The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to ras signaling

Association of MEK1 with p21ras.GMPPNP is dependent on B-Raf

We have previously reported that immobilized p21ras forms a GMPPNP-dependent complex with a MEK activity. Furthermore, the association of the MEK activity was found to be independent of the presence of Raf-1. We have extended those observations to show that MEK1 is the MEK activity previously described to associate with immobilized p21ras.GMPPNP. The association between MEK1 and immobilized p21ras.GMPPNP increased its specific activity towards p42MAPK. We detected the specific association of B-Raf with immobilized p21ras.GMPPNP. In contrast to Raf-1-immunodepleted lysates, preclearance of the cytosolic B-Raf significantly reduced, by 96%, the amount of MEK1 activity associated with immobilized p21ras.GMPPNP. The decrease in MEK1 activity correlated with complete loss in the binding of both B-Raf and MEK1 proteins with immobilized p21ras.GMPPNP. These data suggest that the p21ras.GMPPNP-dependent activation of MEK1 in brain extracts is dependent on the presence of the B-Raf protein kinase.

Activation of mitogen-activated protein kinase and arachidonate release via two G protein-coupled receptors expressed in the rat hippocampus

Platelet-activating factor and somatostatin receptors, two G protein-coupled receptors expressed in the rat hippocampus, were analyzed for the downstream signaling pathways in Chinese hamster ovary cells stably expressing each receptor. Ligand stimulation to each CHO cell line induced (1) inhibition of forskolin-induced accumulation of cAMP, (2) arachidonate release, and (3) activation of mitogen-activated protein kinase and MAP kinase kinase. In contrast, inositol phosphate breakdown was seen only in the PAF-stimulated CHO cells. The induction of these signals accompanied no detectable Ras activation. Suppression of the signals by pertussis toxin was almost complete for the somatostatin receptor but partial for the PAF receptor, suggesting that the somatostatin receptor couples only with PTX-sensitive G protein, while the PAF receptor couples with both PTX-sensitive and -insensitive G proteins. A model of G protein-mediated signaling pathways was proposed in which the signals from Gi and those from Gq converge at MAP kinase kinase and lead to arachidonate release. The present system using CHO cells is useful for analyzing signaling pathways from G proteins to MAP kinase kinase and will thereby provide clues for understanding the mechanisms underlying the physiological and pathological events mediated by PAF, somatostatin, and other G protein-coupled receptors in the central nervous system and other tissues.

Solution structure and ligand-binding site of the carboxy-terminal SH3 domain of GRB2

BACKGROUND

Growth factor receptor-bound protein 2 (GRB2) is an adaptor protein with three Src homology (SH) domains in the order SH3-SH2-SH3. Both SH3 domains of GRB2 are necessary for interaction with the protein Son of sevenless (Sos), which acts as a Ras activator. Thus, GRB2 mediates signal transduction from growth factor receptors to Ras and is thought to be a key molecule in signal transduction.

RESULTS

The three-dimensional structure of the carboxy-terminal SH3 domain of GRB2 (GRB2 C-SH3) was determined by NMR spectroscopy. The SH3 structure consists of six beta-strands arranged in two beta-sheets that are packed together perpendicularly with two additional beta-strands forming the third beta-sheet. GRB2 C-SH3 is very similar to SH3 domains from other proteins. The binding site of the ligand peptide (VPP-PVPPRRR) derived from the Sos protein was mapped on the GRB2 C-SH3 domain indirectly using 1H and 15N chemical shift changes, and directly using several intermolecular nuclear Overhauser effects.

CONCLUSIONS

Despite the structural similarity among the known SH3 domains, the sequence alignment and the secondary structure assignments differ. We therefore propose a standard description of the SH3 structures to facilitate comparison of individual SH3 domains, based on their three-dimensional structures. The binding site of the ligand peptide on GRB2 C-SH3 is in good agreement with those found in other SH3 domains.

Association of MEK1 with p21ras.GMPPNP is dependent on B-Raf

We have previously reported that immobilized p21ras forms a GMPPNP-dependent complex with a MEK activity. Furthermore, the association of the MEK activity was found to be independent of the presence of Raf-1. We have extended those observations to show that MEK1 is the MEK activity previously described to associate with immobilized p21ras.GMPPNP. The association between MEK1 and immobilized p21ras.GMPPNP increased its specific activity towards p42MAPK. We detected the specific association of B-Raf with immobilized p21ras.GMPPNP. In contrast to Raf-1-immunodepleted lysates, preclearance of the cytosolic B-Raf significantly reduced, by 96%, the amount of MEK1 activity associated with immobilized p21ras.GMPPNP. The decrease in MEK1 activity correlated with complete loss in the binding of both B-Raf and MEK1 proteins with immobilized p21ras.GMPPNP. These data suggest that the p21ras.GMPPNP-dependent activation of MEK1 in brain extracts is dependent on the presence of the B-Raf protein kinase.

B cell antigen receptor cross-linking induces tyrosine phosphorylation and membrane translocation of a multimeric Shc complex that is augmented by CD19 co-ligation

The SH2 domain-containing transforming Shc protein has been implicated in mitogenic signaling via several surface receptors through p21ras. Following tyrosine phosphorylation by either receptor or non-receptor tyrosine kinases, Shc may interact with the adaptor protein Grb2, which is linked to Sos1, a guanine nucleotide exchange factor for human ras. Ligation of the antigen receptor complex on B cells (BCR) is known to activate various intracellular signaling pathways, which may accumulate in mitogenic responses. With respect to the initial steps, the activation of BCR-associated non-receptor tyrosine kinases appears to be indispensible. In this report we show that Shc proteins become tyrosine phosphorylated after BCR ligation on both transformed and normal human B cells. This is accompanied by the association of Shc with Grb2 proteins and a yet unidentified 145-kDa tyrosine phosphorylated protein. Subcellular fractionation revealed that this activation-induced multimeric Shc complex rapidly translocates towards the plasma membrane. Co-ligation of the BCR with the CD19 molecule results in a marked increase of these events, whereas CD19 cross-linking alone does not induce Shc tyrosine phosphorylation or translocation. Thus, in B cells the Shc complex may represent a molecular junction between the BCR and the mitogenic p21ras cascade.

