Introduction

Top Summary Introduction Materials & Methods Results Discussion References

Nicotinic acetylcholine receptors are pentamers of homologous subunits with composition ₂ that form a ring around a central channel (Galzi and Changeux, 1994; Karlin and Akabas, 1995). Two of its five subunit interfaces, and , form binding sites for the neurotransmitter acetylcholine. These two binding interfaces are not identical in their affinities for agonists and competitive antagonists (Damle and Karlin, 1978; Neubig and Cohen, 1979; Weiland and Taylor, 1979; Sine and Taylor, 1981). Because the  subunit is common to each binding interface, differences in affinity are attributed to the contributions of the  and  subunits (Blount and Merlie, 1989; Petersen and Cohen, 1990, Sine and Claudio, 1991).

Recent studies showed that certain -conotoxins, 12-14-amino acid disulfide-linked peptides isolated from venom of cone snails (Myers et al., 1991, 1993), bind with unusual selectivity to one of the two ligand binding sites on mouse and Torpedo californica receptors (Kreienkamp et al., 1994; Hann et al., 1994; Utkin et al., 1994; Groebe et al., 1995; Sine et al., 1995a). -Conotoxin M1 binds with high affinity to the site of the mouse receptor (K_D = 0.5 nM), whereas it binds five orders of magnitude less tightly to the site (K_D = 20 mM) (Kreienkamp et al., 1994; Sine et al., 1995a). -Conotoxin Ml is unique in that its degree of selectivity is greater than for any known ligand, and its site preference is opposite to that of curariform antagonists, which bind more tightly to the site (Blount and Merlie, 1989; Petersen and Cohen, 1990; Sine and Claudio, 1991).

The high degree of sequence identity between the  and  subunits suggests that the polypeptide chains of the two subunits fold into similar basic scaffolds. Thus residues in equivalent positions of the linear sequence are predicted to occupy similar positions in three dimensional space. This idea is supported by the striking finding that three residues in equivalent positions of the  and  subunits confer virtually all of the selectivity for -conotoxin Ml (Sine et al., 1995a). Residues in the  subunit that stabilize -conotoxin Ml have not been identified, but mutagenesis studies with agonists and antagonists revealed three regions of the  subunit that contribute to the binding interface (Sine et al., 1994; Galzi et al., 1991; Middleton and Cohen, 1991; Tomaselli et al., 1991; O'Leary and White, 1994; O'Leary et al., 1994; Sugiyama et al., 1996). These three regions differ from binding site regions in the  and  subunits, because they are predicted to be located on the opposite face of the subunit. In this study we employ radioiodinated -conotoxin M1 to show directly that its binding requires an intact subunit interface; high affinity binding is found only at the interface. We then examine through residue replacement the relationships between amino acid determinants in the  subunit and those in the  and  subunits that govern -conotoxin M1 binding.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Radioiodination of conotoxin M1. -Conotoxin M1 (American Peptide Co.), 0.75 nmol (1.125 mg), was iodinated at its single tyrosine (Tyr-12) with 0.1 mCi of Na¹²⁵I (Amersham) and 2.5 mg of lactoperoxidase (Sigma) in 100 ml of a 50 mM NaPO₄ buffer, pH 7.5. Free iodide was removed by selective adsorption on a Dowex 1X-8 (Bio-Rad) cationic resin (Marchot et al., 1993). Labeled -conotoxin M1 was stored as a 0.5 mM solution in a 1:1 methanol:50 mM NaPO₄ buffer, pH 7.5, at 20°. Dilute solutions (1 mM) were prepared in 1 mg/ml bovine serum albumin, 50 mM NaPO₄, pH 7.5, stored at 4°, and used within the next 3 weeks. Specific activities of 100 Ci/mmol were achieved, which corresponded to 0.05 atom of iodine incorporated per molecule of -conotoxin M1. The ratio of labeled species was kept low to minimize formation of diiodo--conotoxin M1.

Cell transfections. Human embryonic kidney 293 cells were transfected with cytomegalovirus-based expression vectors containing the respective cDNAs encoding the individual subunits by Ca₃(PO₄)₂ precipitation (Sine, 1993; Kreienkamp et al., 1994). Typically plasmids containing cDNAs encoding , , , and  subunits in the weight ratio 2:1:1:1, and , , and , or , , and  in the ratio of 2:1:2 were transfected. The transfections of cDNAs encoding the above four and three subunits yielded pentameric receptors ₂, ₂₂ or ₂₂ expressed at the cell surface (Sine and Claudio, 1991), whereas transfection of a cDNA encoding  or  subunit and cotransfection of two cDNAs encoding  and  or  and  required permeabilization of the cells to detect monomeric, dimeric, or tetrameric combinations of subunits within the cells (Green and Claudio, 1993; Kreienkamp et al., 1995).

