Cell Growth & Differentiation 3 Identification of Autophosphorylation Sites of HER2/neu R. Hazan, B. Margolis,’ M. Dombalagian, A. Ullrich, consensus tyrosine in the kinase domain is phosphory- A. Zilberstein, and J. Schlessinger2 lated. Site-directed mutagenesis of this site impairs the Rorer Biotechnology, Inc., King of Prussia, Pennsylvania 19406 [R. H., B. ability of autophosphorylation to activate the kinase (7, M., M. D., A. Z., J. 5.], and Max-Planck-Institut f#{252}rBiochemie, 8033 8). In the EGFR, this consensus tyrosine has not been Martinsreid bei MOnchen, Federal Republic of Germany [A. U.] found to be a site of autophosphorylation.4 It has been found that all autophosphorylation sites of the EGFR are concentrated in the carboxyl-terminal domain of the Abstrad receptor outside of the consensus kinase domain (9, 10). HER2 or c-erbB-2 is a putative growth fador receptor In the EGFR, it has been postulated that the autophos- with sequence homology to the epidermal growth fac- phorylation sites act as competitive inhibitors of exoge- tor receptor. It is the human homologue of the rat nous substrate phosphorylation (6, 1 1) and it has been protooncogene neu and may have an important role in demonstrated that mutagenesis of the individual auto- human malignancies such as breast and ovarian can- phosphorylation sites increases the receptor affinity for cers. Like other growth fador receptors, HER2 has in- exogenous substrates in vitro (6). Additionally, cells which trinsic protein tyrosine kinase adivity and undergoes express these mutated receptors demonstrate an in- autophosphorylation. Recently, we have demonstrated creased mitogenic sensitivity to EGF (12). Thus, intrinsic that, similar to the epidermal growth fador receptor, autophosphorylation sites may compete with exogenous all autophosphorylation sites of HER2 are localized in substrate phosphorylation at low levels of receptor acti- the carboxyl terminus of this protein. In the present vation. study, immunopurified HER2 was allowed to autophos- HER2 or c-erbB-2 is the human homologue of the rat phorylate, and tryptic phosphopeptides were gener- protooncogene neu (13-15), which has great sequence ated. After purification of these phosphopeptides by identity with the EGFR but for which the natural ligand high performance liquid chromatography, microsequen- is not known. HER2 is overexpressed in many breast and cing was performed. Utilizing this approach, two au- ovarian cancers and thus may play an important role in tophosphorylation sites were unequivocally identified human carcinogenesis (16). Recently, we have demon- at Y1023 and Y1248. The sequences of two other tyro- strated that like EGFR, all of the autophosphorylation sine phosphorylated tryptic peptides were determined, sites of the HER2 are located in the carboxyl terminus of but the exad site of autophosphorylation could not be the receptor (10). This was demonstrated utilizing a chi- determined because multiple tyrosines were located on meric receptor where the carboxyl terminus of the EGFR each peptide. However, each of these peptides contains was replaced by the carboxyl terminus of the HER2 tyrosines that correspond to major autophosphorylation receptor. It was found that the autophosphorylation sites sites of the epidermal growth fador receptor, suggest- of this EGFR/HER2 chimera were identical to the auto- ing that, in addition to Y1023 and Y1248, Yl 139 and phosphorylation sites of the wild-type HER2 receptor, Yl 222 also serve as autophosphorylation sites of HER2. indicating that the carboxyl terminus contains all of the autophosphorylation sites. In the present study, we ex- Introdudion tend this work by purifying and sequencing the tryptic peptides that contain the autophosphorylation sites of Many growth factors bind to cell surface receptors with the HER2 receptor. We have unequivocally identified intrinsic protein tyrosine kinase activity (reviewed in Refs. two autophosphorylation sites by direct sequencing and 1-3). Upon ligand binding, the intracellular tyrosine ki- demonstrated that two phosphopeptides contain tyro- nase is activated, resulting in tyrosine phosphorylation of sine residues which correspond to two major autophos- substrates. One of the first substrates to be phosphory- phorylation sites of EGFR. It appears, therefore, that, lated is tyrosine residues in the carboxyl terminus of the except for one site, autophosphorylation sites of the receptor, resulting in receptor autophosphorylation (re- HER2 receptor are similar to those found in the EGFR. viewed in Refs. 1 -3). The exact role of receptor auto- phosphorylation in growth factor signal transduction is still being investigated and appears to vary for different Results receptors. For example, autophosphorylation has been The HER2 receptor was immunoprecipitated and auto- demonstrated to increase the Vmax of the insulin receptor phosphorylated in vitro and then purified by SDS-poly- kinase (4), but not the EGFR3 kinase (5, 6). In the insulin acrylamide gel electrophoresis. The gel slice with the receptor (and several nonreceptor tyrosine kinases) a HER2 receptor contained approximately 1 nmol of phos- phate, as determined by Cerenkov counting. Fig. 1 dem- onstrates the elution pattern of the phosphorylated tryp- tic peptides prepared and separated by reverse phase Received 9/25/89. HPLC at pH 2.0. Pools of fractions 7-9 (peptide A), 1 1 1 Fellow of the Medical Research Council of Canada. 