INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The epidermal growth factor (EGF) receptor (EGFR) (also designated ErbB-1) is a member of the ErbB family of ligand-activated tyrosine kinase receptors, which play a central role in the proliferation, differentiation, and/or oncogenesis of epithelial cells, neural cells, and fibroblasts (82). A plethora of biological responses are triggered by the interaction of EGF, or one of its homologues (29), with the extracellular domain of the EGFR. Upon ligand binding, the kinase domains are activated by homo- and/or heterodimerization of EGFR family members (31, 67). The activated receptor kinase then autophosphorylates C-terminal tyrosines and transphosphorylates intracellular substrates (reviewed in reference 11). The C-terminal phosphotyrosine residues can bind to particular cytoplasmic proteins which have been proposed as a means of amplifying mitogenic signalling from ligand-receptor complex (55, 67).

The ShcGRB-2Son of Sevenless (Sos)Rasmitogen-activated protein kinase (MAPK) cascade (reviewed in reference 5) has been proposed to be the major mitogenic signalling pathway initiated by the EGFR family of kinases. Shc proteins are phosphorylated rapidly on tyrosine following EGF binding to EGFR and associate with the phosphorylated EGFR via their SH2 domains (56); tyrosine-phosphorylated Shc binds in turn to the SH2 domain of GRB-2 (61), resulting in the relocation of the GRB-2-Sos complex from the cytosol to the plasma membrane (44), where Sos stimulates the exchange of GDP for GTP on Ras, converting it to its active state (reviewed in reference 9). The GTP-bound form of Ras leads to activation of a protein kinase cascade mediated by the serine/threonine kinase Raf (79), the dual-specificity tyrosine/threonine kinase MAPK kinase (MEK) (52), MAPKs (also known as extracellular regulated kinases Erk-1 and Erk-2) (40, 79), and eventually AP-1 transcriptional activity (35). While activation of the Ras/MAPK pathway appears to be necessary for the proliferative response to growth factors (8, 37), recent studies have suggested that other, Ras-independent pathways also need to be initiated before cells will respond mitogenically to EGF or platelet-derived growth factor (3, 7) and for transition through the cell cycle (41).

EGFR mutants have been used extensively for defining and evaluating EGF-mediated signalling pathways (4, 13, 26, 27, 72). However, these studies have been performed with cells expressing at least one endogenous EGFR family member; ligand-induced association of EGFR (ErbB) family members with each other (heterodimerization) and the resulting cross-kinase activation and phosphorylation have made it impossible to distinguish between the contributions of endogenous and mutant EGFRs to EGF-induced mitogenic signalling. Interesting conclusions can be drawn when considering the biological responses from EGFR mutants. For example, EGFR mutants lacking all of the phosphorylation sites proposed as binding sites for accessory signalling molecules are still able to phosphorylate Shc and stimulate EGF-dependent mitogenesis (27, 64). One interpretation of these data is that C-terminal phosphorylation is neither necessary nor important for stimulating mitogenesis. However, these studies were performed with cells where the mutant EGFR could potentially heterodimerize and signal by phosphorylating a heterodimer partner (ErbB-2, -3, or -4), leading to the docking of signal-transducing proteins such as Shc and GRB-2. Similarly, in experiments evaluating kinase-negative EGFRs (K721 mutants), EGF stimulated the phosphorylation of Shc and activation of the MAPK pathway (69, 80). Signalling from these kinase-negative EGFRs presumably occurs via heterodimerization with ErbB-2 but does not result in a mitogenic signal. Other mutations in the EGFR kinase domain (D813 and V741G) (14, 22), which also abolish EGFR kinase activity, can induce mitogenesis when expressed in fibroblasts. Again, it is unclear whether the mitogenic signal is delivered solely by the defective receptors or is transduced by ligand-induced heterodimerization and stimulation of an endogenous, kinase-active partner, such as ErbB-2. Furthermore, mice expressing either a kinase-defective EGFR, such as the wa-2 receptor (22, 45) or the dominant-negative CD533 EGFR (51), have a mild phenotype of wavy hair and premature eye opening. In contrast, in most mouse strains, the complete loss of EGFR or ErbB-2 expression through targeted gene disruption results in embryonic lethality (39, 74). Two alternative explanations are possible for such different phenotypes. The EGFR mutants may be truly kinase negative but, by oligomerizing with other family members through their intact extracellular domain, could signal through an active partner, or the mutants may retain a latent tyrosine kinase activity sufficient for cell survival and/or proliferation. Thus, the signalling mechanisms intrinsic to the EGFR or its mutants can be assessed accurately only in the absence of both endogenous EGFRs and the other ErbB proteins, i.e., ErbB-2, -3, and -4.

In a previous study (77) we have utilized the murine pro-B cell line BaF/3 (54), which lacks all EGFR family members, to analyze the expression and biochemical properties of EGFR mutants. We have shown that mutations in the -helix C of the EGFR (V741G and Y740F) reduce profoundly the tyrosine kinase activity of the isolated receptor in cell-free assays but do not prevent EGFR tyrosine phosphorylation when cells are stimulated with EGF. The "in vivo" phosphorylation of the receptors appears to be mediated by a cytosolic tyrosine kinase and not by residual kinase activity intrinsic to the mutant receptors. In this study we have addressed the biological significance of EGFR phosphorylation in the absence of EGFR kinase activity. We report here the effects of point mutations in the kinase domain of the EGFR on the biochemical and biological responses to EGF of BaF/3 cells expressing EGFR or intracellular domain mutants of EGFR.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Generation of cell lines expressing EGFRs. The construction of the EGFR mutants and the establishment of BaF/3 cell lines expressing wild-type (WT) and mutant EGFRs are described in a previous paper (77).

