JAK/STAT3-dependent activation of the RalGDS/Ral pathway in M1 mouse myeloid leukemia cells.

The Ras-related GTPase (Ral) is converted to the GTP-bound form by Ral guanine nucleotide dissociation stimulator (RalGDS), a putative effector protein of Ras. Recently, it was proven that Ral regulates c-Src activity and subsequent phosphorylation of its substrate, STAT3. Here, we show that STAT3 inversely regulates activation of Ral through induction of expression of RalGDS. To identify new leukemia inhibitory factor-induced genes, we have performed representational difference analysis using M1 mouse myeloid leukemia cells and cloned RalGDS. The expression of RalGDS and subsequent activation of RalA were clearly suppressed by a dominant negative form of STAT3 and a JAK inhibitor, JAB/SOCS1/SSI-1, indicating that RalGDS/RalA signaling requires the activation of the JAK/STAT3 pathway. An experiment using a Ras inhibitor demonstrated that full activation of RalA also requires activation of Ras. These results suggest a novel cross-talk between JAK/STAT3 and the Ras/RalGDS/Ral signaling pathways through gp130.

STAT3 have been characterized, most of which belong to genes regulating cell cycle progression and/or anti-apoptosis. For instance, cyclin D1/D2/D3/A, cdc25A, c-myc, pim-1, bcl-2, and bcl-x are up-regulated by STAT3 activation (2)(3)(4)(5). In contrast, in some cases, STAT3 contributes to cell cycle arrest by inducing cyclin-dependent kinase inhibitors such as p19 INK4D and p21 CIP1 (6,7). Their expression is directly or indirectly controlled by STAT3. However, as it is difficult to explain the molecular mechanisms of the diverse signaling effects of STAT3 by any combination of these already identified genes, further identification of STAT3-regulated genes is required.
RalGDS was initially identified as a gene homologous to the CDC25 family and found to stimulate the release of guanine nucleotide from Ral GTPase (8). A distinct family of Ras-related GTPases, Ral, consists of two highly similar proteins, RalA and RalB (85% identical) (9). The function of Ral proteins has yet to be further investigated, but recent reports suggest their crucial role in cellular transformation, differentiation, vesicle transport, and receptor endocytosis (10 -12). Recently, Goi et al. (13) demonstrated that Ral is essential for STAT3 tyrosine phosphorylation induced by epidermal growth factor or IL-6 and that Src kinase activation by Ral plays a key role in this process.
To identify the genes involved in signal transduction via gp130, we have performed representational difference analysis (RDA) using M1 cells, which differentiate into macrophages upon STAT3 activation by LIF/IL-6. Here we demonstrate that RalA is activated by the JAK/STAT3/RalGDS pathway, and its activity is also modulated by activation of Ras in M1 cells.
Northern Blot Analysis-0.5 g of mRNA or 5 g of total RNA was electrophoresed on a 1% agarose/formaldehyde gel and blotted to Hybond-N nylon filter (Amersham Pharmacia Biotech). The filter was hybridized with 32 P-labeled cDNA probes in rapid hybridization buffer (Amersham Pharmacia Biotech).
Immunoblotting Analysis-Whole cell extracts were prepared by lysing cells in Laemmli sample buffer. An equivalent amount of cell lysate (50 g) was electrophoresed on SDS-polyacrylamide gels, transferred to polyvinylidene difluoride membrane (Immobilon, Millipore), stained with antibodies, and detected for signals with the ECL system (Amersham Pharmacia Biotech).
Ral and Ras Activation Assay-The activity of Ras and Ral was * This work was supported by grants-in-aid for COE Research from the Ministry of Education, Science and Culture of Japan and for Scientific Research of Japan Society for promotion of Science. 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.

