Acetylcholine Receptor

Acetylcholine (ACh) is a widely distributed neurotransmitter in both the peripheral and central nervous systems (1). It is synthesized by the enzyme choline acetyltransferase within the cholinergic nerve terminal, where it is packaged into cholinergic vesicles. Arrival of an action potential at a cholinergic nerve terminal triggers quantal release of ACh by fusion of these vesicles with the presynaptic membrane and concomitant exocytosis of their contents into the synaptic cleft. The ACh released diffuses across the synapse to the postsynaptic membrane, where it activates its target, the acetylcholine receptor (AChR). Termination of transmission at cholinergic synapses involves rapid hydrolysis of ACh, by the enzyme acetylcholinesterase (AChE), which hydrolyzes it to choline and acetic acid (2).

The early observations of Dale (3) showed in various pharmacological preparations that ACh evoked responses similar to those evoked by either nicotine or muscarine. This provided the basis for the grouping of all AChRs into two families, nicotinic (nAChRs) and muscarinic (mAChRs). Dale's pioneering classification, based on the action of these two plant alkaloids, is still valid and used, although subtypes have been recognized and defined within each of the two families, both on the basis of specificity with respect to binding of agonists and antagonists and, in recent years, on the basis of gene cloning. Two additional plant alkaloids, d-tubocurarine and atropine, also provide useful tools by serving as specific antagonists of nAChRs and mAChRs, respectively (4).

nAChRs are ligand-gated ion channels, and their activation, by ACh or other agonists, causes a rapid change in ion permeability of the membrane in which they are embedded. mAChRs are members of the family of G-protein-coupled receptors and produce much slower responses, either excitatory or inhibitory, via their corresponding second messengers (5). Initial molecular cloning studies provided the primary structures of both nAChRs (6) and mAChRs (7), revealing that they belong to distinct families of proteins that share neither sequence identity nor a similar fold. As will be discussed in detail below, nAChRs are composed of pentamers of one or more subunits that display substantial sequence homology and, most likely, similar overall folds and transmembrane topologies. Muscarinic receptors are glycoproteins of molecular weight ~80,000 which, as already mentioned, belong to the family of G-protein-coupled receptors (8). Just as for other members of this family, hydrophobicity plots predict seven transmembrane sequences (8). The nAChR belongs to a larger family of ligand-gated ion channels, with which it too appears to share an overall fold, a common pattern of transmembrane sequences and other structural similarities (9, 10; see text below).

The recognition that the large electric organs of electric fish such as the electric eel, Electrophorus electricus, and the electric ray, Torpedo sp., provide a highly enriched and homogeneous preparation of cholinergic synapses (11) led to their use as an experimental model and also as the source of choice for purification and characterization of the nAChR, of AChE, and of homogeneous cholinergic synaptosomes and synaptic vesicles (2). The fact that the cholinergic synapse at the neuromuscular junction was the parallel system of choice for studying the electrophysiological characteristics of synaptic transmission (12), as well as many ultrastructural and developmental parameters (for recent reviews see Refs. 13-15), meant that many of the structural data garnered by use of the electric organ model system could be directly correlated with the functional data obtained with the neuromuscular preparation (2). Thus, the nAChR of electric organ closely resembles that of muscle (16), and the molecular forms of AChE present in electric organ tissue are homologous to those found at muscle endplates (17, 18).

The finding that injection into rabbits of purified nAChR elicited an autoimmune condition closely resembling the human muscle disease myasthenia gravis (19) was followed by detection of circulating antibodies to the nAChR in patients suffering from the disease. This opened up a fertile area of research in which clinical aspects were closely coupled to advances in basic research on the nAChR (20).

The cholinergic synapse has thus served as the prototypic synapse for understanding many fundamental aspects of synaptic transmission at chemical synapses. Even now, when improved electrophysiological techniques, and approaches such as Ca²⁺ imaging, allow better access to the CNS, and when genetic engineering allows expression of the receptor of choice in a system like the Xenopus oocyte that is amenable to user-friendly patch-clamp techniques, as well as in sufficient amounts to carry out structural characterization, the neuromuscular junction and electric organ tissue still occupy a position at the center of the stage.

 1. The Nicotinic Acetylcholine Receptor

1.1.  Identification and Characterization 

As early as 1955, it was suggested by Nachmansohn (21) that the nAChR might be a protein in which a conformational change elicited by the neurotransmitter ACh could induce a change in permeability that would trigger the postsynaptic response, and it was in his laboratory, in the 1960s, that the first evidence was presented that the nAChR is indeed a protein. Thus, Karlin and Bartels (22), using the electric eel electroplaque preparation, showed that the response of the nAChR to carbamylcholine could be modified reversibly by reagents that either modified thiol groups or reduced disulfide bonds. Subsequently, Karlin and co-workers (23, 24) and Changeux et al. (25), also in the Nachmansohn laboratory, demonstrated in situ affinity labeling of the nAChR in the intact electroplaque. O'Brien and co-workers (26, 27), using specific binding of the reversible ligand muscarone and homogenates of electric organ tissue of Torpedo, were able to obtain a good estimate ~2) nmol/g tissue) of the number of nicotinic binding sites.

