Src and Pyk2 Mediate G-protein-coupled Receptor Activation of Epidermal Growth Factor Receptor (EGFR) but Are Not Required for Coupling to the Mitogen-activated Protein (MAP) Kinase Signaling Cascade*

The epidermal growth factor receptor (EGFR) and the non-receptor protein tyrosine kinases Src and Pyk2 have been implicated in linking a variety of G-protein-coupled receptors (GPCR) to the mitogen-activated protein (MAP) kinase signaling cascade. In this report we apply a genetic strategy using cells isolated from Src-, Pyk2-, or EGFR-deficient mice to explore the roles played by these protein tyrosine kinases in GPCR-induced activation of EGFR, Pyk2, and MAP kinase. We show that Src kinases are critical for activation of Pyk2 in response to GPCR-stimulation and that Pyk2 and Src are essential for GPCR-induced tyrosine phosphorylation of EGFR. By contrast, Pyk2, Src, and EGFR are dispensable for GPCR-induced activation of MAP kinase. Moreover, GPCR-induced MAP kinase activation is normal in fibroblasts deficient in both Src and Pyk2 (Src−/−Pyk2−/− cells) as well as in fibroblasts deficient in all three Src kinases expressed in these cells (Src−/−Yes−/−Fyn−/− cells). Finally, experiments are presented demonstrating that, upon stimulation of GPCR, activated Pyk2 forms a complex with Src, which in turn phosphorylates EGFR directly. These experiments reveal a role for Src kinases in Pyk2 activation and a role for Pyk2 and Src in tyrosine phosphorylation of EGFR following GPCR stimulation. In addition, EGFR, Src family kinases, and Pyk2 are not required for linking GPCRs with the MAP kinase signaling cascade.

The protein tyrosine kinases Src, Pyk2, and epidermal growth factor receptor (EGFR) 1 have been implicated as intermediates in signaling networks that couple G-protein-coupled receptors (GPCRs) with the Ras/MAP kinase signaling cascade (1)(2)(3)(4). Al-though the mechanisms underlying these pathways are not fully defined, at least EGFR activation by GPCRs was proposed to be mediated by intracellular (1,4) and extracellular (5) processes. Many ligands acting via GPCRs are known to elicit a mitogenic response in a variety of cell types; MAP kinases appear to be a critical component of these growth-promoting pathways (6). It has been shown that GPCR-mediated activation of MAP kinases occurs via pertussis toxin-sensitive and -insensitive processes as well as by Ras-dependent and Ras-independent mechanisms (2,4,7,8). Because many GPCRs couple to more than one G-protein subtype, their activation will initiate simultaneous stimulation of multiple effector systems. In fibroblasts, however, lysophosphatidic acid (LPA)-induced activation of the MAP kinase pathway is mediated solely by G i -regulated Ras activation (9,10). Several protein tyrosine kinases have been implicated as intermediates between G i and Ras/MAP kinase activation, including Src, Pyk2, and EGFR (1)(2)(3)(4)11). It has been shown that EGFR, Pyk2, and Src family kinases are rapidly and transiently activated by various GPCRs. In PC12 cells, activated Pyk2 binds to Src, association-mediated through binding of the Src homology 2 domain of Src to pY402 of Pyk2 (11). Experiments employing a dominant interfering EGFR mutant or pretreatment of cells with inhibitors of the protein tyrosine kinase of EGFR suggested that EGFR plays a role in GPCR-induced mitogenic response (12)(13)(14). It has been proposed that Src kinases play a crucial role in activation of EGF receptors by GPCR stimulation (14,15). Moreover, it has been demonstrated that inhibition of Src kinase activity impairs LPA and ␤ 2 -adrenergic receptor-mediated activation of MAP kinase (16).
We have employed a genetic strategy to explore the role played by Src, Pyk2, and EGFR in GPCR-induced activation of EGFR and MAP kinase. Using cells isolated from Src-or Pyk2deficient mice we show that Src kinases are critical for activation of Pyk2 and that Src and Pyk2 are essential for GPCRinduced activation of EGFR. However, these kinases are dispensable for GPCR-induced activation of MAP kinase in mouse embryonic fibroblasts. Moreover, GPCR-induced MAP kinase activation is normal in fibroblasts deficient in both Src and Pyk2 as well as in fibroblasts deficient in all three Src kinases expressed in these cells (Src, Yes, and Fyn). Using fibroblasts isolated from EGFR-deficient mice, we showed that EGFR activation is dispensable for GPCR-induced activation of MAP kinase. The experiments described in this report reveal an intracellular mechanism for activation of Pyk2 and EGFR by Src kinases following GPCR stimulation and demonstrate that EGFR, Src, and Pyk2 kinases are not required for linking GPCRs with the MAP kinase signaling cascade.

