C-erbB and the epidermal growth-factor receptor: a molecule with dual identity

Grb2 regulation of the actin-based cytoskeleton is required for ligand-independent EGF receptor-mediated oncogenesis

Mutations within members of the EGF/ErbB receptor family frequently release the oncogenic potential of these receptors, resulting in the activation of downstream signaling events independent of ligand regulatory constraints. We previously have demonstrated that the signal transduction events originating from S3-v-ErbB, a ligand-independent, oncogenic EGF receptor mutant, are qualitatively distinct from the ligand-dependent mitogenic signaling pathways associated with the wild-type EGF receptor. Specifically, expression of S3-v-ErbB in primary fibroblasts results in anchorage-independent growth, increased invasive potential, and the formation of a transformation-specific phosphoprotein signaling complex, all in a Ras-independent manner. Here we demonstrate the transformation-specific interaction between two components of this complex: the adaptor protein Grb2 and the cytoskeletal regulatory protein caldesmon. This interaction is mediated via both the amino-terminal SH3 and central SH2 domains of Grb2, and the amino-terminal (myosin-binding) domain of caldesmon. Expression of a dominant-negative Grb2 deletion mutant, which lacks the carboxy-terminal SH3 domain, in fibroblasts expressing S3-v-ErbB results in a reduction in phosphoprotein complex formation, the loss of anchorage-independent growth, and a reduction in invasive potential. Together, these results demonstrate a Ras-independent role for Grb2 in modulating cytoskeletal function during ligand-independent EGF receptor-mediated transformation, and provide further support for the hypothesis that ligand-independent oncogenic signaling is qualitatively distinct from ligand-dependent mitogenic signaling by the EGF receptor.

A transformation-associated complex involving tyrosine kinase signal adapter proteins and caldesmon links v-erbB signaling to actin stress fiber disassembly

The avian erythroblastosis viral oncogene (v-erbB) encodes a receptor tyrosine kinase that possesses sarcomagenic and leukemogenic potential. We have expressed transforming and nontransforming mutants of v-erbB in fibroblasts to detect transformation-associated signal transduction events. Coimmunoprecipitation and affinity chromatography have been used to identify a transformation-associated, tyrosine phosphorylated, multiprotein complex. This complex consists of Src homologous collagen protein (Shc), growth factor receptor binding protein 2 (Grb2), son of sevenless (Sos), and a novel tyrosine phosphorylated form of the cytoskeletal regulatory protein caldesmon. Immunofluorescence localization studies further reveal that, in contrast to the distribution of caldesmon along actin stress fibers in normal fibroblasts, caldesmon colocalizes with Shc in plasma membrane blebs in transformed fibroblasts. This colocalization of caldesmon and Shc correlates with actin stress fiber disassembly and v-erbB-mediated transformation. The tyrosine phosphorylation of caldesmon, and its association with the Shc-Grb2-Sos signaling complex directly links tyrosine kinase oncogenic signaling events with cytoskeletal regulatory processes, and may define one mechanism regulating actin stress fiber disassembly in transformed cells.

A minor tyrosine phosphorylation site located within the CAIN domain plays a critical role in regulating tissue-specific transformation by erbB kinase

Avian c-erbB encodes a protein that is homologous to the human epidermal growth factor receptor. Truncation of the amino-terminal, ligand-binding domain of this receptor results in an oncogene product which is a potent inducing agent for erythroleukemias but not fibrosarcomas in chickens. Here we show that mutation of a single tyrosine residue, p5, in the carboxyl terminus of the erbB oncogene product allows it to become sarcomagenic in vivo and to transform fibroblasts in vitro. Mutations of other autophosphorylation sites do not generate comparable effects. The increased transforming activity of the p5 mutant is accompanied by an elevated level of mitogen-activated protein kinase phosphorylation. By analogy to the human epidermal growth factor receptor, p5 is a minor autophosphorylation site and is located in a domain known to be involved in regulating calcium influx and receptor internalization (CAIN domain). This area of the erbB product has been found to be repeatedly deleted in various sarcomagenic avian erythroblastosis virus isolates. We precisely deleted the CAIN domain and also made point mutations of the acidic residues within the CAIN domain. In both cases, fibroblast-transforming potential is activated. We interpret these data to mean that p5 and its surrounding region negatively regulate fibroblast-transforming and sarcomagenic potential. To our knowledge, this represents the first point mutation of an autophosphorylation site that activates erbB oncogenicity.