A major breakthrough in the identification of the nAChR, its localization, and its molecular characterization came about by use of (a) the neurotoxin a-bungarotoxin (aBgt), a 74-residue polypeptide chain purified by Lee and co-workers from the venom of the banded krait, Bungarus multicinctus (28), and (b) the homologous a-neurotoxins purified from cobra venom (29). Lee's group presented electrophysiological and pharmacological evidence that these neurotoxins acted as very high-affinity (dissociation constant in the pico-to-femtomolar range) antagonists of the nAChR (30). 125¹²⁵I-aBgt was developed as a powerful tool for localization and quantification of the nAChR by high-resolution autoradiography (31). It was shown to be present as a dense quasi-crystalline array at the top of the folds of the postsynaptic membrane of skeletal muscle, opposite the putative ACh-release sites on the presynaptic membrane, with a density very similar to that in purified preparations of postsynaptic membranes obtained from Torpedo electroplax. A further step forward was taken when it was shown that binding of ¹²⁵I-aBgt was retained after solubilization in nonionic detergents, such as Triton X-100 and cholate, and that a single toxin-binding component, migrating at ~9 S, could be identified on sucrose gradient sedimentation velocity centrifugation (32, 33). Furthermore, the receptor so solubilized could be purified by affinity chromatography using resins to which either the a-neurotoxin (34) or a quaternary nitrogen ligand (35) had been attached. This permitted characterization of the nAChR as a multisubunit protein. Thus it was shown that it was a pentamer, of molecular weight ~250,000 (36), that contains four polypeptide subunits, a, b, g, and d, with apparent molecular weights of 39,000, 48,000, 58,000, and 64,000 and a stoichiometry of a2bgd. In the electron microscope, such purified preparations displayed a rosette-like appearance (37), closely resembling similar, highly organized arrays of rosettes observed in purified postsynaptic membrane preparations (38). This led to the working hypothesis, still currently accepted, that the five subunits lie in the plane of the plasma membrane, surrounding the ion channel running through the center. A fifth polypeptide, the 43-kDa polypeptide (now known as rapsyn), which was found to be weakly associated with the pentamer (39), is believed to be involved in nAChR-clustering and anchoring to the cytoskeleton at the synapse (13). In skeletal muscle the g subunit is present in the nAChR of embryonic muscle, but is replaced by the  subunit in adult muscle (40).

Studies on model systems were important in establishing the role of the nAChR in signal transduction at the cholinergic synapse. Thus the microsac preparation (41) was used to show that ACh could induce a permeability change in such nAChR-enriched vesicles. Furthermore, prolonged exposure to high concentrations of ACh was shown to produce a closed state, with high affinity for ACh (42), equivalent to the desensitized state first described in skeletal muscle by Katz and Thesleff (43) .The availability of purified native receptor permitted reconstitution studies into liposomes and into lipid bilayers, permitting recording of single channels induced in the presence of ACh (44, 45).

The purified pentamer thus contained not only the ACh-binding site, but also the ion channel and the transducing elements involved in activation and desensitization (16).

Affinity labeling studies by Karlin and co-workers (46), using a radioactive reagent, established that the disulfide bond that was susceptible to reducing reagents, as originally demonstrated by Karlin and Bartels (22), was located on the a subunit, in proximity to the binding site for ACh and various quaternary ligands.

An important step forward was the finding of Raftery et al. (47) that the NH2-terminal sequences of all four subunits display substantial sequence homology. When, not long after, the subunits were cloned (6), it was found that such homology extends throughout the whole polypeptide chain, and the receptor pentamer can be viewed as displaying pseudo-fivefold symmetry. One issue that still remains open is the arrangement of the subunits around the lumen. This has been approached primarily by electron microscopy using subunit-specific antibodies, with supplementary information coming from use of affinity labels directed toward the ACh-binding site, which label both the a subunit and an adjacent subunit (see text below). As many as 12 permutations are possible, but it is generally accepted that the two a-subunits are not adjacent to each other (48), and discussion focused on which of the b-, g-, and d-subunits is flanked on both sides by the a-subunit. Karlin et al. (49) proposed that it is the g-subunit that lies between the two a-subunits. Although Kubalek et al. (50( proposed that the b-subunit might occupy this position, the assignment of Karlin and co-workers is generally favored (for detailed discussions see Refs. 10 and 51).

Advances in cloning and expression, taken together with development of the patch-clamp technique, had a dramatic influence on research on the nAChR, just as they did on research on other receptors and on ion channels. In the case of the nAChR, however, the availability of large amounts of highly purified receptor from Torpedo electric organ resulted in fruitful synergy between these new techniques and the techniques of protein chemistry and structural biology.

In general, research has proceeded on two fronts, one concerned with the structure of the ACh-binding site and the other with that of the ion channel, with the eventual objective of understanding the physical mechanism by which ligand-binding causes channel opening. In the following, the overall topology of the receptor and of the individual subunits will first be reviewed briefly. The current status of our knowledge of the ACh-binding site, and then of the ion channel, will be summarized, followed by a discussion of the recent structural work from the laboratory of Unwin, which is beginning to give us a first glimpse of how the receptor may be functioning. For a number of recent reviews that cover these issues in more detail than is possible here, see Refs. 9, 10, 16, and 52, as well as the recent paper of Unwin (53), which summarizes the current status of the structural work.

