Activation of the Mitogen-activated Protein Kinase Pathway by the Erythropoietin Receptor*

The erythropoietin receptor (EpoR) belongs to the cy- tokine receptor family, members of which lack a tyrosine kinase domain. Recent studies, however, have shown that a cytoplasmic tyrosine kinase, JAK2, inter-acts with the cytoplasmic domain of the EpoR and becomes activated upon binding of Epo to the receptor. Epo has also been shown to stimulate activation of Ras and Raf-1. The present studies were undertaken to ex- amine the possible involvement of Epo-induced tyrosine phosphorylation in activation of the Radmitogen-acti-vated protein kinase ( M A P kinase) pathway and to de- termine its significance on the growth signaling from the EpoR. In an interleukin (IL)-3-dependent cell line expressing the transfected wild-type EpoR, Epo, or IL-3 induced tyrosine phosphorylation of Shc and its associa- tion with Grb2. These cytokines also induced tyrosine phosphorylation and activation of MAP kinase isoforms ERKl and E m . A mutant EpoR with a carboxyl-termi-nal deletion of 108 amino acids (H mutant), which is mitogenically functional but lacks tyrosine phosphoryl- ation sites in the carboxyl-terminal region, showed

The erythropoietin receptor (EpoR) belongs to the cytokine receptor family, members of which lack a tyrosine kinase domain. Recent studies, however, have shown that a cytoplasmic tyrosine kinase, JAK2, interacts with the cytoplasmic domain of the EpoR and becomes activated upon binding of Epo to the receptor. Epo has also been shown to stimulate activation of Ras and Raf-1. The present studies were undertaken to examine the possible involvement of Epo-induced tyrosine phosphorylation in activation of the Radmitogen-activated protein kinase ( M A P kinase) pathway and to determine its significance on the growth signaling from the EpoR. In an interleukin (IL)-3-dependent cell line expressing the transfected wild-type EpoR, Epo, or IL-3 induced tyrosine phosphorylation of Shc and its association with Grb2. These cytokines also induced tyrosine phosphorylation and activation of MAP kinase isoforms ERKl and E m . A mutant EpoR with a carboxyl-terminal deletion of 108 amino acids (H mutant), which is mitogenically functional but lacks tyrosine phosphorylation sites in the carboxyl-terminal region, showed markedly diminished abilities to induce tyrosine phosphorylation of Shc and to phosphorylate and activate MAP kinases. A mutant receptor (PM4 mutant) inactivated by a point mutation, T r p 2 * 2 to Arg, which abrogates the interaction with JAK2, failed to induce any effect on Shc or MAP kinases. In cells expressing a mutant EpoR that is constitutively activated by a point mutation, Arg12' to Cys, in the extracellular portion of the receptor, neither tyrosine phosphorylation of Shc nor activation of MAP kinases by phosphorylation was detectable without stimulation with Epo or IL-3. These results suggest that the carboxyl-terminal region of EpoR may play a crucial role in activation of MAP kinases through the Ras signaling pathway which may be activated by tyrosine phosphorylation of Shc and its association with Grb2. The activation of MAP kinases, however, failed to correlate with the mitogenic activity of mutant EpoRs and thus may not be required for growth signaling from the EpoR.
Erythropoietin (Epo)' is a hematopoietic growth factor that regulates the growth and differentiation of erythroid progeni-* This work was supported in part by grants from the Ministry of Education, Science and Culture of Japan, from the Uehara Memorial Foundation, and from the Yamanouch1 Foundation for Research on Metabolic Disorders. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. Tel.: 81-3-3813-6111 (ext. 3218); Fax: 81-3-3818-0448. The abbreviations used are: Epo, erythropoietin; EpoR, erythropoietin receptor; IL, interleukin; SH2, Src homology 2; SH3, Src homology tor cells (1,2). The receptor for Epo (EpoR) (3,4) belongs to the cytokine receptor family which includes receptors for most of the hematopoietic growth factors (5, 6). The members of this family characteristically have 4 positionally conserved cysteines and a Trp-Ser-X-TlpSer (WSXWS) motif in the extracellular domain. Except for a short region showing a limited sequence similarity (7-101, the intracellular region of these receptors is quite diverse and lacks any motifs that would indicate potential function, such as the tyrosine kinase domain. Most of the hematopoietic growth factors, however, induce tyrosine phosphorylation of cellular proteins (3,6). Furthermore, introduction of a variety of activated tyrosine kinases has been shown t o abrogate dependence of hematopoietic cell lines on these hematopoietic growth factors (11). These observations have led to a hypothesis that tyrosine phosphorylation plays a critical role in the growth signaling from the hematopoietic cytokine receptors.
