Structural organization of nicotinic acetylcholine receptors

Nicotinic acetylcholine receptor of the electric ray Torpedo is the most comprehensively characterized neurotransmitter receptor. It consists of five subunits (alpha2beta gammadelta) amino acid sequences of which were determined by cDNA cloning and sequencing. The shape and size of the receptor were determined by electron cryomicroscopy. It has two agonist/competitive antagonist binding sites which are located between subunits near the membrane surface. The receptor ion channel is formed by five transmembrane helices (M2) of all five subunits. The position of the binding site for noncompetitive ion channel blockers was found by photoaffinity labelling and site-directed mutagenesis. The intrinsic feature of the receptor structure is the position of the agonist/competitive antagonist binding sites in close vicinity to the ion channel spanning the bilayer membrane. This peculiarity may substantially enhance allosteric transitions transforming the ligand binding into the channel opening and physiological response. Muscle nicotinic acetylcholine receptors from birds and mammals are also pentaoligomers consisting of four different subunits (alpha2beta gammadelta or alpha2beta epsilondelta) with high homology to the Torpedo receptor. Apparently, the pentaoligomeric structure is the main feature of all nicotinic, both muscle and neuronal, receptors. However, the neuronal receptors are formed only by two subunit types (alpha and beta) or are even pentahomomers (alpha7 neuronal receptors). All nicotinic receptors are ligand-gated ion channel, the properties of the channels being essentially determined by amino acid residues forming M2 transmembrane fragments.     

Neuronal nicotinic acetylcholine receptors expressed in Xenopus oocytes have a pentameric quaternary structure

We have determined the subunit stoichiometry of chicken neuronal nicotinic acetylcholine receptors expressed in Xenopus oocytes by quantitation of the amount of radioactivity in individual subunits of [35S] methionine-labeled receptors. The chicken neuronal nicotinic acetylcholine receptor appears to be a pentamer of two alpha 4 acetylcholine-binding subunits and three beta 2 structural subunits. We also show that these expressed receptors bind L-[3H]nicotine with high affinity, are transported to the surface of the oocyte outer membrane, and cosediment on sucrose gradients with acetylcholine receptors isolated from chicken brain. Using this unique and generally applicable method of determining subunit stoichiometry of receptors expressed in oocytes, we obtained the expected (alpha 1) 2 beta 1 gamma delta stoichiometry for muscle-type acetylcholine receptors assembled from coexpression of either Torpedo alpha 1 or human alpha 1 subunits, with Torpedo beta 1, gamma, and delta subunits.     

Purification of acetylcholine receptors from the muscle of Electrophorus electricus

Muscle from the electric eel Electrophorus electricus contains acetylcholine receptors at 50 times the concentration of normal mammalian muscle and fully one-tenth the concentration of receptors in its electric organ tissue. Receptor is organized much more diffusely over the surface of Electrophorus muscle cells than is the case in normally innervated mammalian skeletal muscle. Receptor was purified from Electrophorus muscle by affinity chromatography on cobra toxin-agarose and found to contain subunits which correspond immunochemically to the alpha, beta, gamma, and delta subunits of receptor from electric organ tissue of Torpedo californica. Receptor purified from Electrophorus muscle appears virtually identical with receptor purified from Electrophorus electric organ tissue.     

Primary structure of delta subunit precursor of calf muscle acetylcholine receptor deduced from cDNA sequence

Clones carrying cDNA sequences for the delta subunit precursor of the acetylcholine receptor from calf skeletal muscle have been isolated. Nucleotide sequence analysis of the cloned cDNA has indicated that this polypeptide consists of 516 amino acids including a hydrophobic prepeptide of 21 amino acids. The delta subunit of the calf muscle acetylcholine receptor, like the alpha, beta and gamma subunits of the same receptor as well as the alpha and gamma subunits of its human counterpart, exhibits structural features common to all four subunits of the Torpedo electroplax receptor, apparently being oriented across the membrane in the same manner as proposed for the fish receptor subunits. The degree of amino acid sequence homology between the calf and Torpedo delta subunits (60%) is comparable to that between the beta subunits (59%) and to that between the gamma subunits (56%), but is lower than that between the alpha subunits of the two species (81%). This suggests that the alpha subunit evolved more slowly than the three other subunits. A dendrogram representing the sequence relatedness among the four subunit precursors of the mammalian and fish acetylcholine receptors has been constructed. Some regions of the delta subunit molecule, including the region containing the putative disulphide bridge and that encompassing the clustered putative transmembrane segments M1, M2 and M3, are relatively well conserved between calf and Torpedo. The relative pattern of regional homology is similar for all four subunit precursors.     

