcholine-binding protein (PDB code 2BR8 (14)) as the template, with programs Modeler 9 version 2 (15) and Sculptor, and/or by docking simulations using the program HEX 5.1 (16). All sequence alignments were generated using ClustalW (17) and were further manually adjusted based on secondary structure alignment. The subunit ␣3 and ␤4 and the pentamers (␣3) 3 (␤4) 2 and (␣3) 2 (␤4) 3 were validated individually using online server Verify3D (18) and Ramachandran plot available from ProFunc data base (19). The solvent accessible area of each interface was calculated using online server CASTp (20). In addition, we further validated our homology models by performing a global docking simulation with ACh. The outcome showed that ACh only docked to the classical agonist binding site in the ␣(ϩ)␤(Ϫ) interface and was unaffected by subunit stoichiometry (supplemental Fig. S1).

Concentration-Response Relationships Obtained for Inhibition of ACh-evoked Currents by Globular and Ribbon Isomers of
AuIB-Concentration-response curves were obtained for the inhibition of rat ␣3␤4 nAChRs expressed in Xenopus oocytes by globular and ribbon isomers of AuIB ␣-conotoxin (Fig. 2). Surprisingly, the two isomers of AuIB demonstrated a marked difference in their efficacy for inhibition of ␣3␤4 nAChRs. AuIB-(globular) at a concentration of 10 M completely inhibited ␣3␤4 nAChRs (95.3 Ϯ 0.9%, n ϭ 6; p Ͻ 0.03), whereas AuIB(ribbon) did not exceed 50% block even at a concentration of 30 M (Fig. 2, A and B). The concentration-response relationships for AuIB(globular) and AuIB(ribbon) gave IC 50 values of FIGURE 2. A, representative superimposed ACh-evoked currents obtained in the absence (Control) and presence of AuIB(ribbon) and AuIB(globular). The time course of ACh application is indicated by the bars above the current traces. B, concentration-response curves for inhibition of rat ␣3␤4 nAChRs expressed in Xenopus oocytes by globular and ribbon isomers of ␣-conotoxin AuIB. Note that maximal block (efficacy) of the ribbon isomer is only ϳ50%.
Comparison of the concentration-response relationships obtained from oocytes with different ␣3␤4 nAChR subunit ratios indicates distinct modes of action of AuIB(ribbon) versus AuIB(globular) (Fig. 3). Shifting ␣3:␤4 ratios did not change the maximal inhibition as well as the other parameters of the AuIB-(globular) concentration-response curve (Fig. 3B, panel i) (Table 1). In contrast, when the subunit ratio changed from 10:1 to 1:10, the concentration-response curve of AuIB(ribbon) flattened out to exhibit no significant inhibition at ␤-dominated oocytes (Fig. 3B, panel ii). These results suggest an intrinsic difference in the mechanisms of action of the two isomers of ␣-conotoxin AuIB.

␣3:␤4 Subunit Ratio Affects AChevoked Current Amplitudes and ACh
Concentration-Response Curves-Varying the ␣3:␤4 subunit ratio in oocytes in either direction had a marked effect on the ACh-induced current amplitude. Application of 50 M ACh elicited the largest currents in oocytes injected at a 1:10 ratio, and the smallest current amplitudes in oocytes injected with a 10:1 ␣3:␤4 ratio. Injecting all three different ratios of RNA in oocytes resulted in a significant difference between the average ACh-evoked current amplitudes in the three   (Fig. 4A). The average amplitude of ACh-evoked currents in the 10:1 ratio injected oocytes was 1.66 Ϯ 0.14 A and 13.74 Ϯ 2.26 A in 1:1, and 23.33 Ϯ 3.48 A in 1:10 (n ϭ 17-21). The current amplitudes in the 10:1 and 1:10 ratio injected oocytes were significantly different from the current