Identification of polypeptides associated with a putative neuronal nicotinic acetylcholine receptor.

Polypeptides involved in the binding of the nicotinic acetylcholine receptor ligand alpha-bungarotoxin (Mr = 8,000) to neuronal membranes were identified by three independent methods: (i) 125I-alpha-bungarotoxin bound to membrane fractions or to monolayer cultures of chick retina was cross-linked to its binding site by using glutaraldehyde, or the photoactivatable bifunctional reagent N-succinimidyl-6-(4'-azido-2'-nitrophenylamino)hexanoate. Electrophoretic analysis of the cross-linked membrane proteins revealed 125I-alpha-bungarotoxin-polypeptide adducts of apparent Mr = 63,000, 43,000, and 33,000. (ii) Affinity purification of the alpha-bungarotoxin binding protein from detergent extracts of [35S]methionine-labeled retina cultures identified one major polypeptide with an Mr = 57,000. (iii) Indirect immunoprecipitation from detergent extracts of [35S]methionine-labeled rat pheochromocytoma cells (PC 12) gave evidence for a specific co-precipitation of alpha-bungarotoxin with three polypeptides (Mr = 57,000, 34,000, and 25,000). The data suggest that polypeptides of Mr - 57,000, 35,000, and 25,000 (+/- 3,000) are located at or close to the alpha-bungarotoxin binding domain of the putative neuronal nicotinic acetylcholine receptor.

Polypeptides involved in the binding of the nicotinic acetylcholine receptor ligand a-bungarotoxin (Mr = 8,000) to neuronal membranes were identified by three independent methods: (i) '"I-a-bungarotoxin bound to membrane fractions or to monolayer cultures of chick retina was cross-linked to its binding site by using glutaraldehyde, or the photoactivatable bifunctional reagent N-succinimidyl-6-(4'-azido-2'-nitrophenylamino)hexanoate. Electrophoretic analysis of the crosslinked membrane proteins revealed '251-a-bungarotoxin-polypeptide adducts of apparent M, = 63,000, 43,000, and 33,000. (ii) Affinity purification of the abungarotoxin binding protein from detergent extracts of [35S]methionine-labeled retina cultures identified one major polypeptide with an M, = 57,000. (iii) Indirect immunoprecipitation from detergent extracts of methionine-labeled rat pheochromocytoma cells (PC 12) gave evidence for a specific co-precipitation of abungarotoxin with three polypeptides (Mr = 57,000, 34,000, and 25,000). The data suggest that polypeptides of M, = 57,000, 35,000, and 25,000 (+3,000) are located at or close to the a-bungarotoxin binding domain of the putative neuronal nicotinic acetylcholine receptor.
The snake venom polypeptide a-bungarotoxin binds with high selectivity and in an almost irreversible fashion to the cholinergic ligand binding site of the nicotinic acetylcholine receptor in fish electric organ and skeletal muscle (1, 2). a-Bungarotoxin binding sites with nicotinic-cholinergic specificity are also present in the peripheral and central nervous systems of vertebrates and invertebrates (reviewed in Refs. 3 and 4). T h e identification of these neuronal a-bungarotoxin binding sites as nicotinic acetylcholine receptors is still controversial, since in most vertebrate preparations a-bungarotoxin fails to inhibit neuronal nicotinic-cholinergic responses (3,4). In the goldfish, toad, and chick visual system, however, a-bungarotoxin can block cholinergic receptor function

(5).'
T h e high molecular weight a-bungarotoxin binding protein of these tissues thus can be classified as a putative neuronal nicotinic acetylcholine receptor (3-9).