Quaternary Structure of the Native GABAA Receptor Determined by Electron Microscopic Image Analysis

Abstract: In the transmitter-gated ion channel class of receptors, the members of which are all believed to be heterooligomers, the number and arrangement of the subunits are only known with any certainty for the nicotinic acetylcholine receptor from Torpedo electric fish. That receptor has been shown to possess a pentameric rosette structure, with five homologous subunits (α₂βγδ) arranged to enclose the central ion channel. The data were obtained by electron image analysis of two-dimensional receptor arrays, which form as a consequence of that receptor's exceptionally high abundance in the Torpedo membranes and are therefore not attainable for other receptors. We have applied another direct approach to determine the quaternary structure of native ionotropic GABA receptors. We have purified those receptors from porcine brain cortex and analysed the rotational symmetry of isolated receptors visualized by electron microscopy. The results show the receptor to have a pentameric structure with a central water-filled pore, which can now be said to be characteristic of the entire superfamily.     

Immunochemical similarities between subunits of acetylcholine receptors from Torpedo, Electrophorus, and mammalian muscle.

Polypeptide chains composing acetylcholine receptors from the electric organs of Torpedo californica and Electrophorus electricus were purified and labeled with 125I. Immunochemical studies with these labeled chains showed that receptor from Electrophorus is composed of three chains corresponding to the alpha, beta, and gamma chains of receptor from Torpedo but lacks a chain corresponding to the delta chain of Torpedo. Experiments suggest that receptor from mammalian muscle contains four groups of antigenic determinants corresponding to all four of the Torpedo chains. Binding of 125I-labeled chains was measured by quantitative immune precipitation and electrophoresis. Antisera to the following immunogens were used: denatured alpha, beta, gamma, and delta chains of Torpedo receptor, native receptor from Torpedo and Electrophorus electric organs and from rat and fetal calf muscle, and human muscle receptor (from autoantisera of patients with myasthenia gravis). The four chains of Torpedo receptor were immunologically distinct from one another and from higher molecular weight chains found in electric organ membranes. Antibodies to these chains reacted very efficiently with native Torpedo receptor, but the reverse was not true. Antibodies to native receptor from Torpedo and Electrophorus reacted slightly with each of the chains of the corresponding receptor. However, cross-reaction between chains and antibodies to any native receptor was most obviuos with the alpha chain of Torpedo or the corresponding alpha' chain of Electrophorus. Antiserum to alpha chains exhibited higher titer aginst receptor from denervated rat muscle. Antibodies from myasthenia gravis patients did not cross-react detectably with 125I-labeled chains from electric organ receptors. Most interspecies cross-reaction occurred at conformationally dependent determinants whose subunit localization could not be determined by reaction with the denatured chains.     

Purification and characterization of a nicotinic acetylcholine receptor from chick brain

Immunohistochemical studies have previously shown that both the chick brain and chick ciliary ganglion neurons contain a component which shares antigenic determinants with the main immunogenic region of the nicotinic acetylcholine receptor from electric organ and skeletal muscle. Here we describe the purification and initial characterization of this putative neuronal acetylcholine receptor. The component was purified by monoclonal antibody affinity chromatography. The solubilized component sediments on sucrose gradients as a species slightly larger than Torpedo acetylcholine receptor monomers. It was affinity labeled with bromo[3H]acetylcholine. Labeling was prevented by carbachol, but not by alpha-bungarotoxin. Two subunits could be detected in the affinity-purified component, apparent molecular weights 48 000 and 59 000. The 48 000 molecular weight subunit was bound both by a monoclonal antibody directed against the main immunogenic region of electric organ and skeletal muscle acetylcholine receptor and by antisera raised against the alpha subunit of Torpedo receptor. Evidence suggests that there are two alpha subunits in the brain component. Antisera from rats immunized with the purified brain component exhibited little or no cross-reactivity with Torpedo electric organ or chick muscle acetylcholine receptor. One antiserum did, however, specifically bind to all four subunits of Torpedo receptor. Experiments to be described elsewhere (J. Stollberg et al., unpublished results) show that antisera to the purified brain component specifically inhibit the electrophysiological function of acetylcholine receptors in chick ciliary ganglion neurons without inhibiting the function of acetylcholine receptors in chick muscle cells. All of these properties suggest that this component is a neuronal nicotinic acetylcholine receptor with limited structural homology to muscle nicotinic acetylcholine receptor.     