RESULTS
cDNA libraries were prepared from M1 cells treated with LIF for 1 h and untreated M1 cells and were applied for RDA as a tester and a driver, respectively (17,18). After three rounds of subtraction, we sequenced 85 DNA fragments. Fourteen fragments were part of the RalGDS gene, corresponding to regions 3098 -3327 (GenBank TM accession number L07924). To confirm that RalGDS is induced upon LIF stimulation, we examined its expression by Northern blot analysis. RalGDS mRNA was obviously induced 1 h after LIF stimulation and down-regulated to basal level at 9 h (Fig. 1A). Stimulation with IL-6 also increased RalGDS mRNA at a similar level (Fig. 1B). While there exist other Ral-specific guanine nucleotide exchange factors (RalGEF) such as Rgl and Rlf (19,20), the expression of none of the two genes was up-regulated upon LIF stimulation (Fig. 1C). Then we examined whether IL-6 would induce RalGDS expression not only in M1 cells but also in primary cells. As shown in Fig. 1D, stimulation of IL-6 plus sIL-6R clearly induced mRNA from RalGDS in murine splenocytes. The induction was, however, weaker compared with that observed in M1 cells.
To investigate whether the induction of RalGDS depends on STAT3 activation, we analyzed its induction in two cell lines: M1/Y705F that expresses a dominant-negative form of STAT3 (Y705F) and M1/JAB that expresses a JAK inhibitor, JAB/ SOCS1/SSI-1 (21). When stimulated with LIF, both cell lines showed much weaker phosphorylation of tyrosine on 705 of STAT3 (Tyr-705) than parental M1 cells ( Fig. 2A). Corresponding to the levels of phosphorylation of Tyr-705, the induction of RalGDS mRNA was clearly reduced in both M1/Y705F and M1/JAB (Fig. 2C, lanes 4 and 6). These data strongly suggest that the induction of RalGDS is dependent on the activation of the JAK/STAT3 signaling pathway.
In addition to the JAK/STAT3 pathway, LIF/IL-6 also activates the SHP-2/Grb2/MAP kinase pathway through gp130 (2,22). Thus, we also investigated the possibility that the induction of RalGDS might require the activity of MAP kinase. As shown in Fig. 2B, MAP kinase was activated in M1 cells upon LIF stimulation, and a MEK inhibitor, PD98059, inhibited activation. However, PD98059 had no effect on induction of RalGDS mRNA (Fig. 2C, lane 8), indicating that the SHP-2/ Grb2/MAP kinase pathway is dispensable for the up-regulation of RalGDS expression. To investigate whether this process requires de novo protein synthesis, M1 cells were stimulated in the presence of cycloheximide (CHX). As shown in Fig. 2D, addition of CHX did not affect RalGDS expression, suggesting that RalGDS might be directly regulated by STAT3.
Next, we examined whether the enhanced expression of Ral-GDS would bring about the activation of Ral in response to LIF stimulation. Activation of Ral was evaluated by in vitro binding to GST⅐RIP1, one of the effector molecules of Ral (23). As shown in Fig. 3A, while the activity of RalB was nearly unchanged, that of RalA increased 1 h after stimulation and returned to the basal level after 12 h. RalA was also activated by IL-6 (data not shown). The time course of RalA activation appears to correspond to that of RalGDS expression. Interestingly, activation of RalA was detected in neither M1/Y705F nor M1/JAB, (Fig. 3B, lanes 2 and 4). On the other hand, PD98059 had no effect on RalA activation (Fig. 3B, lane  6). These data strongly indicated that activation of RalA could be downstream of the JAK/STAT3 signaling pathway rather than via a MAP kinase pathway. In addition, a protein synthesis inhibitor, CHX, diminished the activation of RalA, which indicated that de novo protein synthesis is required for the activation of RalA (Fig. 3B, lane 8).
