Box 3-independent Signaling Mechanisms Are Involved in Leukemia Inhibitory Factor Receptor α- and gp130-mediated Stimulation of Mitogen-activated Protein Kinase

Chimeric receptors containing the entire or various cytoplasmic domains of either gp130 or leukemia inhibitory factor receptor α (LIFR) were used to identify signaling molecules and regions of these polypeptides required for the stimulation of mitogen-activated protein kinase (MAPK). Coexpression of dominant-negative Jak2 inhibited chimeric receptor-stimulated MAPK activity by ∼70%, while expression of dominant-negative Ras completely blocked MAPK activation by either receptor polypeptide. Deletion analysis identified a 24-amino acid region of gp130 that was necessary for maximal stimulation of MAPK, and contained box 3 (positions 120–129) and a consensus tyrosine binding motif (Tyr-118) for the protein-tyrosine phosphatase, SHP2. Expression of receptors lacking this region or of chimeric gp130(Y118F) point mutants inhibited MAPK activity by ∼55%, suggesting that Tyr-118, but not box 3, was required during activation of MAPK by gp130. Similarly, expression of chimeric LIFR constructs lacking box 3 maximally stimulated MAPK activity, while those lacking Tyr-115, a putative SHP2 binding site, inhibited stimulation of MAPK by this polypeptide. Our results demonstrate that gp130 and LIFR stimulate MAPK activity through box 3-independent mechanisms involving: (i) effects at Tyr-118 and Tyr-115, respectively, for maximal stimulation of MAPK activity and (ii) a Jak/Tyk-dependent pathway that, together with Tyr-118- or Tyr-115-generated signals, converges at the level of Ras during activation of MAPK by cytokine.

LIF 1 is a multifunctional cytokine, which elicits a variety of biochemical, physiological, and morphological alterations in cellular homeostasis, including: (i) murine M1 leukemic cell differentiation; (ii) the conversion of sympathethic neurons from noradrenergic to cholinergic phenotypes; (iii) embryonic stem cell, DA-1a myeloid cell, megakaryocyte, and myoblast proliferation; (iv) acute phase protein synthesis in hepatocytes; (v) elevation of circulating platelet levels; and (vi) inhibition of adipogenesis and lipoprotein lipase activity in mature adipocytes (Refs. 1 and 2, and references therein). LIF and its structural relatives ciliary neurotrophic factor, oncostatin M, and the interleukins 6 and 11 constitute a distinct subgroup of cytokines (3) whose members bind to one or more low affinity receptor binding subunits, which associate with the shared receptor subunit gp130 to initiate transmembrane signaling (1,2). Intracellular signaling pathways responsible for LIF action utilize the activation of both Tyr and Ser/Thr protein kinases in mediating cytokine-stimulated gene induction (4,5) and murine M1 leukemic cell differentiation (5). With respect to stimulation of protein-tyrosine kinases, both subunits of activated heterodimeric LIF receptor complexes, namely LIFR and gp130, have been shown to associate with and stimulate all known members of the Jak/Tyk family of nonreceptor proteintyrosine kinases (6). Indeed, stimulation of Jak/Tyk proteintyrosine kinase activity results in the rapid tyrosine phosphorylation and activation of the latent cytosolic transcription factors, STAT1, STAT3, and STAT5 (7,8), and the initiation of gene transcription at sis-inducible elements in some LIF-responsive genes (8 -10). The role of protein-tyrosine kinase activation during LIF receptor signaling is further complicated by the findings that stimulation of LIF receptor and/or gp130 has been shown to activate Lyn, Fyn, Hck, Btk, Tec, Fes, and Yes nonreceptor protein-tyrosine kinases (11)(12)(13)(14)(15).
