Introduction

Top Summary Introduction Materials & Methods Results Discussion References

Nicotinic acetylcholine receptors (nAChR) are ligand-gated cation channels found at the vertebrate neuromuscular junction, nerve cell membranes, and electroplaque of the electric fish. The skeletal muscle and the electroplaque receptor is composed of four different subunits (₂) arranged in a rosette forming an integral membrane ion channel (Unwin, 1993). Neuronal nAChRs are pentameric receptors consisting of a homomeric array of one type of subunit (7, 8, or 9) or a heteromeric array of two different subunits (e.g., 24, 32). Sequence comparisons suggest that the homomeric nAChR forms are ancestral to the nicotinic family (Ortells and Lunt, 1995). Although a variety of subtypes are present, the two most abundant neuronal nAChRs in vertebrate brain are receptors containing the 7 subunit and the 42 receptor (Galzi and Changeux, 1996). Because of the importance of these receptors in a variety of normal (Wonnacott, 1997) and pathological processes (Lindstrom, 1997), as well as in the effects of drugs of abuse (Gotti et al., 1997), the development of subtype-specific cholinergic agents is of considerable interest.

Although pharmacological differences between skeletal muscle and ganglionic nAChRs have been long known (Taylor, 1996), the therapeutic potential of exploiting differences in neuronal nAChRs was recognized following the identification of the different neuronal subtypes (Sargent, 1993) and the finding that nicotinic receptors are important in a number of pathological processes (Lindstrom, 1997). The specific combination of  and  subunits can affect the affinity for a particular agonist. For example, the efficacy of cytisine is dictated almost exclusively by the type of  subunit present in a receptor oligomer (Papke et al., 1991). In the case of the homomeric 7, 8, and 9 receptors, pharmacological differences between these subtypes and between versions of these receptors in different species can be used to identify regions of the receptors involved in pharmacological specificity. For example, the rat and chick 7 nAChR, which are more than 87% identical in primary structure, exhibit dramatically different efficacy for the cholinergic agonist 1,1-dimethyl-4-phenylpiperazinium (DMPP). Thus, these pharmacological differences between highly homologous subunits can be used to determine both the portions of the receptor that control affinity and/or efficacy to specific agonists. In particular, a combination of site-directed mutagenesis with structure-activity relationships of cholinergic agonists has the promise of producing highly specific drugs for particular subtypes.

