Determination of the Tyrosine Phosphorylation Sites of the Nicotinic Acetylcholine Receptor*

The peripheral nicotinic acetylcholine receptor (nAChR) is phosphorylated on tyrosine residues in vivo and in vitro at a high stoichiometry. We have previ- ously reported that this tyrosine phosphorylation occurs on the @, y, and 6 subunits of the receptor and is implicated in both the modulation of the function of the receptor and localization of the receptor at the synapse. The specific tyrosine residue of each subunit which is phosphorylated is now identified. The endog- enously phosphorylated nAChR from the electric organ of Torpedo cdifornica was phosphorylated to maximal stoichiometry in vitro exclusively on tyrosine residues as indicated by phosphoamino acid analysis. Two-di-mensional phosphopeptide maps of thermolysin limit digests of the isolated phosphorylated subunits indicated that each subunit is phosphorylated at a single site. To determine the site of tyrosine phosphorylation [-y-32P]ATP following mol [32P]phosphate in vitro. Phosphoamino Acid Analysis-Phosphorylated subunits were acid hydrolyzed and subjected to one-dimensional thin layer electropho- resis essentially as described by Hirano et al. (1988).

The peripheral nicotinic acetylcholine receptor (nAChR) is phosphorylated on tyrosine residues in vivo and in vitro at a high stoichiometry. We have previously reported that this tyrosine phosphorylation occurs on the @, y, and 6 subunits of the receptor and is implicated in both the modulation of the function of the receptor and localization of the receptor at the synapse. The specific tyrosine residue of each subunit which is phosphorylated is now identified. The endogenously phosphorylated nAChR from the electric organ of Torpedo cdifornica was phosphorylated to maximal stoichiometry in vitro exclusively on tyrosine residues as indicated by phosphoamino acid analysis. Two-dimensional phosphopeptide maps of thermolysin limit digests of the isolated phosphorylated subunits indicated that each subunit is phosphorylated at a single site. To determine the site of tyrosine phosphorylation of the 8, y, and 6 subunits, phosphorylated subunits were isolated and digested with trypsin. A single phosphotyrosine containing peptide from each subunit was purified by antiphosphotyrosine antibody affinity chromatography and reverse phase high performance liquid chromatography. The purified phosphopeptides were subjected to sequential Edman degradation and sequence analysis. Comparison of the phosphopeptide sequence data with the deduced amino acid sequence of each subunit indicated that Tyr-355 of @, Tyr-364 of y, and Tyr-372 of 6 are the sites of in vitro and in vivo tyrosine phosphorylation of the nAChR. Identification of these sites should facilitate further studies of the role of tyrosine phosphorylation in the regulation of receptor function.
The nicotinic acetylcholine receptor (nAChR)' is a ligand-* This work was supported in part by Grant NS24418 from the National Institutes of Health and the Howard Hughes Medical Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a Medical Scientist Training Program Fellowship. $$ To whom correspondence should be addressed The Johns Hopkins University School of Medicine, Howard Hughes Medical Institute, 725 N. Wolfe St., 900 PCTB, Baltimore, MD 21205-2185. Tel.: 301-955-4050;Fax: 301-955-4857. The abbreviations used are: nAChR, nicotinic acetylcholine receptor; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; HPLC, high performance liquid chromatography; PTH, phenylthiohydantoin; GABA, 7-aminobutyric acid. gated ion channel that mediates signal transduction at the postsynaptic membrane of cholinergic synapses such as the neuromuscular junction and the electroplaque of electric fish. It is a pentameric complex composed of four homologous transmembrane subunits in the stoichiometry ap(3y6 (Galzi et al., 1991;Changeux et al., 1984). Recent studies have demonstrated that the nAChR is regulated in vitro and in vivo by serine and tyrosine phosphorylation (Huganir and Miles, 1989). Postsynaptic membranes of the Torpedo electric organ, rich in nAChR, contain endogenous protein kinases which phosphorylate the receptor as well as protein phosphatases which dephosphorylate the receptor (Gordon et al., 1977;Teichberg et al., 1977;Huganir and Greengard, 1983;Huganir et al., 1984;Mei and Huganir, 1991). Endogenous CAMP-dependent protein kinase phosphorylates the y and 6 subunits, while protein kinase C phosphorylates the 6 subunit (Huganir and Greengard, 1983;Safran et al., 1987). In addition, an unidentified protein tyrosine kinase has been demonstrated to specifically phosphorylate the (3, y, and 6 subunits to a high stoichiometry (Huganir et al., 1984;Hopfield et al., 1988).