Activation of c-fos gene expression by a kinase-deficient epidermal growth factor receptor

The intrinsic tyrosine kinase activity of the epidermal growth factor receptor (EGFR) has been shown to be responsible for many of the pleiotropic intracellular effects resulting from ligand stimulation [W.S. Chen, C.S. Lazar, M. Poenie, R.Y. Tsien, G.N. Gill, and M.G. Rosenfeld, Nature (London) 328:820-823, 1987; A.M. Honegger, D. Szapary, A. Schmidt, R. Lyall, E. Van Obberghen, T.J. Dull, A. Ulrich, and J. Schlessinger, Mol. Cell. Biol. 7:4568-4571, 1987]. Recently, however, it has been shown that addition of ligand to cells expressing kinase-defective EGFR mutants can result in the phosphorylation of mitogen-activated protein kinase (R. Campos-González and J.R. Glenney, Jr., J. Biol. Chem. 267:14535-14538, 1992; E. Selva, D.L. Raden, and R.J. Davis, J. Biol. Chem. 268:2250-2254, 1993), as well as stimulation of DNA synthesis (K.J. Coker, J.V. Staros, and C.A. Guyer, Proc. Natl. Acad. Sci. USA 91:6967-6971, 1994). Moreover, mitogen-activated protein kinase has been shown to phosphorylate the transcription factor p62TCF in vitro, leading to enhanced ternary complex formation between p62TCF, p67SRF, and the c-fos serum response element (SRE) [H. Gille, A.D. Sharrocks, and P.E. Shaw, Nature (London) 358:414-417, 1992]. On the basis of these observations, we have investigated the possibility that the intrinsic tyrosine kinase activity of the EGFR may not be necessary for transcriptional activation mediated via p62TCF. Here, we demonstrate that a kinase-defective EGFR mutant can signal ligand-induced expression of c-fos protein and that a significant component of this induction appears to be mediated at the transcriptional level. Investigation of transcriptional activation mediated by the c-fos SRE shows that this response is impaired by mutations in the SRE which eliminate binding of p62(TCF). These data indicate that information inherent in the structure of the EGFR can be accessed by ligand stimulation independent of the receptor's catalytic kinase function.

Activation of c-fos gene expression by a kinase-deficient epidermal growth factor receptor

The intrinsic tyrosine kinase activity of the epidermal growth factor receptor (EGFR) has been shown to be responsible for many of the pleiotropic intracellular effects resulting from ligand stimulation [W.S. Chen, C.S. Lazar, M. Poenie, R.Y. Tsien, G.N. Gill, and M.G. Rosenfeld, Nature (London) 328:820-823, 1987; A.M. Honegger, D. Szapary, A. Schmidt, R. Lyall, E. Van Obberghen, T.J. Dull, A. Ulrich, and J. Schlessinger, Mol. Cell. Biol. 7:4568-4571, 1987]. Recently, however, it has been shown that addition of ligand to cells expressing kinase-defective EGFR mutants can result in the phosphorylation of mitogen-activated protein kinase (R. Campos-González and J.R. Glenney, Jr., J. Biol. Chem. 267:14535-14538, 1992; E. Selva, D.L. Raden, and R.J. Davis, J. Biol. Chem. 268:2250-2254, 1993), as well as stimulation of DNA synthesis (K.J. Coker, J.V. Staros, and C.A. Guyer, Proc. Natl. Acad. Sci. USA 91:6967-6971, 1994). Moreover, mitogen-activated protein kinase has been shown to phosphorylate the transcription factor p62TCF in vitro, leading to enhanced ternary complex formation between p62TCF, p67SRF, and the c-fos serum response element (SRE) [H. Gille, A.D. Sharrocks, and P.E. Shaw, Nature (London) 358:414-417, 1992]. On the basis of these observations, we have investigated the possibility that the intrinsic tyrosine kinase activity of the EGFR may not be necessary for transcriptional activation mediated via p62TCF. Here, we demonstrate that a kinase-defective EGFR mutant can signal ligand-induced expression of c-fos protein and that a significant component of this induction appears to be mediated at the transcriptional level. Investigation of transcriptional activation mediated by the c-fos SRE shows that this response is impaired by mutations in the SRE which eliminate binding of p62(TCF). These data indicate that information inherent in the structure of the EGFR can be accessed by ligand stimulation independent of the receptor's catalytic kinase function.