amplitudes obtained for the 1:1 ratio injected oocytes (p Ͻ 0.001, 10:1; p Ͻ 0.05 1:10, unpaired t test) (Fig. 4A). In a series of experiments on 10:1 ratio injected oocytes, 2.5 mM ACh elicited current amplitudes similar to those observed for 1:10 injected oocytes (data not shown) indicating that the smaller current observed in response to 50 M ACh in 10:1 oocytes was not due to reduced nAChR expression. We hypothesized that the sensitivity to ACh (EC 50 ) may account for different current amplitudes observed for the three different subunit ratios. To test this possibility, ACh sensitivity of oocytes injected with different ␣3:␤4 subunit ratios was examined. The ACh concentration-response curves obtained for 10:1 and 1:10 ratio groups were strikingly different with an EC 50 of 176.9 Ϯ 23.9 M (n ϭ 4 -7) for 10:1 ratio injected oocytes and 22.56 Ϯ 1.41 M (n ϭ 3-8) for the 1:10 oocyte group (Fig. 4B). The Hill slope was also different between the two ratio groups (2.1 Ϯ 0.5 for 10:1 and 1.00 Ϯ 0.1 for 1:10) suggesting that the 10:1 ␣3␤4 nAChR may bind 2 ACh molecules, whereas the 1:10 ␣3␤4 nAChR binds a single agonist molecule (Fig. 4B). The ␤-dominated nAChRs (1:10) are high sensitivity receptors and ␣-dominated nAChRs (10:1) are low sensitivity ␣3␤4 nicotinic receptors. The ACh concentrationresponse curve for the 1:1 injected oocytes was fit with an EC 50 of 92.8 Ϯ 6.6 M and n H ϭ 1.3 Ϯ 0.1 (n ϭ 3-8), which was between the values obtained for 1:10 and 10:1 ratios (Fig. 4B). The parameters obtained from the ACh concentration-response curves for the three subunit ratios are summarized in Table 2. These findings indicate that the different stoichiometries of ␣3␤4 nAChRs expressed in oocytes underlie the differences in EC 50 , Hill slope, and ACh-evoked current amplitudes observed for the three groups.
Inhibition of Low Sensitivity ␣3␤4 nAChRs (10:1) by AuIB(Ribbon), but Not AuIB(Globular), Is Dependent on ACh Concentration-The ACh concentration-response curves determined for the three different ␣3:␤4 ratios indicate that 50 M ACh used in our initial experiments is below the EC 50 for the 10:1 ratio and close to saturation for the 1:10 ratio injected oocytes (Fig. 5). The difference in the EC 50 may account for the differential inhibitory effect of AuIB(ribbon) tested on all three injection groups. The ACh dependence of the inhibition of 10:1 ratio ␣3␤4 nAChRs by both AuIB isomers was investigated for three different ACh concentrations: 50, 175 (ϳEC 50 ), and 500 M. Concentration-response curves for AuIB isomers obtained at different ACh concentrations are shown in Fig. 5. The inhibition of ␣3␤4 nAChRs by AuIB(globular) was not strongly dependent on ACh concentration: IC 50 values of 1.1, 0.9, and 2.3 M were obtained for 50, 175, and 500 M ACh, respectively (Table 3). Overall, increasing the agonist concentration produced a small rightward shift of the concentration-response curve only at saturating ACh concentration (500 M) without affecting maximal inhibition achieved. In contrast, the efficacy of ␣3␤4 nAChRs inhibition by AuIB(ribbon) decreased with increasing ACh concentration (Fig. 5B). 10 M AuIB(ribbon) inhibited ␣3␤4 nAChRs by 73.2 Ϯ 3.0% using 50 M ACh, 34.2 Ϯ 2.1% at 175 M, and 22.2 Ϯ 3.1% at 500 M ACh (Fig. 5B). Inhibition of ␣3␤4 nAChRs at a 10:1 ratio by AuIB(ribbon) is sensitive to the ACh concentration, whereas AuIB(globular) is relatively insensitive to the agonist concentration suggesting a competitive mechanism of action for the ribbon isomer of AuIB and non-competitive for the globular isomer.
