Determinants of Agonist Binding Affinity on Neuronal Nicotinic Receptor β Subunits

Experimental Procedures

Materials.

X. laevis frogs were purchased from Nasco (Ft. Atkinson, WI). Care and use of X. laevis frogs in this study has been approved by the University of Miami Animal Research Committee and meets the guidelines of the National Institutes of Health. RNA transcription kits were obtained from Ambion (Austin, TX). [³H]Epibatidine was purchased from PerkinElmer Life Science Products (Boston, MA). Acetylcholine, DMPP, nicotine, gentamicin, HEPES, polyethylenimine, and 3-aminobenzoic acid ethyl ester were purchased from Sigma (St. Louis, MO). Collagenase B was obtained from Roche Molecular Biochemicals (Indianapolis, IN). 934-AH glass microfiber filters were obtained from Whatman (Clifton, NJ).

Expression of Neuronal nAChRs in X. laevisOocytes.

cDNA clones encoding rat α2, β2, and β4 subunits, as well as the β chimeras and β mutants engineered into the pGEMHE expression vector (Liman et al., 1992) were used as template for cRNA production. Chimeric and mutant β subunits were constructed as described previously (Harvey and Luetje, 1996). Our notation for chimeric subunits is to list the subunit from which the N-terminal sequence is derived, followed by the position at which the chimeric joint is made, followed by the subunit from which the C-terminal sequence is derived. For example, the chimeric subunit β2-54-β4 consists of β2 sequence from the N-terminal until residue 54 followed by β4 sequence from residue 54 until the C-terminal. 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 β2-T59K is a β2 subunit in which threonine 59 has been changed to a lysine.

m⁷G(5′)ppp(5′)G-capped cRNA was synthesized in vitro from linearized template cDNA 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 h at room temperature. Oocytes were injected with 20 ng of cRNA encoding various wild-type, chimeric, and mutant subunit combinations in 23 nl of water and incubated at 19°C in modified Barth's saline (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 0.3 mM Ca(NO₃)₂, 0.41 mM CaCl₂, 0.82 mM MgSO₄, 100 ug/ml gentamicin, 15 mM HEPES, pH 7.6) for 2 to 10 days. RNA transcripts encoding each subunit were injected into oocytes at a molar ratio of 1:1.

Preparation of X. laevis Oocyte Homogenates and [³H]Epibatidine Binding Assays.

Crude membrane homogenates were prepared from X. laevis oocytes expressing wild-type, chimeric, and mutant receptors as described previously (Parker et al., 1998). Briefly, 0.25 to 15 oocytes (depending on expression levels) were homogenized per milliliter of buffer (140 mM NaCl, 1.5 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, 25 mM HEPES, pH 7.5, containing freshly added 0.1 mM phenylmethylsulfonyl fluoride), using a model PT 10/35 homogenizer (Brinkmann, Atlanta, GA). Homogenates were centrifuged at 4°C at 2000g for 10 min. The supernatant was removed for use in experiments, avoiding both the surface lipid layer and the pellet. Receptor expression levels averaged 480 fmol/mg protein (16 fmol/oocyte). In previous work (Parker et al., 1998), we determined that the radioligand binding properties of neuronal nAChRs assayed in crude membrane homogenates or more purified membrane preparations were indistinguishable.