Modulation of erbB kinase activity and oncogenic potential by single point mutations in the glycine loop of the catalytic domain

Avian c-erbB is activated to a leukemia oncogene following truncation of its amino-terminal ligand-binding domain by retroviral insertion. The insertionally activated transcripts encode protein products which have constitutive tyrosine kinase activity and can induce erythroleukemia but not sarcomas. We have previously found that a valine-to-isoleucine point mutation at position 157 (V157I mutant) within the tyrosine kinase domain of this truncated erbB can dramatically activate the sarcomagenic potential of the oncogene and increase the kinase activity of this oncoprotein. This mutation lies at position 157 of the insertionally activated c-erbB product, affecting a highly conserved valine residue of the glycine loop involved in ATP binding and phosphate transfer. To investigate the functional importance of this residue in the catalytic activity of kinases, we have introduced at this position, by site-directed mutagenesis, codons representing the remaining 18 amino acid residues. Most of the mutants have diminished activity, with six of them completely devoid of kinase activity, indicating the sensitivity of this region to conformational changes. Some of these mutants displayed increased kinase activity and greater transforming potential in comparison with IA c-erbB, but none had levels as high as those of the V157I mutant. In general, the sarcomagenic potential of the various erbB mutants correlated with their autophosphorylation state and their ability to cause phosphorylation of MAP kinase. However, there are important exceptions such as the V157G mutant, which lacks enhanced autophosphorylation but is highly sarcomagenic. Studies of this and other autophosphorylation site mutants point to the existence of an autophosphorylation-independent pathway in sarcomagenesis. The requirement for leukemogenic potential is much less stringent and correlates with positivity of kinase activity. When the valine-to-isoleucine substitution was put in context of the full-length erbB protein, the mutation relaxed the ligand dependence and had a positive effect on the transforming potential of the full-length c-erbB.

Modulation of erbB kinase activity and oncogenic potential by single point mutations in the glycine loop of the catalytic domain

Avian c-erbB is activated to a leukemia oncogene following truncation of its amino-terminal ligand-binding domain by retroviral insertion. The insertionally activated transcripts encode protein products which have constitutive tyrosine kinase activity and can induce erythroleukemia but not sarcomas. We have previously found that a valine-to-isoleucine point mutation at position 157 (V157I mutant) within the tyrosine kinase domain of this truncated erbB can dramatically activate the sarcomagenic potential of the oncogene and increase the kinase activity of this oncoprotein. This mutation lies at position 157 of the insertionally activated c-erbB product, affecting a highly conserved valine residue of the glycine loop involved in ATP binding and phosphate transfer. To investigate the functional importance of this residue in the catalytic activity of kinases, we have introduced at this position, by site-directed mutagenesis, codons representing the remaining 18 amino acid residues. Most of the mutants have diminished activity, with six of them completely devoid of kinase activity, indicating the sensitivity of this region to conformational changes. Some of these mutants displayed increased kinase activity and greater transforming potential in comparison with IA c-erbB, but none had levels as high as those of the V157I mutant. In general, the sarcomagenic potential of the various erbB mutants correlated with their autophosphorylation state and their ability to cause phosphorylation of MAP kinase. However, there are important exceptions such as the V157G mutant, which lacks enhanced autophosphorylation but is highly sarcomagenic. Studies of this and other autophosphorylation site mutants point to the existence of an autophosphorylation-independent pathway in sarcomagenesis. The requirement for leukemogenic potential is much less stringent and correlates with positivity of kinase activity. When the valine-to-isoleucine substitution was put in context of the full-length erbB protein, the mutation relaxed the ligand dependence and had a positive effect on the transforming potential of the full-length c-erbB.

Differential modulation of plasminogen activator gene expression by oncogene-encoded protein tyrosine kinases

Urokinase-type plasminogen activator (uPA) gene transcription is increased > or = 50-fold in chicken embryo fibroblasts (CEF) following transformation by the protein tyrosine kinase pp60v-src. Protein phosphorylation appears to play a critical role in uPA gene expression in these cells; protein kinase C-activating phorbol esters cooperate with pp60v-src to synergistically increase uPA mRNA, whereas cyclic AMP (cAMP)-dependent protein kinase-activating agents (e.g., 8-bromo cAMP) repress uPA mRNA levels. To explore the relationship between transforming oncogenes and uPA gene expression, uPA mRNA levels were measured in CEF infected with selected avian retroviruses. We report that v-ras and the transforming protein tyrosine kinases v-src, v-yes, and v-ros all increase cellular uPA mRNAs. However, transformation with the protein tyrosine kinase encoded by v-erbB, or the nuclear proteins encoded by v-jun, v-ski, or v-myc, did not increase uPA mRNA detectably. Ras and all of the protein tyrosine kinases analyzed, including the v-erbB product, but none of the nuclear oncoproteins sensitized cells to phorbol ester induction of uPA gene expression. Thus, increased uPA gene expression is not simply a secondary consequence of cell transformation but, rather, is regulated or comodulated by only a subset of oncogene products. Analysis of cells expressing site-directed mutants of pp60v-src showed that the induction of the uPA gene is dependent on protein tyrosine kinase catalytic activity, myristylation, and plasma membrane localization. However, these properties together are not sufficient; an additional feature in the src homology 2 domain is also required. The major sites of serine phosphorylation, serines 12 and 17, and the autophosphorylation site, tyrosine 416, are not essential for uPA gene induction. However, the reduction of uPA mRNA in pp60v-src-transformed cells by 8-bromo cAMP is dependent on tyrosine 416.

