Introduction

Summary Introduction Procedures Results Discussion References

nAChRs are assembled from a family of 11 distinct subunits, 2-9 and 2-4 (1, 2). Similar to what has been shown for muscle-type nAChRs, the ligand binding sites of neuronal nAChRs are complex, with contributions from both the  and non- () subunits. On expression in Xenopus laevis oocytes, each functional subunit combination displays unique pharmacological properties (3-7). These pharmacological differences can be exploited as probes with which to explore the structural determinants of receptor subtype specificity.

Affinity labeling techniques have been used to identify a number of amino acid residues on the , , and  subunits of muscle-type nAChRs (8-16). Many of these residues are conserved among neuronal nAChR subunits and thus may form features of the ligand binding sites common to all nAChRs. Because they are so highly conserved, these residues cannot be responsible for the pharmacological differences that have been observed among different neuronal nAChR subunit combinations. It is the residues that differ among nAChR subunits that must be responsible for this pharmacological diversity.

An approach to identifying residues responsible for pharmacological differences is to construct a series of chimeras of pharmacologically distinct subunits to map critical sequence segments, followed by site-directed mutagenesis to probe the role of individual residues. This technique has been used with success to map residues that determine sensitivity to both agonists and antagonists. Several sequence segments of neuronal nAChR  subunits have been identified that affect sensitivity to agonists and the competitive antagonist neuronal NBT (17). Regions of 2 and 4 have also been identified that determine sensitivity to agonists (18, 19). We used chimeric and mutant  subunits to identify Thr59 of 2 as a major determinant of sensitivity to the competitive antagonists DHE and NBT (20).

Recently, a novel -conotoxin (-CTx-MII) was isolated that is a highly selective antagonist of the 32 subunit combination (21). NBT, under certain conditions, is also selective for 32 receptors (3, 5, 6; but see Ref. 22). Blockade of 32 by -CTx-MII seems to be competitive because the 32 receptor can be protected from -CTx-MII block by DHE, a known competitive antagonist (23). In this study, we were interested in identifying residues on the  and  subunits of neuronal nAChRs that determine sensitivity to this toxin. We constructed and screened a series of  subunit chimeras and  subunit chimeras to identify critical sequence segments. We then used site-directed mutagenesis to identify Lys185 and Ile188 of 3 and Thr59 of 2 as residues important in determination of -CTx-MII sensitivity of the 32 subunit combination.

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Materials. X. laevis frogs were purchased from Nasco (Fort Atkinson, WI). RNA transcription kits were from Ambion (Austin, TX). ACh, atropine, and 3-aminobenzoic acid ethyl ester were from Sigma Chemical (St. Louis, MO). Collagenase B was from Boehringer-Mannheim (Indianapolis, IN). Sequenase 2.0 kits were from United States Biochemical (Cleveland, OH). CloneAmp kits were from GIBCO BRL (Baltimore, MD). -CTx-MII was synthesized, and proper disulfide bond formation was achieved as previously described (21).

Mutagenesis and construction of chimeric receptors. Chimeric and mutant subunits were constructed using the PCR (24). Our notation for chimeric subunits is to list the source of the amino-terminal portion, followed by the residue number in the amino acid sequence where the chimeric joint is made (numbering taken from the mature 3 and 2 subunit sequences), and then followed by the source of the carboxyl-terminal portion. For example, the chimeric subunit 4-216-3 is composed of 4 sequence from the amino terminus until residue 216, after which it is composed of 3 sequence. Our notation for mutant subunits is to list the naturally occurring residue followed by the position of that residue, followed by the change that has been made. For example, the mutant subunit 3,I188K is an 3 subunit in which Ile188 has been changed to a lysine. PCR products were subcloned into the pAMP1 vector using a CloneAmp kit (GIBCO BRL) or into the pCR-Script SK⁺ vector (Stratagene, La Jolla, CA). To minimize the amount of PCR product in the final construct that would have to be sequenced, as much PCR product as possible was replaced with appropriate wild-type sequence using existing restriction sites. Remaining sequence derived from PCR product was confirmed by sequencing with the use of Sequenase 2.0 (United States Biochemical).