1.2. Conformation and Topology 

Analysis of the sequences of the four subunits, obtained on the basis of their complementary DNA sequences, showed that they shared a very similar topology: A long extracellular NH2-terminal domain is followed by three putative transmembrane a-helices, identified on the basis of hydrophobicity, M1–M3, by a cytoplasmic loop containing ~100–150 residues, and finally, by a fourth transmembrane helix, M4, so that the COOH-terminus is believed to be on the extracellular side of the postsynaptic membrane (54). Other ligand-gated ion channels that belong to the same superfamily, including receptors for g-aminobutyric acid, glycine, and serotonin, display a similar topology (9, 10). A number of consensus sites for N-Glycosylation are present in the extracellular domains, giving rise to a sugar content of ~7% (for literature see Ref. 55); and a number of consensus sites for phosphorylation, present on the cytoplasmic loops, presumably fulfill regulatory functions (56).

Photoaffinity labeling, using lipid-soluble probes, has been used to assign the arrangement of the transmembrane sequences in relation to the lipid bilayer and the putative central ion channel (57-59). From such studies, it has been concluded that the M4 sequence, which is also the least conserved, is the most exposed to the lipid environment, with the M1 and M3 sequences displaying more limited exposure to the lipid environment, and M2 facing the lumen of the ion channel. This assignment of M2 as making the principal contribution to the lining of the ion channel is also supported by labeling by channel blockers and by use of the SCAM technique (see text below).

Various spectroscopic techniques show that the nAChR contains substantial amounts of both a-helices and b-sheets (for literature see Ref. 10). In particular, Naumann et al. (60), using Fourier transform infrared (FTIR) spectroscopy on native receptor-rich membranes, have shown a predominance of b structure, 36 to 43%, and an a-helical content of 32 to 33%. Of particular interest are the putative conformations of the membrane-spanning sequences. Görne-Tschelnokow et al. (61) attacked this problem directly, by performing FTIR spectroscopy on membrane preparations from which the extramembrane domains had been shaved by proteolysis. Their data suggested a beta-sheet content of ~40% for the transmembrane sequences. Modeling studies also suggested a substantial b-sheet component (62). Various labeling studies, which attempt to correlate the degree of labeling of residues with orientation toward either the ion-channel lumen or the lipid bilayer, and thus with either an a-helical or a b-strand pitch, yield complex results (see, for example, Refs. 9, 10, 59, and 63); and precise assignments will, most likely, have to await a high-resolution 3D protein structure.

1.3.  ACh-Binding Site 

Identification of the residues comprising the ACh-binding site of the nAChR was approached by affinity labeling. Karlin and co-workers used the same radioactive affinity label that they had used to label the a subunit selectively, a quaternary derivative of maleimide, to identify the residues labeled. Because they had demonstrated that labeling occurred subsequent to reduction of the receptor, it was not surprising that they found that the residues labeled were the adjacent cysteines, Cys192 and Cys193, which apparently form an intrachain disulfide bond in the native nAChR, and which are present only in the a-subunit (64). As predicted, these residues are localized in the sequence that had been assigned to the extracellular domain, and would thus be the natural candidate for the ACh-binding site. The importance of this region for agonist and antagonist binding was confirmed by the observation that short synthetic peptides containing Cys192 and Cys193 (eg, from residue 185 to 196) displayed specific binding to aBgt, albeit with much lower affinity than the native receptor (65). Moreover, the a-subunits of the nAChR of both snakes and of the mongoose, which are resistant to aBgt, display mutations in this region that can provide a structural basis for resistance (66).

 Changeux and co-workers conducted an extensive study of the residues labeled by the tertiary photoaffinity label, DDF (67). In addition to Cys192 and Cys193, this probe labeled primarily aromatic residues, including Tyr93, Trp149, Tyr190 and Tyr198, all, again, in the putative cytoplasmic domain. Use of other labeling agents, by other laboratories, broadly confirmed these assignments and identified additional aromatic residues (for literature see Ref. 68). Thus the ACh-binding pocket of the nAChR can be viewed as an “aromatic basket” (16) in which the quaternary group of ACh interacts with these aromatics via the p electron–cation interactions (69) which X-ray crystallographic studies have shown to play a prominent role within the “aromatic gorge” of AChE (70, 71).

Although the studies discussed above indicate a prominent role for the a-subunit in binding of both agonists and antagonists, numerous labeling studies have revealed contributions of adjacent subunits to ligand-binding. These studies have been critically reviewed by Hucho et al. (10). Taken together, the data clearly indicate that the two ACh-binding sites are at the interfaces between the two a-subunits and the g- and d-subunits. This, in turn, implies structural difference between the two sites that can serve to explain nonequivalence of the two binding sites in terms of affinity for both a-neurotoxins (72) and antagonists (73).