EXPERIMENTAL PROCEDURES
Reagents, Antibodies and Plasmids-LPA, carbachol, and bradykinin were purchased from Sigma. Recombinant human EGF was from Intergen. EGFR kinase inhibitors SU1478/009 and SU1517/002 were provided by Sugen, Inc. Monoclonal anti-Pyk2 antibodies used for immunoblotting and anti-Ras antibodies were obtained from Transduction Laboratories. Polyclonal anti-Pyk2 antibodies that were used for immunoprecipitation were previously described (11). Monoclonal antiphosphotyrosine antibodies (4G10) were from Upstate Biotechnology. Polyclonal anti-EGFR antibodies RK-2 and anti-C antibodies were used for immunoprecipitation and immunoblotting of EGFR (17). Polyclonal antibodies against phospho-MAP kinase were from New England Bio-Labs. Monoclonal antibodies against Src that were used for immunoprecipitation were from Calbiochem. The Pyk2 expression vector we used was previously described (18). The Src expression vector was obtained from J. Sap, New York University, NY. The EGFR expression vector was previously described (17). Kinase negative mutant of EGFR (EGFR-KA) was generated by substituting the codons of Lys-745 with an Ala residue using Stratagene site-directed mutagenesis kit (Stratagene).
Ras Activation Assay-Cells were lysed in a buffer containing 20 mM Hepes, pH 7.4, 1% Nonidet P-40, 150 mM NaCl, 5 mM MgCl 2 , and 10% glycerol. The lysates were incubated with 20 g of GST-Raf(RBD) provided by J. Bos, Utrecht, The Netherlands and were washed three times in lysis buffer. The amount of Ras pulled down was then assessed by immunoblotting with anti-Ras antibodies. In all pull-down experiments, the lysates were precleared using GST. Analysis of Ras binding to GST alone was performed in each assay but was not detected, indicating that Ras binds to GST-RBD in RBD-and stimulus-dependent manners.

RESULTS AND DISCUSSION
Stimulation of mouse embryonic fibroblasts with LPA, bradykinin, or carbachol acting on their cognate GPCRs triggers tyrosine phosphorylation of EGFR ( Fig. 1, A). The stimulation of EGFR tyrosine phosphorylation by these GPCR agonists is similar to the level of EGFR phosphorylation seen after stimulation with 2 ng/ml of EGF (Fig. 1, A). Higher concentrations of EGF caused higher levels of EGFR tyrosine phosphorylation, but the level of MAP kinase activation remained the same; it was similar to that induced by LPA (2.5 M) (Fig. 1, B). It has been proposed that Src kinases play a pivotal role in EGFR activation in response to GPCR stimulation (2, 4, 15). Because we had previously demonstrated that Pyk2 and Src are acti-vated by the same stimuli (11), we examined the effects of Src or Pyk2 deficiencies on LPA-induced activation of EGFR.
Src and Pyk2 Are Essential for LPA-induced Activation of EGF-SrcϪ/Ϫ or Pyk2Ϫ/Ϫ fibroblasts were stimulated with LPA, and lysates from unstimulated or stimulated cells were subjected to immunoprecipitation with anti-EGFR antibodies followed by immunoblotting with antibodies against phosphotyrosine (anti-pTyr). The experiment presented in Fig. 2 demonstrates that both the amplitude and kinetics of LPA-induced activation of EGFR were reduced in SrcϪ/Ϫ or Pyk2Ϫ/Ϫ fibroblasts. We next examined the time course of LPA-induced activation of EGFR in fibroblasts derived from mice deficient in both Src and Pyk2. The experiment presented in Fig. 2, lower right panel shows higher basal phosphorylation of EGFR in SrcϪ/ϪPyk2Ϫ/Ϫ cells, but LPA-induced stimulation of tyrosine phosphorylation of EGFR could not be detected in these fibroblasts. Similar results were obtained with SrcϪ/ϪPyk2Ϫ/Ϫ fibroblasts that were stimulated with bradykinin or carbachol (data not shown). In addition, similar results were obtained with several different primary cultures and with immortalized SrcϪ/ϪPyk2Ϫ/Ϫ fibroblasts (data not shown). These experiments demonstrate that Src and Pyk2 are essential for GPCRinduced activation of EGFR.