Using Epo-dependent cell lines expressing the endogenous or transfected EpoR, we and others have shown that Epo rapidly induces tyrosine phosphorylation of a series of cellular substrates, some of which are also phosphorylated by stimulation with interleukin (1L)-3 (12, 13). We further revealed that a membrane proximal cytoplasmic region, showing a limited homology with the other cytokine receptors, is critical for induction of tyrosine phosphorylation, the expression of a series of immediate-early genes, including c-myc, c-fos, and egr-1, and mitogenesis (10). Upon Epo binding, the carboxyl-terminal region of the EpoR becomes phosphorylated on tyrosine (13,14) and physically associates with the 85-kDa regulatory subunit of phosphatidylinositol 3-kinase (15) but is not required for growth signaling (13,16,17). Very recently, we have revealed that Epo stimulation induces binding of JAK2, a member of the J A K family of cytoplasmic tyrosine kinases, to the functionally critical membrane proximal region of the receptor and activates its kinase activity (18,19). The JAK2 kinase was thus suggested to couple Epo binding to tyrosine phosphorylation. However, little has been known how the signal is transmitted to the downstream signaling elements.
For both receptor and cytoplasmic tyrosine kinases, activation of Ras has been regarded as a critical step in inducing mitogenesis or differentiation (20,211. The proteins Grb2, Shc, and mSosl have been implicated in the control of Ras by tyrosine kinases in a number of different biological systems. Grb2 is an adapter protein that lacks a catalytic domain and is composed of one Src homology 2 (SH2) domain flanked by two Src homology 3 (SH3) domains (22,23). Grb2 binds to a subset of 3; MAP kinases, mitogen-activated protein kinases; ERK, extracellular signal-regulated kinase; MEK, MAP kinase/ERK-activating kinase; MBP, myelin basic protein; GM-CSF, granulocyte-macrophage colonystimulating factor; G-CSF, granulocyte colony-stimulating factor; PAGE, polyacrylamide gel electrophoresis. autophosphorylated receptor tyrosine kinases through its SH2 domain (22,231 and simultaneously associates through its SH3 domains with mSosl, a guanine nucleotide-releasing protein that activates Ras by inducing exchange of GDP for GTP on Ras (24). Grb2 thus repositions mSosl adjacent to Ras, which is located at the plasma membrane. The shc gene encodes three overlapping proteins of 46, 52, and 66 kDa which possess a carboxyl-terminal SH2 domain and a glycine/proline-rich region but no obvious catalytic domain (25). Shc proteins become tyrosine-phosphorylated upon activation of a variety of receptor tyrosine kinases (25-27) and are phosphorylated constitutively in cells transformed by the v-Src or v-Fps tyrosine kinases (28). Shc has been implicated in activation of Ras, because the tyrosine-phosphorylated Shc proteins associate with Grb2 and mSosl(26,29,30). Moreover, overexpression of Shc proteins led to Ras-dependent neurite outgrowth in PC12 cells (26).