In a preceding publication, we have presented a detailed characterization of the a-bungarotoxin binding protein of chick retina neurons maintained in tissue culture (9). Here we * This work was supported by grants from the Deutsche Forschungsgemeinschaft and the Stiftung Volkswagenwerk. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ' H. Rehm, unpublished observations. report an attempt to identify the polypeptide(s) involved in this a-bungarotoxin binding site by using cross-linking techniques and affinity purification. Immunoprecipitation data on the subunit composition of the a-bungarotoxin binding protein of the rat pheochromocytoma cell line PC 12 are also included.
a-Bungarotoxin-derivatized agarose beads were prepared by reacting Affi-Gel 10 (Bio-Rad Laboratories) with a-bungarotoxin to a final substitution of 0.1-0.3 mg of toxin bound/ml of bead volume (10). Antibodies to a-bungarotoxin were obtained by immunizing rabbits repeatedly with sublethal doses (50-100 pg) of the toxin emulsified in Freund's complete adjuvant (11). a-Bungarotoxin antibodies were then purified from the rabbit antisera by sequential ammonium sulfate precipitation, DEAE-Sephadex chromatography, and affinity chromatography on a-bungarotoxin-derivatized agarose (11, 12).
The pheochromocytoma cell line PC 12 (13) was kindly donated by Dr. L. Greene, New York University.
Crude Synaptic Membranes-P2 membrane fractions were prepared from the retina of newly hatched chickens as described previously (9). These membranes were further purified by a modification of Ref. 14. The P2 pellet was resuspended in 5 mM Tris-CI, pH 7.4, and 60% (w/v) sucrose was added to a final concentration of 15% (w/ v). Aliquots of this membrane suspension were layered on top of 10 ml of a 35%:45% (w/v) two-step sucrose gradient. The gradients were centrifuged in a Beckman SW 27 rotor at 26,000 rpm and 4 "C for 60 min. Material obtained from the 15%:35% sucrose interphase was diluted with 3 volumes of 25 mM potassium phosphate buffer, pH 7.4. The purified membranes were then collected by centrifugation at 100,OOo x g for 20 min, frozen in liquid nitrogen, and stored at -70 "C.
Tissue Culture-Monolayer cultures of retina cells from 8-day-old chick embryos were prepared as described previously (9, 15) with the following modifications: 3 X lo7 dissociated cells were plated per polylysine-coated (16) 100-mm plastic dish (Nunc, Roskilde, Denmark). After 1 day in Dulbecco's modified Eagle's medium containing ' The abbreviations used is: EGTA, ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid. 5% (v/v) fetal calf serum and 5% (v/v) horse serum, the cultures received the serum-free N1 medium of Bottenstein et al. (17) in order to minimize the proliferation of non-neuronal cells (16). I'C 12 cells were grown in tissue culture flasks (Nunc) as described by Greene and Tischler (13).
Metabolic labeling of cell proteins was achieved by incubating 4-to 7-day-old cultures for 24-48 h in 4-8 ml of N1 medium containing 20 p~ methionine, 40-120 pCi/ml of ["'SS]methionine. In order to economize on label, the radioactive medium was re-used for up to four experiments. Usually. 10-208 of the radioactivity present in the medium was incorporated during one labeling period.
Cross-linking Experiments-Crude synaptic membrane fractions from chick retina were suspended in 50 mM sodium phosphate buffer, pH 7.4, containing 0.02% (w/v) Triton X-100 and the anti-protease mixture (buffer A). The membranes were incubated with 4 nM I2"Ia-bungarotoxin at 23 "C for 2 h in the presence or absence of 0.1 mM d-tubocurarine (total volume, 2.5 ml; 3.3 mg/ml of protein). T h e incubations were then diluted with 10 volumes of cold buffer A and centrifuged a t 4 "C and 42,000 X g for 20 min. This washing step was repeated once.
For cross-linking with glutaraldehyde, the "'I-toxin-labeled membranes were suspended in half their original volume of 50 mM sodium phosphate buffer, pH 9.0, containing 5 mM EDTA, 3 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride. Glutaraldehyde was added to 0.2ml aliquots of these membranes a t final concentrations of 0.1-10 mM. After 30 min a t 23 "C, 1 ml of 50 mM Tris-CI, pH 7.4, was added, and the membranes were sedimented for 4 min in an Eppendorf 3200 centrifuge. After washing with the Tris-CI buffer, the membranes were collected for electrophoresis.