Multiple cytokines stimulate the binding of a common 145-kilodalton protein to Shc at the Grb2 recognition site of Shc

We recently reported that interleukin-3, Steel factor, and erythropoietin all induce the tyrosine phosphorylation of Shc and its association with Grb2 in hemopoietic cell lines. We have now further characterized the proteins that become associated with Shc following stimulation with these cytokines and found that, in response to all three, the tyrosine-phosphorylated form of Shc binds to common 145- and 52-kDa proteins which also become tyrosine phosphorylated in response to these growth factors. The 145-kDa protein, which appears, from antiphosphotyrosine blots of two-dimensional O'Farrell gels, to exist in four different phosphorylation states following cytokine stimulation (with isoelectric points ranging from 7.2 to 7.8), does not appear to be immunologically related to the beta subunit of the interleukin-3 receptor, c-Kit, BCR, ABL, JAK1, JAK2, Sos1, eps15, or insulin receptor substrate 1 protein. Silver-stained sodium dodecyl sulfate gels indicate that the association of the 145-kDa protein with Shc occurs only after cytokine stimulation and that it can bind to the tyrosine-phosphorylated form of Shc in its non-tyrosine-phosphorylated state. The latter finding, in conjunction with the observations that p145 does not bind, in vitro, to the Src homology 2 (SH2) domain of Shc, that it is not present in anti-Grb2 immunoprecipitates, and that a phosphopeptide which blocks the binding of Shc to the SH2 domain of Grb2 also blocks the binding of Shc to p145, suggests that p145 contains an SH2 domain and competes with Grb2 for the same tyrosine-phosphorylated site on Shc. This implicates p145 as a potential regulator of Ras activity and, perhaps, of other as yet unidentified functions of Shc.

Tyr-716 in the platelet-derived growth factor beta-receptor kinase insert is involved in GRB2 binding and Ras activation

Ligand stimulation of the platelet-derived growth factor (PDGF) beta-receptor leads to activation of its intrinsic tyrosine kinase and autophosphorylation of the intracellular part of the receptor. The autophosphorylated tyrosine residues mediate interactions with downstream signal transduction molecules and thereby initiate different signalling pathways. A pathway leading to activation of the GTP-binding protein Ras involves the adaptor molecule GRB2. Here we show that Tyr-716, a novel autophosphorylation site in the PDGF beta-receptor kinase insert, mediates direct binding of GRB2 in vitro and in vivo. In a panel of mutant PDGF beta-receptors, in which Tyr-716 and the previously known autophosphorylation sites were individually mutated, only PDGFR beta Y716F failed to bind GRB2. Furthermore, a synthetic phosphorylated peptide containing Tyr-716 bound GRB2, and this peptide specifically interrupted the interaction between GRB2 and the wild-type receptor. In addition, the Y716(P) peptide significantly decreased the amount of GTP bound to Ras in response to PDGF in permeabilized fibroblasts as well as in porcine aortic endothelial cells expressing transfected PDGF beta-receptors. The mutant PDGFR beta Y716F still mediated activation of mitogen-activated protein kinases and an increased DNA synthesis in response to PDGF, indicating that multiple signal transduction pathways transduce mitogenic signals from the activated PDGF beta-receptor.

Functional interaction between the epidermal growth factor receptor and c-Src kinase activity

To study the relationship between the tyrosine kinase c-Src and the epidermal growth factor receptor (EGF-R), we used the breast cancer cell line ZR75-1, which was transfected with the EGF-R. The EGF-R transfected cell line expressed 60 times more EGF-R than a control cell line transfected with the empty vector. In the presence of EGF, the EGF-R over-expressing cell line grew much faster than the control cell line. Both cell lines expressed approximately equal amounts of c-Src. However, the cell line over-expressing the EGF-R showed a twofold enhancement of c-Src kinase activity after EGF stimulation. The activation of c-Src kinase by EGF was confirmed in other EGF-R expressing cell types.

Tyr-716 in the platelet-derived growth factor beta-receptor kinase insert is involved in GRB2 binding and Ras activation

Ligand stimulation of the platelet-derived growth factor (PDGF) beta-receptor leads to activation of its intrinsic tyrosine kinase and autophosphorylation of the intracellular part of the receptor. The autophosphorylated tyrosine residues mediate interactions with downstream signal transduction molecules and thereby initiate different signalling pathways. A pathway leading to activation of the GTP-binding protein Ras involves the adaptor molecule GRB2. Here we show that Tyr-716, a novel autophosphorylation site in the PDGF beta-receptor kinase insert, mediates direct binding of GRB2 in vitro and in vivo. In a panel of mutant PDGF beta-receptors, in which Tyr-716 and the previously known autophosphorylation sites were individually mutated, only PDGFR beta Y716F failed to bind GRB2. Furthermore, a synthetic phosphorylated peptide containing Tyr-716 bound GRB2, and this peptide specifically interrupted the interaction between GRB2 and the wild-type receptor. In addition, the Y716(P) peptide significantly decreased the amount of GTP bound to Ras in response to PDGF in permeabilized fibroblasts as well as in porcine aortic endothelial cells expressing transfected PDGF beta-receptors. The mutant PDGFR beta Y716F still mediated activation of mitogen-activated protein kinases and an increased DNA synthesis in response to PDGF, indicating that multiple signal transduction pathways transduce mitogenic signals from the activated PDGF beta-receptor.

A rapid bioassay for platelet-derived growth factor beta-receptor tyrosine kinase function

We have extended a melanophore-based bioassay for G-protein coupled receptors to include the functional expression of the murine platelet-derived growth factor (PDGF) beta-receptor. The homodimeric ligand PDGF-BB induced activation of the transiently expressed receptor in melanophore cells. This led to dose dependent pigment dispersion whereas it did not induce pigment dispersion in wild type cells. The effective concentration of PDGF-BB giving half-maximal pigment dispersion (EC50) was 1nM after 30 minutes exposure. PDGF-AA had no ability to induce pigment dispersion in melanophore cells transiently expressing the beta-PDGF receptor. PDGF-BB-induced pigment dispersion could be blocked by the bis-indolylmaleimide Ro 31-8220 which is an inhibitor of protein kinase C isoenzymes. Functional expression of the PDGF beta-receptor extends the use of the pigment translocation assay to include transmembrane signaling receptor tyrosine kinases. It opens the opportunity for the discovery of potent agonists and antagonists through massive drug screening and investigations of functional ligand-receptor interactions for single transmembrane domain receptors.