The phosphorylation of the receptor by all three of these protein kinases appears to modulate the function of the receptor channel by increasing its rate of rapid desensitization (Huganir et al., 1986;Hopfield et al., 1988;Eusebi et ai, 1985). In addition, recent results have suggested that tyrosine phosphorylation of the receptor may be important for synapse formation. Tyrosine phosphorylation of the nAChR in the postsynaptic membrane has been shown to be dependent on neuronal innervation (Qu et al., 1990). The effect of innervation on phosphorylation appears to be mediated by agrin, a neuronally derived extracellular matrix protein, which has been shown to induce receptor aggregation underneath the nerve terminal (Wallace et al., 1991). These results suggest that tyrosine phosphorylation of the nAChR may play an integral role in the formation of the neuromuscular junction.
The location of the phosphorylation sites for each of the various protein kinases within the protein sequence of the nAChR subunits were originally proposed based on subunit specificity and known consensus sequence preferences of the different protein kinases (Huganir et al., 1984). All of the proposed phosphorylation sites are within the major intracellular loop of each subunit in close proximity to each other. The location of the CAMP-dependent protein kinase phosphorylation sites on the y and 6 subunits have been confirmed by protein sequencing techniques (Yee and Huganir, 1987). Recent studies using synthetic peptide substrates and sitespecific antibodies have strongly suggested that the protein of the Nicotinic Receptor 23785 kinase C sites are also contained within this region (Safran et al., 1987). In this paper, we have determined the tyrosine phosphorylation sites within the 8, y, and 15 subunits of the nAChR which are phosphorylated in vitro and in vivo by the endogenous protein tyrosine kinase. The primary structural requirements of protein tyrosine kinase substrates have begun to be defined by the use of synthetic peptide substrates and the analysis of protein tyrosine kinase autophosphorylation sites (Kemp and Pearson, 1987;Geahlen and Harrison, 1990). However, relatively few sites of physiological substrates have been determined. The identification of the sites of tyrosine phosphorylation of the nAChR may contribute to a better understanding of the structural determinants required for substrate recognition by protein tyrosine kinases. In addition, this may provide a better understanding of the molecular mechanisms involved in the regulation of receptor function by tyrosine phosphorylation.
Tyrosine Phosphorylation of the nAChR-Postsynaptic membranes, rich in the nAChR and in the endogenous protein tyrosine kinase which phosphorylates the receptor, were prepared from the electric organ of 2' . californica as previously described by Sobel et al. (1977), and modified by Qu et al. (1990). The nAChR was tyrosine phosphorylated in vitro by incubating 50 mg of postsynaptic membrane protein at 0.5 mg/ml in 20 mM Tris-HC1, pH 8.0,20 mM MgClZ, 2 mM MnCI2, 1 mM EGTA, 0.5 mM EDTA, 1 mM ouabain, 1 mM Na2V04, 100 p~ dithiothreitol, 1 p~ Walsh peptide, and 20 pg/ml each of antipain, leupeptin, and aprotinin. Phosphorylation was initiated with the addition of ATP: 40 mg of membrane protein were phosphorylated with 200 p~ ATP and 10 mg were phosphorylated with 200 p~ [-y-32P]ATP at 800 cpm/pmol in order to quantitate the in vitro phosphorylation and to follow the phosphorylated residues by Cerenkov counting throughout the experiment. The reaction was stopped after 45 min at 30 "C by placing the samples on ice.
