Ras-dependent activation of c-Jun N-terminal kinase/stress-activated protein kinase in response to interleukin-3 stimulation in hematopoietic BaF3 cells.

Activation of the c-Jun N-terminal kinase (JNK)/stress-activated protein kinase pathway in response to stimulation of the interleukin (IL)-3 or granulocyte-macrophage colony-stimulating factor (GM-CSF) receptor was examined in mouse hematopoietic BaF3-derived cell lines (BaF3-N6 and -V2 cells). Significant increase in the activity of JNK1 was observed within 30 min following IL-3 or GM-CSF stimulation at physiological concentrations. Dominant-negative Ras(S17N), which is conditionally expressed in the presence of isopropyl-1-thio-β-D-galactoside in BaF3-N6 cells, prevented the IL-3 stimulation of JNK1, whereas anisomycin-induced JNK1 activation was unaffected. Furthermore, a deletion mutant of the common β subunit for IL-3 and GM-CSF receptors that consists of only the membrane-proximal region, including box 1 and box 2 motifs, was incapable of facilitating JNK1 activity as well as Ras activation. These results provide evidence that Ras is required for IL-3-stimulated JNK1 activation. We also examined if constitutively active Ras(G12V) alone could stimulate JNK1 activity by using the inducible expression system. Isopropyl-1-thio-β-D-galactoside induction of Ras(G12V) in the BaF3-V2 cell line caused no significant increase in JNK1 activity, which could be activated by IL-3 or anisomycin. On the contrary, the extracellular signal-regulated kinase/mitogen-activated protein kinase pathway was fully activated following Ras(G12V) induction. Together with these results, it seems likely that the Ras protein is indispensable for the IL-3 stimulation of JNK1 although Ras activation by itself is insufficient for JNK1 activation.

Ras-mediated growth promotion in fibroblast cells (3,4). Activation of the ERK/MAPK pathway ultimately results in transcriptional activation of a serum response element-containing promoter, for example, the c-fos promoter, through phosphorylation of Elk-1 (5).
Stimulation with various cytokines, including interleukin (IL)-2, IL-3, IL-5, and granulocyte-macrophage colony-stimulating factor (GM-CSF), triggers Ras activation as evidenced by accumulation of an active GTP-bound form, implying that Ras acts as a signal transducer also in hematopoietic cells (6,7). In the IL-3 signaling system, Raf and ERK/MAPK activation as well as induction of the c-fos promoter are demonstrated to be Ras-dependent (8). In addition to the Ras pathway, several distinct signaling pathways, for instance, the Janus kinase (JAK)/signal transducers and activators of transcription-mediated pathway, are stimulated by the IL-3 receptor (9).
JNK/SAPK is phosphorylated and subsequently activated by a specific kinase, JNKK1/SEK1/MKK4, which is also activated following phosphorylation by other protein kinases, such as MEKK1 (5), Tpl-2 (15), and MLK-3 (16). Although the precise regulatory mechanism of this kinase cascade is still obscure, involvement of Cdc42 and Rac GTP-binding proteins in regulation of the JNK/SAPK pathway has recently been demonstrated using a transfection assay (17)(18)(19). Furthermore, a significant role of a mammalian homologue of the Saccharomyces cerevisiae Ste20 protein, p21-activated kinase, which is directly activated by a GTP-bound form of Cdc42 or Rac (20), has been suggested (18,(21)(22)(23). Another Ste20 homologue, germinal center kinase, is also a candidate for an activator of the JNK/SAPK pathway (24).
In this paper, we show activation of JNK1 following IL-3 treatment of IL-3-dependent hematopoietic cell lines. Conditionally expressed dominant-negative Ras abrogated the activation. Moreover, a deletion mutant of the IL-3 receptor ␤ c subunit that is unable to stimulate the Ras pathway failed to cause JNK1 activation. On the other hand, a constitutively active mutant of Ras was unable to induce JNK1 activation by itself despite its full activity to stimulate ERK2. Thus, we conclude that Ras is required, but not sufficient, for JNK1 activation in IL-3-dependent hematopoietic cells.