Inhibitors of the insulin receptor tyrosine kinase

Insulin is a polypeptide hormone consisting of 51 amino acids. Insulin promotes a variety of anabolic enzymatic pathways and inhibits many catabolic enzymatic pathways involved in energy storage, as well as in synthesis of structural tissue proteins. In addition, insulin serves as a growth factor, modulating mitogenesis, growth and differentiation. Insulin mediates all of its effects by initially binding and activating its specific cell-surface receptor. Conformational changes induced by insulin binding lead to activation of intrinsic receptor tyrosine kinase. Thus, the study of tyrosine kinase inhibitors, whether synthetically produced or purified from microorganisms or humans, has led to elucidation of molecular details of physiological insulin signaling.

The phosphopeptide-binding specificity of Src family SH2 domains

BACKGROUND

Src homology 2 (SH2) domains mediate protein/protein interactions by binding phosphotyrosyl proteins with high specificity. The protein Lck, a Src-like lymphocyte-specific tyrosine kinase which is important in signals involved in T-cell development, contains one such domain. The crystal structure of a complex of the Lck SH2 domain with a high-affinity ligand, pY324, is known. This ligand has the sequence EPQpYEEIPIYL.

RESULTS

We designed and synthesized a series of phosphopeptides with single amino-acid changes in the four residues C-terminal to the phosphotyrosine (pTyr) in pY324. Surprisingly, the Glu one residue C-terminal to the phosphotyrosine (at position pY + 1) is sensitive to substitution, whereas the Ile at position pY + 3 is much less sensitive, accommodating a Glu with only modest loss of binding affinity. Replacement of the Glu and Pro on either side of the Ile had little effect, as predicted. Truncated phosphopeptides that end at position pY + 5 and have only an acetyl group N-terminal to the pTyr bound with only slightly lower affinity than pY324. In addition, naturally occurring phosphopeptide sequences that span a 1,000-fold range in binding affinity for the Lck SH2 domain have been identified.

CONCLUSIONS

The Lck SH2 domain is highly selective for phosphotyrosyl-peptide binding; its specificity is dictated by the first and third residues C-terminal to the pTyr. The unexpected effects of some amino-acid substitutions indicate that the interactions seen between SH2 domains and ligand in the crystal structure may not be identical to those that occur in solution.

Multiple cytokines stimulate the binding of a common 145-kilodalton protein to Shc at the Grb2 recognition site of Shc

We recently reported that interleukin-3, Steel factor, and erythropoietin all induce the tyrosine phosphorylation of Shc and its association with Grb2 in hemopoietic cell lines. We have now further characterized the proteins that become associated with Shc following stimulation with these cytokines and found that, in response to all three, the tyrosine-phosphorylated form of Shc binds to common 145- and 52-kDa proteins which also become tyrosine phosphorylated in response to these growth factors. The 145-kDa protein, which appears, from antiphosphotyrosine blots of two-dimensional O'Farrell gels, to exist in four different phosphorylation states following cytokine stimulation (with isoelectric points ranging from 7.2 to 7.8), does not appear to be immunologically related to the beta subunit of the interleukin-3 receptor, c-Kit, BCR, ABL, JAK1, JAK2, Sos1, eps15, or insulin receptor substrate 1 protein. Silver-stained sodium dodecyl sulfate gels indicate that the association of the 145-kDa protein with Shc occurs only after cytokine stimulation and that it can bind to the tyrosine-phosphorylated form of Shc in its non-tyrosine-phosphorylated state. The latter finding, in conjunction with the observations that p145 does not bind, in vitro, to the Src homology 2 (SH2) domain of Shc, that it is not present in anti-Grb2 immunoprecipitates, and that a phosphopeptide which blocks the binding of Shc to the SH2 domain of Grb2 also blocks the binding of Shc to p145, suggests that p145 contains an SH2 domain and competes with Grb2 for the same tyrosine-phosphorylated site on Shc. This implicates p145 as a potential regulator of Ras activity and, perhaps, of other as yet unidentified functions of Shc.

The sevenmaker gain-of-function mutation in p42 MAP kinase leads to enhanced signalling and reduced sensitivity to dual specificity phosphatase action

A mammalian mutant MAP kinase, D319N ERK2, analogous to Drosophila melanogaster sevenmaker (rlsem) gain-of-function mutation was shown to have an increased sensitivity to low levels of signalling in vivo. However, the mutation does not lead to an elevated basal kinase activity and still requires activation by MAP kinase kinase (MAPKK) as does wild type ERK2. This increased responsiveness seen in vivo is not due to an increased ability to phosphorylate substrates but appears to reflect a reduced sensitivity to a MAP kinase phosphatase CL100.

Membrane targeting of the nucleotide exchange factor Sos is sufficient for activating the Ras signaling pathway

Activation of growth factor receptors results in tyrosine autophosphorylation and recruitment of SH2 domain-containing effectors, including Grb2. Grb2 recruitment mediates activation of the Ras nucleotide exchanger Sos by an unknown mechanism. To examine the role of membrane recruitment, we prepared Sos derivatives containing either myristoylation or farnesylation signals. This resulted in plasma membrane targeting of Sos and stimulation of the Ras signaling pathway, including ERK and AP-1 activities leading to oncogenic transformation. Sos derivatives with nonfunctional myristoylation or farnesylation sequences were inactive. Farnesylation of Sos also activated Ras signaling in yeast. In both mammalian cells and yeast, membrane-targeted Sos derivatives lacking the C-terminal region were considerably more active. Therefore, targeting of Sos to the plasma membrane in the vicinity of Ras appears to be the primary mechanism leading to activation of the Ras pathway. A secondary mechanism could involve relief of the inhibitory effect of the Sos C-terminal region.