The optimal geometric requirements for an agonist of the nAChR are a cationic head (e.g., a quaternary ammonium group) and a hydrogen bond acceptor (e.g., a carbonyl oxygen) separated by 5.9 Å (Beers and Reich, 1970). We have developed a series of cholinergic agonists with channel blocking properties that have proven useful in studying the ion channel of the skeletal muscle nAChR (Carter and Oswald, 1993). These compounds are based on the structure of a synthetic agonist, 1,1-dimethyl-4-acetylpiperazinium iodide (HPIP) (Spivak et al., 1986), which is held by its ring structure in approximately the correct configuration for binding (the quaternary amine is 6.1 Å from the van der Waals extension of the carbonyl oxygen). By adding methylene groups to the acetyl moiety of HPIP (Fig. 1), the off-rate of channel blockade of the skeletal muscle nAChR is decreased and the position of blockade within the channel is modulated (Carter and Oswald, 1993). Activation of the skeletal muscle channel, however, is essentially identical for the entire series. Unlike the skeletal muscle nAChR, the rat and chick forms of the homomeric 7 nAChR are differentially activated by compounds in the piperazinium (PIP) series, and in an analogous fashion, they are also differentially activated by the ganglionic cholinergic agonist DMPP. The PIP ring of DMPP is identical with the PIP series, and the PIP series, in effect, substitutes a carbonyl and aliphatic side chain of varying length for the phenyl ring of DMPP. By focusing on nAChR subtypes that could be formed from a single subunit, we minimized the complexity of the multisubunit composition of the binding site and the species differences. In addition, the efficacy of DMPP for homomeric receptors is well known (Séguéla et al., 1993; Elgoyhen et al., 1994, Gerzanich et al., 1994; Chavez-Noriega et al., 1997), and using this information, we postulated that position 84 of 7 receptors may be a critical determinant of DMPP, and by extension, PIP efficacy. We show here that the PIP compounds can be explicitly "tuned" to activate a receptor with a leucine in position 84 but not one with a glutamine in that same position, suggesting a potential strategy for producing subtype specific cholinergic agonists.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Structures of DMPP and PIP compounds.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Materials. DMPP was purchased from Sigma Chemical Company (St. Louis, MO). HPIP was synthesized as described by Spivak et al. (1986). The longer chain derivatives of HPIP were synthesized in the same manner with the appropriate anhydride substituted (equal number of moles) for acetic anhydride: propionic anhydride for EthylPIP, butyric anhydride for PropylPIP, valeric anhydride for ButylPIP, hexanoic anhydride for PentylPIP, and heptanoic anhydride for HexylPIP (Carter and Oswald, 1993). The syntheses were verified by ¹H NMR spectroscopy at 500 MHz (Varian Unity 500 Spectrometer; Varian Instruments, San Fernando, CA). The rat 7 DNA was provided by Dr. Ron Lukas (Barrow Neurological Inst., Phoenix, AZ), the chick 7 was provided by Dr. Marc Ballivet (Université de Genève, Geneva, Switzerland), and the chick 7 L247T mutant was from Dr. Jean-Luc Galzi (Institut Pasteur, Paris).

Mutagenesis. The overlapping polymerase chain reaction (PCR) method was used to generate single point mutations in the chick 7 and the rat 7 sequence. The overlapping primers used for the chick 7 mutant were 5'-CCTGATGGACAGATTTGGAAG-3' (sense) and 5'-CTTCCAAATCTGTTCATCAGG-3' (antisense) L84Q. For the rat 7, the overlapping primers were 5'-CCAGATGGCCTGATTTGGAA-3' (sense) and 5'-TTCCAAATCAGGCCATCTGG-3' (antisense) Q84L. For the chick 7, the PCR fragments were cut with SacII and BglII and exchanged for wild-type chick 7 in the pOEV vector, and then subcloned into the pCEP4 vector with the restriction enzymes XhoI and KpnI. In rat 7, PCR fragments were cut with KpnI and SacI and exchanged for wild-type rat 7 in the pCEP4 vector. The PCR products containing mutations were selected and verified by DNA sequencing.

Recording from Oocytes. For DNA injection, the cDNAs were subcloned into the oocyte expression vector pCEP4 (Invitrogen, San Diego, CA) using the XhoI and KpnI sites. For RNA injections, cDNA was cloned into the SP65 vector and the AmpliScribe T7 or SP6 High Yield Transcription Kit (Epicentre Technologies, Madison, WI) was used to transcribe RNA. Plasmids were propagated in the Escherichia coli host (DH5 strain) and prepared using a Stratagene Midi Plasmid Quik kit (La Jolla, CA). A 10-nl sample of DNA (1-2 ng of plasmid DNA) was injected into the nucleus 3 to 4 days before recording. In the case of RNA injection, a 50-nl sample (approximately 8 ng of RNA per subunit) was injected into the cytoplasm 2 to 3 days before recording. Two-electrode voltage clamp measurements were made at room temperature in oocyte saline solution with 200 µM CaCl₂ (the low concentration of calcium minimizes the calcium-activated chloride channel activation) using a Turbo Tec 01C amplifier. The voltage electrodes were filled with 3 M KCl and had a resistance of 0.5 to 2 M. The current electrode was filled with 250 mM CsCl, 250 mM CsF, and 100 mM EGTA, pH 7.3. The resistance of the current electrode was between 0.3 and 2 M. Saline was allowed to flow over the oocytes at a rate of 4.5 ml/min in a Lucite chamber with a total volume of 300 µl. Agonist solutions were applied to the oocyte from a blunt pipette (1.5-mm diameter) placed 0.5 mm above the oocyte. In this manner, flows of 9 ml/min could be achieved with good mechanical stability of the oocyte. Judging from the rise time of the current following ACh application to oocytes injected with mouse muscle nAChR subunits (more slowing desensitizing than 7 nAChRs), the solutions were changed within approximately 0.5 s. Agonist application was initiated by a computer-triggered stream-switching valve, and data were collected on-line using software developed in the laboratory.