Activated LIF receptors stimulate several components of the mitogen-activated protein kinase cascade (16 -18), including MAPK kinase, the MAPK isozymes ERK1 and ERK2, and S6 protein kinase activities against both the synthetic S6 peptide (pp90 rsk ) and 40S ribosomes (pp70 S6K ), through activation of PKC-dependent and -independent pathways (16). While a number of studies have clearly identified regions termed boxes 1, 2, and 3 within the cytoplasmic domains of LIFR and gp130 required for STAT activation and the initiation of gene induction, no data currently exist describing regions or residues within these polypeptides responsible for eliciting the stimulation of MAPK following cytokine administration. To address this question, we transiently transfected Cos-7 cells with chimeric receptors containing the entire or various truncated cytoplasmic domains of either LIFR or gp130, and measured their ability to stimulate MAPK. Our data indicate that gp130 and LIFR stimulate the activation of MAPK through a box 3-independent mechanism involving effects at Tyr-118 and Tyr-115, respectively, for maximal stimulation of MAPK activity by cytokine. Both receptor subunits use a Jak/Tyk-dependent * This work was supported by National Institutes of Health Training Grants GM07750 (to W. P. S) and GM07018 (to J. L. B.), by National Institutes of Health Grant NS30410 (to N. M. N), and by a grant from the Muscular Dystrophy Assocation (to N. M. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Materials-G-CSF and the chimeric G-CSFR-gp130 and G-CSFR-LIFR cDNA constructs were generously provided by Bruce Mosley (Immunex Corporation, Seattle, WA). The cDNA constructs encoding HA-tagged hamster ERK1 (HA-ERK1) and dominant-negative Jak2 (DNJak2) were kindly provided by Drs. Henry R. Bourne (University of California, San Francisco) and Andrew S. Kraft (Veterans Administration Medical Center, Birmingham, AL), respectively. The cDNA encoding dominant-negative Ras (DNRas), used previously by this laboratory as 17N-Ras (19) but which is now known to contain Ala at position 17 and is still functionally dominant-negative (20), was provided by Dr. Larry A. Feig (Tufts University). Polyclonal anti-G-CSFR antibodies were provided by Dr. David J. Tweardy (University of Pittsburgh). The monoclonal anti-phosphotyrosine antibodies PY20 and 4G10 were purchased from ICN Pharmaceuticals, Inc. (Costa Mesa, CA) and Upstate Biotechnology, Inc. (Lake Placid, NY), respectively, while protein Gagarose and the monoclonal anti-hemagglutinin antibody 12CA5 were obtained from Boehringer Mannheim. Polyclonal anti-gp130 and anti-LIFR antibodies were produced by immunizing rabbits with glutathione S-transferase fusion proteins containing the cytoplasmic amino acids 740 -896 or 1005-1063 of human gp130 or human LIFR, respectively. EGF was purchased from Upstate Biotechnology Inc., while MBP was obtained from Sigma. DMEM, FBS, and penicillin-streptomycin were obtained from Life Technologies, Inc. All additional supplies or materials were routinely available.
Cell Culture and Transient Transfection-Cos-7 cells were plated onto 10-cm dishes and cultured in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin, and were maintained in a constant atmosphere of 10% CO 2 at 37°C. Upon reaching ϳ60 -70% confluence, the cells were transiently transfected by the calcium phosphate/chloroquine method (22) with 30 g of total cDNA consisting of 10, 5, and 15 g of the cDNAs encoding chimeric receptor, HA-ERK1, and carrier cDNA (empty pDC302 vector), respectively. In some experiments, the effects of coexpression of either DNRas or DNJak2 on chimeric receptorstimulated HA-ERK1 activity was tested by transfection of Cos-7 cells as above in the absence (empty vector cDNA) or presence of either 20 g of DNRas cDNA (35 g of total cDNA) or 30 g of DNJak2 cDNA (40 g of total cDNA with 5 g each of receptor and HA-ERK1 cDNA). Twentyfour h post-transfection, the cells were subcultured onto six-well plates, and ϳ2.5 h after plating were rendered quiescent by overnight incubation at 37°C in serum-free DMEM. Preliminary experiments designed to optimize measurable MAPK activity stimulated by either gp130 or LIFR established that transient transfection of Cos-7 cells with 10 g of chimeric receptor cDNA and 5 g of HA-ERK1 cDNA produced the greatest signal to noise ratios when assayed 48 h after transfection (data not shown).