Injection of in vitro synthesized RNA into X. laevis oocytes. m⁷G(5)ppp(5)G capped cRNA was synthesized in vitro from linearized template DNA encoding the 2, 3, 4, 2, and 4 subunits, as well as the various chimeric and mutant subunits, using an Ambion mMessage mMachine kit. Mature X. laevis frogs were anesthetized by submersion in 0.1% 3-aminobenzoic acid ethyl ester, and oocytes were surgically removed. Follicle cells were removed by treatment with collagenase B for 2 hr at room temperature. Each oocyte was injected with 5-50 ng of cRNA in 50 nl of water and was incubated at 19° in modified Barth's saline (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 0.3 mM CaNO₃, 0.41 mM CaCl₂, 0.82 mM MgSO₄, 100 µg/ml gentamicin, 15 mM HEPES, pH 7.6) for 2-7 days. RNA transcripts encoding each subunit were injected into oocytes at a molar ratio of 1:1.

Electrophysiological recordings. Oocytes were perfused at room temperature (20-25°) in a 300-µl chamber with perfusion solution (115 mM NaCl, 1.8 mM CaCl₂, 2.5 mM KCl, 10 mM HEPES, pH 7.2, 1.0 µM atropine). Perfusion was continuous at a rate of ~20 ml/min. ACh was diluted in perfusion solution, and the oocytes were exposed to ACh for ~10 sec, using a solenoid valve. -CTx-MII sensitivity was tested by comparing ACh-induced current responses before and after the oocytes were incubated for 5 min in perfusion solution containing various concentrations of -CTx-MII and 100 µg/ml bovine serum albumin. The postincubation ACh response is presented as a percentage of the preincubation ACh response. ACh concentrations were at or below the EC₅₀ value for each receptor to avoid extensive desensitization. The slowly reversible nature of -CTx-MII blockade allowed the postincubation ACh response to be measured without coapplication of toxin. Because toxin and ACh would not be in direct competition, the degree of block is not dependent on the concentration of ACh, and the ACh concentration used for each receptor would not need to be equipotent with the ACh concentrations used for other receptors. The IC₅₀ value we obtained for -CTx-MII block of 32 (3.5 nM) was somewhat higher than that obtained previously (0.5 nM) (21). A possible reason for this difference is that in the current study, perfusion was stopped during the -CTx-MII incubations to conserve toxin, whereas in the previous study, toxin was perfused continuously. Static application of toxin results in an apparent decrease in toxin potency at all receptor subtypes.¹ We are unsure why this difference occurs; however, nonspecific adsorption of toxin to the chamber is a possible reason. This difference was not a problem in the current study because all subunit combinations are affected similarly (compare Fig. 1 with Ref. 21, Fig. 5) and because all experiments in the current study involved the same static toxin application protocol.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   -CTx-MII is selective for 32. -CTx-MII inhibition of () 32 (), 42 (), 22, and () 34 neuronal nicotinic receptors expressed in X. laevis oocytes is shown. The response to a concentration of ACh at or below the EC50 value for each receptor after a 5-min incubation with various concentrations of -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three or four oocytes). The 32 data are fit to a Hill equation (see Experimental Procedures), yielding IC50 = 3.5 nM, n = 1.35. Some error bars are obscured by symbols.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   -CTx-MII sensitivity of receptors formed by 3, K185Y, 3, I188K, and 2, T59K. Inhibition by -CTx-MII of () 3, K185Y 2, () 3, I188K 2, and () 3 2, T59K receptors expressed in X. laevis oocytes. The dose-inhibition data for () 32 from Fig. 1 are shown for reference. The response to a concentration of ACh at or below the EC50 value for each receptor after a 5-min incubation with various concentrations of -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three or four oocytes). The data are fit to a Hill equation (see Experimental Procedures). For 3, K185Y 2, IC50 = 12 nM, n = 1.07; for 3, I188K 2, IC50 = 39 nM, n = 0.93; and for 3 2, T59K, IC50 = 14 nM, n = 1.29. Some error bars are obscured by symbols.

