ERKs, extracellular signal-regulated MAP-2 kinases.

A family of protein kinases, known alternatively as microtubule-associated protein-2/myelin basic protein kinases or extracellular signal-regulated kinases, is activated by numerous hormones, growth factors and other extracellular stimuli. At least two members of this family function as intermediate kinases in protein phosphorylation cascades. Their mechanisms of activation may involve autophosphorylation, which occurs on both threonine and tyrosine residues.

An insulin-stimulated protein kinase similar to yeast kinases involved in cell cycle control.

A protein kinase characterized by its ability to phosphorylate microtubule-associated protein-2 (MAP2), is thought to be an early intermediate in an insulin-stimulated phosphorylation cascade and in a variety of other mammalian cell responses to extracellular signals. A complementary DNA that encodes this protein serine-threonine kinase has been cloned, and the protein designated extracellular signal-regulated kinase 1 (ERK1). ERK1 has striking similarity to two protein kinases, KSS1 and FUS3, from yeast. The yeast kinases function in an antagonistic manner to regulate the cell cycle in response to mating factors. Thus, ERK1 and the two yeast kinases constitute a family of evolutionarily conserved enzymes involved in regulating the response of eukaryotic cells to extracellular signals.

Choice of STATs and other substrates specified by modular tyrosine-based motifs in cytokine receptors.

Many members of the cytokine receptor superfamily initiate intracellular signaling by activating members of the Jak family of tyrosine kinases. Activation of the same Jaks by multiple cytokines raises the question of how these cytokines activate distinct intracellular signaling pathways. Selection of particular substrates--the transcriptional activator Stat3 and protein tyrosine phosphatase PTP1D--that characterize responses to the ciliary neurotrophic factor-interleukin-6 cytokine family depended not on which Jak was activated, but was instead determined by specific tyrosine-based motifs in the receptor components--gp130 and LIFR--shared by these cytokines. Further, these tyrosine-based motifs were modular, because addition of a Stat3-specifying motif to another cytokine receptor, that for erythropoietin, caused it to activate Stat3 in a ligand-dependent fashion.

Association and activation of Jak-Tyk kinases by CNTF-LIF-OSM-IL-6 beta receptor components.

A recently defined family of cytokines, consisting of ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), oncostatin M (OSM), and interleukin-6 (IL-6), utilize the Jak-Tyk family of cytoplasmic tyrosine kinases. The beta receptor components for this cytokine family, gp130 and LIF receptor beta, constitutively associate with Jak-Tyk kinases. Activation of these kinases occurs as a result of ligand-induced dimerization of the receptor beta components. Unlike other cytokine receptors studied to date, the receptors for the CNTF cytokine family utilize all known members of the Jak-Tyk family, but induce distinct patterns of Jak-Tyk phosphorylation in different cell lines.

CNTF, FGF, and NGF collaborate to drive the terminal differentiation of MAH cells into postmitotic neurons.

The differentiation of neuronal cell progenitors depends on complex interactions between intrinsic cellular programs and environmental cues. Such interactions have recently been explored using an immortalized sympathoadrenal progenitor cell line, MAH. These studies have revealed that depolarizing conditions, in combination with exposure to FGF, can induce responsiveness to NGF. Here we report that CNTF, which utilizes an intracellular signaling pathway distinct from that of both FGF and NGF, can collaborate with FGF to promote efficiently the differentiation of MAH progenitor cells to a stage remarkably reminiscent of NGF-dependent, postmitotic sympathetic neurons. We also find that similar collaborative interactions can occur during transdifferentiation of normal cultured chromaffin cells into sympathetic neurons.

The tails of two proteins: the scrapie prion protein and the ciliary neurotrophic factor receptor.