HA-ERK1 Immunocomplex Assays-Stimulation of quiescent transfected Cos-7 cells was performed at 37°C in 2 ml of serum-free DMEM containing the appropriate diluent (phosphate-buffered saline) or drug additions as indicated. After 0 -60 min, the cells were washed twice in ice-cold phosphate-buffered saline (3 ml/rinse) and were immediately lysed by addition of 500 l of ice-cold buffer H (16) supplemented with 1% Triton X-100. Cell extracts were solubilized on ice for 60 min and subsequently clarified by centrifugation for 15 min at 4°C. Immunoprecipitation of HA-ERK1 was accomplished by rotating the clarified supernatants in the presence of 5 g of anti-hemagglutinin 12CA5 antibody and 30 l of protein G-agarose for ϳ2 h at 4°C. Immunocomplexes were recovered by brief centrifugation and were subsequently washed (500 l/wash) two times in ice-cold lysis buffer, followed by two washes in ice-cold buffer H. Analysis of immunoprecipitated HA-ERK1 phosphotransferase activity was performed in a final assay volume of 40 l containing 0.4 mg/ml MBP and 25 l of buffer H resuspended immunocomplexes. Reactions, initiated by addition of 10 l of 4 ϫ assay buffer (buffer AB: final concentrations of 25 mM ␤-glycerophosphate, 0.5 mM dithiothreitol, 1.25 mM EGTA, 50 M sodium vanadate, 2 M PKI, 10 M calmidazolium, 10 mM MgCl 2 , and 100 M ATP ([␥-32 P]ATP, ϳ2000 cpm/pmol)), were allowed to proceed for 30 min at 30°C and were subsequently terminated by addition of 4 l of 88% formic acid. Reaction mixtures were clarified by brief centrifugation, and were subsequently spotted onto P-81 phosphocellulose paper (Whatman). Following six washes in 125 mM phosphoric acid, the dried papers were subjected to scintillation counting to determine the specific phosphorylation of MBP, which was calculated by subtracting the radioactivity measured in immunocomplexes obtained from mock-transfected cells (carrier cDNA) from those obtained from HA-ERK1-transfected cells.
To control for differences in transfection efficiency, paired basal and G-CSF-stimulated HA-ERK1 activities were corrected against the MAPK activities stimulated through activation of endogenous EGF receptors present on identically transfected Cos-7 cells; cells were stimulated for 5 min at 37°C with 100 ng/ml EGF. Data are presented as the mean (Ϯ S.E.) activity ratios relative to EGF of basal and G-CSFstimulated MBP phosphorylation. In some experiments, the activity ratios were normalized to HA-ERK1 phosphotransferase activity stimulated by the relevant full-length chimeric receptor.
Immunoprecipitation and Western Blotting of Transiently Expressed Chimeric Receptors-Following initial recovery of HA-ERK1 immunocomplexes by brief centrifugation, selected supernatants were subjected to further immunoprecipitation overnight at 4°C following their transfer to fresh microcentrifuge tubes containing 30 l of protein G-agarose and 3 l of either anti-G-CSFR, -gp130, or -LIFR␣ antibodies as indicated. The resulting immunocomplexes were recovered by brief centrifugation, washed twice in lysis buffer and once in buffer H (500 l/rinse), and following their fractionation through 7% SDS-polyacrylamide gel electrophoresis, proteins Ͼ50 kDa were transferred electrophoretically to Immobilon-P (Millipore). After blockade of nonspecific binding sites by incubation in Tris-buffered saline (10 mM Tris, 150 mM NaCl, pH 8.0) supplemented with 5% bovine serum albumin, the immunoblots were probed initially with a mixture of anti-phosphotyrosine antibodies 4G10 and PY20 (1/2000 dilution each), followed by stripping and subsequent reprobing of the immunoblot with the relevant anti-receptor antibodies (1/750 dilution) as indicated. Proteins were visualized with enhanced chemiluminescence according to the manufacturer's recommendations (Renaissance, NEN Life Science Products).