In cross-linking experiments with N-succinimidyl-6-(4'-azido-2"ni-tropheny1amino)hexanoate. the "'I-toxin-labeled membranes were resuspended in 50 mM sodium phosphate buffer, pH 7.4, containing 5 mM EDTA, 3 mM EGTA. and 0.1 mM phenylmethylsulfonyl fluoride. Aliquots (0.2 ml) were transferred to glass tubes, and N-succinimidyl-6-(4'-azido-2"nitrophenylamino)hexanoate dissolved in dimethyl sulfoxide was added in the dark to final concentrations between 2 p~ and 2 mM. In all experiments, the concentration of dimethyl sulfoxide in the reaction mixture was below 5% (v/v). After 10 min at 4 "C, the tubes were illuminated for 15 min at 4 "C with an Osram HNS 15watt OFR mercury arc lamp from a distance of 3 cm. Then 1 ml of 50 mM Tris-CI, pH 7.4, was added, and the contents of the tubes were transferred to 1.5-ml Eppendorf vials and processed as described for glutaraldehyde-cross-linked membranes.
In cross-linking experiments with intact cells, 7-day-old retina cultures were incubated for 1 h at 37 "C with 8.3 nM ""I-a-bungarotoxin in the presence and absence of 0.1 mM d-tubocurarine (9). After three washes with phosphate-buffered saline, the cultures were covered with 10 mM glutaraldehyde in phosphate-buffered saline adjusted to pH 8.0 (4 "C, 30 min). After washing with phosphate-buffered saline, the cells were incubated with 60 mM nicotine in phosphatebuffered saline for 30 min a t 4 "C in order to remove most noncovalently bound "'I-a-toxin (18). The cells were harvested in 0.1 M Tris-CI buffer, pH 8.8, containing the anti-protease mixture and incubated for 30 min a t 4 "C in 1 ml of 50 mM NaBH., dissolved in 0.1 M Tris-CI, pH 9.6. After two more washes with 0.1 M Tris-CI, pH 8.8, the cells were collected for electrophoresis.
Electrophoresis-Polyacrylamide slab gel electrophoresis in the presence of sodium dodecyl sulfate was performed using the discontinuous buffer system of Laemmli (19). Immediately after preparation, all samples were heated for 5 min a t 95 "C in sample buffer (19) containing 2% (w/v) sodium dodecyl sulfate, 5% (w/v) mercaptoethanol. The samples were stored at -20 "C overnight and electrophoresed on 1.5-mm separating gels containing 10%. (w/v) acrylamide. After staining with Coomassie brilliant blue H-250, destaining, and washing in water, the gel slabs were dried on filter paper under vacuum. Kodak X-Omat XR-5 film was exposed to the dried gels a t -70 "C for 1-5 weeks using an intensifying screen. For the dete;mination of incorporation efficiency in cross-linking experiments, radioactive bands localized by autoradiography were cut out from the stained gels and counted. Slab gels containing ["'S]methionine-labeled proteins were processed for fluorography using the sodium salicylate impregnation procedure (20). Densitometric tracings of autoradiograms were made with a Vitatron TLD 100 densitometer. T h e following molecular weight standards (Bio-Had) were used to estimate the apparent molecular weights of the labeled proteins: phosphorylase b (92,500); bovine serum albumin (66,200); ovalbumin (45,000); carbonic anhydrase (31,000); soybean trypsin inhibitor (21,500); and lysozyme (14,000).