Many proteins with a variety of functions have proven to have glycosylphosphatidylinositol (GPI)-linkages; two members of this family are the scrapie prion protein and the receptor for ciliary neurotrophic factor (CNTF). The scrapie prion protein has two isoforms: PrPC is found in brain cells from normal animals, while PrPSc is an abnormal isoform that is only found in scrapie-infected animals. PrPSc is the only identified component of the prion, an infectious agent that apparently does not contain nucleic acid. Models for how prions replicate require that PrPSc must somehow recruit PrPC and catalyze or stabilize a post-translational event that converts PrPC into PrPSc. Extensive characterization has suggested that this critical post-translational event is probably conformational and not a chemical change. The presence of a GPI anchor on CNTFR alpha is an unusual feature for a molecule that must transmit a signal to the inside of the cell. Recent data have indicated that CNTFR alpha must bind CNTF, then interact with two other "beta" receptor components to initiate signal transduction. Furthermore, we have shown that, unlike the vast majority of receptors, CNTFR alpha can function as a soluble molecule to promote CNTF action on cells that contain the two beta components, but do not themselves express CNTFR alpha. Intriguingly, we have also demonstrated that CNTFR alpha is present in cerebrospinal fluid and blood in vivo, and the release of CNTFR alpha from skeletal muscle is increased by denervation of the muscle. Whether the soluble form is released through GPI-anchor cleavage remains to be determined.

Identification of multiple extracellular signal-regulated kinases (ERKs) with antipeptide antibodies.

A protein kinase characterized by its ability to phosphorylate microtubule-associated protein-2 (MAP2) and myelin basic protein (MBP) is thought to play a pivotal role in the transduction of signals from many receptors in response to their ligands. A kinase with such activity, named extracellular signal-regulated kinase 1 (ERK1), is activated rapidly by numerous extracellular signals, requires phosphorylation on tyrosine to be fully active, and in vitro can activate a kinase (a ribosomal S6 protein kinase) that is downstream in phosphorylation cascades. From the protein sequence predicted by the rat ERK1 cDNA, peptides were synthesized and used to elicit antibodies. The antibodies recognize both ERK1; a closely related kinase, ERK2; and a third novel ERK-related protein. Using these antibodies we have determined that ERK1 and ERK2 are ubiquitously distributed in rat tissues. Both enzymes are expressed most highly in brain and spinal cord as are their mRNAs. The third ERK protein was found in spinal cord and in testes. The antibodies detect ERKs in cell lines from multiple species, including human, mouse, dog, chicken, and frog, in addition to rat, indicating that the kinases are conserved across species. ERK1 and ERK2 have been separated by chromatography on Mono Q. Stimulation by insulin increases the phosphorylation of both kinases on tyrosine residues, as assessed by immunoblotting with phosphotyrosine antibodies, and retards their elution from Mono Q. Each of these ERKs appears to account for a distinct peak of MBP kinase activity. The activity in each peak is diminished by incubation with either phosphatase 2a or CD45. Therefore, both enzymes have similar modes of regulation and appear to contribute to the growth factor-stimulated MAP2/MBP kinase activity measured in cell extracts.

Purification and properties of extracellular signal-regulated kinase 1, an insulin-stimulated microtubule-associated protein 2 kinase.

In rat 1 fibroblasts, insulin has little or no stimulatory effect on the activities of either MAP2 protein kinase or ribosomal protein S6 kinase. In contrast, in rat 1 cells that overexpress the normal human insulin receptor (rat 1 HIRc B; McClain et al. (1987) J. Biol. Chem. 262, 14663-14671), insulin activates both MAP2 and S6 kinase activities close to 5-fold. A MAP2 kinase has been purified from insulin-treated rat 1 HIRc B cells over 6300-fold by chromatography on Q-Sepharose, phenyl-Sepharose, S-Sepharose, phosphocellulose, QAE-Sepharose, UltrogelAcA54, DEAE-cellulose, and a second Q-Sepharose. Its specific activity is approximately 0.8-1 mumol.min-1.mg-1 with MAP2 and 3 mumol.min-1.mg-1 with myelin basic protein. The enzyme preparation contains one major band of Mr = 43,000 upon SDS-polyacrylamide gel electrophoresis, which is immunoblotted by antibodies to phosphotyrosine. A sequence from the 43-kDa band led to the isolation of a cDNA encoding the enzyme, which we have named ERK1 for extracellular signal-regulated kinase (Boulton et al. (1990) Science 249, 64-67).

