Expression of human tyrosine kinase-negative epidermal growth factor receptor amplifies signaling through endogenous murine epidermal growth factor receptor.

Recent findings have suggested that certain ligand-dependent responses to EGF may be propagated in a manner that is not dependent on the intrinsic tyrosine kinase activity of the epidermal growth factor receptor (EGF-R, Campos-Gonzalez, R., and Glenney, J. R., Jr. (1992) J. Biol. Chem. 267, 14535-14538) or, alternatively, that these responses may occur through the interaction of the human tyrosine kinase-deficient EGF-R with an as yet unidentified kinase (Selva, E., Raden, D. L., and Davis, R. J. (1993) J. Biol. Chem. 268, 2250-2254). These conclusions represent a significant departure from our current understanding of signal transduction by receptor tyrosine kinases. Therefore we examined the effect of expression of tyrosine kinase-negative human EGF receptor in murine NIH-3T3-2.2 cells on the EGF-dependent phosphorylation of mitogen-activated protein (MAP-2) kinase. In parental cells (NIH-3T3-2.2) that express low levels of endogenous murine EGF-R, there was no demonstrable EGF-dependent coupling to MAP-2 kinase. In NIH-3T3-2.2 cells transfected with tyrosine kinase-negative human EGF-R, there was unexpected EGF-dependent phosphorylation of MAP-2 kinase. Analysis of the tyrosine kinase-negative human EGF-R in these cells revealed significant tyrosine phosphorylation of the EGF-R. A low level of endogenous murine EGF-R present in these cells were also phosphorylated on tyrosine residues and displayed autokinase activity. Similar results were obtained using an unrelated cell line (B82L cells), in which EGF-dependent phosphorylation of MAP-2 kinase was previously attributed to signal propagation through a tyrosine kinase-negative human EGF-R (Campos-Gonzalez, R., and Glenney, J. R., Jr. (1992) J. Biol. Chem. 267, 14535-14538). Taken together, these results suggest that the tyrosine kinase-negative human EGF-R are able to amplify the response to activation of low levels of endogenous murine EGF-R, thus leading to EGF-dependent phosphorylation of MAP-2 kinase in cells expressing tyrosine kinase-negative human EGF-R.

Recent findings have suggested that certain liganddependent responses to EGF may be propagated in a manner that is not dependent on the intrinsic tyrosine kinase activity of the epidermal growth factor receptor (EGF-R, Campos-Gonzalez, R., and Glenney, J. R., Jr. (1992) J. Biol. Chem. 267,[14535][14536][14537][14538] or, alternatively, that these responses may occur through the interaction of the human tyrosine kinase-deficient EGF-R with an as yet unidentified kinase (Selva, E., Raden, D. L., and Davis, R. J. (1993) J. Biol. Chem. 268,[2250][2251][2252][2253][2254]. These conclusions represent a significant departure from our current understanding of signal transduction by receptor tyrosine kinases. Therefore we examined the effect of expression of tyrosine kinase-negative human EGF receptor in murine NIH-3T3-2.2 cells on the EGF-dependent phosphorylation of mitogen-activated protein (MAP-2) kinase. In parental cells (NIH-3T3-2.2) that express low levels of endogenous murine EGF-R, there was no demonstrable EGF-dependent coupling to MAP-2 kinase. In NIH-3T3-2.2 cells transfected with tyrosine kinasenegative human EGF-R, there was unexpected EGFdependent phosphorylation of MAP-2 kinase. Analysis of the tyrosine kinase-negative human EGF-R in these cells revealed significant tyrosine phosphorylation of the EGF-R. A low level of endogenous murine EGF-R present in these cells were also phosphorylated on tyrosine residues and displayed autokinase activity. Similar results were obtained using an unrelated cell line (B82L cells), in which EGF-dependent phosphorylation of MAP-2 kinase was previously attributed to signal propagation through a tyrosine kinase-negative human EGF-R (Campos-Gonzalez, R., and Glenney, J. R., Jr. (1992) J. Biol. Chem. 267,[14535][14536][14537][14538].