1.4. Ion Channel 

Pharmacological studies on the nAChR revealed a number of noncompetitive antagonists, both natural and synthetic, that blocked the change in permeability elicited by ACh without inhibiting the binding of ACh itself. It was, therefore, suggested that at least some of these compounds act by entering the ion channel and sterically blocking ion movement through it (for literature see Refs. 9, 10, and 16). Based on the currently held view that the ion channel is the hole in the rosette seen in the electron microscope, such “channel blockers” are believed to exert their action by plugging this hole. Some channel blockers, such as chlorpromazine (74) and trimethylphosphonium (75), bind irreversibly upon UV irradiation; and their labeled derivatives can, accordingly, be used to identify putative elements of the channel lining. Rapid-mixing studies, using nAChR-enriched microsacs, showed that exposure to ACh leads, initially, to greatly increased labeling, followed by diminished incorporation which is consistent with desensitization as measured by ion flux measurements (76). Such labeling studies showed dominant labeling of the residues in the M2 transmembrane sequences of all four subunits, suggesting that they make the principal contribution to the lining of the lumen of the ion channel, in the open state of the channel. One such noncompetitive channel blocker, quinacrine azide, was, however, shown to label amino acid residues in M1 (77).

In parallel to such labeling studies on the purified receptor, use of site-directed mutagenesis, combined with expression in Xenopus oocytes (78) and use of the patch-clamp technique, permitted assessment of the contributions of individual amino acid residues to the conductance properties of the ion channel, an approach pioneered by the joint efforts of the Numa and Sakmann laboratories (see, for example, Refs. 79 and 80).

 A third approach, developed more recently, is the substituted-cysteine-accessibility method, abbreviated as SCAM (81), which combines site-directed mutagenesis with chemical modification, so as to identify the amino acid residues lining the ion channel and their involvement in function. Thus, residues believed to line the ion channel (primarily from the M2 transmembrane sequence, but also from M1) are replaced one-by-one by cysteine residues, using site-directed mutagenesis, and expressed in Xenopus oocytes. The reactivity of their thiol groups with thiol reagents bearing negative or positive charges is examined in situ, both in the presence and absence of ACh, thus permitting assessment of their accessibility in both the closed and open states of the channel. Furthermore, these sulfhydryl reagents can be added both intra- and extracellularly, yielding information concerning accessibility of channel-lining residues from both surfaces (63).

These various experimental approaches have led to the identification of residues within the M2 sequences of all five subunits that determine the conductance and selectivity of the nAChR channel. In particular, using nAChR mutants expressed in Xenopus oocytes, three rings of negatively charged residues which may be referred to as the extracellular (or outer), the intermediate, and the cytoplasmic (or inner) ring, have been shown to play important roles in determining channel conductance (80). A ring of polar (serine and threonine) residues, named the central ring, is located above the intermediate ring, forming a constriction that may serve as part of the selectivity filter (82,( 83 . The bulk of the transmembrane sequence of M2 lies between this central ring and the outer ring.

A more detailed topographical picture has been obtained by the SCAM technique, as summarized by Karlin and Akabas (9). Thus, broadly speaking, the residues in M2 of the a-subunit can be fitted to an Amphipathic helix so that, with one exception, all the residues on one face of the helix are labeled in the SCAM protocol; some are equally accessible in the presence and absence of ACh, some are more accessible in its presence, and some are more accessible in its absence. Furthermore, certain residues at the extracellular end of the M1 helix of the a-subunit are also labeled. The pattern of labeling of M2 in the absence of ACh is consistent with an a-helical conformation, except for a short stretch in the middle (aLeu250 to aSer252), while in the presence of ACh it is consistent with an uninterrupted a-helix. The pattern of labeling of M1 cannot be ascribed to a specific conformational motif, although it too is changed in the presence of ACh. Karlin and Akabas (9) hypothesize that a movement of M1 and M2 relative to each other, with a concomitant change in secondary structure, may flip open the gate of the channel. More recently, the SCAM technique has also located the gate, using comparative studies in which labeling was carried out from either the extracellular or cytoplasmic sides (63). In the absence of ACh—that is, in the closed state of the channel—there is a barrier to the sulfhydryl agents, when added to either side, reacting with residues aGly240 and to aThr244. ACh binding removes this barrier, which serves as an activation gate. Residues aGly240, aGlu241, aLys242, and aThr244 line a narrow part of the channel in which this gate is located. It should be noted that the second of these residues, aGlu241, belongs to the inner anionic ring, and aThr244 to the central ring, as defined above (80, 82).