Ciliary neurotrophic factor/leukemia inhibitory factor/interleukin 6/oncostatin M family of cytokines induces tyrosine phosphorylation of a common set of proteins overlapping those induced by other cytokines and growth factors.

Ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), oncostatin M (OSM), and interleukin-6 (IL6) compose a family of distantly related cytokines that initiate signaling by inducing either homodimerization of the "beta" signal transducing receptor component gp130 (in the case of IL6) or heterodimerization between gp130 and the gp130-related LIFR beta (in the case of CNTF, LIF, and OSM); dimerization of beta receptor components in turn activates members of the Jak/Tyk family of receptor-associated tyrosine kinases. Here we report that CNTF, LIF, OSM, and IL6 induce most of the same protein tyrosine phosphorylations, regardless of the cell type assayed or whether they initiate signaling by inducing homo- or heterodimerization of beta components. Although several of the protein tyrosine phosphorylations induced by the CNTF/LIF/OSM/IL6 family of factors may correspond to novel tyrosine kinase targets, we have been able to demonstrate the involvement of known signaling molecules, such as phospholipase C gamma, phosphoinositol 3-kinase, phosphotyrosine phosphatase (PTP1D), pp120, SHC, GRB2, STAT91, Raf-1, and the mitogen-activated protein kinases ERK1 and ERK2, revealing substantial convergence not only between the pathways activated by this cytokine family and other cytokines, but with pathways previously known to be activated only by factors that utilize receptor tyrosine kinases. Our data suggest the beta receptor components can form complexes with some of the signaling proteins identified and may play some role in their recruitment.

ERK3 is a constitutively nuclear protein kinase.

The ERK3 cDNA predicts a protein of 62,000 in size with a C-terminal domain that extends 180 amino acids beyond the conserved core of ERK family protein kinases. Immunoblotting with antibodies raised to recombinant protein and to peptides from the catalytic core and three regions of the C-terminal tail revealed that ERK3 is the expected size and is ubiquitously expressed in a variety of cell lines and tissues. ERK3, unlike the MAP kinases ERK1 and ERK2, is localized in the nucleus in exponentially growing, quiescent, and growth factor-stimulated cells. If the 180 amino acids at its C terminus are deleted, the resulting ERK3 fragment of 45 kDa is still found primarily in the nucleus, indicating that the C terminus is not required for its localization. Recombinant ERK3 expressed in mammalian cells or in bacteria is a protein kinase, as deduced from its capacity to autophosphorylate. Mutation of a conserved residue (Asp171) expected to be involved in catalysis eliminated autophosphorylation. Ser189 of ERK3, which corresponds to Thr183, one of the activating phosphorylation sites of ERK2, is autophosphorylated in vitro and phosphorylated in vivo. Despite marked similarities to ERK1 and ERK2, ERK3 does not phosphorylate typical MAP kinase substrates, indicating that it has distinct functions.

Extracellular signal-regulated kinases in T cells: characterization of human ERK1 and ERK2 cDNAs.

Extracellular signal-regulated kinases 1 and 2 are growth factor-sensitive serine/threonine kinases. cDNAs for both human kinases were isolated and sequenced. The nucleic acid and deduced protein sequences of human extracellular signal-regulated kinase 1 were 88% and 96% identical, respectively, to the homologous rat sequences. The nucleic acid and deduced protein sequences of human extracellular signal-regulated kinase 2 were 90% and 98% identical, respectively, to the corresponding rat sequences. A human extracellular signal-regulated kinase 2 specific probe was used to demonstrate that the mRNA for this kinase was present in T cells and did not change with activation. The deduced protein sequences of both human kinases were greater than 95% identical to two Xenopus kinase sequences, indicating that these enzymes are highly conserved across species.

