Interleukin-3 (IL-3)-induced c-fos Activation Is Modulated by Gab2-Calcineurin Interaction*

Interleukin-3 (IL-3) regulates cell growth by affecting various processes such as cell death, survival, and proliferation. Cues from the external environment are sensed by surface receptors, and complex signaling mechanisms arise within the cells, leading to specific functional outcomes. In this study, we demonstrate that the cytokine IL-3 induces the activation of the Ca2+-dependent phosphatase, calcineurin (Cn). Furthermore Cn dephosphorylates Gab2, resulting in c-fos activation and cell proliferation. We also report that there is a direct interaction between Cn and Gab2 upon IL-3 stimulation, and Akt can regulate this interaction.

The tight control of survival, proliferation, and apoptosis is essential for normal cellular growth and development. A variety of extracellular signals including growth factors and cytokines regulate viability, proliferation, differentiation, and function in bone marrow progenitor cells to maintain normal hematopoiesis. Interleukin-3 (IL-3) 3 is a multipotent hematopoietic growth factor, which acts on progenitor, myeloid, and mast cells to promote cell survival, induce proliferation, and facilitate differentiation.
Upon IL-3 treatment, there is rapid phosphorylation of tyrosine residues in the ␤-subunit of the IL-3 receptor followed by tyrosine phosphorylation of a variety of signaling molecules including Gab2 (Grb2-associated binder-2) (1). Gab2 is a member of a family of pleckstrin homology domain-containing adaptor proteins that includes mammalian Gab-1, Gab-3, and Drosophila DOS (daughter of sevenless). Gab2 has been implicated in a variety of cellular functions including the negative regulation of T-cell receptor signaling (2) and the positive regulation of growth factor, cytokine, and antigen receptor signal-ing (3)(4)(5). Gab2 knock-out mice show tissue-specific defects in mast cell signaling (6) and osteoclast differentiation (7). Gab2 knock-out mice also have defective hematopoiesis (8). It has been reported that Gab2 plays a pivotal role in BCR/ABL-induced transformation (9) and breast cancer (10). Gab2 is one of several genes in a highly amplified locus in some breast cancers (11). Gab2 has also been shown to be a direct target of the transcription factor E2F and an essential effector of E2F-dependent Akt activation during cell cycle progression (12).
Calcium (Ca 2ϩ ) is a universal second messenger, and the temporal and spatial regulation of intracellular Ca 2ϩ enables cells to respond to various stimuli and control cellular processes including proliferation, development, contraction, secretion, and motility (13). Ca 2ϩ can act directly on target proteins, or its effect can be mediated via intracellular Ca 2ϩbinding proteins. A major intracellular modulator for Ca 2ϩ is calmodulin (CaM), which is a highly conserved four EF-handcontaining Ca 2ϩ -binding protein. One of the Ca 2ϩ /CaM-dependent targets is the serine/threonine phosphatase, calcineurin (Cn) (reviewed in Refs. 14 and 15). Although Ca 2ϩ / CaM has been shown to be required for proliferation in both unicellular and multicellular eukaryotes, the nature and regulation of the Ca 2ϩ -dependent pathway, including the role of Cn in regulating cell proliferation, is not clear.
Here we report a novel, cytokine (IL-3)-induced, calciumand Cn-dependent activation of c-fos, modulated by the scaffolding adaptor protein Gab2. We demonstrate that Gab2 directly associates with Cn and regulates c-fos, and this association modulates IL-3-mediated cell growth and proliferation. These results reveal a novel link between cytokine signals and cell growth and proliferation signals mediated via Gab2 and Cn. We also show that the Cn-Gab2 interaction can be regulated by the Akt Kinase.
* 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. 1  Trypan Blue 0.4% was from Invitrogen. The Dual-Luciferase assay kit was purchased from Promega (Madison, WI). Plasmid Constructs-The Gab2 constructs used in these studies have been described previously (1) or were generated by PCR-assisted mutation strategies. Each mutated construct was sequenced to verify the existence of the desired mutation and for the absence of PCR-generated mutations. All the Cn constructs used have been described previously (16). Cn A subunit or mutants were always co-transfected with Cn B subunit to obtain the optimal Cn function in vivo. The Akt and Akt (kinase dead) constructs were a kind gift from Dr. Moses Chao (NYU School of Medicine, NY).