Taken together, these results suggest that the tyrosine kinase-negative human EGF-R are able to amplify the response to activation of low levels of endogenous murine EGF-R, thus leading to EGF-dependent phosphorylation of MAP-2 kinase in cells expressing tyrosine kinase-negative human EGF-R. The human epidermal growth factor receptor (EGF-R)' is comprised of a single 170-kDa glycosylated polypeptide, which forms an extracellular ligand binding domain, an intracellular tyrosine kinase and autophosphorylation domain, and a connecting transmembrane sequence (1,2). EGF-R exemplifies a class of ligand-stimulated receptor tyrosine kinases that serve both as activators and as targets for phosphorylation by kinases involved in mitogenic signaling (1-3). The binding of EGF to the extracellular domain results in an increase in tyrosine kinase activity and autophosphorylation by a mechanism involving receptor dimerization (Ref. 1 and references therein). A compendium of evidence has accumulated suggesting that the tyrosine kinase activity of the EGF-R is essential for many of the biochemical events that follow receptor activation (1,2,4,5). Thus, mutations inhibiting the tyrosine kinase activity of the EGF-R have been reported to block signal transduction by the receptor, thereby leading to the hypothesis that the receptor tyrosine kinase directly mediates the process of signaling by the EGF-R (4,5). However, recent studies have suggested the possibility that certain ligand-dependent responses of receptor tyrosine kinases might occur through a mechanism independent of the intrinsic tyrosine kinase activity of the receptor (6)(7)(8)(9)(10)(11).
One such response is phosphorylation and activation of members of the mitogen-activated protein (MAP) kinase family (also known as ERKs; Ref. 12). These are a family of predominantly serinelthreonine kinases, which are phosphorylated on both tyrosine and serinelthreonine residues and activated in response to occupancy of the EGF-R and other receptor tyrosine kinases, as well as in response to other mitogens (e.g. tumor promoting phorbol esters). In one recent study (6), EGF-dependent phosphorylation of MAP-2 kinase was demonstrated in B82L cells that had been transfected with a human EGF-R defective in intrinsic tyrosine kinase activity by virtue of the substitution of methionine in place of lysine at position 721 in the cytoplasmic ATP binding domain of the receptor. The possibility of a pathway for propagation of a ligand-dependent signal directly through a tyrosine kinase-deficient EGF-R or, alternatively, through an interaction with another kinase potentially has profound implications for our overall understanding of growth factor action and represents a significant departure from previous studies regarding receptor tyrosine kinase signaling. Accordingly, we sought to clarify whether signal propagation resulting in phosphorylation of MAP-2 kinase could result from an 'The abbreviations used are: EGF-R, epidermal growth factor receptor; ECL, enzyme chemiluminescence; EGF, epidermal growth factor; mAb, monoclonal antibody; MAP, mitogen-activated protein; ERK, extracellular regulated kinase; PAGE, polyacrylamide gel electrophoresis; DMEM, Dulbecco's modified Eagle's medium. amplification of the ligand-dependent signal emanating from a small population of endogenous kinase-competent murine receptors as a result of the transfection of kinase-negative human EGF-R.

EXPERIMENTAL PROCEDURES
Materials-EGF (receptor-grade) and anti-phosphotyrosine antibodies 4G10 and 4G10 coupled to agarose were purchased from Upstate Biotechnology (Lake Placid, NY). Protein A coupled to Sepharose CL-4B was from Pharmacia LKB Biotechnology Inc. &"PO4 (1000 mCi/mmol), [-y-"P]ATP (4500 Ci/mmol), and antiphosphotyrosine antibody (PY20) were purchased from ICN. DMEM, calf serum and tissue culture reagents were purchased from Life Technologies, Inc. Antibodies linked to alkaline phosphatase or horseradish peroxidase and prestained molecular weight markers were obtained from Bio-rad. ECL developer was from Amersham Corp.