Enhancement of JNK1 Activity in Response to IL-3 Stimulation in the BaF3-N6 Hematopoietic Cell
Line-As a first step to examine possible involvement of JNK/SAPK in IL-3-stimulated signal transduction pathways, we measured the activity of endogenous JNK1 in IL-3-stimulated as well as unstimulated BaF3-N6 cells. The BaF3-N6 cell line was isolated from mouse IL-3-dependent hematopoietic BaF3 cells as a stable transfectant of an inducible expression system for dominant-negative Ras(S17N) (8). Serum and IL-3-starved BaF3-N6 cells were stimulated with various concentrations of mouse IL-3 for 40 min. Endogenous JNK1 was collected by immunoprecipitation using a specific antibody, and the activity was measured by an in vitro kinase assay using GST-c-Jun(1-223) as a substrate. As shown in Fig. 1, A and B, dose-dependent increase of JNK1 activity was observed, which paralleled the induction of cell growth as measured by a colorimetric assay using 3-(4,5-di-methylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (8). Incubation of the cells without mouse IL-3 for 40 min exhibited no effect on JNK1 activity, indicating that the increased kinase activity was not due to nonspecific stress during the incubation (data not shown). Time course of JNK1 activation following IL-3 treatment is demonstrated in Fig. 1, C and D. An increase in JNK1 activity was measurable at 20 min after stimulation. The activity reached a maximum at 40 min, and remained at plateau for at least another 20 min. Increase of JNK1 activity was rather slower than the case of ERK2 (Fig. 1, E and F). The activity of immunoprecipitated ERK2 was measured by an in vitro kinase assay using myelin basic protein as a substrate, where the activation was already detectable after 5 min stimulation. The observation is well correlated with time course of was immunoprecipitated by a specific antibody (sc-474) from the cell lysate followed by an in vitro kinase assay using GST-c-Jun(1-223) as a substrate. B, quantitation of the data in A. Radioactivity was quantitated by an image analyzer and relative JNK1 activity was shown as mean Ϯ S.E.(n ϭ 3). C, time-dependent phosphorylation of GST-c-Jun(1-223) in response to mouse IL-3. Serum and IL-3-deprived BaF3-N6 cells were stimulated with mouse IL-3 (50 ng/ml) for specified periods. JNK1 was immunoprecipitated by a specific antibody (sc-474) from the cell lysate followed by an in vitro kinase assay using GST-c-Jun(1-223) as a substrate. D, quantitation of the data in C. Radioactivity was quantitated by an image analyzer, and relative JNK1 activity was shown as mean Ϯ S.E. (n ϭ 3). E, time-dependent phosphorylation of myelin basic protein (MBP) in response to mouse IL-3. Serum and IL-3-deprived BaF3-N6 cells were stimulated with mouse IL-3 (50 ng/ ml) for specified periods. ERK2 was immunoprecipitated by a specific antibody (sc-154) from the cell lysate followed by an in vitro kinase assay using myelin basic protein as a substrate. F, quantitation of the data in E. Radioactivity was quantitated by an image analyzer, and relative ERK2 activity was shown as mean Ϯ S.E. (n ϭ 3). hyperphosphorylation described previously (8). Moreover, we observed similar delayed activation of JNK1 compared with ERK2 in another BaF3-derived stable clone. 2 IL-3 Receptor-mediated JNK/SAPK Activation Is Ras-dependent-Since the Ras protein seems to play an important role for regulation of JNK/SAPK activity in a variety of signal transduction pathways, including those initiated by stimulation with epidermal growth factor, nerve growth factor, and G protein-coupled receptor ligands (14,17,18,25), we next assessed the role of Ras in IL-3-stimulated JNK1 activation. As described previously (8), dominant-negative Ras(S17N) was induced in BaF3-N6 cells by incubation with isopropyl-1-thio-␤-D-galactoside (IPTG) for 16 h. The IL-3 stimulation of JNK1 activity was found to be diminished considerably after induction of Ras(S17N), whereas JNK1 activation in response to a translational inhibitor, anisomycin, which has been shown to be Ras-independent in COS7 cells (14), remained unaffected ( Fig. 2A). Likewise, ERK2 stimulation by IL-3 was inhibited by Ras(S17N) (Fig. 2B), which is consistent with our previous results that Ras(S17N) inhibited c-Raf-1 activation as well as ERK2 hyperphosphorylation. These results suggest that both JNK1 and ERK2 pathways are regulated by Ras. Further examination of the assumption that the JNK1 pathway is dependent on Ras was performed by using a mutant receptor for GM-CSF. It has been shown that human GM-CSF is capable of stimulating proliferation of a BaF3-derived transfectant of ␣ and ␤ c subunits of the human GM-CSF receptor (26). ␤ 544 is a mutant of the ␤ c subunit, which is composed of intact extracel-lular and transmembrane domains and a C-terminal truncated cytoplasmic domain (a membrane-proximal region including box 1 and box 2 motifs) (26). The ␤ 544 receptor is incapable of activating the Ras pathway as determined by a lack of Ras⅐GTP formation and impaired increase in ERK2 activity, whereas human GM-CSF can stimulate the Ras pathway to an extent similar to mouse IL-3 in BaF3/␣␤ c cells (Ref. 27 and data not shown). In contrast, the ␤ 544 receptor is considered intact in terms of stimulation of the JAK/signal transducers and activators of transcription pathway because human GM-CSF-induced tyrosine phosphorylation of JAK2 was detected also in BaF3/ ␣␤ 544 cells (Fig. 3A, see also Refs. 28 and 29). Using these transfectants, we evaluated the role of the Ras pathway in JNK1 activation. As demonstrated in Fig. 3B, the ␤ c receptor was fully active in promoting the JNK1 pathway upon human GM-CSF stimulation in contrast to the ␤ 544 receptor, which failed to activate JNK1. Mouse IL-3 could induce JNK1 activation similarly in both cell lines. Taken together with the above results, it is likely that Ras is indispensable for stimulation of the JNK1 pathway.