Purification of Phosphorylated nAChR Subunits-Phosphorylated membranes were centrifuged at 114,000 X g , , at 4 "C for 20 min. The pellet was resuspended in one-half volume of 20 mM Tris-HC1, pH 8.0, 100 mM NaCl, 50 mM KCI, 1 mM EDTA, 1 mM EGTA and then solubilized for 30 min at 4 "C with 1% Triton X-100. Following a 20min centrifugation at 114,000 X g . , at 4 "C, the receptor was purified from the supernatant by acetylcholine affinity chromatography as described previously (Huganir and Racker, 1982). One milligram of pooled 32P-labeled and unlabeled phosphorylated receptor, at a ratio of approximately 1:4, was electrophoresed on two preparative 8% SDS-polyacrylamide gels (Laemmli, 1970), and transferred to nitrocellulose (Towbin et al., 1979). Regions of nitrocellulose containing the phosphorylated @, 7, and d subunits were detected by autoradiography and cut out in strips.
Enzymatic Cleavage of Electroblotted Subunits-Nitrocellulose strips were denatured with 2 M urea and 100 mM NH4HC03, reduced with 1.125 mM dithiothreitol and carboxyamidomethylated with 2.5 mM iodoacetamide as described by Stone et al. (1989). The subunits were then separately digested in situ as described by Aebersold et al. (1987), using a trypsin/subunit ratio of 1:20 (w/w) in 100 mM NH4HCO3 pH 8.2. At the end of the digestion any remaining peptide was eluted from the strips with 5% acetonitrile.
Puri/ication of Tyrosine Phosphopeptides-Approximately 15 nmol of each digested subunit in 0.5 ml was diluted 5-fold in 100 mM Tris-HCI, pH 8.0, and 1 mM phenylmethylsulfonyl fluoride. After removing residual nitrocellulose with a 0.45-pm filter, the peptide samples were boiled for 10 min to completely inactivate the trypsin. The samples were then cooled to room temperature and separately loaded onto 1.0-ml agarose-conjugated antiphosphotyrosine antibody affinity col-umns pre-equilibrated in 50 mM Tris-HC1, pH 7.4, 200 mM NaC1. After washing with 15 column volumes of 200 mM NHdHC03, the phosphotyrosine containing peptides were specifically eluted with 10 mM phenyl phosphate in the same buffer.
Phosphopeptides were repeatedly lyophilized to remove NHdHCO, and subsequently redissolved in 0.06% trifluoroacetic acid prior to purification by reverse phase HPLC. Final purification of samples containing 240-1000 pmol of peptides was accomplished using a Vydac C,, column (2.1 X 150-mm) and a buffer system of 0.06% trifluoroacetic acid and 0.056% trifluoroacetic acid/acetonitrile. The gradient was run from 0.0 to 80.0% acetonitrile at an increase of 0.625%/min with a flow rate of 0.2 ml/min. UV absorbance peaks (214 nm) were collected manually and fractions containing the 32Plabeled phosphopeptides were detected by Cerenkov counting and pooled for sequencing.
Seqwnce Anulysis of Phosphopeptides-Sequence analyses at the Marine Biological Laboratory was performed with a sequenator provided by Porton Instruments and at the Biopolymer Laboratory of the Howard Hughes Medical Institute at the Johns Hopkins University School of Medicine were performed with an Applied Biosystems 477A Protein Sequenator. Samples containing 140-780 pmol of phosphopeptide were subjected to automated Edman degradation and the PTH-derivatives were identified and quantitated by reverse phase HPLC.
Phosphotyrosine Immunoblotting-The stoichiometry of phosphorylation on each of the subunits was determined through quantitative immunoblotting using phosphotyrosine antibodies. The subunits of endogenously phosphorylated receptor and receptor phosphorylated in vitro with nonradioactive ATP were separated using 8% SDS-polyacrylamide gels (Laemmli, 1970) and transferred to nitrocellulose (Towbin et al., 1979). Immunoblots were probed with 1:200 dilution of affinity purified polyclonal antiphosphotyrosine antibodies (Hirano et al., 1988) and '*'I-protein A as described by Jahn et al. (1984). To convert '251-protein labeling of the endogenously phosphorylated nAChR and the nAChR phosphorylated in vitro with nonradioactive ATP to moles of phosphotyrosine/mol of subunit, the immunoblotted subunits were compared to receptor subunits phosphorylated in vitro with [-y-32P]ATP using the following formula: (counts/min of "'I-protein A labeling of subunit phosphorylated in vitro) -(counts/min of '"1-protein A labeling of nonphosphorylated subunit) per mol of [32P]phosphate incorporated in vitro.