RESULTS AND DISCUSSION
The results of several recent studies of LIFR and gp130 have identified a number of cytoplasmic motifs and/or signaling modules that are required for the activation of STAT1, STAT3, and STAT5, and the initiation of gene induction (8,10,21,(23)(24)(25)(26), but there are no reports describing specific regions or amino acid residues within these polypeptides required for cytokine-mediated activation of MAPK activity. There are also no data on the potential roles for either Jak/Tyk protein-tyrosine kinases or Ras during activation of MAPK by these receptor polypeptides. To address these questions, we transiently transfected into Cos-7 cells chimeric cytokine receptors composed of the ligand binding domain of the G-CSFR linked to the entire or various truncated cytoplasmic domains of either gp130 or LIFR, and measured their ability to stimulate the MBP phosphotransferase activity of coexpressed epitopetagged HA-ERK1 (MAPK). Treatment of transfected Cos-7 cells with G-CSF stimulated the MBP phosphotransferase activity contained in HA-ERK1 immunocomplexes obtained from chimeric gp130 or LIFR transfectants in a time-and dose-dependent manner (Fig. 1). Control experiments on Cos-7 transfectants expressing only epitope-tagged MAPK (i.e. no chimeric receptors) demonstrated that G-CSF treatment stimulated MAPK activity to 16.3 Ϯ 6.8% (n ϭ 5, S.D.) of the levels stimulated in Cos-7 cells expressing chimeric gp130 receptors. In response to saturating concentrations of G-CSF, chimeric receptor-mediated activation of MAPK activity was rapid, reaching peak levels at 10 min, and thereafter declining to basal levels by 60 min (Fig. 1A). The magnitude of MAPK activation stimulated by chimeric LIFR was typically 70 -80% of that attained by stimulation of chimeric gp130 (Fig. 1B, data  not shown). Although the reason(s) for this difference is currently unknown, its origins may arise from differences in the efficiency of dimerization induced by G-CSF to produce homodimers of gp130 or LIFR. While homodimers of LIFR are not currently known to exist naturally, signaling through chimeric LIFR is physiologically relevant because: (i) LIFR and gp130 exhibit significant homology throughout their cytoplasmic domains (27), (ii) LIFR substitutes for a second molecule of gp130 during initiation of endogenous LIF receptor signaling (28), (iii) gene induction mediated by chimeric LIFR is indistinguishable from that observed with gp130 (21,25), and (iv) the activation and inactivation kinetics for MAPK following stimulation of chimeric LIFR (and chimeric gp130) are indistinguishable from those stimulated by the endogenous LIF receptors on the surface of 3T3-L1 cells (16,18,29), and on the surface of HeLa, HepG2, WI-26-VA4, IMTLH, SK-HEP-1, and JAR cells (17).
Having shown that chimeric gp130 and LIFR can stimulate MAPK activity in a manner similar to that of endogenous LIF receptors, we tested whether or not Jak/Tyk protein-tyrosine activity was required during activation of MAPK by these receptor polypeptides by determining the effect of overexpression of kinase-inactive DNJak2 on chimeric receptor-stimulated MAPK activity. Coexpression of DNJak was found to: (i) inhibit chimeric receptor-stimulated MAPK activity by ϳ70% ( Fig. 2A), (ii) attenuate cytokine-stimulated tyrosine phosphorylation of these receptor polypeptides (Fig. 2B), and (iii) have no effect on chimeric receptor expression levels (Fig. 2C). These effects are specific for the gp130 and LIFR polypeptides, as expression of DNJak failed to alter the magnitude of MAPK activity stimulated through activation of the endogenous EGF receptors present on identically transfected Cos-7 cells (Fig.  2A). These results are the first to show a link between Jak/Tyk protein-tyrosine kinase activation and the ability of the gp130 and LIFR polypeptides to stimulate MAPK activity. Additionally, the Jak/Tyk-stimulated MAPK pathway activated by either gp130 or LIFR must proceed through Ras, as coexpression of DNRas was found to eliminate completely cytokine-stimulated MAPK activation by these polypeptides (Fig. 2D).
Having established the necessity for stimulation of Jak/Tyk protein-tyrosine kinase and Ras activities during gp130-and LIFR-mediated activation of MAPK, we sought to identify cytoplasmic regions within gp130 and LIFR that are required for the stimulation of MAPK. Expression of chimeric gp130 constructs containing progressive C-terminal truncations identified a 24-amino acid region located between amino acids 109 -133 that was necessary for mediating maximal stimulation of MAPK. Deletion of this region by expression of gp130 constructs truncated to position 109 or beyond significantly reduced cytokine-stimulated MAPK activity by ϳ60 -70% of fulllength gp130-stimulated activities (Fig. 3A). Previous work demonstrated that these chimeric receptor constructs are all expressed at similar levels (21), and Western blotting of immunoprecipitated chimeric gp130 receptors with a combination of anti-G-CSFR and -gp130 antibodies confirmed that the various chimeric gp130 receptor deletion constructs were expressed at similar levels in these experiments (data not shown). Thus, the loss of cytokine-stimulated MAPK activity is not due to differences in receptor expression, but rather to the loss of cytoplasmic residues required for maximal activation of MAPK by gp130. The 24-amino acid region between residues 109 and 133 of gp130 contains two interesting features: (i) the conserved cytokine signaling motif, box 3 (amino acids 120 -129), and (ii) a YXXV tyrosine-containing module located at position 118 that has previously been shown to be necessary for gp130mediated activation of the protein-tyrosine phosphatase SHP2 (24). Because SHP2 is known to function as an upstream regulator of MAPK activity in a variety of systems (30), we mutated Tyr-118 to Phe and determined its effects on gp130mediated activation of MAPK. Full-length chimeric gp130(Y118F) mutants were transiently expressed and tyrosine-phosphorylated in response to agonist in a manner indistinguishable from their wild-type counterparts (Fig. 3B); however, cytokine-stimulated MAPK activity was reduced (by ϳ55%), but not eliminated, in cells expressing gp130(Y118F) mutants (Fig. 3A). The finding that chimeric gp130(109) and gp130(Y118F) exhibited identical phenotypes with respect to stimulation of MAPK suggests that gp130-mediated activation of MAPK occurs independently of box 3, and requires effects at Tyr-118, most likely via tyrosine phosphorylation and subsequent stimulation of SHP2, for maximal activation of MAPK activity by the receptor polypeptide.