Affinity Purification-Affinity purificat'ion of the detergent-solubilized a-bungarotoxin binding protein from [""Slmethionine-labeled cell cultures followed procedures previously reported for the muscle acetylcholine receptor (21). The [:'"S)nethionine-labeled cells from 2-6 cultures were harvested and washed in phosphate-buffered saline. The cells were then homogenized in 0.5-2 ml of 10 mM Tris-CI. pH 7.4, containing 17; (w/v) Triton X-100. 130 mM NaCI, and the antiprotease mixture. After 30 min on ice, the homogenate was centrifuged at 150,000 X g for 30 min. The supernatant was incubated with 20-l00pl of a-bungarotoxin-derivatized agarose on a roller shaker at 4 "C for 3 h. The agarose beads were then collected by centrifugation and extensively washed (four washes with 2 ml of 10 mM Tris-CI, pH 7.4. containing 1% (w/v) Triton X-100. 1 M NaCI, and the anti-protease mixture, and two washes with the same buffer without NaCI). The beads were resuspended in 80 pl of 100 p~ d-tubocurarine dissolved in 10 mM Tris-CI, pH 7.4, containing 1% (w/v) Triton X-100 and the anti-protease mixture, and kept a t 4 "C for 30 min. After centrifugation, the supernatants were collected for electrophoresis.
Immunoprecipitation--Detergent extracts from ["Slmethioninelabeled cultures of retina or I'C 12 were prepared as detailed under affinity purification. Indirect immunoprecipitation of the solubilized a-bungarotoxin-tqged toxin binding site with anti-a-bungarotoxin and fixed S. aureus cells was performed as described by Merlie and Sebbane (11). The immunoprecipitates were then washed with buffer as described in the affinity purification procedure and collected for electrophoresis.

RESULTS
Different methods were used to identify polypeptides associated with the a-bungarotoxin binding site of neuronal membranes.
Cross-linking of '2~5Z-a-Bungarotoxin to Membrane Fractions from Chick Retina-In initial experiments, '"I-a-bungarotoxin was bound to membrane preparations from retina which then were treated with different concentrations of dimethylsuberimidate and glutaraldehyde. Only with the latter reagent, and at concentrations greater than 2 mM, was a significant (10% of the specifically bound toxin) covalent radiolabeling of membrane polypeptides obtained. As shown in nicotinic antagonist d-tubocurarine prevented the labeling of these polypeptides.
Cross-linking of '~.iZ-a-Bungaroto.rin to Monolayer Cultures of Chick Embryo Retina-The findings obtained with membrane fractions were extended in cross-linking experiments using intact cultured retinal neurons. After binding ""Ia-bungarotoxin to cell surface receptors, glutaraldehyde treatment and subsequent reduction with NaRH., yielded radioactive polypeptides of M, = 64,000 and 35,000 upon electrophoresis of the total cellular proteins (Fig. 4, right). These radioactively labeled hands were absent from cells which had been preincubated with '~"1-n-bungarotoxin in the presence of 0.1 mM d-tubocurarine.
Affinity Purification of the a-Bungarotoxin Binding Pro-
In control experiments, it was checked that this affinity support isolated the detergent-solubilized acetylcholine receptor from Torpedo californica in its typical four-subunit composition (M, = 40,000, 50,000, 59,000, and 65.000):' Furthermore, the toxin heads were found to adsorb >gor{ of the Triton X-100-solubilized high affinity a-bungarotoxin binding sites present in detergent extracts from chick retina membranes (data not shown). Up to 50% of these toxin binding sites could he recovered by ligand-specific elution of the heads with nicotine or d-tubocurarine, both of which accelerate dissociation of abungarotoxin from its neuronal binding site (18). Removal of the eluting ligand by dialysis or by chromatography on wheat germ agglutinin-Sepharose gave a >1000-fold purification of the a-bungarotoxin binding protein without significantly changing its ""I-a-bungarotoxin and cholinergic ligand binding properties (Ref. 9 and data not shown). The very low quantities of receptor protein which could he isolated from retina prevented, however, the accurate determination of specific binding activities and the analysis of polypeptide composition by protein staining. The receptor therefore was purified from detergent extracts of ['%]methionine-labeled retina cultures using batchwise adsorption on a-bungarotoxinderivatized agarose. After thorough washing, the resin was briefly incubated with a high concentration of d-tubocurarine. Polyacrylamide gel electrophoresis of the d-tubocurarine eluate in the presence of sodium dodecyl sulfate revealed that >90% of the protein radioactivity was contained in a polypeptide of M, = 57,000 (Fig. 3). The isolation of this polypeptide was prevented by preincubating the detergent extracts with 10 PM a-bungarotoxin before adsorption on the affinity resin.