Activation of p42 MAP kinase and the release of oocytes from cell cycle arrest.

Clam oocytes are arrested naturally at the G2/M border in meiosis and contain an inactive 42 kDa ERK/MAP kinase, p42MAPK. Following fertilization, p42MAPK is rapidly phosphorylated on tyrosine residues and concomitantly activated. Both tyrosine phosphorylation and activation of p42MAPK begin within 2-3 min of fertilization, peak at approximately 15 min, then rapidly decline and disappear around the end of meiosis I. Neither the tyrosine phosphorylated form of p42MAPK nor p42MAPK activity reappears during meiosis II or the succeeding mitotic cell cycles. High doses of molybdate, a potent PTPase inhibitor, block the phosphorylation of p42MAPK and entry into the cell cycle. Lower doses of molybdate delay both p42MAPK phosphorylation and the release from cell cycle arrest, but once cells have re-entered the cell cycle, they continue with near-normal timing. These results argue that the transient activation of p42MAPK at fertilization is a one-time event linked to release from cell cycle arrest. In trying to reconcile this one-time activation of p42MAPK in clam embryos with the recurring, M-phase specific activation of MBP/MAP kinases reported in other systems, we show that cdc2 kinase contributes a major portion of the MBP kinase activity in mitotic extracts. Furthermore, a small fraction of p42MAPK and other related kinases are present in p13suc1-bound material, cautioning against the use of p13suc1 beads for experiments where, in addition to cdc2, the unaccounted presence of other kinase activities could be misleading.

Extracellular signal-regulated kinases in T cells. Anti-CD3 and 4 beta-phorbol 12-myristate 13-acetate-induced phosphorylation and activation.

Extracellular signal-regulated kinases (ERK) 1 and 2 are growth factor- and cytokine-sensitive serine/threonine kinases that are known to phosphorylate microtubule-associated protein 2 and myelin basic protein. The current studies examined whether ERK1 and/or ERK2 was present in T cells and whether they were phosphorylated and activated as a consequence of T cell activation. The data demonstrated that both ERK1 and ERK2 were present in Jurkat cells and peripheral blood T cells. In T cells, ERK2 was more prevalent than ERK1. The concentrations of ERK1 and ERK2 were not altered by stimulating the cells for 16 h with immobilized anti-CD3 mAb or anti-CD3 mAb and phorbol myristate acetate. mAb to CD3 and phorbol myristate acetate stimulated an increase in ERK1 and ERK2 MBP kinase activity. Anti-CD3 mAb triggered an increase their phosphate content which was detectable at 2 min but reached a maximum at 5 min. A portion of the increase in phosphate was caused by an increase in phosphotyrosine. We also examined the rate of ERK2 degradation. ERK2 was stable for up to 36 h, and its degradation was unaffected by the activation state of the cells. The data demonstrate that ERK1 and ERK2 are part of an anti-CD3 mAb-stimulated signal transduction cascade that is downstream of protein kinase C and, therefore, suggest that these kinases play an important role in T cell activation.

An insulin-stimulated ribosomal protein S6 kinase from rabbit liver.

In this report we describe an activated form of S6 protein kinase in rabbits treated acutely with insulin. The major insulin-stimulated activity in rabbit liver is increased 2- to 5-fold compared to material from untreated animals based on DEAE-cellulose profiles. The activity observed in DEAE-cellulose fractions can be separated into a major and a minor peak, each having very similar chromatographic behavior. Chromatography on DEAE-cellulose, S-Sepharose, heptyl-Sepharose, heparin-agarose, and Mono Q results in greater than 20,000-fold purification of the insulin-stimulated enzyme with a 12% recovery. The stimulated activity has chromatographic properties similar to an S6 protein kinase studied previously in 3T3-L1 cells (Cobb, M. H. (1986) J. Biol. Chem. 261, 12994-12999) and other systems. The enzyme purified from insulin-treated animals contains a major band that migrates in sodium dodecyl sulfate-polyacrylamide gels with Mr congruent to 70,000; this band also appears in the control preparation. Treatment of the insulin-stimulated S6 kinase with the catalytic subunit of phosphatase 2a reduces its activity by 97%. The activity of the inactivated S6 kinase is stimulated nearly 5-fold by a 15-min preincubation with partially purified insulin-stimulated microtubule-associated protein-2 kinase.