A B
Recent studies have revealed that tyrosine kinase pathway activates mitogen-activated protein kinases ( M A P kinases), also known as extracellular signal-regulated kinases (ERKs), in a Ras-dependent manner (31, 32). MAP kinases are serine/ threonine kinases that phosphorylate well studied regulatory proteins, including transcription factors, membrane proteins, cytoskeletal elements, and other protein kinases (31, 33). MAP kinases are thus thought to be key intermediate regulatory proteins functioning in signal transduction networks. MAP kinases achieve maximum activity when phosphorylated on both tyrosine and threonine residues by MAP kinase/ERK-activating kinase (MEK), a dual specificity kinase (31, 32). MEK in turn is phosphorylated and thus activated by a serine/ threonine kinase, 32). It is known that activated Ras physically associates with Raf-1 and that phosphorylation is required for activation of Raf-1 (32). However, the mechanism how Ras activates Raf-1 remains to be known.
The cascade of intracellular signal transducing events leading to activation of MAP kinases may also be involved in the EpoR signaling, because Epo also activates Ras (34, 35) and induces phosphorylation and activation of Raf-1 (36). The present studies were undertaken to examine the involvement of Ras-activating proteins and MAP kinases in the EpoR signaling. Epo was found to induce tyrosine phosphorylation of Shc and its association with Grb2 in Epo-stimulated cells. Epo also induced tyrosine phosphorylation and activation of MAP kinases. Among mutant EpoRs, the ability to induce phosphorylation of Shc and its association with Grb2 correlated with the ability to activate MAP kinases. However, activation of MAP kinases failed to correlate with the mitogenic activity of mutant EpoRs.

MATERIALS AND METHODS
Cells and Reagents-A clone of 32D cells, an IL-3-dependent cell line originally isolated from long term bone marrow cultures, has been previously described (37). 32D clones expressing the wild-type or various mutant EpoRs and IL-3-dependent DA3 cells expressing the wild-type EpoR were also described previously (10, 13) and maintained in RPMI 1640 medium supplemented with 10% fetal calf serum and 10% WEHI-3 conditioned medium as a source of IL-3.
An expression plasmid for activated mutant EpoR with an ArglZ9 to Cys mutation was constructed by primer-mediated mutagenesis using the polymerase chain reaction method as described previously (10). Transfection of the plasmid into DA3 cells and isolation of a clone expressing the mutant EpoR were also carried out as described previously (13). In brief, DA3 cells were transfected with 10 pg of the expression plasmid and 1 pg of the pSV2neo plasmid by electroporation and selected in medium containing G418. Six clones were isolated by limiting dilution and confirmed to grow in medium containing neither IL-3 nor Epo. These clones were then analyzed by "'I-Epo binding assays, and the clone that bound the highest radioactivity was used for the subsequent studies.
For measurement of the cell number increase, cells were cultured in 25-cm2 culture flasks a t a density of 1 x 105/ml (5 ml of culture) in 10% fetal calf serum-containing RPMI 1640 medium supplemented with 10% WEHI-3 conditioned medium or with 4 unitdml human recombinant Epo. Culture media were changed every 2 days. Viable cell counts were determined by trypan blue staining.
The preparation and properties of rabbit polyclonal antisera against the cytoplasmic portion of recombinant murine EpoR (38) or against synthetic peptides from JAKl and JAK2 (39) have been described. Antiphosphotyrosine monoclonal antibody (4G10) and rabbit antisera against Shc and ERKl (erkl-CT) were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). A rabbit antiserum against ERK2 and a monoclonal antibody against Grb2 were purchased from Transduction Laboratories (Lexington, KY). Recombinant human Epo was kindly provided by Sankyo Pharmaceutical Co. Ltd. (Tokyo, Japan).
Immunoprecipitation and Immunohlotting-For stimulation with Epo or IL-3, cells were starved for 12 h without IL-3 in complete medium. The cells were then left unstimulated as a negative control or stimulated with a saturating concentration of Epo or IL-3 a t 37 "C for 10 min. The cells were lysed in a lysis buffer containing 1% Triton X-100, 20 mM Tris-HCI (pH 7.5), 150 mhl NaCI, 1 mM EDTA, 100 PM sodium orthovanadate, 1 m M phenylmethylsulfonyl fluoride, and 10 pg/ml each of aprotinin and leupeptin. For immunoprecipitation of MAP kinases, 1 x 10' cells were lysed in 100 p1 of boiling lysis buffer containing 1% SDS and 10 mM Tris-HCI (pH 7.4) and boiled for another 5 min. The lysate was then diluted 10-fold with the Triton X-100 lysis buffer and denatured MAP kinases were immunoprecipitated with anti-MAP kinase antibodies.