1.5.  Structural Studies 

Despite the extensive studies performed on the characterization of the ACh-binding site and of the ion channel, the question with which all structural and functional studies on the nAChR are ultimately concerned is how binding of ACh is transduced into a permeability change in the ion channel. This is ultimately a problem best addressed by a structural biology approach. In spite of attempts from many laboratories to obtain nAChR crystals that would diffract X-rays (for a published example, see Ref. 84), this has not yet been achieved. Our current structural knowledge is, therefore, derived from the work of Unwin (for a recent summary see Ref. 53). As mentioned above, the postsynaptic membrane of Torpedo electric organ contains densely packed, partially crystalline arrays of nAChR (38). When isolated, such membranes have a natural propensity to recrystallize in tubular form (85, 86). Since the early 1980s, Unwin has been using electron microscopy to probe the structure of the nAChR from Torpedo marmorata and is now able to compare the receptor in its open and closed states at 9 Å resolution (53). The receptor is revealed as an elongated structure, ~125 Å long; the extracellular portion of each subunit extends ~60 Å above the membrane surface, and the intracellular portions extend ~20 Å from the cytoplasmic surface. A density underlying each receptor most likely represents the cytoskeletal protein rapsyn (see text above), normally present in 1:1 stoichiometry with the receptor (87). A view from above shows the five subunits arranged around a fivefold axis of pseudosymmetry, as predicted. The opening of the putative ion channel in the center is about 20 Å at its entrance, narrowing sharply at the membrane surface. At the current resolution, it is possible to begin to glimpse some elements of secondary structure. Thus in each subunit, in a region about 30 Å away from the extracellular surface, a group of three short rods can be detected, presumably a-helices. The two subunits identified as a-subunits (50) contain cavities shaped by these rods, which may correspond to the ACh-binding pockets. The narrowest portion, corresponding to the pore, appears to be shaped by five bent a-helical rods, each contributed by one of the subunits (Fig. 1, left). The bends in the rods, near the middle of the membrane, are the parts closest to the axis of the pore and may represent the gate. In order to visualize the open state of the nAChR, Berriman and Unwin (88) devised a rapid-freezing device, designed to trap the open state, after application of ACh, by preventing the transition to the desensitized state. Comparison of the ACh-activated receptor with the nonactivated form revealed only slight changes (Fig. 1, right).  Concerted twisting motions appeared to occur within the rods lining the putative ACh-binding pockets, which were consistent with the requirement that two ACh molecules, binding simultaneously to both a-subunits, are required to open the channel. These localized disturbances were associated with small rotations extending along the axis of all the subunits, and with a switching of the helices within the pore from the bent shape seen in the nonactivated receptor to a tapered, more open configuration. If it is assumed that the threonine side chains of the central ring, which are believed to serve as the pore (see text above, along with Ref. 83) face into the pore, a minimum diameter of 10 Å can be estimated, comparable to the diameter obtained by use of various organic cations (89). Such a concerted rotation of the transmembrane sequences surrounding the lumen of the putative ion channel is consistent with the data accumulated by the various studies using site-directed mutagenesis and labeling, some of which were discussed above (see, in particular, Refs. 59 and 63).

Figure 1. Simplified diagram of the nAChR in its closed (left) and open (right) states, as suggested by the structural results. An ACh molecule first enters a binding site in one of the a-subunits (rectangle), but significant displacements are blocked by the neighboring subunit, lying between the two a-subunits. Interaction of a second ACh with the other site then attempts to draw the neighboring subunit out of the way. A concerted localized displacement thereby takes place, initiating small rotations of the subunits along the shaft to the membrane. The rotations disrupt the gate by disrupting the association of a-helices around the pore in the closed state (left), switching over to an alternate configuration in which a widened polar pathway for ion movement is created. (From Ref. 53, with permission) .

Obviously, when higher-resolution images become available, a more detailed mechanistic description will be feasible, but the conceptual stage appears to have been set.

1.6. Neuronal Nicotinic Receptors 

Already in the late 1970s, it was proposed, on the basis of 125I-aBgt-binding studies, that nAChRs were present not only in skeletal muscle, but also in the central nervous system (90, 91). These studies were not the subject of widespread attention, primarily due to the fact that functional correlates were lacking. When, in the mid-1980s, cloning techniques revealed the presence in brain of an nAChR gene family homologous to, but clearly distinct from, those of electric organ and muscle, and widely distributed in different brain areas (92, 93), it became obvious that nAChRs, like muscarinic receptors, must have important functions in brain. For many years, however, the lack of specific agonists and antagonists for a given neuronal nAChR subtype severely hampered the identification of functional nAChRs in various brain areas. These difficulties were further aggravated by the unusually fast kinetics of inactivation of some nAChR subtypes, compared to their peripheral counterpart (for literature, see Ref. 94). Development of techniques permitting rapid application and removal of agonists was necessary to overcome this limitation (see, for example, Refs. 95 and 96).

The cloning approach revealed the existence of a large number of AChR receptor subunits in the brain. In contrast to muscle and electric organ, however, these appear to fall into only two categories, a and b. By now nine a-subunits (97) and nine b-subunits have been described in vertebrate brain (98), and multiple nAChR subunits have also been described in invertebrates such as Caenorhabditis elegans (99) and Drosophila (100).

As discussed by Sargent (98), the various subtypes of subunits are expressed differentially in one brain region or another. The most recently discovered a9-subunit, for example, has a pattern of expression restricted to cochlear hair cells (97). The pharmacology of neuronal nicotinic receptors in relation to such topics as nicotine addiction (101) and Alzheimer's disease (102) are currently topics of intensive investigation; so too is their involvement in brain plasticity (103) and in behavior, learning, and memory (104, 105). A recent development is the observation that choline serves as a selective agonist for a7 nAChRs in rat hippocampal neurons (106), in line with earlier reports for a7 ectopically expressed in oocytes (107, 108). Because a7 belongs to what appears to be the evolutionarily oldest group of nAChrs (109), it is feasible that choline, rather than ACh, was the primeval transmitter for cholinergic receptors.