Constitutively Active Ras(G12V) Is Insufficient to Induce JNK1 Activation-BaF3-V2 cells grow in an IL-3-dependent manner, and produce constitutively active Ras(G12V) following IPTG induction (8). The induced Ras(G12V) can stimulate c-Raf-1 kinase activity, ERK2 hyperphosphorylation, as well as transcriptional activation of the c-fos promoter (Ref. 8 and data not shown). To examine the effect of Ras(G12V) expression on the induction of JNK1 activation, we first measured JNK1 activation in response to mouse IL-3 and anisomycin. Both IL-3 and anisomycin induced JNK1 activation to a similar extent as in the case of BaF3-N6 cells ( Figs. 2A and 4A). Time course of the activation also was similar to BaF3-N6 cells as illustrated in Figs. 1D and 4B. Next, the activity of ERK2 and JNK1 was quantitated in parallel, and the results are described in Fig.  4C. No JNK1 activation was observed in response to IPTG induction of Ras(G12V), while ERK2 activity was markedly 2 K. Tago, Y. Kaziro, and T. Satoh, unpublished results.

FIG. 2. Inhibition of IL-3-induced JNK1 and ERK2 activation by Ras(S17N) in BaF3-N6 cells. A, inhibition of IL-3-induced JNK1
activation by Ras(S17N). BaF3-N6 cells were incubated with or without IPTG (5 mM) for 16 h. During the last 3 h, the cells were cultured in medium deprived of serum and IL-3. The cells were stimulated with mouse IL-3 (50 ng/ml) or anisomycin (20 g/ml) for the indicated periods, and the activity of immunoprecipitated JNK1 was measured. Data are shown as mean Ϯ S.E. (n ϭ 3). B, inhibition of IL-3-induced ERK2 activation by Ras(S17N). BaF3-N6 cells were incubated with or without IPTG (5 mM) for 16 h. During the last 3 h, the cells were cultured in medium deprived of serum and IL-3. The cells were stimulated with mouse IL-3 (50 ng/ml) or phorbol 12-myristate 13-acetate (PMA) (200 ng/ml) for the indicated periods, and the activity of immunoprecipitated ERK2 was measured. Data are shown as mean Ϯ S.E. (n ϭ 3). augmented in accordance with Ras(G12V) accumulation. The results suggest that Ras(G12V) by itself is insufficient for activation of JNK1, where an additional signal may be required.

DISCUSSION
It has been demonstrated that a variety of cytokines, including IL-2, IL-3, IL-5, GM-CSF, and erythropoietin can trigger Ras activation in hematopoietic cell lines (6,7). Although the results suggest a significant role of Ras in cytokine-mediated intracellular signal transduction, it still remains to be solved whether Ras is implicated in growth-promoting or differentiation-inducing signaling pathways. In IL-3 signal transduction, analyses using a series of deletion mutants of the receptor ␤ c subunit have shown that the Ras pathway is responsible for prevention of apoptotic cell death (30). Moreover, conditional expression of a dominant-negative Ras mutant exhibited no inhibitory effect on IL-3-dependent proliferation of BaF3 cells (8). Likewise, the Ras pathway does not seem to be necessary for IL-2 induction of cell proliferation (31)(32)(33). On the other hand, in another hematopoietic cell line 32Dcl3, Ras was required for cell growth in response to IL-3, but not for granulocyte colony-stimulating factor-induced differentiation (34). As for granulocyte colony-stimulating factor receptor-mediated signaling, the Ras pathway was found to be associated with the proliferative response in myeloid leukemia cell lines, such as NFS-60 (35). Taken together, the role of Ras in cells of the hematopoietic lineage is still controversial, and thus, further investigation of Ras-regulated signaling pathways is required for a final conclusion.
In addition to the Ras protein, members of the Cdc42/Rac/ Rho GTP-binding protein family have recently been shown to be important for cellular responses, particularly for the regulation of cytoskeletal organization, in various kinds of mammalian cells (36 -39). They are also implicated in T-cell polarization (40), cell cycle progression (19), cell motility (41), and morphogenesis of neuronal cells (42). Furthermore, their dominant-negative-type mutants prevent Ras-induced transformation of fibroblast cells, suggesting a significant role at a point downstream of Ras (43,44). However, the role of Cdc42, Rac, and Rho proteins in cytokine-mediated signal transduction remains to be defined.