For immunoprecipitation, a relevant antibody was added to the lysates along with protein A-Sepharose beads and incubated for 4 h a t 4 "C. The beads were washed extensively, and the proteins bound to the beads were resolved by SDS-PAGE, transferred to polyvinylidene difluoride membranes, immunoblotted with the indicated antibody, and developed by the enhanced chemiluminescence (ECL) system (Amersham Corp.). Aliquots of the cell lysates were also subjected to immunoblotting after directly mixed with equal volumes of 2 x Laemmli's sample buffer and heated a t 100 "C for 5 min. For reprobing with a different antibody, the membranes were treated at 50 "C for 30 min with stripping buffer containing 100 mh1 2-mercaptoethanol, 2% SDS, and 62.5 mM Tris-HC1 (pH 6.7).
For determination of the stoichiometry of tyrosine phosphorylation of Shc and MAP kinases, 2 x 10' cells stimulated with 75 unitdm1 Epo for 10 min were lysed in 1 ml of the lysis buffer. After clarification, 50 pl of the lysate was mixed with the equal volume of 2 x SDS buffer and boiled for 5 min, whereas 500 pl of the lysate was immunoprecipitated with 10 pg of the anti-phosphotyrosine 4G10 antibody conjugated with agarose beads and subsequently eluted with 50 pl of 1 x SDS buffer by boiling for 5 min. The whole cell lysate and the immunoprecipitate were diluted -E p l U

200-
appropriately and varying amounts of the samples were subjected to immunoblot analysis using antibody against phosphotyrosine, Shc, or MAP kinases.
Kinase Assays of Anti-MAP Kinase Immunoprecipitates in Myelin Basic Protein fMBP)-containing Gels after SDS-PAGE-Determination of MAP kinase activity in MBP-containing gel was carried out essentially as described (40). In brief, anti-ERK1 immunoprecipitates were prepared as described above and electrophoresed on an SDS-12% polyacrylamide gel containing 0.5 mg/ml MBP. After removing SDS with buffer containing 20% 2-propano1, the gel was denatured with 6 M guanidine HCI and then renatured in a 0.04% Tween 40-containing buffer. Phosphorylation of MBP was carried out by incubating the gel a t 22 "C for 1 h in 40 mM HEPES (pH 7.51, 0.1 mM EGTA, 20 mM MgCI,, 20 p~ ATP, and 25 pCi of [Y-:'~P]ATP. After incubation, the gel was washed extensively with 5% trichloroacetic acid, 1% pyrophosphate solution, dried, and subjected to autoradiography. The radioactivity of phosphorylated MBP was also quantified by Bio-Imaging Analyzer BAS2000 (Fuji Film, Tokyo).

RESULTS
Epo Induces Tyrosine Phosphorylation of Shc and Its Association with Grb2-To explore the signal transduction pathways from the EpoR, we first examined whether Epo stimulation induces tyrosine phosphorylation of Shc, which has been implicated in coupling receptor tyrosine kinases to the Ras signaling pathways. As shown in Fig. lA, Epo stimulation induced tyrosine phosphorylation of 150-, 130-, 97-, 92-, 72-, 70-, 52-, and 49-kDa proteins in the 32D cells expressing the wildtype EpoR (32DlEpoR-Wt). IL-3 stimulation induced an almost identical pattern of tyrosine phosphorylation except that the 72-kDa protein was not phosphorylated after IL-3 stimulation. In accordance with our previous results (13,18,19), anti-phosphotyrosine blotting of immunoprecipitates obtained with relevant antisera showed that the 72-and 130-kDa substrates were the EpoR and JAK2, respectively, whereas it was revealed that JAKl was not phosphorylated after stimulation with these cytokines in the 32D cells examined.