Cell Culture and Stimulation-NIH-3T3 clone 2.2 cells and K721A cells were generously provided by Dr J. Schlessinger (Department of Pharmacology, New York University Medical Center, New York, NY) and have been described previously (4,13). NIH-3T3-2.2 cells were shown to express undetectable amounts of endogenous murine EGF-R as assessed by ligand binding (4,13). K721A cells were derived from NIH-3T3-2.2 cells and were transfected with a mutated construct of the human EGF-R such that lysine 721 in the ATP binding fold was replaced with alanine, thus eliminating its tyrosine kinase activity (4,13). All cell types were maintained in DMEM containing 25 mM glucose supplemented with 50 units/ml penicillin, 50 pg/ml streptomycin, and 10% calf serum (Life Technologies, Inc.) in 5% COz in air at 37 "C. Cells were plated in DMEM supplemented with 25 mM glucose and 10% calf serum at a density of 300,000 cells/lOcm dish. At 48 h the medium was replaced with DMEM containing 25 mM glucose and 0.5% calf serum. Experiments were conducted after 72 h of culture, at which time the cells were 90% confluent (14)(15)(16). Cells were washed three times in serum-free DMEM containing 25 mM glucose supplemented with 2 mg/ml bovine serum albumin. Following a 3-min stimulation with EGF (100 nM) or vehicle (50 p~ acetic acid), cells were solubilized in SDS sample buffer containing 200 p~ sodium orthovanadate and analyzed for phosphotyrosinecontaining proteins or MAP-2 kinase as detailed below. Parental B82L cells, B82L cells transfected with wild type human EGF-R, and B82L cells transfected with M721 tyrosine kinase-negative mutant construct of the human EGF-R (M721 cells) were kindly provided by Dr. H. Elsholtz (C. H. Best Institute, Toronto, Canada) from Dr. G. Gill (University of California, San Diego, CA). These cells were grown in DMEM plus 10% calf serum as previously described (5,17) and experiments performed during passages 1-10 in DMEM.
Immunoprecipitation-Following stimulation, phosphotyrosinecontaining proteins were immunoprecipitated using the anti-phosphotyrosine antibody 4G10 coupled to agarose (Upstate Biotechnology). Briefly cells were lysed on ice in 1 ml of 20 mM Tris, pH 8.0, 150 mM NaCl, 10% glycerol, and 1% Nonidet p-40 containing, 1 mM phenylmethylsulfonyl fluoride, 0.15 unit/ml aprotinin, and 1 mM sodium orthovanadate (lysis buffer). Insoluble material was removed by centrifugation at 4 "C for 10 min at 10,000 X g. Supernatants containing equal amounts of protein (18) were transferred to fresh tubes and immunoprecipitated with 4G10 (5 pg) anti-phosphotyrosine antibodies coupled to agarose by rocking at 4 "C for 4 h. The antiphosphotyrosine immunoprecipitates were collected by centrifugation for 15 s at 10,OOO X g and washed three times with lysis buffer. The immunoprecipitates were next washed five times in lysis buffer containing 0.5% Nonidet p-40 and finally with lysis buffer without Nonidet p-40. Samples were then heated in SDS sample buffer, separated by electrophoresis, and analyzed by immunoblotting. Immunoprecipitation analysis of the EGF-R with either mAb 108 or RK2 generously provided by Dr. J. Schlessinger was performed as described by Margolis et al. (19) with the modifications noted below. Where indicated, 1 mM sodium orthovanadate replaced the phosphatase inhibitor mixture of p-nitrophenyl phosphate, sodium fluoride, sodium pyrophosphate, and 200 p~ sodium orthovanadate. When immunoprecipitation was performed using a polyclonal antibody, all lysates were precleared by incubation for 90 min with protein A-Sepharose coated with normal rabbit IgG. Lysates were then used in immunoprecipitation following removal of the protein A-Sepharose antibody complex by centrifugation.
Immunostaining-Immunoprecipitated proteins were separated, analyzed by SDS-PAGE, transferred to Immobilon, and immuno-stained with the anti-phosphotyrosine antibody (4G10 provided by Drs. B. Druker and T. Roberts, Dana-Farber Cancer Institute or PY20 from ICN), a polyclonal antibody to MAP-2 kinase (antibody 691 generously provided by Dr. M. Cobb, University of Texas, Southwestern Medical Center, Dallas, T X Ref. 20) or a polyclonal antibody to the human EGF-R raised to the peptide sequence 984-996 (RK2 provided by Dr. J. Schlessinger,Ref. 21). Blots were incubated with goat anti-rabbit antibody linked to alkaline phosphatase and developed as reported by Hack et al. (14) or with goat anti-rabbit antibody linked to horseradish peroxidase followed by ECL development. In order to reduce background, blots developed in ECL were washed extensively in TBS buffer containing 0.3% Tween 20.