Protein phosphorylation of nicotinic acetylcholine receptors

The nicotinic acetylcholine receptor (nAcChR) is a ligand-gated ion channel found in the postsynaptic membranes of electric organs, at the neuromuscular junction, and at nicotinic cholinergic synapses of the mammalian central and peripheral nervous system. The nAcChR from Torpedo electric organ and mammalian muscle is the most well-characterized neurotransmitter receptor in biology. It has been shown to be comprised of five homologous (two identicle) protein subunits (alpha 2 beta gamma delta) that form both the ion channel and the neurotransmitter receptor. The nAcChR has been purified and reconstituted into lipid vesicles with retention of ion channel function and the primary structure of all four protein subunits has been determined. Protein phosphorylation is a major posttranslational modification known to regulate protein function. The Torpedo nAcChR was first shown to be regulated by phosphorylation by the discovery that postsynaptic membranes contain protein kinases that phosphorylate the nAcChR. Phosphorylation of the nAcChR has since been shown to be regulated by the cAMP-dependent protein kinase, protein kinase C, and a tyrosine-specific protein kinase. Phosphorylation of the nAcChR by cAMP-dependent protein kinase has been shown to increase the rate of nAcChR desensitization, the process by which the nAcChR becomes inactivated in the continued presence of agonist. In cultured muscle cells, phosphorylation of the nAcChR has been shown to be regulated by cAMP-dependent protein kinase, a Ca2+-sensitive protein kinase, and a tyrosine-specific protein kinase. Stimulation of the cAMP-dependent protein kinase in muscle also increases the rate of nAcChR desensitization and correlates well with the increase in nAcChR phosphorylation. The AcChR represents a model system for how receptors and ion channels are regulated by second messengers and protein phosphorylation.     

Reconstitution of acetylcholine receptor function in lipid vesicles of defined composition

The effect of specific lipids on the functional properties of the acetylcholine receptor were examined in reconstituted membranes prepared from purified Torpedo californica acetylcholine receptor and various defined lipids. Cholesterol and negatively charged lipids greatly enhanced the ion influx response of the vesicles as measured by the effect of a receptor agonist on cation translocation across the vesicles. Part of the lipid-dependent effects could be attributed to alterations in the average size of the vesicles. All lipid mixtures used permitted complete incorporation of receptor and retention of ligand binding properties. Quantitative differences in ion flux properties suggest a modulating role for specific lipids in acetylcholine receptor function.     

Identification of exposed and buried determinants of the membrane-bound acetylcholine receptor from Torpedo californica

The ability of five rabbit anti-acetylcholine receptor antisera to recognize the membrane-bound receptor from Torpedo californica has been investigated. Two antisera, raised against affinity-purified native receptor, react extensively with purified receptor-rich membrane vesicles. Since the membrane vesicles are impermeable to macromolecules and are oriented right side out, these two antisera recognize predominantly extracellular determinants. Two antisera against sodium dodecyl sulfate denatured receptor and one against purified delta subunit react poorly with the membrane-bound receptor. Only 10-20% of the determinants recognized by these antisera are accessible to antibodies when the receptor is membrane bound. Many of the latent sites can be exposed by permeabilizing the vesicles with saponin, by alkaline extraction of the membranes to remove peripheral proteins, or by a combination of these two treatments. These treatments neither solubilize the receptors nor interfere with their ability to undergo agonist-induced affinity changes. Subunit analysis of the sites on the membrane-bound receptor that are accessible to antibodies indicates that the alpha, beta, and delta chains possess extracellular determinants. Buried sites are present on all four of the subunits. Saponin permeabilization makes latent sites accessible on alpha and delta while alkaline extraction uncovers determinants on alpha, gamma, and delta. Treatment of membranes by both procedures reveals sites on beta, gamma, and delta that are not uncovered by either treatment alone. This study, in conjunction with results from other laboratories demonstrating that the gamma chain is extracellularly exposed, suggests that all four subunits are transmembrane proteins.     