Neuronal nAChRs have not yet been purified or expressed in large amounts. Thus, essentially all our knowledge of their functional properties comes from studies in which subunits were expressed singly or in combination, most frequently by injection into Xenopus oocytes. Such expression studies were augmented, as for Torpedo and muscle nAChRs, by site-directed mutagenesis. These studies have been reviewed in detail by Sargent (98), and only a few points will be discussed here briefly.

Presence of an a-subunit is necessary to obtain a functional receptor; and in the case of the a7-subunit, injection of it alone into oocytes is sufficient for it to form functional oligomers (110). Changeux, Bertrand, and co-workers took advantage of this experimental system to perform expression, combined with site-directed mutagenesis, which showed that the ligand-gated ion channel produced by the a7-subunit closely resembles that of the pentameric muscle and Torpedo receptors (see, for example, Ref. 111 and the review in Ref. 16). Direct experimental evidence for a pentameric stoichiometry was provided for the chick nAChR produced in oocytes by expression of the a4 and b2 subunits with radioactive label incorporated into them. The stoichiometry of the subunits in the purified receptor was consistent with an a4₂b2₃ pentamer (112). This does not mean, however, that expression of the mRNAs for any mixture of a-subunits, with or without a b-subunit, will produce a functional oligomer. Thus far, only a7 has been shown to produce functional oligomers containing only a subunits. Furthermore, a recent study of Yu and Role (113) has shown that a5 can participate in the function of an ACh-gated ion channel only if it is coexpressed with another a- and a b-subunit. The structural basis for such restrictions with respect to assembly and/or function are, at this stage, unknown. Patrick and co-workers (114) observed, however, that a Cyclophilin is required for expression of functional homo-oligomeric, but not hetero-oligomeric, nAChRs. They raise the possibility that the a-subunits in the homo-pentamer may not assume identical folds. Cyclophilins may thus play a critical role in the maturation of such homo-oligomers, acting directly or indirectly as prolyl cis/trans Isomerases or as molecular chaperones (115).

References

1. P. Taylor (1990) In The Pharmacological Basis of Therapeutics (A. G. Gilman, T. W. Rall, A. S. Nies, and P. Taylor, eds.), 8th Edition, Pergamon Press, New York, pp. 122–130. 

2. V. P. Whittaker (1992) The Cholinergic Neuron and its Target: The Electromotor Innervation of the Electric Ray Torpedo as a Model, Birkhäuser, Boston. 

3. H. H. Dale (1914) J. Pharmacol. Exp. Ther. 6, 146–190. 

4.R. J. Lefkowitz, B. B. Hoffman, and P. Taylor (1990) In The Pharmacological Basis of Therapeutics (A. G. Gilman, T. W. Rall, A. S. Nies, and P. Taylor, eds.), 8th ed., Pergamon Press, New York, pp. 84–121. 