It has been reported that another GTPase molecule, Ras, is necessary for the activation of Ral, possibly by the recruitment of RalGDS to the cytoplasmic membrane (24 -26). Thus, we examined the activity of Ras in M1 cells by in vitro binding to the GST⅐Raf fusion protein. Ras was already activated 0.5 h after stimulation with LIF, and the activation continued to 3 h (Fig. 4A). To examine whether Ras activity is required for full activation of RalA, we utilized FTI-277, a specific farnesylation inhibitor. After 24 h of treatment with FTI-277, M1 cells were stimulated with LIF and then subjected to the Ras or RalA activation assay. As shown in the lower panel of Fig. 4Ba, FTI-277 converted about 50% of Ras into a non-farnesylated form (27), followed by decreased activation of both Ras and RalA (Fig. 4B, a and b). On the other hand, FTI-277 did not suppress the induction of RalGDS (Fig. 4Bc). These results indicate that activation of RalA by LIF are also modulated by the activity of Ras. Some evidence suggests the existence of other signaling molecules such as Rap GTPases to activate the RalGDS/Ral pathway (11), but the activation of Rap1 was not observed in M1 cells (data not shown). DISCUSSION IL-6 family cytokines transduce signals into the cell by binding to their cognate receptors, followed by dimerization and activation of JAK kinases, which subsequently phosphorylate tyrosine residues of gp130 to recruit signal-transducing molecules, most notably STAT1/STAT3 and SHP-2. Tyrosine-phosphorylated STAT3 enters the nucleus, where it regulates the transcription of a variety of genes. On the other hand, SHP-2 interacts with Grb2 and Gab1 and mediates signals to ERK MAP kinase. Thus gp130 generates two distinct major signaling pathways (1).
In this study, we have demonstrated that RalGDS mRNA, one of the effector molecules of Ras (24), is clearly induced by LIF/IL-6 in M1 cells. This induction appears rather specific to RalGDS among the RalGEF because Rgl or Rlf was not induced. Moreover, we have shown that induction of RalGDS is definitely regulated by JAK/STAT3 signaling and that the activity of RalA is dependent on the transcription of RalGDS and on the gp130/Ras pathway, which is likely to involve an SHP-2/Grb2/Sos cascade (22) (Fig. 5). As far as we know, this is the first report that a small G protein is activated by transcriptional induction of a guanine nucleotide exchange factor.
The finding of an increase of RalGDS expression in IL-6stimulated splenocytes may provide evidence for a physiological involvement of RalGDS induction in gp130 signaling. However, its induction by STAT3 activation seemed to be cell-type specific. Whereas a murine lymphoid subline, DA1.a, expressed RalGDS mRNA upon LIF stimulation, G-CSF, which activates STAT3, did not induce its expression in a murine myeloid cell line, 32D (data not shown). These results suggest that the induction of RalGDS might require some events other than STAT3 activation.
Goi et al. (13) reported that activated Ral regulates the phosphorylation of STAT3 through activation of Src kinase. Together with their observation, our results provide the possibility that STAT3 and Ral regulate each other to exert biological functions. Indeed, some studies have shown that Ral and STAT3 exert similar biological functions. Introduction of the catalytic domain of RalGEF, Rgr, into PC12 cells inhibited NGF-induced neurite outgrowth whereas a dominant-negative form of Ral accelerated it. Thus the RalGEF/Ral pathway is likely to have a negative effect on the NGF-induced differentiation of PC12 cells (28). Interestingly, Ihara et al. (29) found that STAT3 also negatively regulates the neurite outgrowth of PC12 cells. Recent studies provided the evidence that STAT3 is a critical signaling molecule involved in v-Src transformation (30,31) and that a constitutively active form of STAT3 itself was sufficient for oncogenesis (4). Urano et al. (25) reported that Ral enhances the transforming activities of other oncogenes, while it does not induce oncogenic transformation on its own. Further studies should elucidate the biological function of possible cross-talk between STAT3 and Ral.
Taken together, the data we present here indicate that LIF/ IL-6 activates RalA by a novel combination of JAK/STAT3/ RalGDS and Ras pathways in M1 cells and will shed new light on the elucidation of signal transduction of gp130.