Evidence for insulin-dependent activation of S6 and microtubule-associated protein-2 kinases via a human insulin receptor/v-ros hybrid.

The abilities of a series of six mutants of the human insulin receptor, an insulin receptor/v-ros hybrid (IR-ros) and the P68gag-ros transforming protein to stimulate S6 protein kinase have been assessed. Insulin receptor mutants in which either 1 or 2 tyrosine residues have been replaced with phenylalanine (YF1, YF3) have lost some or all of the capacity to mediate the activation of S6 kinase in response to insulin. None of the four mutants that contain deletions (spBam, spBamYF3, iBgl, T-t) elicit an insulin-dependent stimulation of S6 kinase. A previous study of the IRros hybrid receptor demonstrated that it was unable to cause either insulin-stimulated thymidine incorporation or glucose uptake (Ellis, L., Morgan, D. O., Jong, S.-M., Wang, L.-H., Roth, R. A., and Rutter, W. J. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 5101-5105). In contrast, the IRros chimera appears to mediate the activation of S6 protein kinase by insulin. In further evaluating the biological activities of the IRros hybrid, we have examined its effects on a microtubule-associated protein-2 (MAP2) kinase that is thought to be an early target in the cascade of reactions leading to increased S6 phosphorylation (Sturgill, T. W., Ray, L. B., Erickson, E., and Maller, J. L. (1988) Nature 334, 715-718). We find that the IRros receptor stimulates the MAP2 protein kinase from 3- to 6-fold in insulin-treated cells, conferring more than a 30-fold increase in the insulin sensitivity of MAP2 kinase activation.

STAT3 activation by cytokines utilizing gp130 and related transducers involves a secondary modification requiring an H7-sensitive kinase.

Ciliary neurotrophic factor, oncostatin M, leukemia-inhibitory factor, and interleukin 6 are related cytokines that initiate signaling by homodimerizing the signal-transducing receptor component gp130 or by heterodimerizing gp130 with a gp130-related receptor component. Receptor dimerization in turn activates receptor-associated kinases of the Jak/Tyk family, resulting in the rapid tyrosine phosphorylation of several intracellular proteins, including those of two members of the signal transducers and activators of transcription (STAT) family--STAT1 and STAT3. Here we show that all cytokines that utilize gp130 sequentially induce two distinct forms of STAT3 in all responding cells examined, with the two forms apparently differing because of a time-dependent secondary serine/threonine phosphorylation involving an H7-sensitive kinase. While both STAT3 forms bind DNA and translocate to the nucleus, the striking time-dependent progression from one form to the other implies other important functional differences between the two forms. Granulocyte colony-stimulating factor, which utilizes a receptor highly related to gp130, also induces these two forms of STAT3. In contrast to a number of other cytokines and growth factors, all cytokines using gp130 and related signal transducers consistently and preferentially induce the two forms of STAT3 as compared with STAT1; this characteristic STAT activation pattern is seen regardless of which Jak/Tyk kinases are used in a particular response, consistent with the notion that the receptor components themselves are the primary determinants of which STATs are activated.

CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gp130.