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Differential Efficacy of Agonists on Chick and Rat 7 Homomeric Acetylcholine Receptors. DMPP and the PIP series of cholinergic agonists (Fig. 1) were used to activate both rat and chick 7 homomeric acetylcholine receptors. As shown in Fig. 2A, all were able to activate the rat 7 nAChR. Activation and desensitization by ButylPIP was considerably slower than the other PIP derivatives and the dose-response curve was bell shaped with a maximum near 1 mM (Fig. 5A), similar to other agonists that can both activate the receptor and block the ion channel. The reason for these differences in kinetics is not clear, although in skeletal muscle nAChR (a receptor for which channel blockade by ButylPIP has been demonstrated; Carter and Oswald, 1993), similar slow kinetics were observed (R. Vazquez, unpublished results). As reported previously (Gerzanich et al., 1994), DMPP exhibits a dramatically decreased efficacy on the chick 7 nicotinic receptor (3% of the current observed with saturating concentrations of ACh; Fig. 2B). In the case of the PIP series, the first three compounds in the series (HPIP, EthylPIP, and PropylPIP) were activators of the chick 7. However, ButylPIP, which has one more methylene group than PropylPIP, was completely inactive.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Responses of rat 7 (A) and chick 7 nAChRs (B) to the PIP compounds and DMPP. For each oocyte, the ACh control (1 mM) is shown as is the corresponding calibration. All compounds were applied at 1 mM and are representative of four determinations for the PIP series and three determinations for DMPP (A), and eight determinations for the PIP series and six determinations for DMPP (B). The response to 1 mM ACh in the presence of 1 mM ButylPIP was 24 ± 4% of the response to 1 mM ACh alone. Unless otherwise noted, oocytes were clamped at 120 mV in this and all subsequent figures.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   A, response to 1 mM ACh and ButylPIP of the chick L247T mutant 7 receptor (representative of eight determinations). B, responses of the chick wild-type 7 homomeric receptor to 1 mM ACh and 1 mM ButylPIP with the oocyte clamped at +60 mV. EDTA (100 nl of 8 mM) was injected into the oocyte 15 min before recording (representative of five determinations).

Identification of an Amino Acid Partially Controlling Affinity for ACh, DMPP, and PIP Compounds. A sequence alignment of nAChR  subunits revealed very high homology between the chick and rat 7 nAChRs in the N-terminal extracellullar domain. However, a number of differences exist that may explain the differential efficacy of DMPP and ButylPIP. By comparing the efficacy of DMPP for all 7, 8, and 9 homomeric nAChRs, a glutamine in position 84 was strongly correlated with high efficacy for DMPP and a leucine in this position was correlated with low efficacy (Fig. 4). Other differences between the chick and rat 7 subunits showed no correlation when compared across 7, 8, and 9 sequences for which activation by DMPP had been reported. The L84Q mutant of chick 7 and the Q84L mutant of rat 7 were constructed to determine whether this site at least partially determines the efficacy for DMPP and the PIP compounds. As shown in Fig. 5A, both mutants exhibited a moderate change in affinity for ACh. In the case of the rat Q84L mutation, the EC₅₀ for ACh increased (260 µM for the mutant versus 150 µM for wild type). The chick L84Q mutation exhibited a decreased EC₅₀ for ACh (180 µM for the mutant versus 270 µM for wild type). Thus, although the mutations result in modest symmetrical changes in the affinity for acetylcholine, they do not produce a major change in the normal function of the two receptors.