5. D. A. Brown et al. (1997) Life Sci. 60, 1137–1144. 

6. S. Numa et al. (1983) Cold Spring Harbor Symp. Quant. Biol. 48, 57–69. 

7. T. Kubo et al. (1986) Nature 323, 411–416. 

8. J. Wess et al. (1997) Life Sci. 60, 1007–1014. 

9. A. Karlin and M. H. Akabas (1995) Neuron 15, 1231–1244. 

10. F. Hucho, V. I. Tsetlin, and J. Machold (1996) Eur. J. Biochem. 239, 539–557. 

11. D. Nachmansohn (1959) Chemical and Molecular Basis of Nerve Activity, Academic Press, New York. 

12. B. Katz (1966) Nerve, Muscle and Synapse, McGraw-Hill, New York. 

13. S. C. Froehner (1993) Annu. Rev. Neurosci. 16, 347–368. 

14. Z.-Z. Wang et al. (1996) Cold Spring Harbor Symp. Quant. Biol. 61, 363–371. 

15. J. R. Sanes (1997) Curr. Opin. Neurobiol. 7, 93–100. 

16. J.-P. Changeux (1995) Biochem. Soc. Trans. 23, 195–205.

17. I. Silman and A. H. Futerman (1987) Eur. J. Biochem. 170, 11–22. 

18. J. Massoulié et al. (1993) Prog. Neurobiol. 41, 31–91. 

19. J. Patrick and J. Lindstrom (1973) Science 180, 871–872. 

20. A. Vincent et al. (1995) J. Physiol. Paris 89, 129–136. 

21. D. Nachmansohn (1955) Harvey Lect. 39, 57–99. 

22. A. Karlin and E. Bartels (1966) Biochim. Biophys. Acta 126, 525–535. 

23. A. Karlin and M. Winnik (1968) Proc. Natl. Acad. Sci. USA 60, 668–674. 

24. I. Silman and A. Karlin (1968) Science 164, 1420–1421. 

25. J.-P. Changeux, T. Podleski, and L. Wofsy (1967) Proc. Natl. Acad. Sci. USA 58, 2063–2070. 

26. R. D. O''Brien and L. P. Gilmour (1969) Proc. Natl. Acad. Sci. USA 63, 496–503. 

27. R. D. O''Brien, L. P. Gilmour, and M. E. Eldefrawi (1970) Proc. Natl. Acad. Sci. USA 65438–445,  .

28. C. C. Chang and C. Y. Lee (1963) Arch. Int. Pharmacodyn. 144, 241–257. 

29. C. C. Chang and C. Y. Lee (1966) Brit. J. Pharmacol. 28, 172–181. 

30. C. Y. Lee (1972) Annu. Rev. Pharmacol. 12, 265–286. 

31. H. C. Fertuck and M. M. Salpeter (1976) J. Cell Biol. 69, 144–158. 

32. J.-P. Changeux, M. Kasai, and C. Y. Lee (1970) Proc. Natl. Acad. Sci. USA 67, 1241–1247. 

33.  R. Miledi, P. Molinoff, and L. T. Potter (1971) Nature 229, 554–557.

34. E. Karlsson, E. Heilbronn, and L. Widlund (1972) FEBS Lett. 28, 107–111. 

35. A. Karlin and D. A. Cowburn (1973) Proc. Natl. Acad. Sci. USA 70, 3636–3640. 

36. J. A. Reynolds and A. Karlin (1978) Biochemistry 17, 2035–2038. 

37. J. Cartaud et al. (1973) FEBS Lett. 33, 109–113. 

38. J. E. Heuser and S. R. Salpeter (1979) J. Cell Biol. 82, 150–173. 

39. R. R. Neubig, E. K. Krodel, N. D. Boyd, and J. B. Cohen (1979) Proc. Natl. Acad. Sci. USA.  690–76,694 

40. M. Mishina et al. (1986) Nature 321, 406–411. 

41. M. Kasai and J. P. Changeux (1971) J. Membr. Biol. 6, 1–80. 

42. M. Weber, T. David-Pfeuty, and J. P. Changeux (1975) Proc. Natl. Acad. Sci. USA 72, 3443-3447 .

43. B. Katz and S. Thesleff (1957) J. Physiol. 138, 63–80. 

44. N. Nelson, R. Anholt, J. Lindstrom, and M. Montal (1980) Proc. Natl. Acad. Sci. USA 77. 3057–3061,

45. G. Boheim et al. (1981) Proc. Natl. Acad. Sci. USA 78, 3586–3590. 

46. C. L. Weill, M. G. McNamee, and A. Karlin (1974) Biochem. Biophys. Res. Commun. 61,. 997-1003

47. M. A. Raftery, M. W. Hunkapiller, C. D. Strader, and L. E. Hood (1980) Science 208, 14541456- .

48. D. S. Wise, J. Wall, and A. Karlin (1981) J. Biol. Chem. 256, 12624–12627. 

49. A. Karlin et al. (1983) J. Biol. Chem. 258, 6678–6681. 

50. E. Kubalek, S. Ralston, J. Lindstrom, and N. Unwin (1987) J. Cell Biol. 105, 9–18. 

51. P. C. Kearney et al. (1996) Neuron 17, 1221–1229. 

52. H. A. Lester (1997) Harvey Lect., 91, 79–98. 

53. N. Unwin (1998) J. Struct. Biol. 121, 181–190. 

54. J.-L. Popot and J.-P. Changeux (1984) Physiol. Rev. 64, 1162–1239. 

55. H. Shoji et al. (1992) Eur. J. Biochem. 207, 631–641. 

56. R. L. Huganir and P. Greengard (1987) Trends Pharmacol. Sci. 8, 472–477.

57. J. Giraudat, C. Montecucco, R. Bisson, and J.-P. Changeux (1985) Biochemistry 24, 31213127– .

58. M. P. Blanton and J. B. Cohen (1994) Biochemistry 33, 2859–2872. 

59. M. P. Blanton et al. (1998) J. Biol. Chem. 273, 8659–8668. 

60. D. Naumann, C. Schultz, U. Görne-Tschelnokow, and F. Hucho (1993) Biochemistry 32 . 3162–3168,