Assembly of Torpedo acetylcholine receptors in Xenopus oocytes

To study pathways by which acetylcholine receptor (AChR) subunits might assemble, Torpedo alpha subunits were expressed in Xenopus oocytes alone or in combination with beta, gamma, or delta subunits. The maturation of the conformation of the main immunogenic region (MIR) on alpha subunits was measured by binding of mAbs and the maturation of the conformation of the AChR binding site on alpha subunits was measured by binding of alpha-bungarotoxin (alpha Bgt) and cholinergic ligands. The size of subunits and subunit complexes was assayed by sedimentation on sucrose gradients. It is generally accepted that native AChRs have the subunit composition alpha 2 beta gamma delta. Torpedo alpha subunits expressed alone resulted in an amorphous range of complexes with little affinity for alpha Bgt or mAbs to the MIR, rather than in a unique 5S monomeric assembly intermediate species. A previously recognized temperature-dependent failure in alpha subunit maturation may cause instability of the monomeric assembly intermediate and accumulation of aggregated denatured alpha subunits. Coexpression of alpha with beta subunits also resulted in an amorphous range of complexes. However, coexpression of alpha subunits with gamma or delta subunits resulted in the efficient formation of 6.5S alpha gamma or alpha delta complexes with high affinity for mAbs to the MIR, alpha Bgt, and small cholinergic ligands. These alpha gamma and alpha delta subunit pairs may represent normal assembly intermediates in which Torpedo alpha is stabilized and matured in conformation. Coexpression of alpha, gamma, and delta efficiently formed 8.8S complexes, whereas complexes containing alpha beta and gamma or alpha beta and delta subunits are formed less efficiently. Assembly of beta subunits with complexes containing alpha gamma and delta subunits may normally be a rate-limiting step in assembly of AChRs.     

Functional immobilisation of the nicotinic acetylcholine receptor in tethered lipid membranes

The nicotinic acetylcholine receptor from Torpedo was immobilised in tethered membranes. Surface plasmon resonance was used to quantify the binding of ligands and antibodies to the receptor. The orientation and structural integrity of the surface-reconstituted receptor was probed using monoclonal antibodies, demonstrating that approximately 65% of the receptors present their ligand-binding site towards the lumen of the flow cell and that at least 85% of these receptors are structurally intact. The conformation of the receptor in tethered membranes was investigated with Fourier transform infrared spectroscopy and found to be practically identical to that of receptors reconstituted in lipid vesicles. The affinity of small receptor ligands was determined in a competition assay against a monoclonal antibody directed against the ligand-binding site which yielded dissociation constants in agreement with radioligand binding assays. The presented method for the functional immobilisation of the nicotinic acetylcholine receptor in tethered membranes might be generally applicable to other membrane proteins.     

Photolabeling the Torpedo nicotinic acetylcholine receptor with 4-azido-2,3,5,6-tetrafluorobenzoylcholine, a partial agonist

The interactions of a photoreactive analogue of benzoylcholine, 4-azido-2,3,5,6-tetrafluorobenzoylcholine (APFBzcholine), with nicotinic acetylcholine receptors (nAChRs) were studied using electrophysiology and photolabeling. APFBzcholine acted as a low-efficacy partial agonist, eliciting maximal responses that were 0.3 and 0.1% of that of acetylcholine for embryonic mouse and Torpedo nAChRs expressed in Xenopus oocytes, respectively. Equilibrium binding studies of [3H]APFBzcholine with nAChR-rich membranes from Torpedo electric organ revealed equal affinities (K(eq) = 12 microM) for the two agonist binding sites. Upon UV irradiation at 254 nm, [3H]APFBzcholine was photoincorporated into the nAChR alpha, gamma, and delta subunits in an agonist-inhibitable manner. Photolabeled amino acids in the agonist binding sites were identified by Edman degradation of isolated, labeled subunit fragments. [3H]APFBzcholine photolabeled gammaLeu-109/deltaLeu-111, gammaTyr-111, and gammaTyr-117 in binding site segment E as well as alphaTyr-198 in alpha subunit binding site segment C. The observed pattern of photolabeling is examined in relation to the predicted orientation of the azide when APFBzcholine is docked in the agonist binding site of a homology model of the nAChR extracellular domain based upon the structure of the snail acetylcholine binding protein.     