Shc, Grb2, Sos1, and a 150-kilodalton tyrosine-phosphorylated protein form complexes with Fms in hematopoietic cells

Fms, the macrophage colony-stimulating factor (M-CSF) receptor, is normally expressed in myeloid cells and initiates signals for both growth and development along the monocyte/macrophage lineage. We have examined Fms signal transduction pathways in the murine myeloid progenitor cell line FDC-P1. M-CSF stimulation of FDC-P1 cells expressing exogenous Fms resulted in tyrosine phosphorylation of a variety of cellular proteins in addition to Fms. M-CSF stimulation also resulted in Fms association with two of these tyrosine-phosphorylated proteins, one of which was identified as the 55-kDa Shc, which is shown in other systems to be involved in growth stimulation, and the other was a previously uncharacterized 150-kDa protein (p150). Fms also formed complexes with Grb2 and Sos1, and neither contained phosphotyrosine. Whereas both Grb2 and Sos1 complexed with Fms only after M-CSF stimulation, the amount of Sos1 complexed with Grb2 was not M-CSF dependent. Shc coimmunoprecipitated Sos1, Grb2, and tyrosine-phosphorylated p150, while Grb2 immunoprecipitates contained mainly phosphorylated p150, Fms, Shc, and Sos1. Shc interacted with tyrosine-phosphorylated p150 via its SH2 domain, and the Grb2 SH2 domain likewise bound tyrosine-phosphorylated Fms and p150. Analysis of Fms mutated at each of four tyrosine autophosphorylation sites indicated that none of these sites dramatically affected p150 phosphorylation or its association with Shc and Grb2. M-CSF stimulation of fibroblast cell lines expressing exogenous murine Fms did not phosphorylate p150, and this protein was not detected either in cell lysates or in Grb2 or Shc immunoprecipitates. The p150 protein is not related to known signal transduction molecules and may be myeloid cell specific. These results suggest that M-CSF stimulation of myeloid cells could activate Ras through the nucleotide exchange factor Sos1 by Grb2 binding to either Fms, Shc, or p150 and that Fms signal transduction in myeloid cells differs from that in fibroblasts.

Shc, Grb2, Sos1, and a 150-kilodalton tyrosine-phosphorylated protein form complexes with Fms in hematopoietic cells

Fms, the macrophage colony-stimulating factor (M-CSF) receptor, is normally expressed in myeloid cells and initiates signals for both growth and development along the monocyte/macrophage lineage. We have examined Fms signal transduction pathways in the murine myeloid progenitor cell line FDC-P1. M-CSF stimulation of FDC-P1 cells expressing exogenous Fms resulted in tyrosine phosphorylation of a variety of cellular proteins in addition to Fms. M-CSF stimulation also resulted in Fms association with two of these tyrosine-phosphorylated proteins, one of which was identified as the 55-kDa Shc, which is shown in other systems to be involved in growth stimulation, and the other was a previously uncharacterized 150-kDa protein (p150). Fms also formed complexes with Grb2 and Sos1, and neither contained phosphotyrosine. Whereas both Grb2 and Sos1 complexed with Fms only after M-CSF stimulation, the amount of Sos1 complexed with Grb2 was not M-CSF dependent. Shc coimmunoprecipitated Sos1, Grb2, and tyrosine-phosphorylated p150, while Grb2 immunoprecipitates contained mainly phosphorylated p150, Fms, Shc, and Sos1. Shc interacted with tyrosine-phosphorylated p150 via its SH2 domain, and the Grb2 SH2 domain likewise bound tyrosine-phosphorylated Fms and p150. Analysis of Fms mutated at each of four tyrosine autophosphorylation sites indicated that none of these sites dramatically affected p150 phosphorylation or its association with Shc and Grb2. M-CSF stimulation of fibroblast cell lines expressing exogenous murine Fms did not phosphorylate p150, and this protein was not detected either in cell lysates or in Grb2 or Shc immunoprecipitates. The p150 protein is not related to known signal transduction molecules and may be myeloid cell specific. These results suggest that M-CSF stimulation of myeloid cells could activate Ras through the nucleotide exchange factor Sos1 by Grb2 binding to either Fms, Shc, or p150 and that Fms signal transduction in myeloid cells differs from that in fibroblasts.

The role of p21ras in CD28 signal transduction: triggering of CD28 with antibodies, but not the ligand B7-1, activates p21ras

CD28 is a 44-kD homodimer expressed on the surface of the majority of human T cells that provides an important costimulus for T cell activation. The biochemical basis of the CD28 accessory signals is poorly understood. Triggering of the T cell antigen receptor (TCR) activates the p21ras proteins. Here we show that ligation of CD28 by a monoclonal antibody (mAb) also stimulates p21ras and induces Ras-dependent events such as stimulation of the microtubule-associated protein (MAP) kinase ERK2 and hyperphosphorylation of Raf-1. One physiological ligand for CD28 is the molecule B7-1. In contrast to the effect of CD28 mAb, the present studies show that interactions between CD28 and B7-1 do not stimulate p21ras signaling pathways. Two substrates for TCR-regulated protein tyrosine kinases (PTKs) have been implicated in p21ras activation in T cells: p95vav and a 36-kD protein that associates with a complex of Grb2 and the Ras exchange protein Sos. Triggering CD28 with both antibodies and B7-1 activates cellular PTKs, and we have exploited the differences between antibodies and B7-1 for p21ras activation in an attempt to identify critical PTK-controlled events for Ras activation in T cells. The data show that antibodies against TCR or CD28 induce tyrosine phosphorylation of both Vav and p36. B7-1 also induces Vav tyrosine phosphorylation but has no apparent effect on tyrosine phosphorylation of the Grb2-associated p36 protein. The intensity of the Vav tyrosine phosphorylation is greater in B7-1 than in TCR-stimulated cells. Moreover the kinetics of Vav tyrosine phosphorylation is prolonged in the B7-1-stimulated cells. These studies show that for CD28 signaling, the activation of p21ras correlates more closely with p36 tyrosine phosphorylation than with Vav tyrosine phosphorylation. However, the experiments demonstrate that Vav is a major substrate for B7-activated PTKs and hence could be important in CD28 signal transduction pathway.