In Vitro Kinase Assay-Cells from 15-cm dishes were grown to 90% confluence and stimulated with equal volumes of either vehicle (50 pM acetic acid, final) or EGF (100 nM, final) as described above. Following stimulation, cells were solubilized in 1 ml of lysis buffer containing 10% glycerol, 1% Triton X-100, 10 mM EGTA, 50 mM NaF, 10 pg/ml leupeptin, 25 mM Hepes, pH 7.4, and 150 mM NaCl. Insoluble material was removed by centrifugation at 100,000 X g in a Beckman tabletop ultracentrifuge for 30 min at 4 "C. Supernatants were then equalized for protein (18), and 1.0 ml containing the indicated amounts of protein was added to 20 p1 of protein A-Sepharose prebound to a monoclonal anti-EGF-R antibody mAb 108 or RK2 as previously reported (19). Following incubation for 90 min at 4 "C, the immunoprecipitates were washed twice in lysis buffer and once in HNTG (consisting of 50 mM, Hepes, pH 7.4, 10% glycerol, 150 mM NaCl, 0.1% Triton X-100, and 5 mM MgCl,). The immunoprecipitates were then incubated for 15 min at 22 'C in 0.5 ml of HNTG containing 100 nM EGF. The beads were then pelleted and resuspended in 30 pl of HNTG containing 100 nM EGF at 4 "C. Autophosphorylation was initiated by the addition of 10 pl of 20 p~ [-y-"P]ATP (50 Ci/mmol). Following incubation for 15 min, phosphorylation was terminated with 1 ml of HNTG containing 10 mM EDTA and without MgC12. The immunoprecipitates were recovered by centrifugation and solubilized in 100 pl of SDS sample buffer by heating for 5 min in a boiling water bath. Solubilized samples were analyzed by SDS-PAGE in a 6.0% gel, dried, and exposed to Kodak X-Omat XAR film. Following autoradiography the gel was subjected to phosphor image analysis (see below).
Quuntitation of Phosphorylated EGF Receptors-The dried gel was exposed to a photostimulable storage phosphor imaging plate in a Molecular Dynamics Inc. (Sunnyvale, CA) exposure cassette for 2 h at room temperature and scanned in a Molecular Dynamics 400A PhosphorImager using a 10-milliwatt helium-neon laser. Luminescence at 390 nm was collected, digitized, and stored on disk. A digital image was generated, and the bands of interest were quantified using Molecular Dynamics Image Quant software. Storage phosphor imaging plates have a large linear dynamic range and a high sensitivity that allows accurate quantitative measurements (22).

RESULTS
The pattern of EGF-dependent phosphorylation in parental NIH-3T3-2.2 and K721A cells is shown in Fig. la. No evident EGF-dependent tyrosine phosphorylation was detected in NIH-3T3-2.2 cells, consistent with a deficiency of active endogenous murine EGF-R. In contrast, EGF-dependent phosphorylation of a 42-kDa protein was evident in the K721A cells. Similar results were obtained using several different anti-phosphotyrosine antibodies including 4G10 and PY20; moreover, the phosphorylation of the 42-kDa protein was also evident following stimulation by phorbol myristate acetate (data not shown). This suggested that the 42-kDa protein might be a MAP-2 kinase. Accordingly, K721A cells were stimulated with vehicle or EGF, the phosphotyrosinecontaining proteins were immunoprecipitated from cell lysates using anti-phosphotyrosine antibodies, and the immunoprecipitates were immunoblotted with an antibody to MAP-2 kinase. As shown in Fig. lb, MAP-2 kinase was phosphorylated on tyrosine residues in an EGF-dependent manner in K721A cells, The faster migrating band corresponds to the 42-kDa band described in Fig. la and likely represents ERK2 (12). In addition, a more slowly migrating 45-kDa band was also resolved and was phosphorylated in an In order to further explore the mechanism of phosphorylation of MAP-2 kinase in the tyrosine kinase-deficient K721A cells, we evaluated the state of phosphorylation of the human EGF-R present in these cells. Accordingly, immunoprecipitates of the human EGF receptors from cells treated with vehicle or EGF were analyzed by immunoblotting with either an anti-phosphotyrosine antibody or an anti-EGF-R anti-body. As shown in Fig. 2, tyrosine phosphorylation of the human EGF-R in the K721A cells was indeed evident. The level of EGF-dependent phosphorylation of the tyrosine kinase-deficient EGF-R in the K721A cells was approximately 23% of the EGF-dependent phosphorylation using the same number of HER cells transfected with wild type EGF-R. We also analyzed another murine-derived cell line (M721) expressing tyrosine kinase-negative receptors, in which EGFdependent phosphorylation of MAP-2 kinase has been described (6). As shown in Fig. 2, this cell line also displayed EGF-dependent phosphorylation of the EGF-R, which was approximately 10% of the EGF-dependent phosphorylation of the corresponding wild type receptor from the same number of cells.