Activation of EGFR Is Not Essential for GPCR-induced Stimulation of MAP Kinase-If EGFR plays a critical role in the coupling of GPCRs to MAP kinase activation (12)(13)(14) and Src and Pyk2 are essential for tyrosine phosphorylation of EGFR, one would expect GPCR-induced activation of MAP kinase to be altered in SrcϪ/Ϫ or Pyk2Ϫ/Ϫ fibroblasts and particularly in SrcϪ/ϪPyk2Ϫ/Ϫ fibroblasts. However, LPA-stimulated MAP kinase activation was readily detected in cells deficient in either Src (20) or Pyk2 and in cells deficient in both protein tyrosine kinases (Fig. 3A). In these experiments the cells were stimulated with LPA for 3 min and analyzed for MAP kinase activation using antibodies that recognize the phosphorylated (activated) form of MAP kinase. We next compared the kinetics of LPA-induced MAP kinase activation in SrcϪ/Ϫ, Pyk2Ϫ/Ϫ, or SrcϪ/ϪPyk2Ϫ/Ϫ cells with the kinetics of LPA-induced MAP kinase activation in wild type fibroblasts. Again, no detectable differences in the kinetics of MAP kinase stimulation were observed (Fig. 3B). Interestingly, LPA-stimulated MAP kinase activation was also detected in SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts, cells deficient in all Src family members that are expressed in these cells (21) (Fig. 3C, left panel). Moreover, LPAinduced activation of Ras and MAP kinase in SrcϪ/ϪYesϪ/ ϪFynϪ/Ϫ fibroblasts was inhibited by pertussis toxin (Fig. 3C,  right panel), indicating that MAP kinase activation in these cells is mediated by G i as it is in wild type cells (9,10). These experiments demonstrate that LPA-induced MAP kinase activation is normal in fibroblasts deficient in all Src kinases and that Pyk2 and EGFR activation are not essential for MAP kinase response.
To address the role of EGFR in LPA-induced MAPK activation more directly, we compared LPA-stimulation of MAP kinase in wild type fibroblasts to fibroblasts derived from EGFRϪ/Ϫ mice. In this experiment, wild type or EGFRϪ/Ϫ fibroblasts were stimulated with LPA for different times. Lysates from unstimulated or stimulated cells were analyzed for MAP kinase activation with antibodies that recognize activated, phosphorylated MAP kinase. These experiments showed no difference in the kinetics or amplitude of LPA-induced MAP kinase activation of wild type and EGFRϪ/Ϫ fibroblasts at the indicated time points (Fig. 4A) and up to 90 min of LPA stimulation (data not shown). Surprisingly, pretreatment of wild type fibroblasts with two different inhibitors that block the tyrosine kinase activity of EGFR but not Src did not reduce EGF or LPA-induced MAP kinase activation (Fig. 4B). It is noteworthy that the tyrosine kinase inhibitors that were used in this experiment also block the tyrosine kinase activity of erbB2 and erbB4 (data not shown), demonstrating that erbB2 and erbB4 do not compensate for the loss of EGFR and that these receptors are not involved in this process. Taken together, these experiments demonstrate that activation of EGFR is not essential for GPCR-induced MAP kinase activation.
We have previously shown that stimulation of PC12 cells with LPA leads to tyrosine phosphorylation of Pyk2, which in turn recruits Src via its Src homology 2 domain (11). The experiment presented in Fig. 5A shows that LPA stimulation leads to complex formation between endogenous Src and Pyk2 in fibroblasts. Moreover, in accordance with the data reported previously for different cell types (9,14,22), Src and EGFR form a complex in LPA-stimulated fibroblasts, as shown in Fig.  5A. However, we have detected association between EGFR and Src in lysates from unstimulated Pyk2Ϫ/Ϫ fibroblasts. It is possible that Src forms a complex with EGFR in unstimulated Pyk2Ϫ/Ϫ fibroblasts because of the higher basal activity of EGFR in Pyk2Ϫ/Ϫ cells. To further analyze the nature of the association between Src-Pyk2 complex and EGFR, 293 cells were transfected with expression vectors that direct the synthesis of EGFR and EGFR-KA (a kinase negative EGFR point mutant at Lys-745) together with expression vectors for Src or Pyk2. The experiment presented in Fig. 5B shows that, in transfected 293 cells, the kinase negative mutant of EGFR is tyrosine-phosphorylated by Src but not by Pyk2.