Antibodies and growth factors. Sheep polyclonal anti-EGFR antibodies targeted to the intracellular domain of EGFR, antiphosphotyrosine mouse monoclonal 4G10 antibodies conjugated to protein A-Sepharose beads, and monoclonal antiphosphotyrosine 4G10 were purchased from Upstate Biotechnology, Inc. (Lake Placid, N.Y.). Rabbit polyclonal anti-Shc antibodies were obtained from Transduction Laboratories (Lexington, Ky.). Anti-Ras rat monoclonal Y13-259 antibodies (24) and anti-EGFR monoclonal 528 antibodies (25) were produced and purified from the hybridoma culture medium in our laboratory. Anti-GRB-2 rabbit polyclonal antibody was raised in our laboratory against a glutathione S-transferase (GST)-GRB-2 fusion protein. Horseradish peroxidase-conjugated secondary (goat anti-rabbit and rabbit anti-mouse) antibodies were from Bio-Rad (Richmond, Calif.), and rabbit anti-sheep secondary antibody was from DAKO-Immunoblot (Carpinteria, Calif.). Protein A- and protein G-Sepharose-conjugated beads were obtained from AMRAD Pharmacia Biotech (Melbourne, Australia). Anti-B-Raf antibodies (C-19) and anti-Erk1 antibodies (C-16) were purchased from Santa Cruz Biotechnology (Santa Cruz, Calif.). Anti-phospho-p44/42 MAPK antibody was purchased from New England BioLabs. EGF was purified from mouse submaxillary glands as described previously (10).

Inhibitors. MAPK inhibitor PD98059 (18) was purchased from Calbiochem (Alexandria, New South Wales, Australia). JAK kinase inhibitor AG490 (48) and Src kinase inhibitor PP1 (30) were gifts from Alexander Levitzki. LY294002 (76) was obtained from Sigma (St. Louis, Mo.).

Cells and cell culture. Parental BaF/3 and BaF/3 cell lines expressing the different EGFRs were maintained routinely in RPMI 1640 (GIBCO BRL) supplemented with 10% fetal calf serum (FCS) (GIBCO BRL) and 10% WEHI-3B cell line conditioned medium (15) as a source of interleukin-3 (IL-3). All cell lines were grown at 37°C in an air-CO₂ (95%-5%) atmosphere. Before EGF treatment, cells were transferred to RPMI 1640 without additions and left for at least 4 h to induce quiescence.

Immunoprecipitations and immunoblotting. After treatment with EGF (100 ng/ml, unless otherwise indicated), cells were washed with ice-cold phosphate-buffered saline (PBS), pH 7.5. Whole-cell lysates were prepared in extraction buffer (30 mM HEPES, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride [PMSF], 10² U of Trasylol ml¹, and 100 µM sodium orthovanadate [pH 7.5]). The lysates were cleared by centrifugation (15,000 × g, 15 min), and the appropriate antibodies were added. After 60 min at room temperature or overnight incubation at 4°C, immune complexes were collected by treatment with protein A-Sepharose beads or protein G-Sepharose beads (Pharmacia) for 45 min at 4°C. Tyrosine-phosphorylated proteins were immunopurified with antiphosphotyrosine antibody conjugated to agarose beads (4G10-agarose; 5 µl of a 50% slurry) for 3 h at 4°C. Immunoprecipitated proteins were resolved by discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (38) in polyacrylamide gels (1.5-mm thickness) as indicated and transferred onto an Immobilon-P membrane (Millipore Corp. Bedford, Mass.) by electroblotting. Residual binding sites on the membranes were blocked with blocking buffer (PBS [pH 7.5], 3% bovine serum albumin, 0.02% Tween 20 [BDH Laboratory Supplies, Poole, United Kingdom]) for 2 h at room temperature or overnight at 4°C. Blots were then rinsed with wash buffer (PBS [pH 7.5], 0.1% bovine serum albumin, 0.02% Tween 20), incubated with primary antibody diluted in wash buffer for 1 h at room temperature, washed again in wash buffer, and developed with horseradish peroxidase-conjugated second antibody for 30 min at room temperature. After washing, blots were incubated with enhanced chemiluminescence substrate solution (ECL; Amersham Corp., Aylesbury, United Kingdom) and exposed to Kodak-X-Omat film (Eastman Kodak Company) to visualize immunoreactive bands. In some experiments, bound antibodies were stripped with stripping buffer (62.5 mM Tris-HCl [pH 6.7], 2% SDS, 100 mM 2-mercaptoethanol) at 50°C for 15 min, and the membranes were reprobed with another primary antibody.

Quantitation of Western blots and autoradiograms. Films obtained from exposure of the immunoblots were scanned on a Molecular Dynamics scanning densitometer, and radioactive gels were scanned directly by using a Molecular Dynamics PhosphorImager. In both cases, volume integration of the bands was performed with ImageQuant (Molecular Dynamics, Sunnyvale, Calif.) according to the manufacturer's instructions.

Proliferation assay. BaF/3 cells were washed twice with 40 ml of RPMI 1640 medium containing FCS (10%, vol/vol) and suspended in RPMI 1640 medium containing 10% FCS and 0.02% (vol/vol) WEHI-3B conditioned medium (medium A). This concentration of IL-3 did not stimulate proliferation. Cells were seeded at approximately 10⁵/ml in 24-well plates and cultured with or without EGF for up to 7 days (see figure legends for details of specific experiments). Viable cell numbers were determined by trypan blue exclusion and phase-contrast microscopy.

Survival assay. Cells were washed twice with a large volume (40 ml) of RPMI 1640 containing FCS (10%, vol/vol). Cells were resuspended at approximately 10⁵/ml in RPMI 1640-10% FCS without additions (minimal medium), with 100 ng of EGF per ml, or with 10% (vol/vol) WEHI-3B conditioned medium and incubated for 3 days. On the fourth day, cells from each protocol were washed twice in RPMI 1640-10% FCS and transferred to an equal volume of RPMI 1640-10% FCS-10% WEHI-3B conditioned medium. Viable cell numbers were determined daily for 7 days by trypan blue exclusion.