Similar to gp130, expression of chimeric LIFR constructs lacking box 3 (amino acids 136 -145) retained their ability to maximally stimulate MAPK (Fig. 3C). Previous work has demonstrated that these chimeric receptor constructs are expressed at similar levels (21), and Western blot analyses confirmed that they were expressed here at concentrations comparable to wildtype expression levels (data not shown). Truncation of chimeric LIFR from position 121 to 111 significantly reduced cytokinestimulated MAPK activity (Fig. 3C), suggesting that this 10amino acid region is required for maximal stimulation of MAPK by LIFR. Interestingly, this region also contains a putative SHP2 tyrosine binding site located at position Tyr-115. Although attempts to demonstrate LIFR-mediated activation of SHP2 have been unsuccessful (24), expression of full-length chimeric LIFR(Y115F) mutants was found to attenuate (by ϳ53%), but not eliminate, cytokine-stimulated MAPK activity (Fig. 3C). Similar to G-CSFR-gp130(Y118F), G-CSFR-LIFR(Y115F) mutants were expressed and tyrosine phosphorylated in response to agonist in a manner indistinguishable from their wild-type counterparts (Fig. 3D), suggesting that LIFR-mediated activation of MAPK is also independent of box 3 and occurs through effects at the putative SHP2 consensus binding site, Tyr-115.
The results presented here show that gp130 and LIFR stim-ulate the activation of MAPK through: (i) Jak/Tyk proteintyrosine kinase-and Ras-dependent pathways and (ii) box 3-independent mechanisms involving effects at Tyr-118 and Tyr-115, respectively, for maximal stimulation of MAPK activity by cytokine. Previous work by Stahl and colleagues (24) showed that tyrosine phosphorylation of SHP2 stimulated by gp130 occurred through the tyrosine-based YXXV motif located at position 118 of the receptor polypeptide. Their results, coupled with those in the current paper, suggest that tyrosine phosphorylation at position 118 of gp130 leads to the phosphorylation and activation of SHP2 and the subsequent activation of MAPK. Although they were unable to show LIFR-stimulated tyrosine phosphorylation of SHP2 (24), we find that mutation of the putative SHP2 binding motif (i.e. Tyr-115) within LIFR was capable of significantly inhibiting MAPK activation in response to cytokine, suggesting that tyrosine phosphorylation of Tyr-115 within LIFR, like Tyr-118 within gp130, elicits stimulation of MAPK through the activation of SHP2. Although the reason(s) for the discrepancy between their findings and those in the present study remain to be determined fully, we have found recently that coexpression of dominant-negative SHP2 blocks ϳ70% of chimeric gp130-and LIFR-stimulated MAPK activity in Cos-7 cells (31). Thus, it is likely that stim- ulation of MAPK by gp130 and LIFR is mediated in part through activation of SHP2.
In light of this result, we were surprised to find that expression of the G-CSFR-gp130(Y118F) and G-CSFR-LIFR(Y115F) mutants failed to completely eliminate the ability of these receptor polypeptides to stimulate MAPK activity. Because cytokine-stimulated Jak/Tyk activity would be expected to lie upstream of Tyr-118 and Tyr-115 phosphorylation, expression of chimeric gp130(Y118F) or LIFR(Y115F) constructs would be predicted to produce inhibitions of cytokine-stimulated MAPK activity equaling or surpassing those produced by expression of DNJak. The finding that gp130(Y118F) and LIFR(Y115F) mutants signal more effectively to MAPK than cells expressing wild-type receptors with DNJak, suggests the presence of an additional membrane proximal signaling pathway(s) that, together with effects at Tyr-118 and Tyr-115 within gp130 and LIFR, respectively, converge at the level of Ras for maximal activation of MAPK by these polypeptides.