Immunoprecipitation of a-Rungarotoxin-binding Polypeptides from Detergent Extracts of [,'"'SJMethionine-hbeled
Cultures of the Pheochromocytoma Cell Line P C 12-The isolation of polypeptides involved in a-bungarotoxin binding from ["%]methionine-labeled cells was also attempted using an indirect immunoprecipitation procedure (21). In this technique, the toxin binding site was labeled with a-bungarotoxin and subsequently precipitated with antitoxin and S. aureus (21). With detergent extracts from retina cultures, the amount of radiolabel precipitated was too low to allow the detection of any receptor-specific bands (data not shown). We therefore used the rat pheochromocytoma cell line PC 12 which, since it proliferates, enables a very high specific activity radiolabeling of i t s cellular proteins. Immunoprecipitates from PC 12 extracts incubated with 10 nM a-bungarotoxin, antitoxin, and S. aureus contained several polypeptides, three of which were absent when the precipitation was performed in the presence of a large excess of a-bungarotoxin over antitoxin (Fig. 4, left).
The apparent molecular weights of these bands were 57,000, 34,000, and 25,000. T h e M , = 57,000 band co-migrated with the polypeptide isolated from retina cultures by affinity purification.

DISCUSSION
In this paper, several different techniques were used to identify the polypeptide(s) involved in the binding of a-bungarotoxin to the putative nicotinic acetylcholine receptor of retina. Using affinity purification on a-bungarotoxin-derivatized agarose, a polypeptide of M , = 57,000 was isolated from detergent extracts of ["%]methionine-labeled retina cultures. Different lines of evidence indicate that this polypeptide contains the a-bungarotoxin and cholinergic ligand binding site (6,7,9) of the retinal a-bungarotoxin receptor: (i) >90% of the protein radioactivity in the eluate from the affinity resin was contained in this polypeptide, and its isolation was prevented when the purification was performed in the presence of a n excess of a-bungarotoxin; (ii) the a-toxin and cholinergic ligand binding functions of the receptor were recovered in the eluates from a-bungarotoxin-derivatized agarose after removal of the eluting ligand d-tubocurarine.
The participation of a M , = 57,000 (+3,000) polypeptide in a-toxin binding to chick retina was confirmed by cross-linking of '"I-labeled a-bungarotoxin to synaptosomal membrane preparations or cultured cells. With glutaraldehyde and Nsuccinimidyl-6-(4'-azido-2'-nitrophenylamino)hexanoate as cross-linkers, labeled polypeptide species of about M , = 63,000, 43,000, and 33,000 were found upon polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. Assuming that each radiolabeled polypeptide contained one covalently bound toxin molecule, these molecular weights correspond to receptor polypeptides of M , = 55,000, 35,000, and 25,000 (+3,000). The band labeled by cross-linking of M , = 63,000 thus corroborates the relevance of the M , = 57,000 polypeptide isolated by affinity purification. T h e origin of the two other polypeptides identified by cross-linking is not clear. These polypeptides may represent receptor subunits which were lost during the incubation and washing procedures used in the affinity purification. Alternatively, they may not be integral subunits of the putative neuronal acetylcholine receptor, or they may have resulted from proteolysis of the polypeptide of M , = 57,000. In view of the fact that cross-linking was performed within a few hours and that protease inhibitors and chelating agents were included, the last possibility appears, however, less likely. Also, a M , = 35,000 polypeptide was detected in the cross-linking experiments with intact retinal cells in culture. We therefore suggest that the lower molecular weight bands described here are not artifactual, but located in close proximity to the retinal a-bungarotoxin binding site.