Anti 49-kDa proteins in cells stimulated with Epo or IL-3 (Fig. lA ).
These proteins were identified as Shc by reprobing with anti-Shc (Fig. 1B). In addition, tyrosine phosphorylated 210-and 150-kDa proteins were coimmunoprecipitated with Shc after Epo or IL-3 stimulation.
To address the functional significance of tyrosine phosphorylation of Shc, we next examined previously characterized 32D clones expressing various mutant EpoRs (10, 13). 32DEpoR-H cells express the truncated H mutant lacking the carboxylterminal 108 amino acids. Although this mutant lacks the tyrosine phosphorylation sites in the carboxyl-terminal region and fails to associate with PI 3-kinase (151, it retains the abilities to activate JAK2 (18,19) and to transduce a mitogenic signal (10, 13). The mitogenic response of 32D/EpoR-H, which was selected by the G418 resistance, to Epo was thus comparable with that to IL-3 as determined by ["]thymidine incorporation (10) or by the growth curves (Fig. 2). 32DEpoR-PM4 cells express the PM4 mutant, which contains a mutation, Trp2*' to Arg (W282R), in a membrane proximal region of the cytoplasmic domain that shows homology with other members of the cytokine receptor superfamily (10). Having lost the ability to associates with JAK2 to activate its kinase activity (19), the PM4 mutant fails to induce tyrosine phosphorylation of cellular substrates, including the receptor itself, and to elicit a mitogenic response (10, 13). 32DEpoR-PM4, thus, failed to grow in response to Epo, as shown in Fig. 2.
The abilities of these mutant EpoRs to induce tyrosine phosphorylation of cellular proteins were first examined by antiphosphotyrosine blotting of total cell lysates. As shown in Fig.  3A, Epo induced tyrosine phosphorylation of p130 (JAK2) and p92 in 32DEpoR-H cells. In most of the repeated experiments, the tyrosine phosphorylation of JAK2 in these cells was enhanced as compared with that in the 32DEpoR-Wt cells, whereas p70 showed variable degrees of tyrosine phosphorylation in response to Epo. On the other hand, the tyrosine phosphorylation of p52 and p49, which correspond to Shc, after Epo stimulation in 32DEpoR-H was constantly and remarkably diminished as compared with that observed after IL-3 stimulation in this cell line or after Epo stimulation in 32DEpoR-Wt. The PM4 mutant failed to induce any detectable tyrosine phosphorylation of cellular substrates in response to Epo (Fig. 3A).
The tyrosine phosphorylation of Shc in these cell lines was then directly examined by anti-phosphotyrosine blotting of anti-Shc immunoprecipitates. Results shown in Fig. 3B confirmed that the Epo-induced tyrosine phosphorylation of Shc was severely impaired or abolished in 32DEpoR-H or 32D/ EpoR-PM4, respectively.
Because tyrosine-phosphorylated Shc has been shown to activate the Ras signaling pathway by physically associating with the Grb2-mSosl complex through the SH2 domain of Grb2, the anti-Shc immunoprecipitates were reprobed with anti-Grb2 to examine the association of Grb2 with Shc in Epo-stimulated cells. As shown in Fig. 30, in 32DEpoR-Wt, Grb2 was shown to coimmunoprecipitate with Shc after stimulation with Epo or IL-3. 32DEpoR-H cells also showed the Epo-induced association of Grb2 with Shc. However, the amount of Grb2 associated with Shc was markedly decreased in accordance with the decrease in tyrosine phosphorylation of Shc. Epo failed to induce the association of Grb2 with Shc in 32DEpoR-PM4 (Fig. 30).