Ciliary neurotrophic factor (CNTF) has a variety of actions within the nervous system. While some of the actions of leukemia inhibitory factor (LIF) on neurons resemble those of CNTF, LIF also has broad actions outside of the nervous system that in many cases mimic those of interleukin-6 (IL-6). Comparison of the tyrosine phosphorylations and gene activations induced by CNTF and LIF in neuron cell lines reveals that they are indistinguishable and also very similar to signaling events that characterize LIF and IL-6 responses in hematopoietic cells. We provide a basis for the overlapping actions of these three factors by demonstrating that the shared CNTF and LIF signaling pathways involve the IL-6 signal transducing receptor component gp130. Thus, the receptor system for CNTF is surprisingly unlike those used by the nerve growth factor family of neurotrophic factors, but is instead related to those used by a subclass of hematopoietic cytokines.

Ciliary neurotrophic factor induces down-regulation of its receptor and desensitization of signal transduction pathways in vivo: non-equivalence with pharmacological activity.

Despite the widespread use of polypeptide growth factors as pharmacological agents, little is known about the extent to which these molecules regulate their cognate cell surface receptors and signal transduction pathways in vivo. We have addressed this issue with respect to the neurotrophic molecule ciliary neurotrophic factor (CNTF). Administration of CNTF in vivo resulted in modest decreases in levels of CNTFRalpha mRNA and protein in skeletal muscle. CNTF causes the rapid tyrosine phosphorylation of LIFRbeta and gp130 and the induction of the immediate-early gene, tis11; injection of CNTF 3-7 h after an initial exposure failed to re-stimulate these immediate-early responses, suggesting a biochemical desensitization to CNTF not accounted for by decreased receptor protein. To determine whether the desensitization of immediate-early responses caused by CNTF resulted in a functional desensitization, we compared the efficacy of multiple daily injections versus a single daily dose of CNTF in preventing the denervation-induced atrophy of skeletal muscle. Surprisingly, injections of CNTF every 6 h, which falls within the putative refractory period for biochemical responses, resulted in efficacy equal to or greater than injections once daily. These results suggest that although much of the CNTF signal transduction machinery is down-regulated with frequent CNTF dosing, biological signals continue to be recognized and interpreted by the cell.

Microtubule-associated protein 2 kinases, ERK1 and ERK2, undergo autophosphorylation on both tyrosine and threonine residues: implications for their mechanism of activation.

Microtubule-associated protein 2 kinase (MAP kinase), which exists in several forms, is a protein serine/threonine kinase that participates in a growth factor-activated protein kinase cascade in which it activates a ribosomal protein S6 kinase (pp90rsk) while being regulated itself by a cytoplasmic factor (MAP kinase activator). Experiments with recombinant MAP kinase, ERK2, purified from Escherichia coli in a nonactivated form revealed a self-catalyzed phosphate incorporation into both tyrosine and threonine residues. Another MAP kinase, ERK1, purified from insulin-stimulated cells also autophosphorylated on tyrosine and threonine residues. Autophosphorylation of ERK2 correlated with its autoactivation, although both autophosphorylation and autoactivation were slow compared to that occurring in the presence of MAP kinase activator. Therefore, we propose that autophosphorylation is probably involved in the MAP kinase activation process in vitro, but it may not be sufficient for full activation. The specificity toward tyrosine and threonine residues indicates that the MAP kinases ERK1 and ERK2 are members of a group of kinases with specificity for tyrosine as well as serine and threonine residues.

Neurotrophic factors, their receptors, and the signal transduction pathways they activate.

Our studies of the spatiotemporal availability of neurotrophic factors, coupled with tagged ligand binding assays that identify cell bearing receptors for these factors, should lead toward defining the physiological roles of these molecules in the animal. The use of the tagged ligands to identify factor-responsive cell lines has also provided new model systems for the examination of ligand-receptor interactions, as well as for the study of the subsequent induction of intracellular response pathways. To obtain insights into such intracellular pathways, we have molecularly cloned genes encoding a family of serine-threonine protein kinases, most closely related to kinases involved in the yeast response to pheromones. These kinases may be crucial regulators of early steps in the response of mammalian cells to neurotrophic factors as well as other extracellular signals.