To avoid problems with ligand depletion during saturation experiments, the reaction volumes varied at different epibatidine concentrations. During characterization of the chimeric receptors, final reaction volumes of 0.5 ml were used for epibatidine concentrations between 2 and 5 nM; 1-ml volumes were used for concentrations between 500 pM and 1 nM; 2 ml volumes were used for epibatidine concentrations between 15 and 250 pM; and 5 ml volumes were used for concentrations below 15 pM epibatidine. During saturation analysis of mutant receptors, the concentration range was between 1.95 pM and 4 nM; the volumes of these reactions were kept constant at 1 ml, which was sufficient to avoid significant ligand depletion at each concentration. In competition studies, 500 pM [³H]epibatidine was used for analysis of all chimeric and mutant receptors. Reaction volumes of 0.5 ml were sufficient to avoid ligand depletion in the competition studies for the concentrations of [³H]epibatidine and competitors used. Both competition and saturation experiments contained between 10 and 25 fmol of receptor per reaction tube. For reactions involving ACh, the homogenate was preincubated for 30 min with 200 nM diisopropylfluorophosphate, a cholinesterase inhibitor, prior to the addition of ligands. Reactions were initiated by the addition of oocyte homogenate and were incubated at 25°C in a shaking water bath for 3.5 to 4.0 h. The reactions were stopped by filtration onto glass fiber filters pretreated with 0.1% polyethylenimine (Whatman 934-AH) using a model M-24 harvester (Brandel, Gaithersburg, MD). Nonspecific binding was determined in parallel reactions containing either 100 nM epibatidine (Figs.1 and 2) or 100 μM DMPP (Figs. 3 and 4). Nonspecific binding was between 10 and 15% of the total binding at [³H]epibatidine concentrations near theK_D value and did not exceed 45% at the highest radioligand concentration.

Data Analysis.

Data from saturation experiments were analyzed using the equation: B = (B_max ·Lⁿ)/(K_Dⁿ+ Lⁿ), where Bis the binding at free ligand concentration L,B_max is the maximum specific binding,K_D is the equilibrium dissociation constant, and n is the Hill coefficient. Values forB_max,K_D, and n were calculated by nonlinear regression. IC₅₀ values were derived using the equation: B =B_o/[1 + (I/IC₅₀)ⁿ], where B is ligand bound at competitor concentrationI, B_o is binding in the absence of competitor, IC₅₀ is the concentration of ligand that reduces the specific binding by one-half, andn is the Hill coefficient.K_I values were calculated using the Cheng and Prusoff equation: K_I = IC₅₀/[1 + ([L]/K_D)] (Cheng and Prusoff, 1973). Because of the variation in receptor expression level from day to day after injection of the oocytes and among oocyte batches, all results were normalized as the percentage of maximum specific binding. Prism software (GraphPad, San Diego, CA) was used to fit the data and to assess statistical significance using a two-tailed unpaired t test.

Results

Region 54–63 of β Subunits is Critical for Determining Agonist Affinity.

Previously, we found that receptors containing the β2 subunit had consistently higher affinities for agonists than did receptors containing the β4 subunit (Parker et al., 1998). Amino acid residues responsible for these differences in agonist affinity might be located within one of the three loops of the complementary component (D, E, and F) or within some previously unidentified sequence segment of β subunits. To determine which region should be examined in detail, we expressed receptors in X. laevis oocytes using a series of chimeric β subunits. The chimeric subunits consisted of portions of the β2 and β4 subunits. The structure of these chimeras is presented diagrammatically in Fig. 1B. In our previous work (Parker et al., 1998), the largest differences in agonist affinity were observed when the β subunits were coexpressed with the α2 subunit. For this reason, all chimeric and mutant β subunits in the current study are coexpressed with the α2 subunit.

Substitution of the first 133 N-terminal residues of the β2 subunit with the same portion of the β4 subunit yields a chimera (β4-133-β2) containing the D and E loops of β4 and the F loop of β2. Receptors formed by this chimera displayed a β4-like affinity for [³H]epibatidine of 85 pM (Fig. 1C). This result suggests that at least some of the residues involved in determining agonist affinity are located within region 1–133, possibly within the D or E loops. To further subdivide this region, we examined a chimera formed at position 80 (β2-80-β4) containing the D loop of β2 and the E and F loops of β4. This chimera formed receptors with a β2-like affinity for [³H]epibatidine of 12 pM (Fig. 1C), suggesting that critical residues lie within region 1–80 and possibly within the D loop. Although a β2-54-β4 chimera formed receptors with a β4-like affinity for [³H]epibatidine of 75 pM, a β2-63-β4 chimera formed receptors with a β2-like affinity for [³H]epibatidine of 9 pM (Fig. 1, A and C). We conclude from these results that residues within the D loop (residues 54–63) of β subunits are critical determinants of [³H]epibatidine affinity.K_D and apparent Hill coefficient (n_H) values for receptors formed by wild-type and chimeric subunits are provided in Table1.