Differential modulation of plasminogen activator gene expression by oncogene-encoded protein tyrosine kinases

Urokinase-type plasminogen activator (uPA) gene transcription is increased > or = 50-fold in chicken embryo fibroblasts (CEF) following transformation by the protein tyrosine kinase pp60v-src. Protein phosphorylation appears to play a critical role in uPA gene expression in these cells; protein kinase C-activating phorbol esters cooperate with pp60v-src to synergistically increase uPA mRNA, whereas cyclic AMP (cAMP)-dependent protein kinase-activating agents (e.g., 8-bromo cAMP) repress uPA mRNA levels. To explore the relationship between transforming oncogenes and uPA gene expression, uPA mRNA levels were measured in CEF infected with selected avian retroviruses. We report that v-ras and the transforming protein tyrosine kinases v-src, v-yes, and v-ros all increase cellular uPA mRNAs. However, transformation with the protein tyrosine kinase encoded by v-erbB, or the nuclear proteins encoded by v-jun, v-ski, or v-myc, did not increase uPA mRNA detectably. Ras and all of the protein tyrosine kinases analyzed, including the v-erbB product, but none of the nuclear oncoproteins sensitized cells to phorbol ester induction of uPA gene expression. Thus, increased uPA gene expression is not simply a secondary consequence of cell transformation but, rather, is regulated or comodulated by only a subset of oncogene products. Analysis of cells expressing site-directed mutants of pp60v-src showed that the induction of the uPA gene is dependent on protein tyrosine kinase catalytic activity, myristylation, and plasma membrane localization. However, these properties together are not sufficient; an additional feature in the src homology 2 domain is also required. The major sites of serine phosphorylation, serines 12 and 17, and the autophosphorylation site, tyrosine 416, are not essential for uPA gene induction. However, the reduction of uPA mRNA in pp60v-src-transformed cells by 8-bromo cAMP is dependent on tyrosine 416.

Alteration of the kinetic properties of the epidermal growth factor receptor tyrosine kinase by basic proteins

Previous studies have shown that lysine- and arginine-rich proteins can enhance the activity of tyrosine and serine/threonine protein kinases. However, the kinetics and mechanism of this activation are not fully understood. Therefore we investigated the ability of poly(amino acids) and the arginine-rich protein, protamine, to alter the kinetic properties of epidermal growth factor (EGF) receptor protein-tyrosine kinase activity using immunoaffinity-purified receptor isolated from human epidermoid carcinoma (A431) cells. Poly(L-lysine), poly(L-arginine) and protamine stimulated EGF receptor kinase activity by 3-5-fold at non-saturating doses of ATP and peptide substrate, while poly(L-glutamate) had no effect. Initial kinetic studies demonstrated an increase in the maximum velocity and a decrease in the apparent Km for the peptide substrate angiotensin II in the presence of the basic effectors. Further analysis of the kinetic mechanism by product inhibition revealed that protamine altered the pattern of ADP inhibition towards the peptide substrate but not towards ATP. The change was indicative of the receptor's ability to form an enzyme-angiotensin II-ADP ternary complex in the presence of protamine but not in its absence. In addition, the basic effectors had a substantially decreased influence on the kinase activity of a C-terminally truncated form of the EGF receptor. Thus the changes in kinase activity may be partially mediated by the C-terminal region of the receptor, which contains the sites of receptor self-phosphorylation. These results suggest that the basic domains of proteins can interact with the EGF receptor to induce changes in its kinetic properties, especially with regard to reactant recognition and binding.