nrosine Phosphorylation and Activation of MAP Kinases
Induced by Epo Stimulation-Recent studies have revealed that activation of MAP kinases by a variety of receptor tyrosine kinases is mediated by Ras (31,32). Since Epo has been shown to stimulate the activation of Ras (34, 35), we next examined whether Epo stimulation induces tyrosine phosphorylation and activation of MAP kinases, which are activated by phosphoryl- ation on both tyrosine and serinekhreonine residues (31, 32). To examine tyrosine phosphorylation of MAP kinases, immunoprecipitates obtained with anti-ERK1 (erk 1-CT), which also recognizes ERK2, were examined by anti-phosphotyrosine blotting. As shown in Fig. 4A, 42-and 44-kDa species of MAP kinases, which should correspond to ERK2 and ERK1, respectively, were found to be tyrosine-phosphorylated after Epo or IL-3 stimulation in 32DEpoR-Wt. However, tyrosine phosphorylation of MAP kinases induced by Epo stimulation was remarkably reduced in 32DEpoR-H (Fig. 4A). Reprobing of the membrane with anti-ERK1 confirmed that the 42-and 44-kDa species of tyrosine-phosphorylated proteins are directly recognized by anti-ERK1 and demonstrated equal loading of samples (Fig. 4A, center panel). The anti-ERK1 immunoprecipitates were then subjected to the MAP kinase assay in MBP-  containing gel. As shown in Fig. 4A (lower panel), Epo stimulation was found to stimulate the kinase activities of both ERKl and ERK2. However, although the effect of Epo on the MAP kinase activity was about half as much as that of IL-3 in 32DIEpoR-Wt, Epo showed only a marginal effect on the MAP kinase activity in 32DIEpoR-H (Fig. 4, A and B ). In 32DIEpoR-PM4 cells, Epo failed to induce any detectable tyrosine phosphorylation or activation of MAP kinases (data not shown). Stoichiometry of the Tyrosine Phosphorylation of Shc and MAP Kinases-We next tried to determine what fraction of Shc or MAP kinases is tyrosine-phosphorylated in response to Epo stimulation. For this purpose, phosphotyrosyl proteins were immunoprecipitated with the 4G10 antibody from Epo-stimulated 32DIEpoR-Wt cell lysate and, along with varying amounts of whole cell lysate, subjected to immunoblot analyses. First, the efficiency of immunoprecipitation with 4G10 was evaluated by immunoblotting with this antibody. The amount of tyrosine-phosphorylated Shc or MAP kinases in the 4G10 immunoprecipitate was then estimated by immunoblotting with anti-Shc or anti-MAP kinase, respectively. As shown in Fig. 5, the efficiency of immunoprecipitation with 4G10 was rather low and varied significantly with each phosphotyrosyl proteins, which may be a t least partly because some phosphorylated tyrosine residues are involved in intramolecular or intermolecular interaction with the SH2 domains and thus may not bind with 4G10. Densitometric analysis of the results shown in Fig. 5, representative of three repeated experiments, revealed that the efficiency of 4G10 immunoprecipitation of tyrosine-phosphorylated Shc or ERK-1 was -8 or -5%, respectively, whereas -3 or -0.5% of the total cellular Shc or ERK-1, respectively, was shown to be present in the 4G10 immunoprecipitates. From these results, it was calculated that -38 or -10% of total cellular Shc or ERK-1, respectively, undergoes tyrosine phosphorylation in response to Epo in 32DIEpoR-Wt cells. Thus, Epo induced significant but not near-stoichiometric tyrosine phosphorylation of Shc and MAP kinases. The rather low stoichiometry of tyrosine phosphorylation, particularly of MAP kinases, is compatible with our results that only a barely detectable activity of MAP kinases was observed in Epo-stimulated 32DEpoR-H cells, in which Epo-induced tyrosine phos- phorylation of Shc and MAP kinases is markedly diminished as compared with that in 32DIEpoR-Wt.