The EGF-dependent phosphorylation of the human EGF-R and of MAP-2 kinase in cells expressing tyrosine kinasenegative human EGF-R was unexpected. Although the parental NIH-3T3-2.2 clone chosen had been reported to be deficient in endogenous murine EGF-R on the basis of ligand binding experiments, the absence of immunodetectable EGF-R by immunostaining of cell lysates, and the absence of EGFdependent signaling responses, nevertheless we considered the possibility of the presence of a small number of endogenous murine EGF-R that might be capable of participating in signal propagation following transfection with a large number of kinase-negative human EGF-R. Accordingly, lysates from EGF-stimulated NIH-3T3-2.2 cells were immunoprecipitated with anti-EGF-R antibody RK2 (21), which recognizes both the human and murine EGF-R, and the immunoprecipitates were immunoblotted with anti-phosphotyrosine antibodies. As shown in Fig. 3, using this experimental approach, endogenous EGF-R were indeed detected in NIH-3T3-2.2 cells, which are the parental cells for the K721A cell line. Similarly, endogenous EGF-R were also detected (although to a lesser extent) in the B82L cells, which are the parental cell line for M721 cells. There was an EGF-dependent increase in the detection of endogenous phosphorylated EGF-R in both NIH-3T3-2.2 and B82L cells. Thus receptors from NIH-3T3-2.2 cells gave an optical density reading of 0.008/mg of protein (basal) and 0.141/mg of protein (EGF), while the B82L cells gave an optical density reading of 0.020/mg of protein (basal) and 0.034/mg of protein (EGF).
We next sought to ascertain whether in K721A cells, evidence could be obtained for the presence of endogenous murine EGF-R, which could participate in EGF-dependent signal propagation. Lysates from K721A cells were immunoprecipitated with anti-EGF-R antibodies mAb 108 or RK2. Immunoprecipitated proteins were then separated, transferred to Immobilon, and immunostained with RK2 anti-EGF-R antibody. Fig. 4a demonstrates the presence of only one band with an electrophoretic mobility of 170 kDa corresponding to the human EGF-R in the mAb 108 immunoprecipitate, but the presence of two bands, one at 170 kDa and another at a slightly higher electrophoretic mobility, in the RK2 antibody immunoprecipitate. Since RK2 antibody is able to detect both murine and human EGF-R while mAb 108 preferentially detects human EGF-R (23), these results suggested that the slower migrating band in the RK2 antibody immunoprecipitate corresponded to endogenous murine EGF-R.
We next sought to determine the pattern of phosphorylation of the different populations of EGF-R in the K721A cells. Vehicle or EGF-treated cells were lysed and the lysates immunoprecipitated with either antibody RK2 or mAb 108, and then immunoprecipitated proteins immunoblotted with antiphoephotyrosine antihodies. Fig, 4h shows a difference in the appearance of the phosphotyrosine-containing band corre-  (-) or EGF (+) for 3 min at 37 "C. Following cell lysis in buffer containing 1 mM sodium orthovanadate, EGF-R were immunoprecipitated with mAb 108 (monoclonal anti-EGF-R antibody) and analyzed by SDS-PAGE using a 7.5% separating gel. Following transfer to Immobilon, immunoprecipitated proteins were immunostained with the anti-phosphotyrosine antibody (a-P-tyr, 4G10) 2 (NZH2.2) and B82L cells from 15-cm dishes and were stimulated with EGF (100 nM) for 3 min at 37 "C. Following cell lysis in buffer containing 1 mM sodium orthovanadate EGF-R were immunoprecipitated from NIH-3T3-2.2 cell lysates (7 mg of protein) and B82L cell lysates (16 mg of protein) with a polyclonal anti-EGF-R antibody (RK2). The immunoprecipitated proteins were analyzed by SDS-PAGE in a 7.5% gel, transferred to Immobilon, and immunostained with the anti-phosphotyrosine antibody (4G10) and detected with horseradish peroxidase-conjugated secondary antibody followed by ECL development. The immunostained EGF-R was scanned in an LKB Ultroscan and the band intensity normalized with respect to total protein. The NIH-3T3-2.2 cells gave an optical density reading of 0.141/mg of total protein, and the B82L cells gave an optical density reading of 0.034/mg of total protein. The blot shown is representative of three separate experiments.