Effects of inhibitors on cell proliferation and survival. Proliferation assays were performed as described above, except for the addition of the following inhibitors: PD98059 (MAPK inhibitor) (18) LY294002 (phosphatidylinositol-3-kinase [PI-3-kinase] inhibitor) (76), PP-1 (Src family kinase inhibitor) (30), and AG490 (JAK-2 kinase inhibitor) (48). All inhibitors were dissolved in dimethyl sulfoxide (DMSO) and stored at 20°C. Serial dilutions of each inhibitor in DMSO were prepared from this stock, and an equal volume of each was added to the cells at the time of incubation with EGF. Control cultures received an equivalent volume of DMSO. Viable cell numbers were determined by dye exclusion after 3 days of incubation.

Ras activation assay. Cells were starved in serum-free RPMI 1640 medium for 4 h to induce quiescence. [-³²P]GTP (25 µCi) was introduced into the cells by electroporation (400 mV, 25 µF, two pulses in a 2-min interval) in 100 µl of electroporation buffer (140 mM KCl, 10 mM HEPES, 1 mM MgCl₂, 1 mM EGTA, 1 mM glucose, 193 µM CaCl₂). Cells were resuspended, incubated on ice for 1 min, and then stimulated with 100 ng of EGF per ml for 5 min at 37°C. Stimulation was terminated by addition of 900 µl of ice-cold Ras lysis buffer (0.5% [vol/vol] Triton X-100, 25 mM Tris-HCl [pH 7.5], 137 mM NaCl, 5 mM MgCl₂, 5 mM KCl, 1 mM sodium phosphate buffer [pH 7.4], 0.7 mM CaCl₂) for 10 min in the presence of 40 µl of Y13-259 anti-Ras antibodies. Lysates were cleared by centrifugation (15,000 × g, 15 min, 4°C). Protein G-Sepharose beads (30 µl) were added after a further 10 min of incubation on ice, and tubes were rotated at 4°C for 20 min. Beads were washed three times with Ras lysis buffer; after the first two washes they were transferred to fresh tubes to minimize background radioactivity. Ras protein was eluted from the beads by the addition of 15 µl of 1 M KH₂PO₄ (pH 3.4) and heated for 3 min at 95°C. GTP and GDP were then separated by thin-layer chromatography as described by Zhang et al. (82) and quantitated by autoradiography with ImageQuant.

B-Raf activation. Cells (10⁷/ml) were incubated for 4 h in RPMI 1640 to induce quiescence and resuspended in RPMI 1640 containing 200 µM sodium pervanadate and 0.1% bovine serum albumin in triplicate tubes. Tubes received either RPMI 1640 (control), WEHI-3B conditioned medium (to 10% [vol/vol]), or EGF (100 ng/ml in RPMI 1640), and incubation was continued for 30 min at 25°C. Cells were collected and then lysed on ice for 45 min in 1 ml of ice-cold extraction buffer containing PMSF (1 mM), leupeptin (10 µg/ml), aprotinin (100 U/ml), microcystin LR (0.5 µg/ml), sodium pervanadate (50 µM), -glycerophosphate (100 mM), and NaF (50 mM). Insoluble material was pelleted by centrifugation, and the soluble extracts were incubated with anti-B-Raf antibodies overnight at 4°C. Immunocomplexes were collected by addition of protein A-Sepharose beads (a 50% [vol/vol] slurry). For the determination of B-Raf activity, the immunoprecipitates were washed three times in extraction buffer and twice in kinase buffer (20 mM HEPES, 10 mM MgCl₂, 10 mM dithiothreitol) and were resuspended in kinase buffer containing protease inhibitors (10 µg of leupeptin per ml, 100 U of aprotinin per ml, and 1 mM PMSF), phosphatase inhibitors (50 mM NaF, 100 mM -glycerophosphate, 50 mM sodium pervanadate, and 0.5 µg of microcystin LR per ml), 5 µg of the B-Raf substrate GST-MAPKK (kinase negative and mutated at both threonine MAPK phosphorylation sites) (2) per reaction mixture, and 10 µCi of [-³²P]ATP per reaction mixture. Samples were incubated at 30°C for 30 min, and then the supernatants, containing phosphorylated GST-MEK, were separated from the B-Raf-antibody-protein A complex by centrifugation and analyzed by SDS-7.5% PAGE followed by autoradiography. The pellets, containing the B-Raf immunoprecipitates, were also separated by SDS-PAGE, followed by Western blotting to allow quantitation of B-Raf proteins.

MAPK activation. Cell stimulation with EGF was performed as described for the B-Raf kinase activation assay. Approximately 5 × 10⁵ cells from each treatment were lysed directly in Laemmli's sample buffer and analyzed by SDS-PAGE on 10% gels, followed by immunoblotting with anti-phospho-MAPK antibodies (New England BioLabs). For MAPK inhibition experiments, cells were preincubated with media containing different concentrations of PD98059 in DMSO for 30 min prior to the addition of EGF or with control medium. Immunoprecipitation was carried out with anti-Erk-1 antibodies and protein A-Sepharose. Erk-1 kinase activity was assayed by using myelin basic protein (MBP) (Sigma, Castle Hill, New South Wales, Australia) (5 µg/reaction mixture) as a substrate in the presence of 10 µCi of [-³²P]ATP per reaction mixture for 30 min at 30°C. The degree of MBP phosphorylation was determined by separation by SDS-12.5% PAGE followed by autoradiography.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

EGFR mutants. Schematic representations of the human WT EGFR and of the four EGFR mutants are shown in Fig. 1. V741G, Y740F, and K721R each have a single point mutation at the specified residue. V741G is the human homologue of the mouse waved-2 EGFR and transduces a weak mitogenic signal in fibroblasts (22). V741G and Y740F have negligible kinase activity (22, 45, 77), suggesting that the -helix C in the EGFR kinase domain is important for the activation of the EGFR kinase. Mutations of K721 of the EGFR abolish its kinase activity (50) by creating a nonproductive binding mode for ATP (60). CT957 represents an EGFR which is truncated at amino acid 957 and is therefore missing the five major autophosphorylation sites. This mutant has an intact kinase domain but is not expected to bind adapter molecules or signal transducers such as Shc and GRB-2.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   Schematic representation of human WT and mutant EGFR. cys, cysteine-rich regions; TM, transmembrane domain; p, major autophosphorylation sites.