2 To whom requests for reprints should be addressed, at Rorer Bio- and 12 (peptide B), 25 and 26 (peptide C), and 27-29 technology, Inc., 680 Allendale Rd., King of Prussia, PA 19406. 3 The abbreviations used are: EGFR, epidermal growth factor receptor; EGF, epidermal growth factor; SDS, sodium dodecyl sulfate; HPLC, high performance liquid chromatography; PLC-y, phospholipase C-y. 4 5. Bianchi and J. schlessinger, unpublished observation. 4 HER2 Autophosphorylation Sites Discussion We have utilized standard tryptic peptide purification and sequencing to identify the in vitro autophosphory- B. lation sites of the HER2 protein. Two sites were unequiv- ocally assigned as Y1023 and Y1248. Two other sites 4, 4, were localized to specific tryptic peptides, but the exact residue could not be assigned due to multiple tyrosines C in these peptides. However, tyrosine residues located in these later peptides correspond to major autophosphor- ylation sites of the EGFR, suggesting that they also serve i as autophosphorylation sites. It is noteworthy that the autophosphorylated HER2 protein was generated by an in vitro reaction, and it is possible that additional or other sites are phosphorylated in vivo. However, quantitative comparisons of phosphopeptide maps of EGFR and 30 40 50 HER2 after either in vitro or in vivo autophosphorylation S ACETONITRILE indicated that similar sites were phosphorylated for each of the two receptors under both in vitro and in vivo Fig. 1. HPLC separation of tryptic phosphopeptides from HER2. Auto- conditions. It is likely, therefore, that the in vitro auto- phosphorylated HER2 protein from transfected NIH/3T3 cells (inset) and phosphorylation sites described in this study represent HPLC separation of phosphopeptides. HER2 receptors were immunopre- the same sites which are phosphorylated in living cells. cipitated from transfected NIH/3T3 cells and allowed to autophosphor- These results are consistent with our previous results ylate in the presence of [.‘2PJATP The phosphorylated receptor was purified by 5D5-polyacrylamide gel electrophoresis, and the gel slice demonstrating that all autophosphorylation sites of HER2 containing the receptor was excised. After washing with 10% methanol, are localized in the carboxyl terminus of the receptor. the gel slice was dried and treated three times with 5 ig trypsin. The EGFR and HER2 are highly homologous in the kinase tryptic peptides were applied to a C, reverse phase column and eluted with increasing acetonitrile concentration at pH 2.0 as described in domain but have only 35% sequence identity in the Materials and Methods. Fractions of 0.5 ml were collected, and Cer- carboxyl terminus (11, 12). Nonetheless, three ofthe four enkov counts were determined. Arrows, the major fractions that were tyrosine autophosphorylation sites of EGFR are also con- subjected to further purification. served. All of these HER2 tyrosine residues were located on phosphorylated tryptic peptides, strongly suggesting that these sites of autophosphorylation are highly con- (peptide D) were selected as representing the major served between these two receptor molecules. However, peaks and were further purified in preparation for se- one site (Y1086) is phosphorylated in EGFR but not in quencing. HER2, and one site (Y1023) is phosphorylated in HER2 Fig. 2 represents the elution pattern of radioactive but not in EGFR (Fig. 3). peptides in the second HPLC purification at pH 6.5. The The role of autophosphorylation in HER2 signal trans- largest radioactive peak from each fraction was directly duction is unknown. In EGFR, tyrosine autophosphory- applied to the gas phase sequencer, and the sequences lation sites have been shown to act as competitive inhib- obtained are shown in Fig. 2. Tryptic peptides B and C itors of exogenous substrate phosphorylation (6), but this contain only one tyrosine each, Y1248 and Y1023, re- may not be their only function. The exact location of the spectively, indicating that these must be sites of auto- autophosphorylation sites in the carboxyl terminus would phosphorylation.