	Results

Summary Introduction Procedures Results Discussion References

Determinants of -CTx-MII sensitivity on 3 lie within sequence segments 121-181 and 181-195. -CTx-MII is highly selective for the 32 subunit combination (Fig. 1, ; IC₅₀ = 3.5 nM), which is in agreement with the results of Cartier et al. (21). Both the 3 subunit and the 2 subunit are required for high sensitivity to -CTx-MII. Receptors containing a different  subunit (42 or 22) or a different  subunit (34) are 200-fold less sensitive to -CTx-MII than the 32 receptor. This high degree of selectivity makes -CTx-MII a promising probe for investigation of the structure of the antagonist binding sites on neuronal nicotinic receptors. We selected an -CTx-MII concentration of 50 nM as a test dose for screening chimeric and mutant subunits. This toxin concentration almost completely blocks 32 (postincubation response = 1.8 ± 0.5% of control) but has no effect on the 42, 22, or 34 receptors.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   -CTx-MII sensitivity of receptors formed by chimeric  subunits. A, Chimeras constructed from the 3 and 4 subunits and coexpressed in X. laevis oocytes with 2. B, Chimeras constructed from the 3 and 2 subunits and coexpressed in X. laevis oocytes with 2. Current in response to an ACh concentration at or below the EC50 value for each receptor after a 5-min incubation with 50 nM -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three or four separate oocytes). Significantly different from 2 are 2-121-3, 2-181-3, 3-195-2, and 3-215-2 (p < 0.001). Significantly different from 3 are 4-175-3, 4-183-3, 4-195-3, 4-216-3, 3-183-4, 2-181-3, 2-195-3, and 2-215-3 (p < 0.001) and 2-121-3 (p < 0.05). Significantly different from 4 are 4-175-3, 4-183-3, 4-195-3, 3-183-4, and 3-216-4 (p < 0.001). Some error bars are too small to appear.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   -CTx-MII sensitivity of receptors formed by 2-181-3-195-2. Inhibition by -CTx-MII of 2-181-3-195-2 2 receptors () expressed in X. laevis oocytes. The dose-inhibition data for () 32 and () 22 from Fig. 1 are shown for reference. The response to a concentration of ACh at or below the EC50 value for each receptor after a 5-min incubation with various concentrations of -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three or four oocytes). The 2-181-3-195-2 2 data are fit to a Hill equation (see Experimental Procedures), yielding IC50 = 91.6 nM, n = 1.19. Some error bars are obscured by symbols.

Lys185 and Ile188 of 3 are determinants of -CTx-MII sensitivity. Sequence segment 181-195 of 3 is substantially divergent from 2 and 4 sequence, differing at 8 of 15 residues (Fig. 4A). To determine which of these residues are important for toxin sensitivity, we constructed mutants of 3 in which one or two residues in 3 were changed to what occurs in 2. Receptors formed by these mutants were screened for loss of sensitivity to 50 nM -CTx-MII (Fig. 4B). The mutations K180N, P182T, Y184T, H186N, E187S, N191D, and E194A caused no significant loss in toxin sensitivity. Only the mutants 3,K185Y and 3,I188K formed receptors that were significantly less sensitive to toxin blockade than receptors formed by wild-type 3, identifying Lys185 and Ile188 as determinants of -CTx-MII sensitivity.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Lys185 and Ile188 of 3 are important for -CTx-MII sensitivity. A, Amino acid alignment of region 180-195 of 2, 3, and 4. , Residues in 2 or 4 that differ from 3. B, -CTx-MII sensitivity of 3 mutants coexpressed with 2. Current in response to an ACh concentration at or below the EC50 value for each receptor after a 5-min incubation with 50 nM -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three to four separate oocytes). Significantly different from 3 are 3, K180N, P182T (p < 0.05), 3, K185Y (p < 0.001), and 3, I188K (p < 0.001). Some error bars are too small to appear.

Determinants of -CTx-MII sensitivity on 2 lie within sequence segments 1-54, 54-63, and 63-80. The identity of the  subunit is also critical to determining sensitivity to -CTx-MII. This is clear in Fig. 1, in which 32 is completely blocked by 50 nM -CTx-MII, whereas 34 is blocked only slightly by 500 nM -CTx-MII (postincubation response = 92.8 ± 2.3% of control). To map residues on 2 that contribute to -CTx-MII sensitivity, we tested receptors formed by a series of chimeric and mutant  subunits (20) in combination with the 3 subunit. Replacement of the first 54 residues of 2 with 4 sequence (4-54-2) had little effect on toxin blockade (Fig. 6), whereas replacement of the first 103 residues of 2 with 4 sequence (4-103-2) resulted in a complete loss of sensitivity to 50 nM -CTx-MII (postincubation response = 99.1 ± 9.7% of control). These results suggest that residues within segment 54-103 are critical to -CTx-MII sensitivity. To map these residues more closely, we replaced 4 sequence with 2 sequence to determine which sequence segments are required to confer toxin sensitivity. Replacement of the first 54 residues of 4 with 2 sequence conferred some toxin sensitivity, with the 2-54-4 subunit forming receptors partially blocked by 50 nM -CTx-MII (postincubation response = 72.2 ± 2.2% of control). The addition of the first 63 residues of 2 (2-63-4) conferred a larger portion of toxin sensitivity (postincubation response = 18.0 ± 6.6% of control). Receptors formed by 2-80-4 were as sensitive to toxin blockade as wild-type 32, suggesting that an additional determinant lies between residues 63 and 80. Thus, 2 contributes several determinants to the -CTx-MII sensitivity of 32, lying within sequence segments 1-54, 54-63, and 63-80.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   -CTx-MII sensitivity of receptors formed by chimeric  subunits. Chimeras constructed from the 2 and 4 subunits are coexpressed with the 3 subunit. Current in response to an ACh concentration at or below the EC50 value for each receptor after a 5-min incubation with 50 nM -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three or four separate oocytes). Significantly different from 2 are 4-103-2 (p < 0.001) and 2-54-4 and 2-63-4 (p < 0.01). Significantly different form 4 are 4-54-2, 2-63-4, and 2-80-4 (p < 0.001) and 2-54-4 (p < 0.05). Some error bars are too small to appear.