Tyrosine phosphorylation of EGFRs in BaF/3 cells. We have shown previously that none of the kinase-defective EGFR mutants autophosphorylates or phosphorylates a peptide substrate in cell-free kinase assays, due to a profound suppression of the phosphotransfer activity (77). However, when BaF/3 cells expressing the V741G or Y740F EGFR are stimulated with EGF, the helix C mutants become phosphorylated on tyrosine residues to levels equivalent to those for the WT receptor (77) (Fig. 2). Since EGF-dependent tyrosine phosphorylation of these mutants does not occur in isolated plasma membranes, we proposed that the phosphorylation is due to the action of a cytosolic tyrosine kinase (77). However, activation of a cytosolic kinase and/or phosphorylation of the mutant receptors may not compensate sufficiently for the defects in their EGFR kinase activity, so that mitogenic signalling may be impaired; therefore, we have investigated the biological responses of BaF/3 cells expressing the WT or mutant receptors to EGF.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Tyrosine phosphorylation of WT and mutant EGFRs in intact BaF/3 cells. Serum- and IL-3-starved BaF/3 cells were stimulated with EGF (100 ng/ml) for 10 min at 37°C. (A) Tyrosine phosphorylation of the EGFR in control () and EGF-stimulated (+) cells was determined by immunoprecipitation of cellular lysates with anti-EGFR antibody 528 and detection with antiphosphotyrosine antibodies after SDS-PAGE and transfer to Immobilon membranes. (B) The blots were stripped and reprobed with an anti-EGFR antibody directed to the C terminus of the receptor protein. This antibody reacts less strongly with the phosphorylated protein.

Mitogenic response to EGF or BaF/3 cells expressing WT and mutant EGFRs. Previous reports indicate that WT EGFRs are at least partly functional in BaF/3 cells: EGF stimulates DNA synthesis in these cells, but the cells arrest in S phase and do not complete cell division (70). We have been able to optimize the culture conditions so that a proliferation assay, rather than DNA incorporation, could be used to test the ability of the EGFR mutants to deliver a mitogenic signal in response to EGF (Fig. 3). In preliminary experiments we confirmed that when BaF/3 cells expressing the WT EGFR are arrested in G₀ by IL-3 starvation, they do not proliferate significantly in response to addition of EGF (not shown). However, when limiting amounts of IL-3 (insufficient to induce proliferation of the parental BaF/3 cells [Fig. 3A]) were added to the cultures, the WT EGFR-expressing cells proliferated in response to EGF (Fig. 3B). In contrast parental BaF/3 cells do not proliferate in response to EGF under these conditions (data not shown). We then utilized these conditions to investigate the proliferative potential of BaF/3 transfectants expressing mutant EGFRs.

View larger version (14K): [in this window] [in a new window]   FIG. 3.   Mitogenic stimulation of BaF/3 cell lines expressing EGFRs. Cells were washed extensively before addition of medium and growth factors. Cell proliferation was measured by viable cell count at the times indicated. (A) Titration of WEHI-3B (also called D) cell line conditioned medium on parental BaF/3 cells. Viable cell numbers were determined at day 5. (B) BaF/3 cells expressing WT EGFR were cultured under the indicated conditions for up to 7 days. Control medium was RPMI 1640-10% FCS supplemented with 0.02% (vol/vol) WEHI-3B conditioned medium (CM). (C and D) BaF/3 cells expressing the indicated EGFRs were incubated in control medium containing increasing concentrations of EGF. Cell numbers were determined at day 5. All experiments were performed in triplicate. The results (means and standard deviations) are representative of those from five independent experiments.

EGF-mediated survival of EGFR-expressing BaF/3 cells. In the course of these experiments we noted that despite the absence of proliferation, cultures of cells expressing the kinase-impaired EGFR mutants V741G and Y740F always contained more viable cells in the presence of EGF than in the control medium. In the absence of IL-3, the viability of the parental BaF/3 cell line is not influenced by EGF. Thus, via a receptor-mediated process, EGF affords the V741G and Y740F EGFR-expressing BaF/3 cells some protection from the apoptotic death that follows IL-3 deprivation. We tested this hypothesis further by using the survivability assay described by Fridell et al. (23). This assay measures the functional survival of cells when EGF is substituted for IL-3 in the cultures, as detected by the ability of the cells to resume proliferation once returned to IL-3 (Fig. 4). Cells were cultured for 4 days in minimal medium alone or supplemented with IL-3 or EGF; on day 4 the cells were washed briefly and reseeded in an equal volume of complete growth medium (RPMI 1640, 10% FCS, 10% WEHI-3B conditioned medium). Viable cell numbers were determined daily. Parental BaF/3 and K721R EGFR-expressing BaF/3 cells died rapidly in minimal medium and in medium supplemented with EGF; however, the viability of cells expressing the WT, V741G, Y740F, or CT957 EGFR was maintained by EGF (Fig. 4) even in the absence of cell proliferation. Therefore, we conclude that, in BaF/3 cells, EGF stimulation of receptors with an impaired kinase activity can support survival, while an intact EGFR kinase domain is necessary for mitogenic signalling.

View larger version (15K): [in this window] [in a new window]   FIG. 4.   EGF-mediated survival after IL-3 withdrawal. Cells were incubated in RPMI 1640-10% FCS with no additions (open squares), in medium containing 50 ng of EGF per ml (closed squares), or in medium supplemented with 10% WEHI-3B conditioned medium (IL-3) (open circles). At day 4 the cells were collected by centrifugation and reseeded in complete growth medium (IL-3) (dashed line). Viable cell counts were determined each day up to day 7.