Phosphoamino Acid Analysis-Phosphorylated subunits were acid hydrolyzed and subjected to one-dimensional thin layer electrophoresis essentially as described by Hirano et al. (1988).
Two-dimensional Phosphopeptide Maps-Phosphorylated subunits were digested with thermolysin (0.15 mg/ml), applied to thin layer chromatography plates, and subjected to electrophoresis in the first dimension followed by ascending chromatography in the second dimension (Huganir et al., 1984).
Protein Determination-Protein concentration was determined by the bicinchoninic acid protein assay kit purchased from Pierce Chemical Co.

RESULTS
Postsynaptic membrane preparations isolated from T. californica electric organ are enriched in the nAChR and contain high levels of protein tyrosine kinase activity Huganir e t al., 1984). The endogenous protein tyrosine kinase specifically phosphorylates the nAChR on the p, y, and 6 subunits in vitro and in vivo (Fig. 1). Under the conditions described the in vitro phosphorylation occurs exclusively on tyrosine residues (Fig. 2). Two-dimensional peptide maps of the 8, y, and 6 subunits digested with thermolysin showed that each contained a single distinct phosphopeptide (Fig. 3). I n vitro phosphorylation occurs to a final maximal stoichiometry of 0.43-0.60 mol of 32P incorporated/mol of subunit.
To determine the level of endogenous tyrosine phosphorylation of the nAChR, a quantitative immunoblotting technique with an antiphosphotyrosine antibody was used. The AChR is phosphorylated in vivo to a stoichiometry of 0.1-0.5 mol of phosphate/mol of subunit depending on the preparation. Thus, in vitro phosphorylation, together with the initial level of endogenous tyrosine phosphorylation of the receptor, resulted in a final stoichiometry of 0.86 mol of phosphate/mol   of p subunit, 0.41 mol of phosphate/mol of y subunit, and 0.88 mol of phosphate/mol of 6 subunit (Fig. 1).
To determine the site of tyrosine phosphorylation on each subunit, 1 mg of purified tyrosine-phosphorylated nAChR was prepared. Phosphorylated subunits of the purified receptor were separated by SDS-PAGE and transferred to nitrocellulose. The subunits were then subjected to limit digestion in situ with trypsin. Nearly 100% of the phosphopeptides were recovered from the nitrocellulose as judged by Cerenkov counting.
The phosphotyrosine containing peptide of each subunit was purified by affinity chromatography on an antiphosphotyrosine antibody column, followed by HPLC using a CIR column. After inactivation of the trypsin, digested subunits were individually applied to an agarose-conjugated antiphosphotyrosine antibody affinity column. After removing other tryptic peptides in the wash, 7040% of the B and 6 phosphopeptides were specifically recovered in the phenyl phosphate eluate. The antiphosphotyrosine antibodies appear to have a lower affinity for the y subunit phosphopeptide, which bound weakly to the affinity column and eluted during the 15-column volume wash. The washes containing y phosphopeptide, as detected by Cerenkov counting, were repeatedly reapplied to phopeptide. The 6 phosphopeptide was purified from other tryptic peptides on an agarose-conjugated antiphosphotyrosine antibody affinity column, eluted with phenyl phosphate, and isolated bv HPLC using a CIR column. The single major peak of UV absorbance cneluted with >90% of the radioactivity (eluting at 17% acetonitrile). The fraction corresponding to this peak was collected and suhjected to sequence analysis. the column and the final fractions containing the phosphopeptide were pooled.
The phosphopeptides were then separated from the phenyl phosphate and trace levels of other tryptic peptides by reverse phase chromatography. There was one major peak of UV absorbance for each subunit which corresponded to the peak of radioactivity as shown for the 6 subunit in Fig. 4. Similar column profiles were seen with the /3 and y phosphopeptides.
The presence of a single phosphotyrosine containing peptide in each subunit suggests that the in uitro and in uiuo tvrosine phosphorylation sites are the same. For the subunit, 67% of the radioactivity from the antiphosphotyrosine antibodv column eluted in a major peak at 17% acetonitrile on HPLC. Seventy-four percent of the radioactivity from the y subunit were recovered a t 14% acetonitrile and 71?h of the radioactivity from the 6 subunit were recovered at 17% acetonitrile.