Insulin receptor substrate-1 variants in non-insulin-dependent diabetes

Insulin receptor substrate-1 (IRS-1) plays an important role in insulin-stimulated signaling mechanisms. Therefore, we investigated the frequency and clinical significance of variants in the coding region of this gene in patients with non-insulin-dependent diabetes (NIDDM). Initial screening included a population-based sample of 40 Finnish patients with typical NIDDM. Applying single strand conformation polymorphism analysis the following amino acid substitutions were found among the 40 NIDDM patients: Gly818-Arg, Ser892Gly, and Gly971Arg. The first two variants have not been previously reported. Additional samples of 72 patients with NIDDM and 104 healthy control subjects with completely normal oral glucose tolerance test and a negative family history of diabetes were screened. The most common polymorphism was the Gly971Arg substitution which was found in 11 (9.8%) of 112 NIDDM patients and in 9 (8.7%) of 104 control subjects. The Gly818Arg substitution was found in 2 (1.8%) of NIDDM patients and in 2 (1.9%) of control subjects, and the Ser892Gly substitution was found in 3 (2.7%) NIDDM patients and in 1 (1.0%) control subject. The Gly971Arg substitution was not associated with an impairment in insulin secretion capacity (estimated by insulin responses in an oral glucose tolerance test or by the hyperglycemic clamp) or insulin action (estimated by the euglycemic clamp). Of the three amino acid substitutions observed Ser892Gly is the most interesting one since it abolishes one of the potential serine phosphorylation sites (SPGE) which is located immediately NH2-terminal to the only SH2 binding site of growth factor receptor-bound protein (GRB2), and thus could potentially influence some aspects of signal transduction and metabolic response to insulin.

Beta-actin mRNA localization is regulated by signal transduction mechanisms

Beta-actin mRNA is localized in the leading lamellae of chicken embryo fibroblasts (CEFs) (Lawrence, J., and R. Singer. 1986. Cell. 45:407-415), close to where actin polymerization in the lamellipodia drives cellular motility. During serum starvation beta-actin mRNA becomes diffuse and non-localized. Addition of FCS induces a rapid (within 2-5 min) redistribution of beta-actin mRNA into the leading lamellae. A similar redistribution was seen with PDGF, a fibroblast chemotactic factor. PDGF-induced beta-actin mRNA redistribution was inhibited by the tyrosine kinase inhibitor herbimycin, indicating that this process requires intact tyrosine kinase activity, similar to actin filament polymerization and chemotaxis. Lysophosphatidic acid, which has been shown to rapidly induce actin stress fiber formation (Ridley, A., and A. Hall. 1992. Cell. 790:389-399), also increases peripheral beta-actin mRNA localization within minutes. This suggests that actin polymerization and mRNA localization may be regulated by similar signaling pathways. Additionally, activators or inhibitors of kinase A or C can also delocalize steady-state beta-actin mRNA in cells grown in serum, and can inhibit the serum induction of peripherally localized beta-actin mRNA in serum-starved CEFs. These data show that physiologically relevant extracellular factors operating through a signal transduction pathway can regulate spatial sites of actin protein synthesis, which may in turn affect cellular polarity and motility.

Membrane-targeting potentiates guanine nucleotide exchange factor CDC25 and SOS1 activation of Ras transforming activity

Growth factor-triggered activation of Ras proteins is believed to be mediated by guanine nucleotide exchange factors (CDC25/GRF and SOS1/2) that promote formation of the active Ras GTP-bound state. Although the mechanism(s) of guanine nucleotide exchange factor regulation is unclear, recent studies suggest that translocation of SOS1 to the plasma membrane, where Ras is located, might be responsible for Ras activation. To evaluate this model, we generated constructs that encode the catalytic domains of human CDC25 or mouse SOS1, either alone (designated cCDC25 and cSOS1, respectively) or terminating in the carboxyl-terminal CAAX membrane-targeting sequence from K-Ras4B (designated cCDC25-CAAX and cSOS1-CAAX, respectively; in CAAX, C is Cys, A is an aliphatic amino acid, and X is Ser or Met). We then compared the transforming potential of cCDC25 and cSOS1 with their membrane-targeted counterparts. We observed that addition of the Ras plasma membrane-targeting sequence to the catalytic domains of CDC25 and SOS1 greatly enhanced their focus-forming activity (10- to 50-fold) in NIH 3T3 transfection assays. Similarly, we observed that the membrane-targeted versions showed a 5- to 10-fold enhanced ability to induce transcriptional activation from the Ets/AP-1 Ras-responsive element. Furthermore, whereas cells that stably expressed cCDC25 or cSOS1 exhibited the same morphologies as untransformed NIH 3T3 cells, cells expressing cCDC25-CAAX or cSOS1-CAAX displayed transformed morphologies that were indistinguishable from the elongated and refractile morphology of oncogenic Ras-transformed cells. Thus, these results suggest that membrane translocation alone is sufficient to potentiate guanine nucleotide exchange factor activation of Ras.

Hierarchy of binding sites for Grb2 and Shc on the epidermal growth factor receptor

We analyzed the binding site(s) for Grb2 on the epidermal growth factor (EGF) receptor (EGFR), using cell lines overexpressing EGFRs containing various point and deletion mutations in the carboxy-terminal tail. Results of co-immunoprecipitation experiments suggest that phosphotyrosines Y-1068 and Y-1173 mediate the binding of Grb2 to the EGFR. Competition experiments with synthetic phosphopeptides corresponding to known autophosphorylation sites on the EGFR demonstrated that phosphopeptides containing Y-1068, and to a lesser extent Y-1086, were able to inhibit the binding of Grb2 to the EGFR, while a Y-1173 peptide did not. These findings were confirmed by using a dephosphorylation protection assay and by measuring the dissociation constants of Grb2's SH2 domain to tyrosine-phosphorylated peptides, using real-time biospecific interaction analysis (BIAcore). From these studies, we concluded that Grb2 binds directly to the EGFR at Y-1068, to a lesser extent at Y-1086, and indirectly at Y-1173. Since Grb2 also binds Shc after EGF stimulation, we investigated whether Y-1173 is a binding site for the SH2 domain of Shc on the EGFR. Both competition experiments with synthetic phosphopeptides and dephosphorylation protection analysis demonstrated that Y-1173 and Y-992 are major and minor binding sites, respectively, for Shc on the EGFR. However, other phosphorylation sites in the carboxy-terminal tail of the EGFR are able to compensate for the loss of the main binding sites for Shc. These analyses reveal a hierarchy of interactions between Grb2 and Shc with the EGFR and indicate that Grb2 can bind the tyrosine-phosphorylated EGFR directly, as well as indirectly via Shc.