Shc phosphorylation by EGFR mutants in BaF/3 cells. The Ras/MAPK pathway (9) has been proposed as the main mitogenic signalling pathway triggered by activation of the EGFR (12, 28). The lack of mitogenic signalling by the mutant EGFRs could therefore be due to their inability to activate this pathway, which is initiated by the tyrosine phosphorylation of Shc. We compared the abilities of WT and mutant EGFRs to induce Shc phosphorylation in BaF/3 cells following EGF stimulation (Fig. 5A, upper panels). Immunodetection of the Shc protein on the same blot shows that comparable levels of Shc were immunopurified from all cell lines (Fig. 5A, lower panels). EGF induced strong tyrosine phosphorylation of Shc in cells expressing the WT EGFR, while tyrosine phosphorylation of the Shc proteins in the parental BaF/3 cell line or in cells expressing the K721R EGFR was undetectable. K721 mutants have been shown in a previous report to induce Shc phosphorylation in an EGF-dependent manner; the authors proposed that substrate phosphorylation by the kinase-negative receptor may have been mediated by heterodimerization with endogenous ErbB-2 (80). Our results confirm that in the absence of other ErbB family members, the K721R mutant is indeed incapable of phosphorylating Shc, and heterodimerization with a kinase-active EGFR family member is likely to be responsible for Shc phosphorylation in fibroblasts. The CT957 EGFR mutant mediated Shc phosphorylation, but at a reduced level. CT957 is missing the autophosphorylation sites to which the SH2 domain and phosphotyrosine-binding domains of Shc would normally bind; efficient EGF-dependent phosphorylation of Shc appears to be favored by a stable association between Shc and the EGFR. In contrast, Shc phosphorylations by the -helix C mutants and the WT EGFR were comparable, at least at the high concentration of EGF (16 nM) used in this experiment. To determine whether the EGFR mutants could also mediate Shc phosphorylation at physiological doses of EGF and whether there is a correlation between Shc phosphorylation and mitogenic response, we analyzed the tyrosine phosphorylation of Shc in response to different concentrations of EGF (Fig. 5B). The level of Shc phosphorylation at each concentration of EGF was then compared with EGFR occupancy and EGF-dependent mitogenic responses (Fig. 5C and D). For each cell line the intensity of Shc phosphorylation at 10 nM EGF was taken as maximal, since at this concentration >90% of the receptors are occupied; this is true even for V741G EGFR-expressing BaF/3 cells, which do not have high-affinity EGF binding sites (77). The number of EGFRs occupied for each cell line at each concentration of EGF was determined from the formula ([L]/[L] + Kd₁) × R₁ + ([L]/[L] + Kd₂) × R₂, where [L] is the EGF concentration, Kd₁ and Kd₂ are the equilibrium binding constants, and R₁ and R₂ are the number of high-affinity and low-affinity receptors per cell, respectively. Shc phosphorylation was then plotted against the number of EGFRs occupied for each EGF concentration (Fig. 5C): clearly, occupancy of as few as 10,000 to 20,000 EGFRs is sufficient to achieve high levels of Shc phosphorylation, particularly in the case of the -helix C mutants. Finally, we compared fractional receptor occupancy to Shc phosphorylation and mitogenic activity for the WT EGFR (Fig. 5D). Mitogenesis and occupancy of high-affinity sites are correlated, as are Shc phosphorylation and low-affinity receptor occupancy; however, there is no apparent correlation between the extent of EGF-induced Shc phosphorylation and the mitogenic responses to EGF. Taken together, these results show that the lack of mitogenic signalling by the -helix C mutants of the EGFR is not due to their inability to phosphorylate Shc. Furthermore, mitogenic signalling is not impaired when Shc phosphorylation is reduced, as is the case for the CT957 EGFR.

View larger version (44K): [in this window] [in a new window]   FIG. 5.   Tyrosine phosphorylation of Shc by mutant EGFRs. (A) Quiescent cells (107/cell line) were incubated in RPMI 1640 medium containing sodium pervanadate (200 µM) with or without EGF (100 ng/ml) for 10 min at room temperature. Cells were lysed in detergent and immunoprecipitated with anti-Shc antibodies and protein A-Sepharose. The immunoprecipitates were separated by SDS-10% PAGE and transferred to an Immobilon-P membrane. Immunoblots were probed with antiphosphotyrosine antibodies (upper panels), stripped, and reprobed with anti-Shc antibodies (lower panels). The reactive proteins were visualized by ECL. (B) Quiescent cells were stimulated with various concentrations of EGF as described for panel A. The amount of tyrosine-phosphorylated Shc was determined by immunoprecipitation with 4G10 antiphosphotyrosine beads and detection with polyclonal anti-Shc antibodies. (C) The intensity of Shc phosphorylation, quantitated in ImageQuant and expressed as percentage of the maximum, is plotted against the number of WT or mutant EGFRs occupied at each concentration of EGF. Receptor occupancy was calculated for each cell line as described in the text. Closed circles, WT; open circles, Y740F; open squares, V741G. (D) Relationship between fractional receptor occupancy (left y axis), Shc phosphorylation, and mitogenic response to EGF (both expressed as percentage of the maximum) (right y axis) for BaF/3 cells expressing WT EGFR. Closed squares, high-affinity receptor occupancy; open squares, low-affinity receptor occupancy; closed circles, mitogenic response; open circles, Shc phosphorylation.

Association of Shc with GRB-2 and p145. Signal transduction from Shc proteins to the Ras oncogene product is thought to be mediated by the association of tyrosine-phosphorylated Shc with GRB-2 via the GRB-2 SH2 domain (44, 61). Tyr 317 is the major Shc tyrosine phosphorylation site in cells, and this is a high-affinity binding site for GRB-2 (62). In order to exclude the possibility that Shc may be phosphorylated on different sites after stimulation of particular EGFR mutants, and to determine whether our system is capable of activating the Shc-GRB-2-Sos-Ras pathway, the association between Shc proteins and GRB-2 was examined in the cell lines in which Shc tyrosine phosphorylation had been observed. Lysates from cells stimulated with EGF or with control medium were immunoprecipitated with anti-Shc antibodies, and immunoprecipitates were analyzed by Western blotting with antiphosphotyrosine (Fig. 6A) or anti-GRB-2 (Fig. 6B) antibodies. In all cell lines examined, the GRB-2 protein copurified with tyrosine-phosphorylated Shc. The Shc-GRB-2 association was EGF dependent and occurred only when Shc was phosphorylated on tyrosine. Interestingly, although the level of Shc phosphorylation in BaF/3 cells expressing CT957 EGFR was reduced, the amount of GRB-2 coprecipitating with Shc from CT957 EGFR-expressing cells was similar to that for the WT EGFR-expressing cells.