Cells Expressing a Constitutively Activated EpoR Mutant Show Epo-independent Growth without Tyrosine Phosphorylation of Shc or Activation of MAP
Kinases-Previously, a single point mutation, resulting in an Arg to Cys change a t residue 129 of the extracellular domain of EpoR, has been shown to activate the receptor independent of Epo binding (41,421. The R129C mutant was confirmed to abrogate the factor-dependence of DA3 cells, because six subclones of transfectants, selected by the G418 resistance due to the cotransfected pSV2neo plasmid, grew in the absence of IL-3 or Epo; the clone used in this study grew comparably in medium with or without Epo (Fig. 6). We examined the tyrosine phosphorylation states of Shc and MAP kinases as well as the activity of MAP kinases in a DA3 clone that expresses this mutant receptor and thus grows without added growth factors. As shown in Fig. 7, antiphosphotyrosine and anti-Grb2 blotting of anti-Shc immunoprecipitates showed that tyrosine phosphorylation of Shc and its association with Grb2 was dependent on Epo stimulation in these cells showing the Epo-independent growth. Tyrosine phosphorylation and activation of MAP kinases were also dependent on stimulation with Epo or IL-3 (Fig. 8). These cells were thus found to grow without showing tyrosine phosphorylation of Shc or activation of MAP kinases in growth factordeficient medium.

DISCUSSION
The present studies demonstrated that Shc and MAP kinases are among the substrates of tyrosine phosphorylation induced by Epo stimulation. In addition, Epo induced physical association of Shc with Grb2 and activated the catalytic activity of MAP kinases. However, Epo did not have any effect on these signaling molecules in cells expressing the PM4 mutant EpoR, which has a point mutation, Trp282 to Arg, abolishing the ability of EpoR to couple with JAK2 to transduce a mitogenic signal. The effects of Epo on Shc and MAP kinases were markedly diminished in cells expressing the mitogenically functional, truncated H mutant, which has lost tyrosine phosphorylation sites in the carboxyl-terminal region. In cells expressing the mutant EpoR that is constitutively activated by a point mutation, Arg'" to Cys, tyrosine phosphorylation of Shc and its association with Grb2 as well as tyrosine phosphorylation and activation of MAP kinases were not observed without stimula- examined. A tyrosine-phosphorylated protein of 145 kDa (43,44) or 140 kDa (45), which should correspond to the 15O-kDa tion with Epo or IL-3. Thus, in cells expressing these mutant protein in the present studies, were also found to be associated EpoRs, activation of MAP kinases correlated with tyrosine with Shc in cells stimulated with Epo or IL-3. Although the phosphorylation of Shc and its association with Grb2. Taken identity of this species remains unknown, we found in a previtogether with previous reports showing that Epo activates Ras ous study that this 150-kDa protein coimmunoprecipitated and Raf-1 (34-361, the present results suggest that Shc and with the EpoR or JAK2 in digitonin lysates and underwent Grb2 may mediate activation of the Ras/MAP kinases pathway tyrosine phosphorylation in vitro (19). It is thus possible that from the EpoR and that the carboxyl-terminal region of the the EpoR-JAK2 complex physically associates with the 140receptor may play a major role in activation of this pathway. kDa protein, Shc, and Grb2, although the association may be Activation of MAP kinases, however, may not be required for the transduction of growth signal from the EpoR, because the activity of MAP kinases did not correlate with proliferation of cells expressing the mutant EpoRs.
Recent studies also demonstrated that Epo or IL-3 induces tyrosine phosphorylation of Shc and its association with Grb2 (43)(44)(45). Damen et al. (43) also observed that tyrosine-phosphorylated Shc associates with the EpoR in a human M07 cell line expressing transfected murine EpoR. However, association of transient or unstable.