Src Kinases Are Critical for GPCR-induced Stimulation of Pyk2-We have next tested whether LPA-induced activation of Pyk2 is dependent upon Src kinases. In this experiment wild type or SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts were stimulated with LPA, and lysates from stimulated or unstimulated cells were subjected to immunoprecipitation with anti-Pyk2 anti- Embryonic fibroblasts derived from wild type (WT), SrcϪ/Ϫ, Pyk2Ϫ/Ϫ, or Pyk2Ϫ/ ϪSrcϪ/Ϫ mutant mice were starved for 48 h and either left untreated or stimulated with LPA (2.5 M) or EGF (2 nM) for the indicated times, lysed, and subjected to SDS-PAGE after immunoprecipitation (IP) with anti-EGFR antibodies followed by immunoblotting with anti-phosphotyrosine antibodies (pY). The filters were stripped and reblotted with anti-EGFR antibodies. All filters were developed simultaneously for quantitative comparison of the results. Position and size (kDa) of a standard protein marker is indicated on the right.
bodies followed by immunoblotting with anti-pTyr antibodies. The experiment presented in Fig. 6 shows that LPA failed to stimulate Pyk2 activation in SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts, demonstrating that Src kinases are crucial for LPAinduced activation of Pyk2. Similar results were obtained with SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts stimulated with bradykinin, carbachol, or angiotensin II (data not shown). However, LPA-induced activation of Pyk2 was normal in SrcϪ/Ϫ fibroblasts, indicating that either Yes or Fyn are able to compensate for the loss of Src in these cells. Indeed, stimulation of SrcϪ/ ϪFynϪ/Ϫ fibroblasts with LPA resulted in normal activation of Pyk2, confirming that Yes is capable in mediating LPA-induced Pyk2 activation. Moreover, LPA-induced Pyk2 activation is completely restored in SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts by tivation was analyzed by immunoblotting of total lysates using antiphospho-MAPK antibodies. All filters were developed simultaneously for quantitative comparison of the results. Positions and sizes (kDa) of standard protein markers are indicated on the right.

FIG. 3. Activation of Src or Pyk2 is not essential for LPAinduced MAP kinase activation in fibroblasts.
A, embryonic fibroblasts derived from wild type (WT), SrcϪ/Ϫ, Pyk2Ϫ/Ϫ, or Pyk2Ϫ/ ϪSrcϪ/Ϫ mutant mice were starved for 48 h and either left untreated or stimulated with LPA (2.5 M) for 3 min, lysed, subjected to SDS-PAGE, and immunoblotted with antibodies that recognize the activated form of MAP kinase (pMAPK). The filter was stripped and reblotted with anti-ERK1 antibodies as a loading control. All filters were developed simultaneously for quantitative comparison of the results. B, embryonic fibroblasts derived from SrcϪ/Ϫ, Pyk2Ϫ/Ϫ, or Pyk2Ϫ/ ϪSrcϪ/Ϫ mutant mice were starved for 48 h and either left untreated or stimulated with EGF (2 nM) or LPA (2.5 M) for the indicated times, lysed, subjected to SDS-PAGE, and immunoblotted with antibodies that recognize the activated form of MAP kinase (pMAPK). The same filters were reblotted with anti-ERK1 antibodies. All filters were developed simultaneously for quantitative comparison of the results. C, LPAinduced pertussis toxin-sensitive activation of Ras and MAP kinase in SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts. Mouse embryonic fibroblasts derived from wild type mice (WT) or SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ mice were cultured in serum-free medium overnight in the absence or presence of 100 ng/ml pertussis toxin. Cells were then stimulated with LPA (2.5 M) for 3 min, and Ras activation was assessed by GST-RBD pull-down from cell lysates as described under "Experimental Procedures." MAPK ac-

FIG. 4. EGFR activation is not essential for LPA-induced MAP kinase activation in fibroblasts.