In this report, we provide evidence that the JNK1 pathway is activated upon IL-3 stimulation, where Ras may be implicated as a regulator. Furthermore, we have obtained data showing that dominant-negative mSos1 that is capable of interfering with IL-3-induced Ras activation could block JNK1 activation upon IL-3 treatment. 2 Collectively, it is conceivable that a Ras-dependent JNK1 pathway exists within BaF3 cells, which may play an important role in IL-3-induced cellular responses.
Although ERK/MAPK and JNK/SAPK pathways converge at the point of transcriptional activation by the AP-1 complex (5), the precise physiological role of each pathway is still obscure. Some growth factor signals preferentially stimulate the ERK/ MAPK pathway, while others stimulate only the JNK/SAPK pathway. Growth factors like epidermal growth factor as well as IL-3, as illustrated in this paper, are able to activate both. Additionally, in the case of T-cell activation signaling, for example, two apparently independent signals from the T-cell receptor and the CD28 costimulatory receptor are necessary to augment JNK/SAPK activity, whereas ERK/MAPK is fully activated in response to T-cell receptor stimulation alone (45). In PC12 pheochromocytoma cells, ERK/MAPK is responsible for nerve growth factor-induced differentiation and proliferation, whereas the JNK/SAPK pathway is modulated by a signal of nerve growth factor withdrawal, which causes apoptotic cell death (46). Taking into consideration these findings, it is possible to speculate that JNK/SAPK may be responsible for cellular responses distinct from those regulated by ERK/MAPK in cytokine receptor-mediated signal transduction.
We have shown that IL-3-induced JNK1 activation occurs in a Ras-dependent manner, which is also the case in other types of cells, including PC12, HeLa, NIH 3T3, and COS7 cells (14,17,18,25,47). However, at present, the mechanism of JNK/ SAPK activation through Ras has not been fully understood. It may be possible that MEKK1, which binds directly to a GTPbound form of Ras (48), is activated by Ras, thereby stimulating a kinase cascade comprising of JNKK1/SEK1/MKK4 and JNK/ SAPK (5). Ras and MEKK1-dependent JNK/SAPK activation was, in fact, observed in HeLa cells (25). Instead, Rac may be involved between Ras and the JNK/SAPK cascade because Rac is thought to act as a downstream element of Ras (43,44), and regulate the JNK/SAPK pathway (17)(18)(19). These possibilities are currently under examination.
As shown in Fig. 4, Ras(G12V) is incapable of promoting JNK1 activity by itself in contrast to ERK2, which is strongly activated by Ras(G12V) alone in BaF3-V2 cells (see also Ref. 8). In HeLa and NIH 3T3 cell lines, an activated form of Ras is able to induce JNK/SAPK activation significantly, but less than an activated Cdc42 or Rac protein (18), whereas, in COS7 cells, v-Ras induces only a slight increase of JNK/SAPK activity (17). Another paper has reported that an activated Ras induced only a partial increase in JNK/SAPK activity, which could be further enhanced by irradiation of UV light (47). Furthermore, as FIG. 4. Ras(G12V)-induced activation of ERK2, but not of JNK1, in BaF3-V2 cells. A, IL-3 and anisomycin-induced JNK1 activation. Serum and IL-3-deprived BaF3-V2 cells were stimulated with mouse IL-3 (50 ng/ml) or anisomycin (20 g/ml) for the indicated periods, and the activity of immunoprecipitated JNK1 was measured. Data are shown as mean Ϯ S.E. (n ϭ 3). B, time course of IL-3-induced JNK1 activation. Serum and IL-3-deprived BaF3-V2 cells were stimulated with mouse IL-3 (50 ng/ml) for specified periods, and the activity of immunoprecipitated JNK1 was measured. Data are shown as mean Ϯ S.E. (n ϭ 3). C, Ras(G12V)-induced activation of ERK2, but not of JNK1. BaF3-V2 cells were incubated with IPTG (5 mM) for specified periods. During the last 3 h, the cells were cultured in medium deprived of serum and IL-3. The cells were left unstimulated (E, Ç) or stimulated with mouse IL-3 (50 ng/ml) for 40 min (q, å), and the activity of immunoprecipitated JNK1 (q, E) and ERK2 (Ç, å) was measured. Data are shown as mean Ϯ S.E. (n ϭ 3). described above, stimulation of the T-cell receptor, which is sufficient to induce Ras activation in T-lymphocytes, fails to activate JNK/SAPK, suggesting the activated Ras protein is unable to stimulate JNK/SAPK by itself. Taken together with these observations, a second signal, besides a signal from Ras, may be required for full activation of JNK/SAPK, especially in hematopoietic and lymphoid cells. The elucidation of the biochemical mechanism by which Ras regulates the JNK/SAPK pathway will contribute to our understanding of the role of this pathway in growth and differentiation signaling in hematopoietic cells.