The peak from the reverse phase HPLC purification for each tryptic phosphopeptide was judged to be pure by peptide sequencing. Sequence analysis of the peptides identified the residues shown in Table I. The sequence obtained for the / 3 subunit tryptic phosphopeptide corresponds uniquely to residues 351-358 of the deduced amino acid sequence (Noda et al., 1983a). The y subunit sequence corresponds to residues 361-367 and the 6 subunit sequence corresponds to residues 369-376 (Fig. 5) (Claudio et al., 1983;Noda et al., 1983aNoda et al., , 1983b. All of these regions lie within the major intracellular loop of each subunit and adjacent to sites for serine phosphorylation. Phosphoryl groups bind tightly to the treated glass-fiber filter used in the sequence cartridge and therefore the PTH-derivative of phosphotyrosine does not appear in any of the sequenator cycles. From comparison of the peptide sequences with the sequences deduced from cDNA, it is clear that the blank cycles in each subunit correspond to the position of the phosphorylated tyrosine residue. We conclude then that the tyrosine phosphorylation site in the p subunit is Y-355, the y subunit is Y-364, and the 6 subunit is Y-372.
The yield of PTH-derivatives obtained from each cycle of the sequential degradation was quantified. The initial yield of each peptide was larger than the yield predicted from the calculation based on the in vitro phosphorylation stoichiometry. This suggests that the endogenously phosphorylated peptides co-purified with the in vitro phosphorylated peptides and is additional evidence that the in uiuo and the in vitro phosphorylation sites are identical.

TABLE I
Amino acid sequence of tryptic phosphopeptides of the @, y, and 6 subunits Tryptic phosphopeptides purified by affinity chromatography and HPLC were subjected to automated Edman degradation. PTH-derivatives were identified and quantitated by reverse phase HPLC. The PTH-derivative of phosphotyrosine did not elute from the treated glass-fiber filter used in the sequence cartridge and its position is identified by a blank cycle.

I S R A N D E Y F I R K m GABAAR,ylsubunit
FIG. 5. Identified and proposed sites of tyrosine phosphorylation in the major intracellular loops of ligand-gated ion channels. The sequences surrounding the sites of tyrosine phosphorylation on the @, y , and 6 subunits of the nAChR which are located in each of the subunits' major intracellular loops are similar to potential tyrosine phosphorylation sites in homologous domains of the GABAA and glycine receptor subunits.

DISCUSSION
This paper has described the determination of the sites of tyrosine phosphorylation on the p, y, and 6 subunits of the nAChR. The receptor was phosphorylated selectively on tyrosine residues, purified over an acetylcholine affinity column, and its subunits isolated by SDS-PAGE. The phosphorylated p, y, and 6 subunits were individually digested and shown to each contain only one phosphopeptide after trypsin or thermolysin cleavage. The tryptic phosphopeptides were purified over an agarose-conjugated antiphosphotyrosine affinity column and then by reverse phase HPLC. The sequences obtained from the purified phosphopeptides were: Ala-Asn-Asp-Glu-Tyr(P)-Phe-Ile-Arg for the p subunit, Ala-Glu-Glu-Tyr(P)-Ile-Leu-Lys for the y subunit, and Ala-Gln-Glu-Tyr(P)-Phe-Asn-Ile-Lys for the 6 subunit. These sequences correspond uniquely to residues 351-358 of the / 3 subunit, 361-367 of the y subunit, and 369-376 of the 6 subunit, which are all homologous regions within the major intracellular loops of the integral membrane subunits (Noda et al., 1983a(Noda et al., , 1983b.
The sites for in vivo and in uitro tyrosine phosphorylation of the nAChR are identical. The receptor used in this study was phosphorylated on tyrosine residues in uiuo at approximately 1 mol of phosphate/mol of receptor as determined.by quantitative immunoblotting. In addition to this endogenous level of phosphorylation, the receptor was phosphorylated i n uitro to a final level of 2 mol of phosphate/mol of receptor.