A, time course of LPA-induced MAP kinase activation in wild type or EGFRϪ/Ϫ fibroblasts. Mouse embryonic fibroblasts derived from wild type (WT) or EGFRϪ/Ϫ mutant mice were starved for 48 h and either left untreated or stimulated with LPA (2.5 M) for the indicated times, lysed, subjected to SDS-PAGE, and immunoblotted with antibodies that recognize the activated form of MAP kinase (pMAPK). The filters were stripped and reblotted with anti-ERK1 antibodies. B, inhibitors of EGFR do not influence EGF-or LPA-induced MAP kinase activation in wild type fibroblasts. Mouse embryonic fibroblasts isolated from wild type mice were starved for 48 h and either left untreated or pre-treated with EGFR tyrosine kinase inhibitors SU1478/009 or SU1517/002 (1 M) for 20 min, stimulated with EGF (2 nM) or LPA (2.5 M) for 3 min, lysed, subjected to SDS-PAGE directly (Total lysate), and analyzed with antibodies that recognize the activated form of MAP kinase (pMAPK). Lysates were also subjected to immunoprecipitation (IP) with anti-EGFR or anti-Src antibodies and analyzed with anti-phosphotyrosine antibodies (pY). The filters were stripped and reblotted with anti-ERK1, anti-EGFR, or anti-Src antibodies, respectively. Positions and sizes (kDa) of standard protein markers are indicated on the right. ectopic expression of Src but not by a kinase negative Src mutant demonstrating that expression of a single active Src kinase is sufficient for linking LPA stimulation with Pyk2 activation.

CONCLUSIONS
Previous studies performed in PC12 cells using dominant negative mutants of Pyk2 (11) and studies performed in Rat-1 fibroblasts using dominant interfering mutants of EGFR or Src (12,23) demonstrated that EGFR, Src, and Pyk2 play a role in linking GPCR activation with the MAP kinase signaling pathway. Experiments performed in embryonic fibroblasts derived from EGFRϪ/Ϫ, SrcϪ/Ϫ, Pyk2Ϫ/Ϫ, or SrcϪ/ϪPyk2Ϫ/Ϫ mice demonstrate that both Src and Pyk2 are essential for GPCRinduced tyrosine phosphorylation of EGFR. However, together with EGFR, they are dispensable for coupling to the MAP kinase signaling cascade. Moreover, GPCR-induced MAP kinase stimulation is normal in fibroblasts deficient in Src, Yes, and Fyn. Taken together, these studies show that different signaling networks may couple GPCRs with the Ras/MAP kinase signaling pathway in different cell types. In fibroblasts, different experimental strategies have produced conflicting results as to the role played by EGFR, Src, and Pyk2 in GPCRinduced MAP kinase activation. Experiments with dominant interfering EGFR and Src mutants could be misleading because vast overexpression of mutant proteins necessary for inhibition of MAP kinase may interfere with the action of other proteins involved in linking GPCR activation with the MAP kinase signaling cascade. Certain EGFR kinase inhibitors are certainly not sufficiently specific, and most likely block the action of protein tyrosine kinases other than EGFR. In addition, studies using a genetic approach are not always easy to interpret; deficiency in a protein tyrosine kinase may be compensated for by another member of the same family of enzymes. We have addressed this issue for Src kinases by using fibroblasts deficient in all Src kinases expressed in these cells. Our results clearly demonstrate that GPCR-induced activation of MAP kinase is normal even in SrcϪ/ϪYesϪ/ϪFynϪ/Ϫ fibroblasts. However, a single Src kinase is sufficient for mediating LPA-induced stimulation of Pyk2.
Taken together, these studies suggest that establishment of intracellular links between signaling proteins cannot be based upon the sole use of inhibitors or expression of dominant negative mutant proteins. Genetic tools should be used for establishing links between signaling pathways, for testing hypotheses, and confirming conclusions drawn from experiments based upon indirect approaches. Finally, an alternative and intriguing possibility is that the link between GPCR activation and the MAP kinase signaling cascade is mediated by several parallel signaling pathways. Inhibition of one of the pathways by a drug or mutation may lead to a bypass or rerouting of the information flow via alternative pathway(s) for coupling GPCR-activation with the MAP kinase signaling cascade.