Inhibition of High Sensitivity ␣3␤4 nAChRs by AuIB Isomers at Different ACh Concentrations-The effect of ACh concentration on the inhibition of ␤4-dominant oocytes (1:10 ratio) by AuIB isomers was examined. AuIB(globular) inhibition of ␤4-dominant nAChRs was tested for three different ACh con-   (Fig. 6A). Halfmaximal inhibitory AuIB(globular) concentrations were 1.24, 3.0, and 5.8 M for 5 M, 50 M, and 5 mM ACh, respectively (Table 3). Because AuIB(ribbon) inhibition is competitive and ACh concentration dependent in LS nAChRs, there is a possibility that in HS inhibition is masked due to lower ACh EC 50 for HS; however, AuIB(ribbon) was not active in 1:10 ratio injected oocytes even at low ACh concentrations (3 M) (Fig. 6B, inset). These data suggest that the lack of activity of AuIB(ribbon) at 1:10 nAChRs may be due to dependence of the inhibition on ␣3␤4 nAChR stoichiometry, probably, because of a modified ACh binding site at 1:10 ␣3␤4 nAChRs. Effects of different ACh concentrations on concentration-response curve parameters for the AuIB isomers obtained in 10:1 and 1:10 oocytes are summarized in Table 3. AuIB(Globular) Inhibition of ␣3␤4 nAChRs Is Voltage Independent and Non-competitive-AuIB(globular) inhibition of ␣3␤4 nAChRs, in contrast to AuIB(ribbon), is not significantly impaired by subunit stoichiometry or agonist concentration.
The mechanism of nAChR inhibition by AuIB(globular) is consistent with non-competitive receptor antagonism. To further examine mechanism(s) of action of AuIB-(globular), ACh concentration-response curves were determined in the presence of different concentrations of AuIB(globular) at 1:10 ratio injected oocytes where the ribbon isomer is inactive. Increasing concentrations of AuIB(globular) did not induce a parallel shift of the ACh concentration-response curve (see Fig. 7A) and even saturating ACh concentrations did not impair inhibition of HS ␣3␤4 nAChRs by AuIB(globular) when it was applied at concentrations Ͼ1 M. A similar result was obtained in 10:1 injected oocytes (data not shown). Thus, inhibition of ␣3␤4 nAChRs by AuIB(globular) was insurmountable with increasing agonist concentration, which is characteristic of a non-competitive inhibition. Given that some antagonists of neuronal nAChRs have been shown to be voltage-dependent channel pore blockers (for review, see Ref. 21), the voltage dependence of AuIB-(globular) (3 M) inhibition of 1:10 ratio injected oocytes was examined at three different holding potentials: Ϫ120, Ϫ80, and Ϫ45 mV (Fig. 7B). Average inhibition by AuIB(globular) did not differ significantly at the different membrane potentials; 42.7 Ϯ 7.0, 50.5 Ϯ 4.2, and 46.8 Ϯ 10.9% at Ϫ120, Ϫ80, Ϫ45 mV, respectively (p Ͼ 0.3 for Ϫ120 versus Ϫ80 mV and p Ͼ 0.7 for Ϫ45 versus Ϫ80 mV, unpaired t test) indicating that AuIB-(globular) blockage of ␣3␤4 nAChRs is voltage-independent. These data strongly suggest that the two isomers of ␣-conotoxin AuIB inhibit ␣3␤4 nAChRs differentially dependent on receptor subunit stoichiometry.