Comparative effects of hepatocyte growth factor and epidermal growth factor on motility, morphology, mitogenesis, and signal transduction of primary rat hepatocytes

Hepatocyte growth factor (HGF) and epidermal growth factor (EGF) are major hepatocyte mitogens, but HGF, also known as scatter factor (SF), has also been shown as a potent motogen for epithelial and endothelial cells. The mechanisms by which HGF is a stronger motogen compared to other mitogens are not understood. Here we report a comparative study of the effect of the two growth factors on cultured primary rat hepatocytes regarding their differential effects on morphology, mitogenicity, and motility as well as the phosphorylation of cytoskeletal-associated proteins. Using three different motility assays, both HGF and EGF increased the motility of hepatocytes, but HGF consistently elicited a significantly greater motility response than EGF. Additionally, HGF induced a more flattened, highly spread morphology compared to EGF. To examine if HGF and EGF phosphorylated different cytoskeletal elements as signal transduction targets in view of the observed variation in morphology and motility, primary cultures of 32P-loaded rat hepatocytes were stimulated by either HGF or EGF for up to 60 min. Both mitogens rapidly stimulated four isoforms of MAP kinase with similar kinetics and also rapidly facilitated the phosphorylation of cytoskeletal-associated F-actin. Two cytoskeletal-associated proteins, however, were observed to undergo rapid phosphorylation by HGF and not EGF during the time points described. One protein of 28 kDa was observed to become phosphorylated fivefold over controls, while the EGF-stimulated cells showed only a slight increase in the phosphorylation of this protein. Another protein with an apparent mwt of 42 kDa was phosphorylated 20-fold at 1 min and remained phosphorylated over 50-fold over control up to the 60 min time point. This protein was observed to become phosphorylated by EGF only after 10 min, and to a lesser extent (20-fold). Taken together, the data suggest that HGF and EGF stimulate divergent as well as redundant signal transduction pathways in the hepatocyte cytoskeleton, and this may result in unique HGF- or EGF-specific motility, morphology, and mitogenicity in hepatocytes.

Signal transduction and cell fate specification during Caenorhabditis elegans vulval development

A receptor tyrosine kinase/Ras signaling pathway controls the specification of vulval cell fates in Caenorhabditis elegans. Recently, C. elegans genes encoding proteins with similarity to mammalian Raf (lin-45), mitogen-activated protein kinase (mpk-1/sur-1), and an HNF-3 transcription factor (lin-31) have been identified and shown to act downstream of let-60 (ras) in this pathway. These genetically identified gene products bridge the gap between signal transduction at the plasma membrane and the control of cell fate specification in the nucleus.

Hierarchy of binding sites for Grb2 and Shc on the epidermal growth factor receptor

We analyzed the binding site(s) for Grb2 on the epidermal growth factor (EGF) receptor (EGFR), using cell lines overexpressing EGFRs containing various point and deletion mutations in the carboxy-terminal tail. Results of co-immunoprecipitation experiments suggest that phosphotyrosines Y-1068 and Y-1173 mediate the binding of Grb2 to the EGFR. Competition experiments with synthetic phosphopeptides corresponding to known autophosphorylation sites on the EGFR demonstrated that phosphopeptides containing Y-1068, and to a lesser extent Y-1086, were able to inhibit the binding of Grb2 to the EGFR, while a Y-1173 peptide did not. These findings were confirmed by using a dephosphorylation protection assay and by measuring the dissociation constants of Grb2's SH2 domain to tyrosine-phosphorylated peptides, using real-time biospecific interaction analysis (BIAcore). From these studies, we concluded that Grb2 binds directly to the EGFR at Y-1068, to a lesser extent at Y-1086, and indirectly at Y-1173. Since Grb2 also binds Shc after EGF stimulation, we investigated whether Y-1173 is a binding site for the SH2 domain of Shc on the EGFR. Both competition experiments with synthetic phosphopeptides and dephosphorylation protection analysis demonstrated that Y-1173 and Y-992 are major and minor binding sites, respectively, for Shc on the EGFR. However, other phosphorylation sites in the carboxy-terminal tail of the EGFR are able to compensate for the loss of the main binding sites for Shc. These analyses reveal a hierarchy of interactions between Grb2 and Shc with the EGFR and indicate that Grb2 can bind the tyrosine-phosphorylated EGFR directly, as well as indirectly via Shc.

HIV-1 Nef leads to inhibition or activation of T cells depending on its intracellular localization

Nef of primate lentiviruses is required for viremia and progression to AIDS in monkeys. Negative, positive, and no effects of Nef have also been reported on viral replication in cells. To reconcile these observations, we expressed a hybrid CD8-Nef protein in Jurkat cells. Two opposite phenotypes were found, which depended on the intracellular localization of Nef. Expressed in the cytoplasm or on the cell surface, the chimera inhibited or activated early signaling events from the T cell antigen receptor. Activated Jurkat cells died by apoptosis, and only cells with mutated nef genes expressing truncated Nefs survived, which rendered Nef nonfunctional. These mutations paralleled those in other viral strains passaged in vitro. Not only do these positional effects of Nef reconcile diverse phenotypes of Nef and suggest a role for its N-terminal myristylation, but they also explain effects of Nef in HIV infection and progression to AIDS.