However, a single tryptic phosphopeptide was obtained from the affinity column for each subunit as judged by the coelution of single peaks of absorption and radioactivity observed during reverse phase chromatography. Based on initial yields from the sequencing runs, the amount of phosphopeptide recovery was greater than that estimated from values obtained using the i n vitro phosphorylation stoichiometry. These observations indicate that tryptic phosphopeptides from receptor phosphorylated i n uiuo co-purified with those derived from in uitro phosphorylation and were identical by sequence analysis.
The sequence surrounding the tyrosine phosphorylation sites of the nAChR subunits are similar to those of most of the other known substrates of protein tyrosine kinases, with acidic amino acids immediately preceding the tyrosine residue (Kemp and Pearson, 1990;Geahlen and Harrison, 1990). Acidic amino acids adjacent to the phosphorylated tyrosine residues are found at the autophosphorylation sites of most protein tyrosine kinases, at the sites of tyrosine phosphorylation of in vivo substrates such as phospholipase C-y and erythrocyte Band 3, and in effective synthetic peptide substrates for protein tyrosine kinases such as poly(Glu,Tyr),,. (Hanks et al., 1988;Kim et aL, 1990;Wahl et al., 1990;Dekowski et al., 1983;Braun et al., 1984). The phosphorylation sites of the three subunits of the nAChR also share additional features, including hydrophobic residues followed by basic residues on the COOH-terminal side of the tyrosine residue. However, these similarities may represent shared features of very homologous proteins rather than primary structural determinants that are important for site recognition by the protein tyrosine kinase. Secondary or tertiary structural determinants may play an important role in substrate recognition by the endogenous protein tyrosine kinase which phosphorylates the nAChR subunits. Short synthetic peptides which lack higher order structural determinants have been reported to be poor substrates for protein tyrosine kinases (Geahlen and Harrison, 1990). Similarly, synthetic peptides corresponding to the region phosphorylated on the nAChR are inefficiently phosphorylated by the endogenous protein tyrosine kinase with K,,, values in the millimolar range (data not shown). Thus, secondary and tertiary structural determinants are likely to be important for the recognition of substrates by protein tyrosine kinases.
The sites of tyrosine phosphorylation of the nAChR are conserved in the nAChR from all species examined except for the y subunit, whose tyrosine phosphorylation site is present only in Torpedo and chicken (Huganir and Miles, 1989). This suggests that tyrosine phosphorylation plays an integral role in the regulation of the function of the receptor. Recent studies indicate that tyrosine phosphorylation of the receptor may be important during development of the neuromuscular junction. Innervation with chick myotubes in culture by ciliary ganglion neurons and innervation of rat muscle during development induce tyrosine phosphorylation of the nAChR (Qu et al., 1990).' This nerve-induced phosphorylation appears to be mediated by agrin, a factor released from the neuron, which is known to cause aggregation of the receptor in myotubes (Wallace, 1986;Nitkin et al., 1987). Agrin has recently been shown to specifically induce tyrosine phosphorylation of the (3 subunit of the nAChR in chick myotubes, suggesting that tyrosine phosphorylation of the nAChR is directly involved in aggregation of the nAChR under the nerve terminal (Wallace et al., 1991).
Phosphorylation may be a common mechanism of modulating ligand-gated ion channels. Similar to the nAChR, other ligand-gated ion channels are multimeric structures composed of homologous subunits with several membrane-spanning domains (Barnard et al., 1987). The GABAA receptor has been shown to be phosphorylated by CAMP-dependent protein kinase and protein kinase C in uitro and the glycine receptor is phosphorylated by protein kinase C in vitro (Kirkness et al., 1989;Browning et al., 1990;Ruiz-Gomez et al., 1991). In addition, the y1 and y2 subunits of the GABAA receptors and the (3 subunit of the glycine receptor have a tyrosine residue surrounded by acidic amino acids in their major intracellular loops (Fig. 5) (Ymer et al., 1990;Pritchett et al., 1989;Grenningloh et al., 1990). The similarity of these sites to the tyrosine phosphorylation sites within the nAChR subunits suggests that these ligand-gated ion channels may also be substrates of protein tyrosine kinases and tyrosine phosphorylation may be a general mechanism in the regulation of neurotransmitter receptor function.