Stoichiometry-dependent Block of ␣3␤4 nAChRs by AuIB Isomers
competitive inhibition of the LS ␣3␤4 nAChRs, which was also observed in our experiments (see Figs. 3 and 5). Interestingly, AuIB(ribbon) docked predominantly to the ␤(ϩ)␣(Ϫ) interface in (␣3) 2 (␤4) 3 (1:10) (Fig. 8A), which may explain its inability to inhibit HS ␣3␤4 nAChRs observed in 1:10 oocytes. Upon a closer examination of the interfaces in both models, we found that the interfaces formed by the fifth subunit of both models ␣5th ((␣3) 3 (␤4) 2 ) and ␤5th ((␣3) 2 (␤4) 3 ) exhibited significant differences in the solvent accessible areas, which is also observed in the Torpedo nAChR structure, where interfaces formed by different combinations of subunits (␣, ␤, ␥, ␦) resulted in variable solvent accessible areas (9). Together with the flexible C-loop observed in AChBP co-crystal structures (10 -13), variations in the solvent accessible area are likely to affect the number of ligand-receptor contacts and hence contribute to ligand selectivity. The solvent accessible area of ␤(ϩ)␣5th(Ϫ) is significantly larger (ϳ36%) than that of ␤(ϩ)␤5th(Ϫ), which resulted in 8 -20% less solvent accessible areas in the remaining four interfaces of (␣3) 3 (␤4) 2 (Fig. 8B). Despite the differences in the solvent accessible area in both stoichiometries, we reason that the flexibility of the AuIB-(ribbon) structure is more likely to adopt the C-loop binding pocket, compared with AuIB(globular), which may also contribute to ligand selectivity. However, the larger solvent accessible area of ␣(ϩ)␤(Ϫ) interfaces observed in (␣3) 2 (␤4) 3 model may cause loss of inter-molecular interactions between the AuIB(ribbon) and the receptor, and hence the C-loop binding pocket is no longer energetically favorable. Together, our homology modeling, docking simulation, and experimental data provide a molecular basis and possible explanation for the antagonist selectivity of stoichiometrically different ␣3␤4 nAChRs, which may be applicable to other hetero-pentameric nAChRs.

DISCUSSION
Here we show that the heterologous expression of different ratios of rat ␣3␤4 nAChR subunits in X. laevis oocytes results in high and low sensitivity nAChRs reflecting different stoichiometries of ␣ and ␤ subunits. The globular and ribbon isomers of ␣-conotoxin AuIB differentially inhibit the high and low sensitive ␣3␤4 nAChRs; AuIB(globular) inhibits both HS and LS nAChRs, whereas the ribbon isomer is active only at LS nAChRs. Furthermore, the isomers exhibit distinct inhibitory mechanisms: AuIB(ribbon) acts as a classical competitive antagonist, whereas AuIB(globular) acts as a non-competitive antagonist.
Previously, the IC 50 of AuIB(ribbon) at heterologously expressed ␣3␤4 nAChRs was estimated to be ϳ27 M based on two concentrations tested and assuming 100% efficacy (6). Investigation of a broader range of AuIB(ribbon) concentrations used here surprisingly revealed that the maximal inhibition observed with the ribbon isomer is limited to only 50%. Furthermore, the IC 50 of 0.77 M for inhibition of oocyte-expressed ␣3␤4 nAChRs by AuIB(ribbon) is less than that of the globular isomer (2.48 M) (see Table 1). We hypothesized that the partial block of ␣3␤4 nAChRs by AuIB(ribbon) may be due to nAChR stoichiometry because ␣3␤4 nAChR likely exists in at least two stoichiometries, differing in their fifth subunit (either ␣ or ␤) as has been demonstrated previously for rat and human ␣4␤2 nAChRs expressed in oocytes (22,23).
In experiments with altered ␣3:␤4 injection ratios, we demonstrated differential inhibitory action of the two AuIB isoforms (Figs. 2 and 3). Blockade by AuIB(ribbon) was dependent on ␣3:␤4 injection ratios such that excess ␣ subunits increased efficacy and excess ␤ subunits abolished inhibition completely. In contrast, concentration-response curves of AuIB(globular) were not significantly affected by altering the ␣3:␤4 ratio (Fig.  2). To our knowledge nAChR stoichiometry-restricted antagonism has not been reported previously for any nAChR antagonists including ␣-conotoxins and it is surprising that only one of the two AuIB isomers is stoichiometry selective.
Competitive versus Non-competitive Antagonism and Implications for SAR-The effect of agonist (ACh) concentration on concentration-response curves obtained for inhibition ␣3␤4 nAChRs by the two AuIB isomers suggested differential inhibitory mechanisms for the globular and ribbon AuIB isoforms (Figs. 6 and 7). In the case of LS ␣3␤4 nAChR, increasing the ACh concentration led to markedly reduced inhibition by AuIB(ribbon), whereas inhibition by AuIB(globular) did not change significantly. Higher ACh concentrations at HS ␣3␤4 nAChRs produced a slight shift of concentration-response curve for AuIB(globular) inhibition at low toxin concentrations without affecting maximal inhibition, suggesting both competitive and non-competitive action of AuIB(globular) at HS with low and high affinity, respectively. AuIB(ribbon) was not active at HS ␣3␤4 nAChRs over a range of ACh concentrations suggesting that AuIB(ribbon) inhibition is ␣3␤4 stoichiometry-dependent, presumably due to differences in ACh binding site properties between the two stoichiometries.