Thr59 of 2 is a determinant of -CTx-MII sensitivity. Previously, we found a major determinant of NBT and DHE sensitivity to reside within segment 54-63. We identified this determinant as Thr59 (20). To determine whether this residue also plays a role in determining -CTx-MII sensitivity, we examined a series of mutant 2 subunits in which each residue that differs between 2 and 4 within segment 54-63 was changed from what occurs in 2 to what occurs in 4 (Fig. 7). The mutations N55S, V56I, and E63T had no effect on toxin sensitivity. Only receptors formed by the mutant 2,T59K were significantly less sensitive to toxin than wild-type 32. Thus, similar to our results for NBT and DHE, Thr59 is involved in determining the -CTx-MII sensitivity of receptors formed by 2. On testing a range of toxin concentrations (Fig. 5), we found that receptors formed by 2,T59K were 4-fold less sensitive to -CTx-MII (IC₅₀ = 14 nM) than 32.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Thr59 of 2 is important for -CTx-MII sensitivity. A, Alignment of 2 and 4 sequences within segment 54-63. , Residues that differ. B, -CTx-MII sensitivity of receptors formed by each of a series of mutant 2 subunits. Current in response to an ACh concentration at or below the EC50 value for each receptor after a 5-min incubation with 50 nM -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three to six separate oocytes). 2, T59K is significantly different from 2 (p < 0.01). Some error bars are too small to appear. C, Installation of aspartate at position 59 had no effect on -CTx-MII sensitivity. Current in response to an ACh concentration at or below the EC50 value for each receptor after a 5-min incubation with 3 nM -CTx-MII is presented as a percentage of the preincubation ACh response (mean ± standard deviation of three separate oocytes).

	Discussion

Summary Introduction Procedures Results Discussion References

We found that determinants of -CTx-MII sensitivity on the 3 subunit reside within sequence segments 121-181 and 181-195. This contrasts with the distribution of determinants for sensitivity to NBT, a peptide neurotoxin isolated from snake venom that has a selectivity for 32 receptors. Major determinants of NBT sensitivity on 3 lie within sequence segments 84-121, 121-181, and 195-215 (17). Only a minor determinant of NBT sensitivity has been found to reside within segment 181-195.² NBT is much larger than -CTx-MII (66 and 16 amino acid residues, respectively), and this may explain why NBT sensitivity requires a larger array of determinants. It is interesting, given that both toxins are competitive antagonists, that residues critical to -CTx-MII sensitivity (181-195) are of only minor importance to NBT sensitivity. Within this region, residues Y190, C192, and C193 have been identified in affinity labeling experiments as part of the ligand binding site of muscle-type nAChRs isolated from electric organs of various Torpedo species (10, 12, 14). These residues are highly conserved in both muscle-type and neuronal-type nAChRs across many species and thus can be thought of as the common features of nAChR binding sites. This conservation rules out these residues in terms of conferring pharmacological differences among nAChR subtypes. It is the residues that differ among subunits that must play this role.

The amino acid residues within sequence segment 181-195 of 3 that determine -CTx-MII sensitivity are Lys185 and Ile188, positioned close to Y190, C192, and C193, the common features of nicotinic binding sites. The 2 and 4 subunits both have a tyrosine at the position analogous to Lys185 of 3. The 3.4-fold loss in -CTx-MII sensitivity that we see with the 3,K185Y mutant may be due to the increase in side-chain volume (32.3 Å³), or to the loss of the positive charge. Lys185 of 3 may be interacting with Glu11 of -CTx-MII, and thus the loss of the positive charge may be a critical factor. At the position analogous to Ile188, 2 has a lysine and 4 has an arginine. The 11-fold loss in -CTx-MII sensitivity with the 3,I188K mutant is unlikely to be due to the modest change in side-chain volume (2.5 Å³). Another possibility is that the introduction of the positive charge might be responsible for the loss in toxin sensitivity through electrostatic repulsion. This seems unlikely because -CTx-MII contains no lysines or arginines (21), although the two histidines could be charged. A third possibility is that loss of the isoleucine and the resulting decrease in the hydrophobic character of the -CTx-MII binding site are responsible for the decrease in sensitivity. These questions can be addressed with more extensive mutational analysis.