Our use of [³H]epibatidine saturation analysis identifies the 54–63 region as important for agonist affinity. However, the difference in [³H]epibatidine affinity between wild-type α2β2 and α2β4 is only 9-fold. To provide additional evidence of the importance of the 54–63 region, we used several agonists that display larger differences in affinity for receptors formed by the wild-type β subunits. In previous work, we found that the affinity of α2β2 and α2β4 receptors for ACh, nicotine, and DMPP differed by 61-, 87-, and 120-fold, respectively. The affinity of these agonists for the various chimeric receptors was determined in competition assays. IC₅₀ andK_I values were then calculated as described under Experimental Procedures. Figure 2 compares the affinity of receptors formed by chimeric and wild-type β subunits for each of these agonists. K_I andn_H values are provided in Table 1.

The results shown in Fig. 2 suggest that determinants of affinity for ACh, nicotine, and DMPP are located, at least in part, within region 54–63. The similarity to the localization of determinants of epibatidine affinity is greatest for ACh affinity. Both β4-133-β2 and β2-54-β4 form receptors with a β4-like affinity for ACh. Addition of the 54–63 region (β2-63-β4) results in a nearly complete transition to a β2-like ACh affinity. The transition in nicotine affinity from β2-54-β4 to β2-63-β4 is not as complete, but the affinity of β2-63-β4-containing receptors for nicotine is within 3-fold of that of wild-type β2. Receptors formed by β2-80-β4 also have an affinity for nicotine lower than that of wild-type receptors, suggesting that additional minor determinants of nicotine affinity may lie C-terminal of residue 80.

The situation presented by the affinity of the various chimeras for DMPP differs from that of epibatidine, ACh, and nicotine. The affinity of β2-54-β4-containing receptors differs from wild-type β4-containing receptors by 6-fold. This contrasts with the affinities of this chimeric receptor for epibatidine, ACh, and nicotine, which differ from those of wild-type β4 by less than 2-fold. This result suggests that a determinant of DMPP affinity lies within the 1–54 region. The transition to β2-like DMPP affinity is complete upon addition of the 54–63 region (β2-63-β4), suggesting that like the results for the other agonists, region 54–63 is involved in determining DMPP affinity.

Multiple Residues within Region 54–63 Determine Agonist Affinity.

Our results indicate that the 54–63 region of the β subunit is a major determinant of epibatidine, ACh, nicotine, and DMPP affinity. An alignment of this region from the β2 and β4 subunits is shown in Fig. 3A. These two subunits differ at positions 55, 56, 59, and 63. One or more of these residues could be responsible for the affinity differences we have mapped to this region.

A series of mutant β2 subunits, in which the residue at position 55, 56, 59, or 63 has been changed to what occurs in β4, were coexpressed with α2, and [³H]epibatidine saturation analysis was performed. Saturation analyses of the mutant receptors were done using slightly different conditions (see Experimental Procedures) than the saturation for the wild-type and chimeric receptors. Therefore, saturation analyses were redone for wild-type α2β2 and α2β4 receptors using the new conditions to provide an accurate comparison with the mutant receptors. Agonist affinities of wild-type receptors determined using the two methods differed by no more than 2-fold. The results of these analyses are displayed in Fig.3. K_D andn_H values are provided in Table2. The N55S and E63T mutations had no effect on epibatidine affinity, and the effects of the V56I and T59K mutations were modest (less than 2-fold).

Similar to our work with the chimeric receptors, we turned to the use of competition analyses with additional agonists to gain a better view of the potential involvement of the various residues in the 54–63 region (Fig. 4). When calculated using theK_D values in Table 2, theK_I values for ACh, nicotine, and DMPP binding to α2β2 and α2β4 differ by 81-, 116-, and 159-fold, respectively. We found that mutation at each of the four positions resulted in a significant decrease in affinity for ACh and nicotine. The N55S, V56I, and E63T mutations lowered the affinity of the resulting receptors for both ACh and nicotine by 3- to 4-fold. The T59K mutation lowered the affinity of the resulting receptor for ACh and nicotine by 8- and 7-fold, respectively. These mutations had no effect on DMPP affinity.