sponding to the EGF-R in the immunoblot derived from the antibody RK2 immunoprecipitate, compared to the immunoblot derived from the mAb 108 immunoprecipitate. A lower molecular weight component corresponding to the human EGF-R was immunoprecipitated by both the RK2 as well as the mAb 108 antibodies, whereas a higher molecular weight component is only evident in the immunoblot derived from the RK2 immunoprecipitate. This difference in the appearance of the bands cannot be attributed to differences in the interaction of the two different immunoprecipitating antibodies with the same protein, since the proteins have been solubilized and separated by SDS-PAGE.
These results were confirmed using an in vitro autokinase assay using immunoprecipitates prepared with antibodies RK2 and 108 (Fig. 4c). Again a difference in the phosphorylation pattern from the RK2 immunoprecipitates was observed when compared to the mAb 108 immunoprecipitate. A lower molecular weight component, which was phosphorylated in an EGF-dependent manner in this autokinase reaction and corresponded to the human EGF-R, was immunoprecipitated by both the RK2 and the mAb 108 antibodies. In contrast, a higher molecular weight component, also phosphorylated in an EGF-dependent manner, was only evident in the immunoprecipitate derived using the RK2 antibody, but not in the mAb 108 immunoprecipitate. The areas within these bands were quantitated using storage phosphor technology and the data analyzed using Image Quant software. This analysis showed a significant EGF-dependent increase in the phosphorylation of the human EGF-R immunoprecipitated with antibodies mAb 108 (PhosphorImager counts based on three separate experiments increased from 3959 f 674 (vehicle, mean f S.D.) to 8176 f 761 (EGF), p < 0.001) and RK2 (PhosphorImager counts increased from 3785 f 334 (vehicle) to 13104 f 761 (EGF), p < 0.001), respectively. We also detected a significant EGF-dependent increase in the phosphorylation of the slower migrating band, which was only detected in the RK2 immunoprecipitate and autokinase reaction (PhosphorImager counts increased from 1023 f 200 (vehicle) to 4880 & 293 (EGF), p < 0.001).

DISCUSSION
In the current study, we have used a clonally expanded NIH-3T3-2.2 cell line, which has been previously reported to be deficient in endogenous EGF-R and therefore has been extensively utilized for the study of transfected normal or mutant human EGF-R (see Ref. 1 and references therein). Indeed, in the current study, we confirmed that in the parental NIH-3T3-2.2 cells, no detectable EGF-dependent signal propagation was evident as had been previously reported (4). However, NIH-3T3-2.2 cells, which have been transfected to express high levels of the tyrosine kinase-negative human EGF-R (K721A cells), did display EGF-dependent tyrosine phosphorylation of MAP-2 kinase as demonstrated in Fig. 1. Initially, these results appeared to be similar to those reported by Campos-Gonzalez and Glenney (6). On the surface, these results support the possibility of a novel mechanism for signal transduction by receptor tyrosine kinases independent of their intrinsic tyrosine kinase activity, which would represent a potentially significant departure from the currently accepted paradigm regarding transmembrane signaling by this class of receptors. However, further studies support an alternative explanation entirely consistent with recent evidence pertain-  (50 p~ acetic acid, lanes 1 and 3 ) or EGF (100 nM, lanes 2 and 4 ) for 3 min at 37 "C. Following cell lysis in buffer containing 50 mM NaF, 10 mg of protein from each lysate was used to immunoprecipitate the EGF-R with RK2 (polyclonal anti-EGF-R antibody, lanes 1 and 2) or mAb 108 (monoclonal anti-EGF-R antibody, lanes 3 and 4 ) . The immunoprecipitated EGF-R were analyzed by SDS-PAGE in a 6% gel and immunostained with a polyclonal anti-EGF-R antibody (RK2) and detected with horseradish peroxidase-conjugated secondary antibody