Association of ¹²⁵I--conotoxin M1 and ¹²⁵I--bungarotoxin with isolated and assembled subunits. Three days after transfection, cells were harvested by gentle agitation in phosphate-buffered saline, pH 7.4, containing 5 mM EDTA. After low speed sedimentation, cells were permeabilized with saponin-containing buffer (10 mM, EDTA, 0.1% bovine serum albumin, and 0.5% saponin in 10 mM NaPO₄, pH 7.4), and then incubated on ice with 5 nM ¹²⁵I--bungarotoxin (specific activity of 8-16 mCi/mg; DuPont-New England Nuclear, Boston, MA) or with ¹²⁵I--conotoxin M1 at the specified concentrations. Cells were then sedimented and washed free of excess unbound ligand with potassium Ringer's buffer. The pellets were solubilized on ice in 1.0% Triton X-100, 150 mM NaCl, 5 mM EDTA, 50 mM Tris·HCl, pH 7.5. After 4 hr, supernatants were layered over 3-30% sucrose gradients containing the same detergent buffer. Layered gradients were centrifuged in a Beckman SW41 rotor at 40,000 rpm for 22 hr at 4°. Fractions were collected and assayed; the S values were determined as previously described (Kreienkamp et al., 1995; Sugiyama et al., 1996).

-Conotoxin competition with ¹²⁵I--Bungarotoxin. Harvested cells were resuspended in potassium Ringer's buffer to measure ligand binding to receptors expressed on the cell surface. Specified concentrations of -conotoxin M1 were added to each aliquot of cell suspension 60 min before measurement of the initial rate of ¹²⁵I--bungarotoxin binding. The fractional reduction in the initial rate corresponds to the fractional occupation of sites by -conotoxin M1 (Sine and Taylor, 1981; Sine et al., 1995a). K_D values for each mutant are given as averages from two separate cell transfections. Sufficient cells were transfected to generate an entire concentration profile with duplicate samples measured at each concentration. The K_Dvalues typically varied by less than 20%.

Site-directed mutagenesis. Mutations of the individual subunits were generated by the method of Kunkel (1985); the entire mutagenic insert was sequenced to verify the mutation and rule out random polymerase errors.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

¹²⁵I--Conotoxin M1 binding to the receptor subunits. When cDNA encoding  subunit is cotransfected with cDNAs encoding either the  or  subunit, the resulting assembled subunits are retained within the cell. To detect association of labeled -conotoxin, with monomeric subunits and the and assembled oligomers, the cells were permeabilized with saponin, incubated with labeled toxin, washed, solubilized, and centrifuged into sucrose density gradients. The gradients resolve free toxin as well as labeled toxin associated with monomeric, dimeric, and tetrameric subunit oligomers (Fig. 1).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Sucrose density gradient profiles of 125I--Bungarotoxin (B1-B3) and 125I--conotoxin M1 (C1-C5) association with nicotinic acetylcholine receptors subunits. HEK cells were transfected with cDNAs encoding the respective subunits,  (B1 and C1) and  (C2) subunits, or two subunit combinations (B2 and C3) or (B3, C4, and C5). After 48 hr, cells were permeabilized with saponin. The respective radioiodinated toxins [~5 nM 125I--bungarotoxin, 20 nM 125I--conotoxin M1 (C1-C4), and 400 nM 125I--conotoxin M1 (C5)] were allowed to bind. Excess toxin was removed by washing and the cells solubilized with Triton X-100. The solubilized material was layered on a 5-30% sucrose gradient, and fractions collected from the top of the tubes and counted. Comparisons between 125I--conotoxin M1 and 125I--bungarotoxin binding for particular subunits were made from the same platings of transfected cells.

Identification of  subunit determinants for -conotoxin M1 binding. Site-directed labeling and mutagenesis studies have defined three linearly distinct regions in the  subunit (regions A-C) that contribute to the ligand binding site (Dennis et al., 1988; Galzi et al., 1991; Middleton and Cohen, 1991; Tomaselli et al., 1991; O'Leary and White, 1992; O'Leary et al., 1994; Sine et al., 1994; Fu and Sine, 1994; Keller et al., 1995). Each region, presumably existing as a loop at the subunit interface, contains conserved aromatic residues, including Y93, W149, Y190, and Yl98 (Fig. 2). Here we mutate residues in these three regions and measure changes in -conotoxin M1 affinity. Most mutations are substitutions for the aromatic residues, but also include substitutions of residues differing between muscle and neuronal  subunits. Binding of -conotoxin M1 was measured by competition against the initial rate of ¹²⁵I--bungarotoxin binding, as described previously (Sine and Taylor, 1981).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Amino acid sequences and studied residue mutations for the three domains in the  subunit which contribute to ligand binding. Bold residues, mutations.