The largest effects on agonist affinity occurred with the T59K mutation. The most obvious change in side chain character with this mutation is the introduction of the positive charge. To further examine the potential role of side chain charge at position 59, we also performed saturation and competition analyses with a α2β2-T59D receptor. The T59D mutation had little effect on epibatidine and DMPP affinities (<2-fold) but decreased ACh and nicotine affinity by 4-fold and 7-fold, respectively.

Discussion

Neuronal nAChRs display a wide range of equilibrium binding affinities for agonists. We demonstrated previously that the primary determinant of agonist affinity for simple heteromeric receptors was the identity of the β subunit, with β2-containing receptors having consistently higher affinities for agonists than β4-containing receptors (Parker et al., 1998). We have now used chimeric and mutant β subunits to demonstrate that several amino acid residues within the region known as the D loop are important in determining agonist binding affinity.

Within the D loop region, the β2 and β4 subunits differ at only four positions. Mutation at each of these positions in β2 resulted in a significant loss in both ACh and nicotine affinity. Although experiments in Fig. 2 suggest that the D loop accounts for a large fraction of the difference in affinity between β2 and β4 receptors, the loss of affinity for each mutant was relatively small. The largest effect for both ACh and nicotine affinity occurred with the T59K mutation and was only 7- to 8-fold. However, when considered together, the effects of the four mutations more than account for the differences in ACh and nicotine affinity between β2- and β4-containing receptors. Each of the four mutations had only a minimal (<2-fold) effect on epibatidine affinity. This is not surprising given the modest difference in epibatidine affinity between β2 and β4 receptors. What was surprising was the lack of effect of any of the mutations on DMPP affinity. The largest difference in agonist affinity between β2 and β4 receptors was seen with DMPP, and results in Fig. 2 clearly indicate the importance of the 54–63 region. However, Fig. 2 also indicates that region 1–54 plays an important role in determining DMPP affinity. Thus, it may be that loss of any one determinant in the D loop, although leaving intact other D loop determinants and region 1–54, is not sufficient to destabilize the high affinity binding of DMPP.

The T59K mutation had the largest effect on ACh and nicotine affinity. The most obvious change in side chain character is the change from a polar hydroxyl group to a positive charge. In previous work, we found that the β2-T59K mutation decreased the sensitivity of the α3β2 receptor to antagonism by neuronal bungarotoxin (Harvey and Luetje, 1996). When a negative charge was introduced at this position (T59D), sensitivity to neuronal bungarotoxin was increased, suggesting that the decreased sensitivity of the T59K mutant was indeed due to the introduction of the positive charge. To determine whether side chain charge at position 59 is also the critical factor in determining agonist affinity, we examined the agonist affinity of receptors formed by β2-T59D. In contrast to the increased neuronal bungarotoxin sensitivity, we found that the T59D mutation decreased ACh and nicotine affinity. This result suggests that some other change in side chain property, such as loss of the hydroxyl, is important in determining ACh and nicotine affinity.