A truncated, secreted form of the epidermal growth factor receptor is encoded by an alternatively spliced transcript in normal rat tissue

Two independent cDNA clones corresponding to a 2.7-kilobase (kb) epidermal growth factor receptor (EGF-R) mRNA were isolated from a rat liver cDNA library. Sequence analysis revealed 100% homology in the external domain when compared with the full-length rat EGF-R nucleotide sequence and 80 to 90% similarity relative to the human EGF-R. However, the 3'-terminal sequence of these clones did not match EGF-R or any other known sequence(s) and was distinct from the 3' end of the 2.8-kb mRNA, which encodes a truncated EGF-R in A431 cells. The deduced amino acid sequence revealed an open reading frame which is homologous to the external domain of the EGF-R but which terminates prior to the transmembrane region. Southern blot analysis of rat genomic DNA indicated that the 3'-terminal sequence of this transcript is derived from the EGF-R gene. Analysis of a genomic clone containing the 3' end of the 2.7-kb transcript revealed that this sequence is present as a discrete exon in the mid-region of the receptor gene in proximity to the exon encoding the transmembrane domain. Introduction of an expression vector containing the truncated EGF-R cDNA into Chinese hamster ovary (CHO) cells led to the expression of a 95-kilodalton protein which was detected in conditioned media, by using antisera directed against the EGF-R. A similarly sized protein was also detected in the media of WB cells, a continuous, nontransformed line of rat hepatic epithelial cells. Northern (RNA blot) analysis established that the truncated receptor is encoded by a 2.7-kb transcript found in normal rat liver. Furthermore, Northern analysis of rat poly(A)+ RNA showed that the 2.7-kb EGF-R transcript is expressed at differing levels in various fetal and adult tissues. These data indicate that alternative splicing of the EGF-R primary transcript yields a 2.7-kb mRNA which codes for a truncated form of the receptor. This receptor is secreted by rat hepatic epithelial cells in culture, which suggests that it may be secreted by normal rat cells or tissues and perhaps serve an as yet unknown physiological function.

A truncated, secreted form of the epidermal growth factor receptor is encoded by an alternatively spliced transcript in normal rat tissue

Two independent cDNA clones corresponding to a 2.7-kilobase (kb) epidermal growth factor receptor (EGF-R) mRNA were isolated from a rat liver cDNA library. Sequence analysis revealed 100% homology in the external domain when compared with the full-length rat EGF-R nucleotide sequence and 80 to 90% similarity relative to the human EGF-R. However, the 3'-terminal sequence of these clones did not match EGF-R or any other known sequence(s) and was distinct from the 3' end of the 2.8-kb mRNA, which encodes a truncated EGF-R in A431 cells. The deduced amino acid sequence revealed an open reading frame which is homologous to the external domain of the EGF-R but which terminates prior to the transmembrane region. Southern blot analysis of rat genomic DNA indicated that the 3'-terminal sequence of this transcript is derived from the EGF-R gene. Analysis of a genomic clone containing the 3' end of the 2.7-kb transcript revealed that this sequence is present as a discrete exon in the mid-region of the receptor gene in proximity to the exon encoding the transmembrane domain. Introduction of an expression vector containing the truncated EGF-R cDNA into Chinese hamster ovary (CHO) cells led to the expression of a 95-kilodalton protein which was detected in conditioned media, by using antisera directed against the EGF-R. A similarly sized protein was also detected in the media of WB cells, a continuous, nontransformed line of rat hepatic epithelial cells. Northern (RNA blot) analysis established that the truncated receptor is encoded by a 2.7-kb transcript found in normal rat liver. Furthermore, Northern analysis of rat poly(A)+ RNA showed that the 2.7-kb EGF-R transcript is expressed at differing levels in various fetal and adult tissues. These data indicate that alternative splicing of the EGF-R primary transcript yields a 2.7-kb mRNA which codes for a truncated form of the receptor. This receptor is secreted by rat hepatic epithelial cells in culture, which suggests that it may be secreted by normal rat cells or tissues and perhaps serve an as yet unknown physiological function.

Disease tropism of c-erbB: effects of carboxyl-terminal tyrosine and internal mutations on tissue-specific transformation

Avian leukosis virus induces erythroleukemia in chickens by proviral insertional mutation of the protooncogene c-erbB. The product of the insertionally activated c-erbB locus lacks the extracellular ligand-binding domain and is strictly leukemogenic. It has previously been demonstrated that the disease spectrum associated with aberrant c-erbB expression can be expanded by structural perturbation of the cytoplasmic domain of this protein. In this report, we use mutagenesis and retroviral vectors to identify specific mutations in the carboxyl-terminal domain of the insertionally activated c-erbB product that are sufficient to activate the sarcomagenic potential of this protein. Interestingly, a point mutation in the kinase domain appears to be sufficient for sarcomagenic activation. However, removal of the terminal tyrosine residue of the c-erbB product, implicated in modulating kinase activity, does not lead to a fully transforming phenotype. These studies suggest that there are multiple ways to activate the fibroblast-transforming potential of the insertionally activated c-erbB product. The conformation of this protein may play a more significant role in oncogenic activation than the phosphorylation status of the putative carboxyl-terminal autophosphorylation site.