It is clear in this and previous studies that neuronal  subunits contribute to the pharmacological properties of neuronal nicotinic receptors. This can be seen clearly by comparing the -CTx-MII sensitivities of 32 and 34 in Fig. 1. The 34 receptor is also much less sensitive to the antagonists NBT and DHE than is the 32 receptor (20). The  subunits are also involved in determining sensitivity to agonists (4). One of the sequence segments of 2 that determines -CTx-MII sensitivity (54-63) is also important in determining NBT and DHE sensitivity (20). Within this region, Thr59 is important for sensitivity to all three competitive antagonists. However, the role that Thr59 plays in determining antagonist sensitivity may differ for these three antagonists. The T59K mutation may reduce NBT sensitivity through electrostatic repulsion because the introduction of a negative charge (T59D) actually results in an increase in NBT sensitivity (20). In contrast, the T59D mutation has no effect on sensitivity to either DHE or -CTx-MII.

Despite a confusing nomenclature, the  subunits of neuronal nAChRs are thought to fulfill a role analogous to that of muscle nicotinic  and  subunits; they pair with  subunits to form ligand binding sites. It is therefore useful to compare our results with those obtained in the mapping of binding site determinants on muscle  and  subunits. Of particular interest are covalent labeling experiments involving the competitive antagonist d-tubocurarine, which demonstrated incorporation of label onto a tryptophan residue of the  and  subunits (residues 55 and 57, respectively) (8). This residue is conserved in the rat neuronal 2 and 4 subunits (position 57, Fig. 7A), located near residue T/K 59 of 2/4 that we have identified as important in determination of sensitivity to -CTx-MII, DHE, and NBT (current study and Ref. 20). Other important residues have also been identified on  and  subunits. Affinity labeling, cross-linking, and mutational analysis identified Asp180 and Glu189 of the  subunit (9, 11). Both 2 and 4 have a glutamate at a position analogous to E189 of the  subunit. Tyr117 of the  subunit (T in ) has been shown to be critically important in determining curare sensitivity (27, 28). This residue is not conserved in either 2 or 4. Also of interest is another -conotoxin from Conus magus (-CTx-MI) that shows a 10,000-fold selectivity for the - binding site over the - binding site of mouse muscle nAChRs (29). Chimeric and mutant  and  subunits were used to identify the residues Ser36/Lys34, Tyr113/Ser111, and Ile178/Phe172 of the / subunits as responsible for most of the difference in binding site affinity. There is little conservation of these determinants in neuronal  subunits, which helps explain the failure of -CTx-MI to antagonize neuronal nAChRs (5, 30).

The fact that both  and  subunits affect the sensitivity of neuronal nAChRs to a wide variety of agonists and antagonists (3-6, 20, 21) suggests that the ligand binding sites of neuronal nAChRs are formed similarly to those of muscle nAChRs, at the interface between  and non- () subunits. For muscle nAChRs, the subunits are thought to be organized within the receptor with a positive face of one subunit associating with a negative face of the next subunit in a rotationally symmetrical fashion (11, 29, 31). In this model, the ligand binding sites are formed by the positive face of an  subunit and the negative face of a  or  subunit. In neuronal nAChRs such as 32, the ligand binding sites may also be formed in this fashion: from the positive face of the  subunit and the negative face of the  subunit. The 7 subunit, which can form homomeric receptors in the X. laevis oocyte expression system, seems to be able to provide both faces of the ligand binding site. To provide the positive face, 7 has the residues that  subunits contribute to the binding site (Tyr93, Trp149, Tyr190, Cys192, Cys193, Tyr198). In addition, 7 has a tryptophan at position 54, analogous to Trp55/57 of /. Mutational analysis of this residue demonstrated involvement in determining sensitivity to both agonists and antagonists, leading to the proposal that 7 contributes both an " component" (i.e., a positive face) and a "non- component" (i.e., a negative face) when forming homomeric receptors (32). For 7 to achieve this, the activity of a prolyl isomerase seems to be required (33). The formation of two binding sites by two pairs of subunits within each receptor suggests that just as for muscle nAChRs, neuronal nAChRs may exist with pharmacologically distinct ligand binding sites within a single receptor. In fact, the existence of neuronal nAChRs containing more than one type of  or  subunit seems to be possible (34-38).