Monoclonal antibodies to cytoplasmic domains of the acetylcholine receptor

Fourteen clonal hybridoma lines that secrete monoclonal antibodies (mabs) to the Torpedo acetylcholine receptor (AChR) have been isolated. When analyzed by an immunoreplica technique, two mabs recognized the alpha subunit, three reacted with the beta subunit, one reacted with the gamma chain, and five recognized the delta subunit. One mab failed to react with any of the subunits using this assay and two mabs recognized determinants found on both the gamma and the delta subunits. These were classified according to their reactivities with the membrane-bound Torpedo AChR. One category is comprised of mabs (including both anti-alpha mabs) that recognize extracellular epitopes. A second classification included mabs that are unable to bind the membrane-associated AChR. The third category is comprised of mabs directed against cytoplasmic epitopes of the AChR. The latter mabs, all of which recognize the gamma or delta subunits or both, bind only slightly to sealed, outside-out Torpedo vesicles. The binding is increased 10-20-fold by either alkaline extraction or treatment of the vesicles with 10 mM lithium diiodosalicylate but not by permeabilization of the vesicles with saponin. Three of the six mabs in this category react with frog muscle endplates but only if the cytoplasmic surface of the membrane is accessible.     

Chromosomes move poleward in anaphase along stationary microtubules that coordinately disassemble from their kinetochore ends

The binding sites on the nicotinic acetylcholine receptor of labels specific for the alpha-, beta-, and delta-subunits were determined by electron image analysis, using tubular crystals of receptors grown from the postsynaptic membranes of Torpedo marmorata electric organ. The labels were alpha-bungarotoxin (which attaches to the acetylcholine binding sites on the pair of alpha-subunits), Fab35 (a monoclonal antibody Fab fragment directed against the main immunogenic region of the alpha-subunit), Fab111 (a monoclonal antibody Fab fragment directed against a cytoplasmic site on the beta-subunit), and wheat germ agglutinin (which binds to N-acetylglucosamine residues on the delta-subunit). These labels, bound to receptors in the crystals, were located by comparing labeled with native structures, averaged in each case over more than 5,000 molecules. From the assignments made, we find that the clockwise arrangement of subunits around the receptor, viewed from the synaptic face, is: alpha, beta, alpha, gamma, and delta; that the main immunogenic region is at (or close to) the side of the alpha-subunit; and that the two acetylcholine binding sites are at the synaptic end of the alpha-subunits, 27-28 A from the central axis and approximately 53 A apart. In the crystal lattice, neighboring molecules are paired so that their delta- and alpha-subunits are juxtaposed, an organization that appears to relate closely to the grouping of receptors in vivo.     

Tryptophan and cystein residues of the acetylcholine receptors of Torpedo species. Relationship to binding of cholinergic ligands.

Several methods were used to analyze for tryptophan in the acetylcholine (ACh) receptors purified from the electric organs of the electric rays, Torpedo californica and Torpedo marmorata. The best value of tryptophan was 2.4 mol %. When excited at 290 nm, both receptors fluoresced with a maximum at 336, but there was no change in the fluorescence emission spectra upon binding of carbamylcholine, d-tubocurarine, ACh, or decamethonium. The free SH content of the Torpedo receptors varied in different preparations, and was highest in that purified from fresh T. californica using deaerated solutions and dialysis under nitrogen, and lowest in that prepared from the aged lyophilized membranes of T. marmorata. The maximum free SH content was 20 nmol/mg of protein or 0.22 mol %, equal to at most 18% of the total cysteic acid residues. Reaction of either 33% or of all the SH residues with p-chloromercuribenzoate reduced maximum ACh binding to the pure receptor prepared from fresh T. californica by only 23%.     

Acetylcholine receptor assembly is stimulated by phosphorylation of its gamma subunit

Different combinations of Torpedo acetylcholine receptor (AChR) subunits stably expressed in mouse fibroblasts were used to establish a role for phosphorylation in AChR biogenesis. When cell lines expressing fully functional AChR complexes (alpha 2 beta gamma delta) were labeled with 32P, only gamma and delta subunits were phosphorylated. Forskolin, which causes a 2- to 3-fold increase in AChR expression by stimulating subunit assembly, increased unassembled gamma phosphorylation, but had little effect on unassembled delta. The forskolin effect on subunit phosphorylation was rapid, significantly preceding its effect on expression. The pivotal role of the gamma subunit was established by treating alpha beta gamma and alpha beta delta cell lines with forskolin and observing increased expression of only alpha beta gamma complexes. This effect was also observed in alpha gamma, but not alpha delta cells. We conclude that the cAMP-induced increase in expression of cell surface AChRs is due to phosphorylation of unassembled gamma subunits, which leads to increased efficiency of assembly of all four subunits.     