Mutation of conserved tyrosines in regions A, B, and C. Mutation of Y93, Y190, and Y198 to aliphatic hydroxyl side chains (S or T) reduces affinity for -conotoxin M1 (Fig. 3 and Table 1). Furthermore, the Y190T and Y198T mutations result in a selective influence for the affinity of the site is affected to much greater extent than the site. Removal of the aromatic hydroxyl at Y151, Y190, and Y198 has no effect on -conotoxin M1 affinity, whereas the Y93F mutation enhances affinity in a site-selective manner, increasing affinity for the site without affecting the site (Fig. 4 and Table 1). Thus, mutation of the four conserved tyrosines affects -conotoxin affinity in a manner similar to other antagonists (O'Leary et al., 1994; Sine et al., 1994). Aliphatic hydroxyl substitutions dramatically reduce affinity for -conotoxin as observed previously for agonists and antagonists, whereas the removal of the hydroxyl group by substitution of phenylalanine either has little influence or enhances affinity (Sine et al., 1994, 1995b; Tomaselli et al., 1991; O'Leary et al., 1994). Interestingly, Y93F increases affinity of -conotoxin (Fig. 4), whereas Y198F enhances d-tubocurarine affinity (Fu and Sine, 1994; Sine et al., 1994).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   -Conotoxin M1 competition with the initial rate of -bungarotoxin association at cell surface nicotinic receptors after transfection of HEK cells with cDNAs encoding the component subunits. Cells were transfected with wild-type or mutant  subunit cDNAs along with , , and  cDNAs. After 48 hr, the cells were incubated with -conotoxin M1 at the specified concentrations, and initial rates of 125I--bungarotoxin association were measured. A, , wild-type  subunit; , Y190T; , Y198T. B, , wild-type  subunit; , D152N. A shows that the mutation in the  subunit affects the high affinity, interface more than the low affinity, interface. B shows a shift by the D152N mutation only at the interface. kobs and kmax are the bimolecular rate constants for 125I--bungarotoxin association in the presence and absence of -conotoxin M1.

View this table: [in this window] [in a new window]   TABLE 1 Dissociation constants for -conotoxin M1 and nicotinic acetylcholine receptor complexes

View larger version (21K): [in this window] [in a new window]   Fig. 4.   -Conotoxin M1 competition with the initial rate of -bungarotoxin association at cell surface nicotinic receptors after cotransfection of the cDNAs encoding the individual subunits. cDNAs encoding wild-type () or mutant V188K (), Y93F ()  subunits along with specified combinations of other subunits were transfected into HEK cells and -conotoxin M1 binding was measured on the cell surface as described in Methods. A, Wild-type  subunits, V188K or Y93F mutant  subunit with , , and  subunits. B, Wild-type  subunits, V188K or Y93F mutant  subunits with  and  subunits. C, Wild-type  subunits or V188K or Y93F mutant subunits with  and  subunits. Dissociation constants were calculated assuming high and low affinity sites of equal population and are listed in Table 1. The assay is identical to that shown in Fig. 3.

Mutation of charged, aromatic, and proline residues in region C. -Conotoxin M1 is amidated on its carboxyl-terminus, lacks negatively charged side chains, but contains three or four positively charged side chains depending on pH. We therefore focused on charged residues in loop C, which differ between muscle and neuronal subunits (Fig. 2). Replacement of positively charged residues with R182V or K185E, or insertion of a positive charge with F189K, fails to affect -conotoxin M1 affinity (Table 1). By contrast, inserting a positive charge with V188K selectively decreases -conotoxin Ml affinity for the site without affecting affinity for the site (Fig. 4 and Table 1). Thus, electrostatic interactions with -conotoxin M1 seem localized to position 188 and specifically affect the site.

Mutations in loop B. Mutations of the conserved W149 and Y151 to phenylalanine do not significantly affect -conotoxin Ml binding (Table 1). The lack of effect of W149F on -conotoxin M1 affinity contrasts with the large reduction in agonist affinity associated with this mutation (Sine et al., 1994). Previous studies showed that D152N introduces a glycosylation consensus site that becomes glycosylated, and further that agonist and antagonist affinities are markedly reduced (Sugiyama et al., 1996). For -conotoxin M1, D152N reduces affinity slightly for the site without affecting the site. Thus, the aromatic residues examined in region B make no contribution to -conotoxin M1 affinity, whereas the negative charge at position 152 enhances its affinity.