61. U. Görne-Tschelnokow et al. (1994) EMBO J. 13, 338–341. 

62. M. O. Ortells and G. G. Lunt (1996) Prot. Eng. 9, 51–59. 

63.  G. G. Wilson and A. Karlin (1998) Neuron 20, 1269–1281. 

64. P. N. Kao et al. (1984) J. Biol. Chem. 259, 11662–11665. 

65. D. Neumann, D. Barchan, M. Fridkin, and S. Fuchs (1986) Proc. Natl. Acad. Sci. USA 83. 9250–9253 ,

66. D. Barchan, M. Ovadia, E. Kochva, and S. Fuchs (1995) Biochemistry 34, 9172–9176. 

67. J. Langenbuch-Cachat et al. (1988) Biochemistry 27, 2337–2345. 

68. J.-P. Changeux, J.-L. Galzi, A. Devillers-Thiéry, and D. Bertrand (1992) Q. Rev. Biophys. 25. 395–432 ,

69. D. A. Dougherty and D. A. Stauffer (1990) Science 250, 1558–1560. 

70. M. Harel et al. (1993) Proc. Natl. Acad. Sci. USA 90, 9031–9035. 

71. M. Harel et al. (1996) J. Am. Chem. Soc. 118, 2340–2346. 

72. A. Maelicke and E. Reich (1976) Cold Spring Harbor Symp. Quant. Biol. 40, 231–235. 

73. S. E. Pedersen and J. B. Cohen (1990) Proc. Natl. Acad. Sci. USA 87, 2785–2789. 

74. T. Heidmann, R. E. Oswald, and J. P. Changeux (1983) Biochemistry 22, 3112–3127. 

75. F. Hucho, W. Oberthür, and F. Lottspeich (1986) FEBS Lett. 205, 137–142. 

76 .T. Heidmann, J. Bernhardt, E. Neumann, and J. P. Changeux (1983b) Biochemistry 22, 54525459 – .

77. M. DiPaola, P. N. Kao, and A. Karlin (1990) J. Biol. Chem. 265, 11017–11029. 

78. E. A. Barnard, R. Miledi, and K. Sumikawa (1982) Proc. R. Soc. Lond. B 215, 241–246. 

79. K. Imoto et al. (1986) Nature 324, 670–674. 

80. K. Imoto et al. (1988) Nature 335, 645–648. 

81. M. H. Akabas, D. A. Stauffer, M. Xu, and A. Karlin (1992) Science 258, 307–310. 

82. K. Imoto et al. (1991) FEBS Lett. 289, 193–200. 

83. A. Villarroel, S. Herlitze, M. Koenen, and B. Sakmann (1991) Proc. R. Soc. B 243, 69–74. 

84. S. Hertling-Jaweed et al. (1988) FEBS Lett. 241, 29–32. 

85.  J. Kistler and R. M. Stroud (1981) Proc. Natl. Acad. Sci. USA 78, 3678–3682. 

86. A. Brisson and P. N. T. Unwin (1984) J. Cell Biol. 99, 1202–1211. 

87.  W. J. LaRochelle and S. C. Froehner (1986) J. Biol. Chem. 261, 5270–5274. 

88. J. A. Berriman and N. Unwin (1994) Ultramicroscopy 56, 241–252. 

89. B. N. Cohen, C. Labarca, N. Davidson, and H. A. Lester (1992) J. Gen. Physiol. 100, 373–400.

90. Z. Vogel and M. Nirenberg (1976) Proc. Natl. Acad. Sci. USA 73, 1806–1810. 

91. J. Patrick and W. B. Stallcup (1977) Proc. Natl. Acad. Sci. USA 74, 4689–4692. 

92. D. Goldman et al. (1987) Cell 48, 965–973. 

93. S. Heinemann et al. (1990) Prog. Brain Res. 86, 195–203. 

94. E. X. Albuquerque et al. (1997) J. Pharmacol. Exp. Ther. 280, 1117–1136. 

95. N. G. Castro and E. X. Albuquerque (1993) Neurosci. Lett. 164, 137–140. 

96. Z.-W. Zhang, S. Vijarayaghavan, and D. K. Berg (1994) Neuron 12, 167–177.

97. A. B. Elgoyhen et al. (1994) Cell 79, 705–715. 

98. P. B. Sargent (1993) Annu. Rev. Neurosci. 16, 403–443. 

99. J. T. Fleming et al. (1997) J. Neurosci. 17, 5843–5857. 

100. E. D. Gundelfinger (1992) Trends Neurosci. 15, 206–211. 

101. J. A. Dani and S. Heinemann (1996) Neuron 16, 905–908. 

102. H. Schröder et al. (1991) Neurobiol. Aging 12, 259–262. 

103. J.-P. Changeux et al. (1996) Cold Spring Harbor Symp. Quant. Biol. 61, 343–362. 

104.  N. Le Novère, M. Zoli, and J. P. Changeux (1996) Eur. J. Neurosci. 8, 2428–2439. 

105. E. D. Levin et al. (1997) Brain Res. Bull. 43, 295–304. 

106. M. Alkondon et al. (1997) Eur. J. Neurosci. 9, 2734–2742. 

107. A. Mandelzys, P. De Koninck, and E. Cooper (1995) J. Neurophysiol. 74, 1212–1221. 

108. R. L. Papke, M. Bencherif, and P. Lippiello (1996) Neurosci. Lett. 213, 201–204. 

109. N. Le Novère and J. P. Changeux (1995) J. Mol. Evol. 40, 155–172.

110. S. Couturier et al. (1990) Neuron 5, 847–856. 

111. D. Bertrand et al. (1993) Proc. Natl. Acad. Sci. USA 90, 6971–6975.

112. R. Anand et al. (1991) J. Biol. Chem. 266, 11192–11198. 

113. C. R. Yu and L. W. Role (1998) J. Physiol. 509, 667–681. 

114. S. A. Helekar, D. Char, S. Neff, and J. Patrick (1994) Neuron 12, 179–189. 

115. S. A. Helekar and J. Patrick (1997) Proc. Natl. Acad. Sci. USA 94, 5432–5437.