All of the work presented in this study involves radioligand binding assays of mammalian neuronal nAChR expressed in X. laevisoocytes. Exogenous expression of mammalian receptors in a nonmammalian system raises concern regarding the accuracy and relevance of the results. Specifically, are the pharmacological properties of neuronal nAChRs that we measure using the oocyte expression system an accurate reflection of the properties that these receptors would have in a mammalian context? We have previously provided evidence that this is the case for several different subtypes of nAChR. The agonist pharmacology of mouse muscle nAChRs expressed in oocytes (Luetje and Patrick, 1991) was similar to the pharmacology of the same receptors natively expressed by BC3H-1 cells (Sine and Steinbach, 1986, 1987). Rat α4β2 receptors expressed in oocytes displayed affinities for agonists and antagonists (Parker et al., 1998) that were similar to those of α4β2 receptors expressed in rat brain (Pabreza et al., 1991). Rat α3β4 receptors expressed in oocytes (Parker et al., 1998) displayed agonist binding affinities that were similar to those of α3β4 receptors expressed by rat trigeminal ganglia neurons (Flores et al., 1996) and to those of rat α3β4 receptors exogenously expressed in HEK 293 cells (Xiao et al., 1998). Rat α2β2 and α2β4 receptors expressed in oocytes (Parker et al., 1998) displayed [³H]epibatidine binding affinities that were similar to those of rat α2β2 and α2β4 receptors exogenously expressed in HEK 293 cells (Xiao et al., 1996). These results suggest that mammalian nAChRs expressed in X. laevis oocytes display accurate pharmacological properties.

The recent publication of the crystal structure of a soluble ACh-binding protein secreted by molluscan glia provides new insight into the structure of the ligand binding site of nAChRs (Brejc et al., 2001). The AChBP is pentameric, has a nicotinic pharmacology, and possesses many of the residues thought to be critical participants in the ligand binding site of nAChRs. Thus, it seems likely that the extracellular domains of neuronal nAChR subunits will conform, at least approximately, to this structure. The “D loop” region of nAChR subunits corresponds to the second β strand of the AChBP structure. Within this region, side chains of residues in AChBP that correspond to N55, T59, and E63 of β2, all extend toward the binding site, providing a clear explanation for why alteration of these residues causes changes in agonist binding affinity. The side chain of the residue in AChBP that corresponds to V56 of β2 faces away from the binding site and into the hydrophobic core of the protein. The V56I mutation, which increases the side chain volume, might alter the position of the D loop region and thus the position of the side chains extending into the binding site.

The current model of nAChR agonist binding site structure consists of three loops of amino acid sequence (A, B, and C) from the α subunit forming a principal component and three loops of sequence (D, E, and F) from the non-α subunit (γ, δ, or ε in muscle nAChRs, β in heteromeric neuronal nAChRs) forming a complementary component (Corringer et al., 2000). The recently published AChBP structure confirms this binding site structure (Brejc et al., 2001). Loop E contains a sequence segment (104–120 in β2, 106–122 in β4) that affects agonist sensitivity (Cohen et al., 1995). A residue within this region (β2F106, β4V108) has been identified as a determinant of substance P sensitivity (Stafford et al., 1998). Loop F contains a glutamate at position 177 of β2 (179 in β4) that is analogous to E183 of γ (E189 of δ), which has been shown to be a determinant of agonist affinity (Czajkowski et al., 1993; Martin et al., 1996). Our results indicate that the D loop contains critical residues that determine differences in agonist binding affinity among neuronal nAChRs.

The muscle nAChR D loop contains a tryptophan residue (γ55/δ57) that is labeled by nicotine (Chiara et al., 1998). This tryptophan is one of a series of conserved “core” residues that have been identified by affinity labeling and mutagenesis within loops A, B, C, and D (Corringer et al., 2000). These core residues are flanked by variable residues that control the pharmacological diversity of nAChRs. In muscle nAChRs, residues H60 of γ and A61 of δ are important in determining affinity for dimethyl-d-tubocurarine (Bren and Sine, 1997), whereas residues E57 of γ and D59 of δ are important in determining carbamylcholine binding affinity (Prince and Sine, 1996). In neuronal nAChRs, loop D contains T59 in β2 (K61 in β4). We previously identified this residue as a critical determinant of sensitivity to the competitive antagonists neuronal bungarotoxin, dihydro-β-erythroidine and α-conotoxin-MII (Harvey and Luetje, 1996; Harvey et al., 1997). Our current results show that several residues in the D loop determine agonist binding affinity, with β2T59/β4K61 having the largest effect on ACh and nicotine affinity. Thus, amino acid residues within the D loop play a critical role in determining the affinity of both muscle and neuronal nAChRs for agonists and competitive antagonists.