Although the mechanism of how the EpoR activates Ras has remained elusive, inhibition of the Epo-induced activation of Ras by tyrosine kinase inhibitors suggested that it may be mediated through tyrosine phosphorylation (34). In a previous study (341, GTPase-activating protein was implicated in the Epo-induced activation of Ras, because Epo induced tyrosine phosphorylation of Ras GTPase-activating protein in a human erythroleukemia cell line (HEL). On the other hand, it was Shc with the EpoR was not observed in DA3 transfectants in demonstrated in the present study and also in previous studies their studies (43). We also failed to observe any significant (43)(44)(45) that Epo stimulation induces tyrosine phosphorylation association of the EpoR with Shc in the 32D or DA3 clones of Shc and its association with Grb2. Since Shc is strongly implicated in the activation of Ras through its interaction with Grb2 and indirectly with mSos1 (24, 26), Epo-induced activation of Ras may be mediated through tyrosine phosphorylation of Shc. In accordance with this idea, in cells expressing the mutant EpoRs, tyrosine phosphorylation of Shc and its association with Grb2 correlated with activation of MAP kinases, which are major targets of the Ras signaling pathway (31, 32). Very recently, we found that Epo also induces tyrosine phosphorylation of Vav (46), which has also been implicated in the Ras signaling pathway (47). However, the tyrosine phosphorylation of Vav induced by Epo did not show correlation with the activation of MAP kinases in cells expressing the various mutant EpoRs (46). Therefore, a role, if any, of Vav in the Epoinduced activation of Ras remains to be known.
Recent studies have demonstrated that activation of MAP kinases is involved in signaling from most of the members of cytokine receptor family. However, the functional significance of MAP kinase activation in regulation of growth and differentiation of hematopoietic cells by these cytokines remains elusive. In fibroblast, the MAP kinase activity was shown to be required for proliferation (48). Correlation of proliferative response with activation of MAP kinases has also been shown in hematopoietic cells stimulated with G-CSF; G-CSF activates MAP kinases in cell lines that proliferate in response to G-CSF, whereas neither G-CSF-induced granulocytic differentiation of 32D cells nor nonproliferative response of mature neutrophils to G-CSF was associated with MAP kinase activation (49). On the other hand, GM-CSF induces tyrosine phosphorylation and activation of MAP kinases in nonproliferative mature neutrophils (50,51). Dissociation between mitogenesis and MAP kinase activation has also been demonstrated in hematopoietic cell lines that proliferate in response to IL-4 without activation of MAP kinases (52). In the present studies, activation of MAP kinases also failed to correlate with mitogenesis, since the truncated H mutant, whose mitogenic function is intact, showed a significant impairment in activation of MAP kinases, whereas cells expressing the constitutively activated EpoR mutant grew in growth factor-deficient medium without showing any detectable MAP kinase activation. These observations raise a possibility that activation of MAP kinases may play a role in transduction of signals other than a mitogenic signal.
The present studies indicated that the membrane distal cytoplasmic region of the EpoR may play a major role in tyrosine phosphorylation of Shc and activation of MAP kinases. Sat0 et a2. (53) similarly reported that the membrane distal cytoplasmic region of the common p subunit of the GM-CSF receptor is essential for induction of tyrosine phosphorylation of Shc and for activation of Ras, Raf-1, and MAP kinases. This region of the receptor was also required for induction of c-fos and c-jun by GM-CSF stimulation. The authors thus suggested that GM-CSF-induced tyrosine phosphorylation of Shc may cause activation of Ras, Raf-1, and MAP kinases and ultimately lead to the expression of c-fos and c-jun. Recent studies have also shown that MAP kinases activate the transcription of c-fos by phosphorylating, directly or indirectly, factors that bind the serum-responsive element in the promoter region of the c-fos gene (31,33). On the other hand, a previous study showed that induction of c-fos expression by Epo stimulation was not impaired in cells expressing the truncated H mutant, which is defective in activating the MAP kinase pathway (10). Intriguing in this regard is a recent study disclosing Epo-induced tyrosine phosphorylation and activation of 91and 84-kDa proteins that translocate from the cytoplasm to the nucleus and form DNA-binding complexes which recognize the sis-inducible element in the c-fos promoter (54). A similar Ras-independent signaling pathway directly activating DNA-binding proteins that recognize the sis-inducible element is also utilized by receptors for various growth factors, cytokines, and interferons (55,56). Importantly, the J A K family of tyrosine kinases, which includes JAK2, is strongly implicated in this pathway (55,561. It is thus tempting to speculate that JAK2, which is activated by the H mutant, may activate a latent transcription factor in the cytoplasm through tyrosine phosphorylation and thus leads to activation of the promoters of the c-fos gene and other genes involved in cellular proliferation.