followed by ECL development. The blot is representative of three separate experiments. Panel b, EGF-induced tyrosine phosphorylation of mouse and human EGF receptors in K721A cells. Cells from 15-cm dishes were stimulated in the presence of vehicle (50 p~ acetic acid, lanes 1 and 3 ) or EGF (100 nM, lanes 2 and 4 ) for 3 min at 37 "C. Following cell lysis in buffer containing 1 mM sodium orthovanadate, 10 mg of protein from each lysate was used to immunoprecipitate the EGF-R with RK2 (polyclonal anti-EGF-R antibody, lanes 1 and 2) or mAb 108 (monoclonal anti-EGF-R antibody, lanes 3 and 4 ) . The immunoprecipitated EGF-R were analyzed by SDS-PAGE in a 6% gel and immunostained with an anti-phosphotyrosine antibody (a-P-tyr, 4G10) and detected with horseradish peroxidase-conjugated secondary antibody followed by ECL development described under "Experimental Procedures." The blot shown is representative of four separate experiments. Panel c, in vitro autophosphorylation of the mouse and human EGF receptors in K721A cells. Cells from 15-cm dishes were stimulated in the presence of vehicle (50 p~ acetic acid, lanes 1 and 3 ) or EGF (100 nM, lanes 2 and 4 ) for 3 min. at 37 "C. Following cell lysis 10 mg of protein from each lysate was used to immunoprecipitate the EGF-R with RK2 (polyclonal anti-EGF-R antibody, lanes 1 and 2) or mAb 108 (monoclonal anti-EGF-R antibody, lanes 3 and 4). The immunoprecipitated EGF-R were used in an in vitro autokinase assay, analyzed by SDS-PAGE in a 6% gel, and exposed to x-ray film for 24 h. The autoradiogram shown is representative of three separate experiments. The gel was exposed to storage phosphor plates for 2 h and scanned, and the bands corresponding to the human and mouse EGF-R were quantitated.
In the case of the EGF receptor, the SH2/SH3 domain-containing protein (Grb-2 physically associates through its SH2 domains with the activated EGF-R (28). Further steps in the signaling cascade include the engagement of a guanylnucleotide exchange protein (SOS), through the SH3 domain of Grb-2 followed by activation of Ras and then a downstream signaling cascade including Raf and MAP kinase/ERK kinase among others (29)(30)(31)(32). Grb-2 is itself not phosphorylated, but tyrosine phosphorylation of the EGF receptor is required for physical association of Grb-2 and initiation of the signaling cascade (28). Accordingly, in the current study, it would be possible for the tyrosine kinase-negative human EGF-R to propagate a signal resulting in MAP-2 kinase phosphorylation so long as there was tyrosine phosphorylation of the tyrosine kinase-negative human EGF-R itself. Indeed, we found EGFdependent phosphorylation of EGF-R in both K721A and M721 cells. These findings are not consistent with the previous report by Campos-Gonzalez and Glenney (6) in which it was reported that although two other high molecular mass proteins (170 and 185 kDa) were tyrosine-phosphorylated in M721 cells, they did not represent the EGF-R when analyzed with monoclonal antibodies to the receptor. However, the results in the current findings are consistent with a more recent report by Selva et al. (33) in which EGF-dependent tyrosine phosphorylation of EGF receptors was observed in Chinese hamster ovary cells that had been transfected with tyrosine kinase-negative human EGF receptors. The discrepancy in the ability to detect tyrosine-phosphorylated EGF receptors in an anti-EGF receptor immunoprecipitate from M721 cells in the current study, as compared to the inability to detect tyrosine-phosphorylated EGF receptors in M721 cells in the study of Campos-Gonzales and Glenney, may be attributed to the use of different antibodies to immunoprecipitate the EGF-R in our (mAb 108) uersw the latter (mAb 74 and cll) studies. We have found that mAb 74 does not recognize the murine EGF-R and that in K721A cells, it does not immunostain the human tyrosine kinase-negative EGF-R, even though the latter undergoes tyrosine phosphorylation?