View larger version (68K): [in this window] [in a new window]   FIG. 6.   EGF-induced association of Shc with p145 and GRB-2 proteins. Quiescent cells (2.4 × 107/cell line) were treated with or without EGF (100 ng/ml) for 15 min at room temperature in the presence of 200 µM sodium pervanadate. Detergent lysates were immunoprecipitated with anti-Shc antibodies and separated by SDS-PAGE with 10% (A) or 15% (B) gels. Proteins were transferred to Immobilon-P membranes and probed with antiphosphotyrosine antibodies (A) or anti-GRB-2 antibodies (B). The results shown are representative of those from three independent experiments.

View larger version (65K): [in this window] [in a new window]   FIG. 7.   Activation of B-Raf in response to EGF. (A) Quiescent cells (6 × 106 per lane) were treated with EGF (100 ng/ml) or control medium for 30 min at room temperature. B-Raf was immunoprecipitated from detergent lysates with anti-B-Raf antibodies. (A) B-Raf kinase activity was assessed by an in vitro kinase assay with kinase-negative MEK as a substrate, followed by SDS-PAGE and autoradiography. (B) The relative intensity of MEK phosphorylation was determined by using ImageQuant and plotted as the percentage of control phosphorylation after correction for the amount of B-Raf protein present in each lane. The bars represent the averages and standard deviations from three separate experiments.

GTP loading of Ras in response to EGF stimulation. The adapter protein GRB-2 exists in a complex with Sos (20), the exchange factor for Ras. Binding of GRB-2-Sos to the tyrosine-phosphorylated Shc leads to the localization of Sos in proximity to Ras and results in the exchange of Ras-GDP to Ras-GTP, the activation of Ras, and the activation of the Raf/MAPK cascade. In order to examine whether mutant EGFRs are capable of activating the Ras/MAPK pathway, GTP loading to Ras protein was studied. The degree of stimulation by EGF of GTP loading, as measured by the total binding of [-³²P]GTP to Ras in electroporated cells, is shown in Table 1. We observed an increased association of GTP with the Ras protein in cell lines expressing WT or mutant EGFRs, with the exception of cells expressing the K721R mutant, which consistently exhibited a decrease in Ras-GTP loading when stimulated with EGF. The WT EGFR- and V741G EGFR-expressing cell lines exhibited comparable levels of Ras stimulation in response to EGF, whereas stimulation in Y740F EGFR- and CT957 EGFR-expressing cells was only just detectable.

Activation of Raf in response to EGF. One of the first events downstream of activated Ras is the activation of at least one member of the Raf family of serine/threonine kinases (34). In mammals the Raf kinase family consists of Raf-1, A-Raf, and B-Raf (17). In preliminary experiments it was clear that EGF did not stimulate the activity of Raf-1 kinase, although the BaF/3 cell line expressed high levels of this protein (data not shown). Recently, it was shown that B-Raf is responsible for Ras-dependent activation of the MAPK pathway in PC12 cells and mammalian (rat and bovine) brain (33, 49, 81) and that B-Raf can physically associate with Ras in an activated state (49). Consequently we assessed the activation of B-Raf by its ability to phosphorylate its substrate, extracellular signal-regulated kinase kinase (MEK), in "in vitro" kinase reactions. Initial experiments showed that B-Raf activation by EGF in BaF/3 cells occurred with a relatively slow time course (50% maximal at 15 min), so we chose an EGF stimulation time of 30 min for these Raf activation experiments. EGF treatment resulted in the activation of B-Raf in all EGFR-expressing BaF/3 cell lines, with the exception of K721R (Fig. 7A). Typically we observed a threefold increase in the phosphorylation of MEK following EGF treatment (Fig. 7B). IL-3 also activates B-Raf in these cells at levels comparable to EGF; however, IL-3-dependent activation of B-Raf occurs with a faster time course (50% maximal at 5 min), and by 30 min the response to IL-3 was only marginally above background in all cell lines (data not shown).

MAPK activation in BaF/3-derived cell lines. Activation of Raf results in activation of the serine/threonine kinases Erk-1 and -2 (MAPKs) via the tyrosine/threonine kinase MEK (6, 68). We assessed the EGF-dependent activation of MAPK by using an antibody that recognizes specifically the activated, phosphorylated form of the protein (Fig. 8). Activated MAPK was detected in all cell lines after, but not before, stimulation with EGF, with the exception of cells expressing the K721R EGFR mutant. Similar results were observed when activation of MAPK was tested directly by measuring its ability to phosphorylate its substrate MBP in vitro (data not shown).

View larger version (30K): [in this window] [in a new window]   FIG. 8.   MAPK activation by EGFR mutants. Quiescent cells were incubated with control medium () or medium containing EGF (100 ng/ml) for 30 min at room temperature. Total cell lysates were analyzed directly by SDS-PAGE and Western blotting with anti-activated MAPK antibodies.