View larger version (51K): [in this window] [in a new window]   Fig. 4.   Alignment of several neuronal nAChR subunits. The boxes indicate the positions of conserved sequences thought to participate in the acetylcholine binding site. DMPP is a full agonist for rat 3 (this includes 32 and 34; Luetje and Patrick, 1991), chick 8, human 7, and rat 7 (Chavez-Noriega et al., 1997; Gerzanich et al., 1994; Séguéla et al., 1993). In these cases, position 84 is Q or K. In the cases of chick 7 (Gerzanich et al., 1994), rat 9 (Elgoyhen et al., 1994), and rat 4 (this includes 42 and 44; Luetje and Patrick, 1991), DMPP is a weak partial agonist activating less than 20% of the maximal current induced by acetylcholine.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   A, Dose-response curves for acetylcholine, DMPP, and ButylPIP. Each point in the dose-response curve to acetylcholine is an average of two to six determinations, with the error bars representing the S.E.M. Lines are a nonlinear least squares fit of the entire data set to the Hill equation for ACh and DMPP. Hill coefficients ranged from 1.3 to 1.55 for ACh and were 1 for DMPP. The fit for ButylPIP included a term for inhibition as well as activation. Also shown are responses of the rat 7 Q84L mutant (B) and chick 7 L84Q mutant (C) to acetylcholine, the PIP compounds, and DMPP. For individual traces, the ACh control (1 mM) is shown as is the corresponding calibration. All compounds in individual traces were applied at 1 mM and are representative of four to five determinations. The response to 1 mM ACh in the presence of 1 mM ButylPIP was 35 ± 3% of the response to 1 mM ACh alone (the trace shown was from a separate oocyte but was normalized to the response to ACh alone in that same oocyte).

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

The acetylcholine binding site on nicotinic acetylcholine receptors seems to consist of a number of highly conserved residues found in all subtypes (Fig. 4 and Galzi and Changeux, 1994). However, some compounds (e.g., DMPP, epibatidine, and anabaseine; Séguéla et al., 1993; Gerzanich et al., 1994, 1995, Kem et al., 1997) can distinguish different subtypes of nAChRs. We report here the role of position 84 in the species specificity of cholinergic agonists. In particular, the addition of one methylene group to PIP-based cholinergic agonist (PropylPIP versus ButylPIP) confers selectivity for the rat versus the chick form of the 7 nAChR and that selectivity is based largely on the amino acid in position 84.