The mitogen activated protein kinase signal transduction pathway: from the cell surface to the nucleus

Activation of the mitogen activated protein kinase (MAPK) plays essential roles in many signal transduction pathways. MAPK has been demonstrated to phosphorylate and regulate numerous cellular proteins, including growth factor receptor, transcription factors, cytoskeletal proteins, phospholipase and other protein kinases. Activation of MAPK requires phosphorylation of both threonine and tyrosine residues, which are catalysed by a single protein kinase known as MAPK kinase or MEK. MEK itself is activated by phosphorylation on two conserved serine residues. Three distinct mammalian Ser/Thr kinases, including Raf, Mos and MEKK (for MEK kinase), have been demonstrated to phosphorylate and activate MEK. The MAP kinase cascade is highly conserved in all eukaryotes and involved in numerous cellular responses. Activation of MAPK is a transient event that is tightly regulated by both kinases and phosphatases. A growth factor induced dual specific phosphatase is likely to play an important role in MAPK regulation.

High-resolution crystal structures of tyrosine kinase SH3 domains complexed with proline-rich peptides

Src-homology 3 (SH3) domains bind to proline-rich motifs in target proteins. We have determined high-resolution crystal structures of the complexes between the SH3 domains of Abl and Fyn tyrosine kinases, and two ten-residue proline-rich peptides derived from the SH3-binding proteins 3BP-1 and 3BP-2. The X-ray data show that the basic mode of binding of both proline-rich peptides is the same. Peptides are bound over their entire length and interact with three major sites on the SH3 molecules by both hydrogen-bonding and van der Waals contacts. Residues 4-10 of the peptide adopt the conformation of a left-handed polyproline helix type II. Binding of the proline at position 2 requires a kink at the non-proline position 3.

Hams and Egls: genetic analysis of cell migration in Caenorhabditis elegans

The analysis of mutations that disrupt egg laying by the Caenorhabditis elegans hermaphrodite has identified genes that are required for the long-range migrations of two cell types, the hermaphrodite-specific neurons and the sex myoblasts. Molecular analysis of some of these genes indicates that transcription factors and signal transduction molecules are necessary for the migrations of these cells.

PC12 cells overexpressing the insulin receptor undergo insulin-dependent neuronal differentiation

BACKGROUND

Stimulation of phaeochromocytoma PC12 cells by nerve growth factor leads to growth arrest and neuronal differentiation, whereas insulin induces various metabolic responses such as metabolism of glucose and lipids. Moreover, both insulin and epidermal growth factor stimulate the proliferation of PC12 cells. In spite of their different biological effects, nerve growth factor, insulin and epidermal growth factor induce very similar early responses in PC12 cells. Stimulation with nerve growth factor leads to the sustained activation and nuclear translocation of mitogen-activated protein (MAP) kinase. By contrast, both insulin and epidermal growth factor induce the transient activation of MAP kinase, without pronounced nuclear translocation of the enzyme. We have investigated whether the differential activation of signaling pathway components can account for the distinct cellular responses to these different growth factors.

RESULTS

By overexpressing insulin receptors in PC12 cells, we observed insulin-dependent neurite outgrowth, similar to that induced by nerve growth factor in both non-transfected and overexpressing cells. Overexpression of insulin receptors in PC12 cells leads to a more pronounced, but similar pattern of insulin-induced tyrosine-phosphorylated proteins in PC12 cells, including enhanced recruitment of Grb2/Sos into a complex with either Shc or IRS1. MAP kinase activation in response to insulin stimulation of cells overexpressing the insulin receptor is similar to MAP kinase activation in response to NGF stimulation of parental or overexpressing PC12 cells: the activation is prolonged and nuclear translocation of the enzyme occurs.

CONCLUSION

The differential subcellular localization and duration of MAP kinase activation induced by insulin and NGF may explain the difference in the biological actions of these two factors on PC12 cells. Our results show that the strength of the signal generated by a receptor with tyrosine kinase activity can influence the downstream signaling pathway, leading to cell differentiation instead of cell proliferation.

Overexpression of ERBB2 in human mammary epithelial cells signals inhibition of transcription of the E-cadherin gene

Overexpression of the ERBB2 receptor in transfectants of a human mammary epithelial cell line (MTSV1-7) is associated with a reduced ability to undergo morphogenesis in vitro and with a decreased level of expression of the E-cadherin and alpha 2 integrin genes. The inhibition of expression of the adhesion molecules has been shown to be at the level of transcription by using nuclear run-on assays and by following transcription of a reporter gene fused to 5' sequences of the E-cadherin gene. To relate the effects on gene transcription to a functional ERBB2 protein, signaling from the receptor was inhibited by the antibody 4D5, which blocks phosphorylation of ERBB2 on tyrosine residues and association of the protein with the GRB2/Sem5 protein. After treatment with the antibody 4D5, the ERBB2 transfectants regain the ability to form three-dimensional structures in collagen gels and the rates of transcription of the genes encoding the E-cadherin and the alpha 2 integrin subunit are restored to the levels seen in MTSV1-7neo cells. These results demonstrate that the inhibition of morphogenesis and transcription of specific adhesion molecules in human mammary epithelial cells can be affected by signals generated by the ERBB2 receptor and suggest a role for ERBB2 overexpression in tumor progression and metastasis.

Identification of Src, Fyn, and Lyn SH3-binding proteins: implications for a function of SH3 domains