Subunit interactions inhibit the binding of beta nerve growth factor to receptors on embryonic chick sensory neurons

The binding of the beta nerve growth factor subunit in the 7 S nerve growth factor complex to the two nerve growth factor receptors on chick embryo dorsal root ganglion cells was investigated. Under conditions where the 7 S nerve growth factor complex is maximally stable (in the presence of excess alpha and gamma subunits and of 20 to 30 microM zinc ion), no binding to either receptor was detectable. The time course of the decrease in the binding of beta nerve growth factor to the receptors upon addition of alpha and gamma subunits and zinc ion paralleled the formation of the 7 S complex. Addition of alpha and gamma subunits and zinc ion to the bisdes-arginine118-nerve growth factor, which does not re-form the 7 S complex, failed to inhibit the binding of the derivative to either receptor. While the alpha subunit alone had no effect on beta nerve growth factor binding, the gamma subunit decreased its binding in proportion to the amount of complex formed between these two subunits, suggesting that the beta . gamma complex, like the 7 S complex, does not bind to nerve growth factor receptors.     

Arrangement of the subunits of the nicotinic acetylcholine receptor of Torpedo californica as determined by alpha-neurotoxin cross-linking

[3H]Methyl-alpha-neurotoxin prereacted with dithiobis(succinimidyl propionate) (DTSP) can be covalently linked to each of the subunits of the nicotinic acetylcholine receptor in membranes from the electric tissue of Torpedo californica. Pronounced changes in the cross-linking pattern are observed upon prior incubation with receptor specific ligands and upon reduction and/or alkylation of the receptor. d-Tubocurarine has been shown to bind to two different sites in receptor-rich membranes. These sites are present in equal numbers but have different affinities [Neubig, R. R., & Cohen, J. B. (1979) Biochemistry 18, 5464-5475; Sine, S., & Taylor, P. (1981) J. Biol. Chem. 256, 6692-6699]. Using d-tubocurarine inhibition of [3H]-methyl-alpha-neurotoxin binding, we demonstrate two inhibitory constants for d-tubocurarine of 67 +/- 21 nM and 4.9 +/- 1.7 microM in unreduced membranes. We utilize the large difference in Ki's to preferentially block toxin cross-linking at the high affinity site for d-tubocurarine. Low concentrations of this competitive antagonist selectively block the cross-linking of toxin to the beta and gamma subunits of the receptor, suggesting that these subunits are located close to the toxin binding site which is also the high-affinity binding site for d-tubocurarine. Reduction of disulfide bonds alters the affinity of the receptor for alpha-neurotoxin. Alterations are also seen in the cross-linking pattern of DTSP-activated [3H]methyl-alpha-neurotoxin to reduced and alkylated membranes in the presence of tubocurarine. The constants for d-tubocurarine inhibition of [3H]methyl-alpha-neurotoxin binding to reduced and alkylated membranes are 172 +/- 52 nM and 2.4 +/- 0.4 microM. The effects of bromoacetylcholine, carbamoylcholine, gallamine, and procaine on the cross-linking pattern are also examined. Our observations are consistent with an arrangement of the subunits in the membrane of alpha beta alpha gamma delta.     

Immunochemical demonstration that amino acids 360-377 of the acetylcholine receptor gamma-subunit are cytoplasmic

Two monoclonal antibodies (mabs) previously prepared against Torpedo acetylcholine receptor are shown to recognize a synthetic nonadecapeptide corresponding to lys360-glu377 of the gamma subunit. The reaction was demonstrated by solid-phase enzyme-linked immunoabsorbent assays, by inhibition of binding of the mabs to receptor, and by immunoprecipitation of the peptide conjugated to bovine serum albumin. Immunogold electron microscopy on isolated postsynaptic membranes from Torpedo showed that both mabs bind to intracellular epitopes on the receptor. These results establish that amino acid residues 360-377 of the receptor gamma-subunit, and probably the analogous region of the delta-subunit, reside on the cytoplasmic side of the membrane. Since the primary structures of all four subunits suggest a common transmembrane arrangement, the corresponding domains of the alpha- and beta-subunits are probably also cytoplasmic.