Correlation between MAPK activation and EGF-dependent survival. From our results it is clear that EGF binding results in the activation of the Shc/Ras/MAPK pathway in BaF/3 cells which express either kinase-active or kinase defective -helix C mutants of the EGFR. With the exception of BaF/3 cells expressing the K721R EGFR, EGF can replace IL-3 as a survival stimulus; however, only in cells expressing the kinase-active receptors does EGF stimulate proliferation. Therefore, we postulated that survival, but not proliferation, is mediated by the activation of the Ras/MAPK pathway. To test this hypothesis, we used a specific inhibitor of MAPK activation, PD98059; the inhibitor prevents the association of MAPK (Erk-1 and Erk-2) with MEK and hence selectively blocks the Ras/MAPK pathway (18). PD98059 was indeed effective in abrogating the stimulation by EGF of MAPK activity in the EGFR-expressing BaF/3 cells (Fig. 9A); in all cell lines MAPK activity remained at background levels when the cells were incubated with EGF in the presence of PD98059. Specifically, at a concentration of 50 µM, PD98059 was equally effective in inhibiting MAPK activation in cells expressing WT and mutant EGFRs. At this concentration PD98059 completely abrogated the EGF-dependent survival of V741G and Y740F EGFR-expressing cells but had no effect on EGF-stimulated proliferation of WT and CT957 EGFR-expressing cells (Fig. 9B). Even when the inhibitor concentration was increased to 100 µM, there was no inhibition of EGF-stimulated proliferation (data not shown). EGF activation of the MAPK signalling cascade is therefore linked to survival rather than proliferation of BaF/3 cells.

View larger version (20K): [in this window] [in a new window]   FIG. 9.   Effects of MAPK inhibition on cell survival and proliferation. (A) Quiescent cells were preincubated in control medium containing PD98059 (50 µM) or carrier only (DMSO) prior to stimulation with EGF. MAPK activation was measured as described in Materials and Methods. The relative intensities of MBP phosphorylation in control cells and cells incubated with EGF in the presence or absence of the inhibitor are shown. (B) Cells were seeded at 105/ml (dashed line) in control medium (RPMI 1640-10% FCS), in medium containing 15 nM EGF, or in medium containing 15 nM EGF and 50 µM PD98059. Viable cell numbers were determined in triplicate after 4 days. Results are the means and standard errors from three replicate experiments.

Use of kinase inhibitors to dissect EGFR signalling pathways. The experiments described above suggest that kinase-active EGFRs (WT and CT957) rely on pathways other than the MAPK pathway to induce proliferation in BaF/3 cells. In an attempt to define these alternative pathways, we used specific inhibitors to selectively block some of the molecules likely to be involved in EGFR signalling, such as Src kinases (7), JAK kinases (71, 80), and PI-3-kinase (32). These experiments were performed with BaF/3 cells expressing the WT (kinase-active) or the V741G (kinase-inactive) EGFR. As shown in Fig. 10, the cellular responses to EGF of the WT and V741G EGFR-expressing BaF/3 cells were affected quite differently by inhibitors of cellular kinases. The EGFR kinase inhibitor AG1478 completely abrogates EGF-induced proliferation in WT EGFR-expressing cells but does not significantly affect EGF-dependent survival of V741G EGFR-expressing cells. This result further strengthens the hypothesis that cellular survival in response to EGF in V741G EGFR-expressing cells is mediated by an associated kinase rather than the EGFR kinase itself. The JAK kinase inhibitor AG490 reduces the proliferation of WT cells to "survival-only" levels and has no effect on survival of V741G cells. Inhibition of the Src family of kinases by PP1 has the opposite effect: proliferation of WT EGFR-expressing BaF/3 cells is unaffected by PP1, while survival of V741G EGFR-expressing cells is abolished completely. Finally, the potent and specific inhibitor of PI-3-kinase LY294002 completely abrogates the proliferation and survival responses to EGF in both cell lines. The inhibitors' effects on the response to IL-3 was identical in WT and V741G EGFR-expressing cells, ruling out the possibility of clonal differences, unrelated to the expression of different EGFR constructs, between these cell lines (Fig. 11). Proliferation in both IL-3 and EGF is strongly inhibited by the JAK kinase inhibitor AG490 (50% inhibitory concentration [IC₅₀], 1 to 3 µM), while survival in EGF is only partially affected even at much higher concentrations of the inhibitor. Conversely, the Src kinase inhibitor PP1 appears to affect selectively EGF-dependent survival (IC₅₀, 1 to 3 µM). WT EGFR-expressing cells grown in EGF rather than IL-3 are slightly more sensitive to inhibition by PP1 (IC₅₀, 30 versus >100 µM); this could be simply a reflection of the suboptimal growth of WT EGFR-expressing cells in EGF, or it could result from a partial dependency of EGF signalling on an Src-related pathway. In our hands, the PI-3-kinase inhibitor LY294002 is effective in blocking growth factor-mediated proliferation and survival, although it is significantly more active on EGF-mediated survival (V741G) than proliferation (WT).

View larger version (32K): [in this window] [in a new window]   FIG. 10.   Effect of kinase inhibitors on EGF-dependent survival and proliferation. Cells were seeded at 1.5 × 105/ml (dashed line) in control medium or in medium containing EGF (15 nM) or EGF plus the stated concentrations of inhibitors. Viable cell numbers were determined after 3 days. All inhibitors were dissolved in DMSO, and the concentration of DMSO was adjusted to 0.5% in all wells. Results are the averages and standard errors from three replicate wells and are representative of those from three separate experiments.

View larger version (30K): [in this window] [in a new window]   FIG. 11.   Titration of kinase inhibitors in cultures containing IL-3 or EGF. WT EGFR-expressing cells (closed circles) and V741G EGFR-expressing cells (open circles) were cultured as described in the legend to Fig. 10 in either medium alone or medium containing EGF (15 nM) or IL-3 (10% WEHI-3B conditioned medium) with or without serial dilutions of the kinase inhibitors. DMSO was kept constant in all wells at 0.5%. Viable cell numbers were determined at day 3 and are presented as percentages of control values (determined from wells containing the stimulus but no inhibitor). The numbers of viable cells in the control cultures were as follows: WT EGFR plus EGF, 3.9 × 105 ± 0.05 × 105/ml; WT EGFR plus IL-3, 10.4 × 105 ± 0.6 × 105/ml; V741G EGFR plus EGF, 1.2 × 105 ± 0.05 × 105/ml; V741G EGFR plus IL-3, 10.4 × 105 ± 0.1 × 105/ml.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

It is important to clarify the confusion presently surrounding the EGFR mitogenic signalling pathways. Although innumerable studies have addressed signalling from the EGFR, almost all of these studies have involved the expression of these receptors in cells where EGFR family members can heterodimerize and signal. It is now clear that many events attributed to EGFR are in fact initiated by heterodimer partners. Reports aimed at dissecting the contributions of the different EGFR family members to EGF signalling in 32D (58) and BaF/3 (59) cells have been published recently. However, even these recent studies have yet to address the requirement for the EGFR kinase in the induction of specific signalling pathways and the relationship of these pathways to the mitogenic effects of the activated EGFR signal.