It was of interest to address the question of how human EGF-R are phosphorylated in an EGF-dependent manner in the K721 and M721 cells in the current study. Both Campos-Gonzalez and Glenney (6) and Selva et al. (33) have suggested that an as yet unidentified kinase may be activated by an EGF-dependent conformational change in the EGF-R and may be the initial step resulting in phosphorylation and activation of MAP-2 kinase. This interpretation is contingent upon the assumption that these cell lines lack endogenous kinase-competent EGF-R. The presence of even a small number of endogenous kinase-competent EGF-R could result in EGF-dependent phosphorylation of the tyrosine kinase-negative human EGF-R and thereby provide a route for amplification of the signal resulting in phosphorylation of MAP-2 kinase. In RT-PCR studies using primers specifically designed to detect murine EGF-R, we found clear evidence for the presence of low but equivalent amounts of murine EGF-R transcript in both NIH-3T3-2.2 as well as K721A cells.3 This led us to seek evidence for immunodetectable murine EGF-R in these cells as well as in M721 cells. Immunoprecipitation with the polyclonal anti-EGF-R antibody RK2, followed by immunostaining with anti-phosphotyrosine antibodies allowed the detection of even the low levels of endogenous EGF-R present in these cells. experimental conditions to separately identify in K721A cells bands of different electrophoretic mobility in immunoprecipitates using the RK2 anti-EGF-R antibody, which recognizes both murine and human EGF-R, whereas only human EGF-R was detected in immunoprecipitates using mAb 108. This led us to hypothesize that EGF-dependent heterodimerization of mouse-human receptors could occur in K721A cells, resulting in the heterologous tyrosine phosphorylation of the tyrosine kinase-negative human EGF-R. This is in agreement with recent reports providing clear evidence for heterodimerization and heterologous transphosphorylation of tyrosine kinase-negative EGF receptors as has been reported for HER-2/neu and the EGF-R (34), and a tyrosine kinase-competent FGF receptor and a tyrosine kinase-negative fibroblast growth factor receptor (35).
The findings in the current study are consistent with the postulate that, in K721A cells, EGF causes heterodimers to form between endogenous murine EGF-R and transfected tyrosine kinase-negative human EGF-R with transphosphorylation of the latter by the former. This results in amplification of the signal by making available a larger number of phosphorylated EGF-R available for interaction with Grb-2 and consequent signal propagation. An alternative formulation is that proposed by Selva et al. (33), who postulate the presence of an as yet unidentified kinase, which is activated by an EGF-dependent conformational change in the kinase negative human EGF-R. Such a postulate predicts that if this kinase is cytosolic, then the EGF dependence of its activation requires that it recognize the state of dimerization of the tyrosine kinase negative human EGF-R, or else the unlikely possibility that a conformational change in the cytoplasmic domain of the monomeric EGF-R can occur as a result of an event occurring on the ligand binding extracellular domain. Alternatively, it would be necessary to postulate that the EGF dependence is the result of the activation of a tyrosine kinase, which is itself an EGF-binding protein, and, hence, by definition, another class of EGF-R. We suggest that it is most likely that the results of these studies by Selva et al. might also be explained by the presence of endogenous hamster EGF-R, a possibility that cannot be thoroughly ruled out until studies are done using the appropriate reagents in these cells. Certainly, the absence of the endogenous EGF-R cannot be conclusively inferred simply on the basis of experiments using either radioligand binding, immunostaining of lysates with anti-EGF-R antibodies, studies of EGF-dependent signal propagation, or even analysis of EGF-R mRNA levels using probes or primers derived from the human EGF-R sequence. These approaches would have suggested that NIH-3T3-2.2 and B82L cells are devoid of endogenous EGF-R, but subsequent studies revealed this not to be the case.
The results of the current study suggest an alternative possibility aside from postulating the presence of an as yet unidentified kinase to explain EGF-dependent signal propagation in cells expressing tyrosine kinase-negative human EGF receptors. However, it should be noted that it is possible that the endogenous tyrosine kinase-competent EGF-R, even if they are present, do not play a role in MAP-2 kinase phosphorylation by the tyrosine kinase-negative EGF-R. In view of recent developments in our understanding of the early steps involved in EGF-R signal propagation, and the presumed importance of this pathway in mitogenic and other responses, further evaluation of signal propagation in cells expressing tyrosine kinase-negative human EGF-R in the absence of endogenous EGF-R is warranted.