Mutagenesis of residues in both the  subunits and non- subunits. Previous studies identified three residues in equivalent positions of the  and  subunits that confer higher affinity of -conotoxin M1 for the over the interface (Sine et al., 1995a). To account for the selective effect of Y93F and V188K for the interface, we reasoned that either -conotoxin M1 affinity is enhanced by interaction between determinants in the  and subunits or -conotoxin M1 orients differently at the and interfaces. Thus we investigated these two possibilities by examining -conotoxin Ml binding to receptors containing mutations in both the  and  subunits.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Interrelationships in -conotoxin M1 dissociation constants for  subunit mutant receptors containing mutations in the  and  subunits. Data are determined from concentration profiles similar to those shown in Figs. 3 and 4. A, Cotransfection of three subunits:  (wild-type or mutant Y93F or V188K), wild-type , either wild-type or mutant , or wild-type  were cotransfected. B, Cotransfection of four subunits:  (wild-type or mutant Y93F or V188K), wild-type , wild-type or mutant , and wild-type  were cotransfected. The dissociation constants determined are shown on a logarithmic plot relative to wild-type  subunits.

View this table: [in this window] [in a new window]   TABLE 2 -Conotoxin MI dissociation constants (KD) for wild-type and mutant nicotinic receptors expressed from three and four subunit combinations KD values are averages from two separate transfections. In each transfection, a complete concentration profile is generated using duplicate sampling at each concentration (compare Figs. 3, 4, and 6). Transfection of /// (2:1:1:1), // (2:1:2), and // (2:1:2) in the specified cDNA ratios yields the above stoichiometry of subunits (Sine and Claudio, 1991; Sine 1993).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Influence of  subunit mutations of Y190T and Y198T on the binding of -conotoxin M1 to receptors of 22 composition containing the three  subunit mutations required to increase affinity. The  subunit mutations (K34S, S111Y, and F172I) substitute three residues found at homologous positions in the  subunit into the  subunit template. By substitution of three residues of the  subunit sequence into a  subunit, high affinity for -conotoxin M1 is conferred as well as the sensitivity to tyrosine substitutions in the  subunit resulting in a loss of affinity.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Subunit contributions to -conotoxin M1 association. Radioiodination to form ¹²⁵I--conotoxin M1 enables one to monitor directly the -conotoxin association with various subunit combinations. High affinity binding where bound ¹²⁵I--conotoxin M1 is retained by the receptor subunits after sedimentation is evident only for the dimer. The same procedure shows -bungarotoxin association with  subunit monomers, dimers, dimers, and tetramers. These findings not only document directly the site selectivity of -conotoxin M1, but also reveal that -conotoxin M1 requires an intact subunit interface for high affinity binding.

Residues conferring -conotoxin M1 selectivity. Previous studies have defined three residues in the  subunit (S36, Y113, and I178), which confer high affinity binding of -conotoxin M1 to the interface (Sine et al., 1995a). When substituted with the corresponding residues in the  subunit (K34, S111, and F172), this triad of mutations in the  template confers an -conotoxin M1 affinity approaching that of the  subunit (Sine et al., 1995a). To account for and interfaces showing nearly a 10,000-fold difference in -conotoxin M1 affinity, either the  subunit diminishes the influence of determinants on the  subunit to -conotoxin binding or determinants on the  subunit are major contributors to the high affinity of -conotoxin M1. In turn, the  subunit could enhance affinity either directly through its own interactions with conotoxin or by altering the conformation of the  subunit. Because the differences in -conotoxin M1 affinity between and interfaces are diminished and actually inverted for the T. californica receptor (Hann et al., 1994; Utkin et al., 1994; Groebe et al., 1995; Sine et al., 1995a) and affinity of the subunit interface is intermediate to the and interfaces in mouse (Sine S, unpublished observations), the , , and  subunits likely contribute to stabilization of the -conotoxin complex both directly and by affecting the conformations of the neighboring  subunit.

Relationship between  subunit mutations and / subunit mutations. We examined whether linkage relationships exist between mutations in the  subunit and those in  and  subunits or whether the determinants on each subunit act independently of each other as would be expected if they bound to different portions of the -conotoxin molecule. A strong linkage relationship is apparent between S36 and I178 on the  subunit, where the individual substitutions, S36K or I178F, has a small or no influence on -conotoxin K_D values, yet when both substitutions are made, a marked loss in affinity or increase in K_D evident (Sine et al., 1995a). Using a mutant cycle analysis (Carter et al., 1984) for mutations at corresponding positions 36 and 178 in the  subunit template, we may set the following cycle: Where Then In this formulation, the dissociation constants K, for -conotoxin and the receptor site are subscripted by the residues at positions 36 and 178. S_1, S₂, I₁, and I₂ designate the ratio of dissociation constants for the species at the base of the arrow relative to the tip of the arrow. The values are represented as absolute values without reference to sign.