This interpretation is also supported by the data obtained with another rapid isolation technique, the toxin-antitoxin immunoprecipitation method of Merlie and Sebbane which has originally been developed for muscle acetylcholine receptor. With detergent extracts from ["%]methionine-labeled retina cultures, we were unable to recover sufficient radioactivity to reveal any specifically precipitated receptor polypeptides. This may be due to an accelerated dissociation of the receptor from the toxin-immunoglobulin complex. With the rat pheochromocytoma cell line I T 12 which possesses a-bungarotoxin binding sites very similar to those present in retina (26), a much higher metabolic labeling with radioactive amino acids of the cellular proteins could be achieved. Immunoprecipitates from detergent extracts of these cells contained in addition to several co-precipitated proteins three polypeptides of M , = 57,000,34,000, and 25,000 whose isolation was prevented by an excess of a-bungarotoxin over antitoxin.
Taken together, our results show that the a-bungarotoxin and cholinergic ligand binding site of the putative neuronal nicotinic acetylcholine receptor in chick and rat is contained in a polypeptide of M , = 57,000 (+3,000). In addition, polypeptides of M , = 35,000 and 25,000 (+3,000) may be associated with the a-bungarotoxin binding domain of the neuronal receptor. These molecular weights are different from those reported for the subunits of the nicotinic acetylcholine receptor of Torpedo electric organ ( M , = 40,000,50,000, 60,OOO, and 65,000; reviewed in Ref. 1). In particular, both the a-toxin and the cholinergic ligand binding functions of this peripheral acetylcholine receptor have been localized on the smallest receptor subunit of M , = 40,000 (1). Similar conclusions have been reached for the acetylcholine receptor of rat skeletal muscle (23). It should, however, be noted that photoaffinity labeling studies with '"I-a-bungarotoxin derivatives gave rather divergent results. For example, in membrane fractions of Torpedo electric organ, the receptor polypeptides of M , = 40,000 and 65,000 were preferentially labeled (27-29). In situ photoaffinity labeling with an a-bungarotoxin derivative of the acetylcholine receptor of rat muscle gave in contrast a prominent radiolabeled band of M , = 63,000 (28). Furthermore, purification of the acetylcholine receptor from rat and chick muscle has revealed a receptor polypeptide of M , = 56,000 (23, 30). This subunit of the muscle receptor has a similar molecular weight to the large polypeptide of the neuronal receptor detected in this study; it is, however, not involved in cholinergic ligand binding (23). From immunological and sequence analysis, it has been suggested that the different subunits of the acetylcholine receptor from fish electric organ may have derived from a common ancestor polypeptide (31.32). One may therefore speculate that during evolution the toxin and acetylcholine binding site of the peripheral and the neuronal receptor have been conserved on polypeptides of different size. This hypothesis is supported by the low, but distinct immunological cross-reactivity of the Torpedo and the putative neuronal acetylcholine receptor (9,

33).
During the course of this work, attempts have been made by others to establish the subunit composition of the &-bungarotoxin binding protein of rat brain (34, 35). So far, t h e results are conflicting. One group obtained only one polypeptide species of M, = 51,000 (34), whereas another study described four subunits of M , = 44,000, 50,000, 56,000, and 62,000 (35). Also, after submission of this manuscript, a report on the isolation of the putative nicotinic acetylcholine receptor from chick optic lobe was published (36). In that paper, only one receptor subunit of M,-= 54,000 was detected after chromatography on a-bungarotoxin-and lentil lectin-derivatized Sepharose and shown to react with a cholinergic affinity reagent. This polypeptide appears to be similar to the M, = 57,000 polypeptide isolated in our affinity purification procedure.