Schild plot analysis showed that increasing AuIB(globular) concentration did not produce a parallel shift of the ACh concentration-response curves and inhibition was clearly not surmountable. Inhibition of HS ␣3␤4 nAChRs by AuIB(globular) was not voltage dependent indicating that the globular isomer is not an open channel blocker (Fig. 7). These findings for AuIB-(globular) are consistent with non-competitive receptor antagonism.
To date, relatively few conotoxins acting as non-competitive nAChR antagonists have been reported. -Conotoxins, are a class of non-competitive nAChR antagonists having six cysteines connectivity (35). ␣-Conotoxins are generally considered to be competitive nAChRs antagonists based on binding studies for a few of them (e.g. Ref. 36). However, ␣-conotoxin ImII blocks ␣7 nAChRs (37,38) and Torpedo californica nAChRs in a non-competitive manner (39), despite differing from ImI (a known competitive blocker) by only three residues. Interestingly, a large dimeric conotoxin VXXXB belonging to a novel class ␣-D-conotoxins was also shown to inhibit ␣7 and ␣3␤2 nAChRs non-competitively (40).
There is mounting evidence showing that nicotinic agents target various sites on the nAChR besides the classical agonist binding site (e.g. Ref. 41). Mutations in the ␣(ϩ)/␤(Ϫ) subunit interface outside the ACh-binding site were reported to affect binding of MII, PnIA, and GID 4/7 ␣-conotoxins to ␣3␤2 nAChRs (42). Non-canonical allosteric sites of nAChRs have been studied for positive allosteric modulators of nAChRs. Positive allosteric modulators galanthamine and morantel have been shown to bind at the non-canonical ␤(ϩ)/␣(Ϫ) interface, which is absent in homomeric receptors (41,43). The ␣4/␣4 interface in (␣4) 3 (␤2) 2 nAChRs has been shown to account for stoichiometry-dependent Zn 2ϩ potentiation (26). In the present study, non-competitive action of AuIB(globular) is unlikely to be at the unique ␣3/␣3 interface of (␣3) 3 (␤4) 2 because this isomer non-competitively inhibited both receptor stoichiometries. This result suggests the presence of a non-competitive site at both ␣3␤4 stoichiometries, which was corroborated by our molecular docking study, which showed that AuIB(globular) likely binds outside the orthosteric site at the ␣3(ϩ)/␤4(Ϫ) interface of both stoichiometries (Fig. 8). Similarly, VXXXB is thought to inhibit nAChR allosterically by binding outside of the ACh-binding pocket in the ␣(ϩ)/␤(Ϫ) interface to stabilize the cleft in a conformation, which does not support agonist binding (40). The allosteric site at the ␤(ϩ)/␣(Ϫ) interface might be limited to positive modulation exclusively. AuIB-(ribbon) docking results, which show binding within the classical ACh binding site of LS ␣3␤4 nAChRs only, are in agreement with the experimental data. Stoichiometry-dependent block by AuIB(ribbon) also implies that the ACh binding site is affected by nature of the fifth subunit in the ␣3␤4 nAChR pentamer and calculations of solvent-accessible areas of ␣3(ϩ)/␤4(Ϫ) interfaces for (␣3) 3 (␤4) 2 and (␣3) 2 (␤4) 3 nAChRs suggest a slightly larger binding pocket at HS ␣3␤4 nAChR, which may underlie the loss of activity at this stoichiometry.
Taken together, our results are consistent with AuIB(globular) acting primarily at a non-competitive site of both ␣3␤4 nAChR stoichiometries, whereas AuIB(ribbon) binds to the classical ACh binding site of the LS nAChR only. The ␣-conotoxin AuIB disulfide isomers exhibit strikingly divergent inhibitory mechanisms of action and are potentially valuable tools for further dissecting ligand-receptor interactions and in the design of novel conopeptides with improved selectivity profile.