The bifurcation in gp130 and LIFR signaling may arise at the level of the substrates for the Jak/Tyk protein-tyrosine kinases themselves. Indeed, like the receptors for growth hormone (32, 33), interferons-␣ and -␥ (33, 34), and the interleukins-2, -4, -7, and -15 (35), activated LIF receptors have been reported to stimulate the tyrosine phoshorylation and association of the insulin receptor substrate-1 (IRS-1) with Jak2 (33). This kinase-IRS-1 association would have the potential of permitting activated LIF receptors to greatly amplify their intracellular signals by increasing both the quantity and the diversity of signaling molecules brought into the proximity of its activated nonreceptor protein-tyrosine kinases. Alternatively, the receptor for interleukin-11, a relative of LIF which also uses gp130 in its activated signaling complex (1,2), has recently been shown to induce the formation of complexes comprising Grb2, Fyn, and Jak2 following agonist administration (36). By analogy, the finding that gp130(Y118F) and LIFR(Y115F) mutants retain the ability to stimulate, albeit to submaximal levels, the activation of MAPK activity could be explained if these polypeptides could form complexes that were comprised of Jak2-IRS-1, Jak2-Grb2-Fyn, or both, and were able to bypass the block produced by removal of SHP2 binding sites within gp130 and LIFR (i.e. still maintain stimulation of SHP2 via Jak2-generated complex formation).
Finally, the results of several recent studies have established that C-terminal Ser phosphorylation of STAT1 and STAT3 by a proline-directed protein kinase, presumably MAPK, is capable of augmenting gene induction mediated by these transcription 100 ng/ml), and were subsequently assayed for cytokine-stimulated MBP phosphotransferase activity in HA-ERK1 immunocomplexes as described under "Experimental Procedures." To control for differences in transfection efficiency, paired basal and G-CSF-stimulated HA-ERK1 activities were corrected against the MAPK activities stimulated through activation of endogenous EGF receptor (100 ng/ml for 5 min at 37°C) present on identically transfected Cos-7 cells. Data are the mean (Ϯ S.E.) activity ratios relative to EGF of basal and G-CSF-stimulated HA-ERK1 phosphotransferase activities normalized to relevant wild-type stimulations of 3-12 separate experiments. B, chimeric G-CSFR-gp130(FL) (1), G-CSFR-gp130(Y118F/FL) (2), or mock-transfected cells (3), or D, G-CSFR-LIFR(FL) (1), G-CSFR-LIFR(Y115F/FL) (2), or mock-transfected cells were immunoprecipitated with anti-gp130 or -LIFR antibodies, respectively, as described under "Experimental Procedures." Immunoblots were probed sequentially with a mixture of anti-phosphotyrosine antibodies, followed by reprobing with the relevant receptor antibodies as indicated. Proteins were subsequently visualized by enhanced chemiluminescence.
factors (9,(37)(38)(39)(40). Expression of G-CSFR-gp130(Y118F) mutants, which inhibit gp130-stimulated activation of MAPK by ϳ55%, did not decrease cytokine-stimulated expression of either vasoactive intestinal peptide-luciferase or c-Fos-luciferase reporter genes. 2 This lack of effect on gene induction could occur if the residual stimulation of MAPK activity by Tyr-118deficient gp130 constructs were sufficient for maximal Ser phosphorylation of STAT1 and STAT3 proteins. Alternatively, Ser phosphorylation of STAT1 and STAT3 proteins might be mediated by a proline-directed protein kinase distinct from MAPK, a possibility supported by our earlier findings in IMR-32 cells demonstrating that induction of VIP and c-Fos by CNTF and LIF were dissociated from stimulation of Ras-coupled pathways (19).
In summary, the results of this study support our earlier findings demonstrating that LIF-mediated activation of MAPK occurs through a bifurcated signaling system involving the integration of multiple signaling inputs, one of which is dependent upon the presence of PKC (16). In addition to our recent results describing the ability of activated LIF receptors to stimulate MAPK activity through both Raf-1-dependent and -independent mechanisms, 3 we now show that maximal stimulation of MAPK activity by gp130 and LIFR is mediated through box 3-independent mechanisms involving: (i) effects at Tyr-118 and Tyr-115, respectively, for maximal stimulation of MAPK activity; and (ii) a Jak/Tyk-dependent pathway that, together with Tyr-118-or Tyr-115-generated signals, converges at the level of Ras during activation of MAPK by cytokine.