The primary aim of this project was to study the early events associated with the signalling ability of WT, C-terminally truncated, or kinase-impaired EGFR mutants in the absence of heterodimer-generated signals. We have expressed these receptors in the murine hemopoietic BaF/3 cell line, which is devoid of other EGFR family members; to our knowledge this is the first study to demonstrate that EGFR can stimulate mitogenesis and to define signalling pathways activated by EGFR mutants in the absence of endogenous ErbB-2, ErbB-3, and ErbB-4. Our present studies have allowed two important signalling questions to be addressed: the relevance of Shc phosphorylation and the Ras/MAPK pathway to proliferation and the importance of receptor-associated signalling complexes to the activation of pathways leading to mitogenesis.

Kinase-active EGFR homodimers are mitogenically competent. Our data show that ligand binding to EGFRs with an active kinase domain results in mitogenic signalling in the absence of other EGFR family members. Recent work with WT EGFR expressed in BaF/3 cells (59) has shown that this receptor does not deliver a proliferative signal unless coexpressed with either ErbB-2 or ErbB-4. However, by changing the culture conditions, we have been able to detect a mitogenic response to EGF in BaF/3 cells expressing only the EGFR. When stimulated with EGF in minimal medium, WT EGFR-expressing BaF/3 cells are delayed in S phase, and this delay is responsible for their limited proliferative response (data not shown). The low concentration of IL-3 in the mitogenic assay used in our experiments is insufficient to trigger either exit from G₀ or initiation of DNA synthesis but in synergy with EGF facilitates the transition from S phase to G₂/M (data not shown). The report by Riese et al. (59) that coexpression of EGFR with ErbB-2 or ErbB-4 allows EGF-dependent proliferation suggests that heterodimer signalling in this cellular system may also be involved in the exit from S phase.

The C terminus of the EGFR is not required for mitogenic signalling. Truncation of the EGFR at residue 957 eliminates all five recognized autophosphorylation sites on the C terminus of the receptor, making it unlikely that adapter proteins containing SH2 or phosphotyrosine-binding domains can physically associate with this mutant receptor. Although association of the Shc protein with the C-terminal phosphotyrosines of EGFR has been proposed as an important step towards Shc phosphorylation (61), our results indicate that an intact C terminus may not be required for EGF-stimulated Shc phosphorylation. Interestingly, Shc is phosphorylated at the normal rate, in response to EGF, in fibroblasts expressing C-terminally truncated EGFRs (27); in these cells Shc phosphorylation could be facilitated by the heterodimerization of the truncated EGFR with ErbB-2 and by the association of Shc with the ErbB-2 C terminus (56). The decreased level of Shc phosphorylation in BaF/3 cells expressing the CT957 EGFR, compared to those expressing the WT EGFR, may be due to the lack of stable complex formation between the receptor and Shc; however, the apparently normal GRB-2 binding and Erk-1 activation suggest that the level of Shc phosphorylation induced by CT957 is sufficient for stimulation of this signalling pathway. In a recent review, Pawson and Scott (55) espouse "recruitment of active signalling molecules into multiprotein signalling networks" as the main coordinator of signalling repertoires. They summarize: "Simply stated, either the enzyme goes to the signal or the signal goes to the enzyme." Undoubtedly this is true, but it is not necessary to form a stable receptor-signalling complex. The full proliferative response to EGF of CT957 EGFR-expressing cells indicates that neither C-terminal autophosphorylation of the EGFR nor the consequent association with SH2-containing proteins is required for mitogenic signalling in BaF/3 cells.

Role of Shc phosphorylation and Ras/MAPK activation in BaF/3 cells. Shc is phosphorylated on Tyr 317 following EGF stimulation, and the phosphorylated Tyr 317 serves as a docking site for GRB-2 (63). Since GRB-2 is directly associated with the Ras guanine nucleotide-releasing factor Sos, phosphorylation of Tyr 317 in Shc has been proposed as an important link in the EGF-induced Ras activation, leading to colocalization of Sos and Ras at the plasma membrane (47). Circumstantial evidence has also linked Shc phosphorylation directly to mitogenesis: constitutive phosphorylation of Shc has been detected in tumor cells (57), and Shc overexpression is tumorigenic in nude mice (56). The correlation between genetic blocks of the Ras/MAPK pathway and differentiation in Drosophila, Caenorhabditis elegans, and even yeast yielded the elements of a complex signalling process capable of both transferring and amplifying a signal from the membrane to the nucleus via a series of protein-protein interactions. However, there are numerous biological responses to growth factors and cytokines: morphogenesis, vesicle secretion, cell movement, contractile processes, membrane turnover, and gene activation associated with differentiation, cell survival, and mitogenesis. In this context the association of the EGFR/Shc/Sos/Ras/MAPK pathway with mitogenesis has been tenuous at best. Our results with the V741G and Y740F mutant EGFRs demonstrate that in BaF/3 cells the tyrosine phosphorylation of Shc in response to EGF is not sufficient to induce mitogenesis. Furthermore, Shc is efficiently phosphorylated in response to EGF in cells expressing kinase-defective EGFR mutants, suggesting that it can be the substrate of an associated kinase as well as of the EGFR kinase itself. Obvious candidates for such a kinase are the Src family of cytoplasmic tyrosine kinases (reviewed in reference 21) and p72 Syk (spleen tyrosine kinase), which are apparently pivotal in coupling antigen receptors and Fc receptors to downstream signalling events (1). Further experiments to identify the putative kinase(s) activated by EGF binding to the EGFR in this experimental system are under way.