Role of Position 84 in ACh Binding Site. Considerable evidence suggests that the binding site for acetylcholine exists at the interface between two subunits. In the skeletal muscle form of the receptor, it is the interfaces between the  and  and the  and  subunits that are thought to be the sites of ACh binding. Four regions (labeled loops A, B, C, and D, see Fig. 4) have been identified, based on photoaffinity labeling and site-directed mutagenesis, as potential contributors to the ACh binding site (Galzi and Changeux, 1994; Arias, 1997). Loops A through C are present in the  subunit of all subtypes. In the homomeric 7 receptors loop D is also on the  subunit (presumably forming a binding site with loops A through C of the adjacent subunit), but this loop resides on the  and  subunits of the skeletal muscle nAChR (Corringer et al., 1995). These loops consist of highly conserved residues that are presumed to be important for proper folding and many of the interactions with agonist. However, species and subtype specificity would result from variable residues rather than highly conserved residues. Loop C contains the vicinyl disulfide bond that is characteristic of nAChR  subunits. Labeling with sulfhydryl affinity reagents suggests that the carbonyl half of ACh is in the vicinity of this loop (Kao et al., 1984). Likewise, cross-linking (Chiara and Cohen, 1997; Dennis et al., 1988) and mutagenesis (Corringer et al., 1995) studies indicate that the quaternary amine interacts with loop D, which is presumably located on an adjacent subunit. Both loops A and B have been labeled with photoaffinity reagents directed to the ACh binding site (Dennis et al., 1988; Galzi et al., 1990; Cohen et al., 1991) and have residues that when mutated affect ACh binding and channel activation (O'Leary et al., 1994; Sine et al., 1994). Position 84 is at the edge of Loop A, near residues labeled by p-(N,N-dimethylamino)benzenediazonium fluoroborate (DDF) (Dennis et al. 1988; Galzi et al., 1990) and acetylcholine mustard (Cohen et al., 1991). Although most studies are consistent with the notion that the quaternary ammonium group faces the D loop and the carbonyl side of the molecule faces the C loop, the exact orientation is not known. A homology model of the Torpedo electroplaque nAChR (Tsigelny et al., 1997) places position 84 at the periphery of the binding site, near the central pore. If one makes the assumption that the carbonyl side of the agonist faces the C loop, then this model places loop D too far from the PIP side chain to interact directly. However, small changes in a few selected backbone torsion angles in the model results in a binding site into which ButylPIP can be docked with the alkyl group near position 84. This would suggest that a direct interaction between ButylPIP and position 84 is a strong possibility. The fact that position 84 is in the vicinity of but not within a conserved loop is consist with the fact that a mutation in this position has small but significant effects on ACh binding but can affect the interactions of larger cholinergic agonists to a much greater extent.

Possible Interaction between PIP Side Chains and Position 84. When position 84 is a leucine, activation by ButylPIP is prevented, but DMPP can still activate the 7, although with lower efficacy. Considering all other neuronal  subunits, it is only those subunits that have a nonpolar amino acid (leucine or methionine) in this position that exhibit a low efficacy for DMPP; DMPP is a full agonist for receptors formed from  subunits with polar or charged amino acids in this position (Fig. 4). Considering the data for the PIP compounds, one possibility is that the additional hydrophobic surface area afforded by the longer alkyl chain in ButylPIP versus PropylPIP allows a hydrophobic interaction with the leucine in position 84. If this is true, then the interaction with L84 would either prevent the formation of the open channel form of the receptor or could potentially produce a much more rapid desensitization. A charged or polar group in this position or a shorter chain length would then prevent this interaction and allow channel opening (or decrease the rate of desensitization).

Implications for Development of Subunit-Specific Drugs. The fact that ButylPIP can clearly distinguish rat from chick 7 nAChR does not in itself demonstrate a subunit-specific drug. However, these two versions of 7 have the advantage of very high sequence identity (87%) and have led to the identification of position 84 as a crucial determinant of activation by ButylPIP. Considering first the homomeric nAChRs, only the 7 receptor from humans has been sequenced. However, in the chick, one could potentially develop a drug that would act as an agonist at the 8 homomeric receptor (Q in position 84) but not the 7 homomeric receptor (L in position 84). Perhaps, more importantly, given the prevalence of 7 and 42 receptors in the human brain, it may be possible to target a drug to 7 (Q in position 84) and not 42 (L in position 84 of the 4 subunit).

Summary. Using a series of systematically varying piperazium compounds and chick 7 nAChR, we identified a transition from full agonist activity (PropylPIP) to no agonist activity (ButylPIP) by the simple addition of one methylene group. In a closely related receptor, the rat 7 nAChR, both PropylPIP and ButylPIP are agonists. Using site-directed mutagenesis, we found that mutation of glutamine 84 to leucine in the rat 7 receptor abolished responses to ButylPIP but not PropylPIP. Likewise, mutating leucine 84 to glutamine in the chick 7 nAChR resulted in channel activation by ButylPIP. This is a clear interaction that structural modifications of an agonist can exploit partially conserved amino acid residues in the binding pocket to develop subunit and species specific cholinergic drugs.