Src homology 3 (SH3) domains mediate protein-protein interactions necessary for the coupling of cellular proteins involved in intracellular signal transduction. We previously established solution-binding conditions that allow affinity isolation of Src SH3-binding proteins from cellular extracts (Z. Weng, J. A. Taylor, C. E. Turner, J. S. Brugge, and C. Seidel-Dugan, J. Biol. Chem. 268:14956-14963, 1993). In this report, we identified three of these proteins: Shc, a signaling protein that couples membrane tyrosine kinases with Ras; p62, a protein which can bind to p21rasGAP; and heterogeneous nuclear ribonucleoprotein K, a pre-mRNA-binding protein. All of these proteins contain proline-rich peptide motifs that could serve as SH3 domain ligands, and the binding of these proteins to the Src SH3 domain was inhibited with a proline-rich Src SH3 peptide ligand. These three proteins, as well as most of the other Src SH3 ligands, also bound to the SH3 domains of the closely related protein tyrosine kinases Fyn and Lyn. However, Src- and Lyn-specific SH3-binding proteins were also detected, suggesting subtle differences in the binding specificity of the SH3 domains from these related proteins. Several Src SH3-binding proteins were phosphorylated in Src-transformed cells. The phosphorylation of these proteins was not detected in cells transformed by a mutant variant of Src lacking the SH3 domain, while there was little change in tyrosine phosphorylation of other Src-induced phosphoproteins. In addition, the coprecipitation of v-Src with two tyrosyl-phosphorylated proteins with M(r)s of 62,000 and 130,000 was inhibited by incubation with a Src SH3 peptide ligand, suggesting that the binding of these substrate proteins is dependent on interactions with the SH3 domain. These results strongly suggest a role for the Src SH3 domain in the recruitment of substrates to this protein tyrosine kinase, either through direct interaction with the SH3 domain or indirectly through interactions with proteins that bind to the SH3 domain.

GRB2 and phospholipase C-gamma 1 associate with a 36- to 38-kilodalton phosphotyrosine protein after T-cell receptor stimulation

GRB2, a 25-kDa protein comprising a single SH2 domain flanked by two SH3 domains, has been implicated in linking receptor protein tyrosine kinases (PTKs) to the Ras pathway by interacting with the guanine nucleotide exchange protein SOS. Previous studies have demonstrated that GRB2 directly interacts with Shc, a proto-oncogene product that is tyrosine phosphorylated upon receptor and nonreceptor PTK activation. In this report, we detected low levels of tyrosine phosphorylation of Shc and induced association with GRB2 upon T-cell receptor (TCR) stimulation. Instead, a prominent 36- to 38-kDa tyrosine phosphoprotein (pp36-38) associated with the SH2 domain of GRB2 and formed a stable complex with GRB2/SOS upon TCR stimulation. Cellular fractionation studies showed that whereas both GRB2 and SOS partitioned to the soluble and particulate fractions, pp36-38 was present exclusively in the particulate fraction. This phosphoprotein had the same apparent mobility in sodium dodecyl sulfate-polyacrylamide gel electrophoresis as the phosphoprotein that associates with phospholipase C-gamma 1 (PLC-gamma 1). Furthermore, following partial immunodepletion of GRB2 and of the associated pp36-38, there was a significant reduction in the amount of the 36-kDa phosphoprotein associated with PLC-gamma 1, suggesting that a trimeric PLC-gamma 1/pp36-38/GRB2 complex could form. In support of this notion, we have also been able to detect low levels of PLC-gamma 1 in GRB2 immunoprecipitates. We suggest that pp36-38 may be a bridging protein, coupling different signalling molecules to cytoplasmic PTKs regulated by the TCR.

Phosphatidylinositol 3-kinase activation is required for insulin stimulation of pp70 S6 kinase, DNA synthesis, and glucose transporter translocation

Phosphatidylinositol 3-kinase (PI 3-kinase) is stimulated by insulin and a variety of growth factors, but its exact role in signal transduction remains unclear. We have used a novel, highly specific inhibitor of PT 3-kinase to dissect the role of this enzyme in insulin action. Treatment of intact 3T3-L1 adipocytes with LY294002 produced a dose-dependent inhibition of insulin-stimulated PI 3-kinase (50% inhibitory concentration, 6 microM) with > 95% reduction in the levels of phosphatidylinositol-3,4,5-trisphosphate without changes in the levels of phosphatidylinositol-4-monophosphate or its derivatives. In parallel, there was a complete inhibition of insulin-stimulated phosphorylation and activation of pp70 S6 kinase. Inhibition of PI 3-kinase also effectively blocked insulin- and serum-stimulated DNA synthesis and insulin-stimulated glucose uptake by inhibiting translocation of GLUT 4 glucose transporters to the plasma membrane. By contrast, LY294002 had no effect on insulin stimulation of mitogen-activated protein kinase or pp90 S6 kinase. Thus, activation of PI 3-kinase plays a critical role in mammalian cells and is required for activation of pp70 S6 kinase and DNA synthesis and certain forms of intracellular vesicular trafficking but not mitogen-activated protein kinase or pp90 S6 kinase activation. These data suggest that PI 3-kinase is not only an important component but also a point of divergence in the insulin signaling network.

Enhancement or inhibition of insulin signaling by insulin receptor substrate 1 is cell context dependent

Insulin treatment of Chinese hamster ovary (CHO) cells expressing high levels of the insulin receptor (CHO/IR cells) activates both c-fos serum response element and activator protein 1 (AP-1) reporter genes approximately 10-fold. In contrast, parental CHO cells display only two- to threefold insulin stimulation of reporter gene activity. Transient transfection of parental CHO cells with an insulin receptor substrate 1 (IRS1) expression plasmid enhanced insulin downstream signaling in a biphasic manner, whereas IRS1 transfection of CHO/IR cells inhibited insulin signaling in a dose-dependent fashion. Further, expression of Grb2 in parental CHO cells had no effect on insulin signaling, whereas Grb2 increased insulin activation of reporter gene expression in CHO/IR cells. These data suggest that the expression levels of various effector molecules can either enhance or inhibit insulin downstream signaling events. To assess the relative effects of various insulin receptor, IRS1, and Grb2 levels on insulin signaling, parental CHO cells were transiently transfected with various combinations of expression plasmids encoding these proteins. Although expression of IRS1 resulted in a biphasic increase of insulin signaling in parental CHO cells, coexpression of IRS1 with the insulin receptor resulted in inhibition of signaling. This inhibition of insulin signaling directly correlated with an increased association of Grb2 with IRS1 and a concomitant sequestration of Grb2 away from Shc. Consistent with the Shc-Grb2 pathway as the major route for insulin-stimulated c-Fos and AP-1 transcriptional activation, the IRS1-mediated inhibition was reversed by transfection with an expression plasmid for Grb2. These data demonstrate that the extent of insulin-stimulated downstream signaling was dependent not only on the levels of individual signaling molecules but also on the formation of multiprotein complexes with specific stoichiometries.