Luciferase Assay-Baf3 cells (10 7 ) were transfected with 10 g of the indicated plasmids at 800 microfarads/250 V using an Invitrogen electroporator. All cells were cotransfected with 30 ng of a Renilla luciferase construct (Promega) to normalize for transfection efficiency. Once the cells were recovered in complete RPMI supplemented with WEHI conditioned medium for 2-3 h, they were spun down, washed, and resuspended in IL-3 deprivation medium lacking IL-3 for 16 -18 h. An aliquot (10 6 cells) was set aside for immunoblot analysis to confirm expression of the transfected proteins. For IL-3 stimulation, the transfected cells were resuspended in 1 ml of deprivation medium and divided into aliquots to be left unstimulated or stimulated with IL-3 (10 ng/ml) for 6 h. Cells were then washed once with phosphate-buffered saline and lysed. Luciferase activities were determined with a Dual-Luciferase kit (Promega) and Berthold Detection Systems Sirius single tube luminometer (Oakridge, TN). Immunoprecipitation and Immunoblotting-Cell lysis, immunoprecipitation, immunoblotting, and detection by enhanced chemiluminescence were performed as described previously (17). Anti-Cn B or anti-HA immunoprecipitations utilized 1 g of antibody/10 7 Baf3 cell equivalents. Briefly, unstimulated or stimulated transfected cells were washed and lysed in buffer containing 1% Triton X-100, 150 mM NaCl, 1 mM CaCl 2 , and a mixture of protease and phosphatase inhibitors. Lysates were spun at 13,000 rpm for 10 min and subjected to immunoprecipitation overnight with the appropriate antibody. Protein A-Sepharose beads and rabbit anti-mouse antibody (Jackson ImmunoResearch, West Grove, PA), when required, were added the next day. The beads were washed extensively with immunoprecipitation wash buffer (20 mM HEPES, 150 mm NaCl, 10% glycerol, 1 mM CaCl 2 , and a mixture of protease and phosphatase inhibitors). Bound proteins were resolved by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with the appropriate antibodies. The enhanced chemiluminescence detection system (Amersham Biosciences) was used to visualize proteins.
Cell Cycle Analysis-Plasmid transfections into Baf3 cells were carried out by Amaxa-based nucleofection following the manufacturer's instructions (Amaxa, Gaithersburg, MD). 10 ng of enhanced green fluorescent protein plasmid (eGFP) (Invitrogen) was also co-transfected with each transfection and used to sort the transfected cells using a MoFlo high speed cell sorter (Cytomation, Fort Collins, CO). The sorted cells were labeled with propidium iodide and analyzed on FACScan (BD Bio-sciences). Cell cycle analysis was done by FlowJo software package (Treestar, Ashland, OR).

RESULTS
IL-3 has long been recognized as a key cytokine required for hematopoietic progenitor cell survival and proliferation. The process of induction of cell death upon withdrawal of IL-3 has been extensively investigated (18). IL-3-induced protection against cell death is thought to be mediated by the activation of the PI3-kinase/Akt signaling pathway. In other reports, calcium has been shown to protect against death and promote cell survival (19). Upon withdrawal of IL-3, [Ca 2ϩ ] i and Cn activity were required for cell survival in an IL-3-dependent cell line (20). There are numerous reports of calcium regulating the immediate early genes (IEGs), including c-fos (21), that are involved in cell survival and proliferation in a cell-specific manner. However, the details of proximal signaling mechanisms initiated by IL-3 resulting in cell survival and proliferation are not well understood. We, therefore, investigated the proximal signals induced by IL-3 in activating c-fos and leading to cell survival and proliferation.
We utilized the IL-3-dependent Pro-B cell line Baf3 to investigate the IL-3-dependent activation of c-fos. We investigated: 1) whether verapamil, an ion channel blocker, could block IL-3 signals; 2) whether a constitutively activated form of Cn that does not require activation by Ca 2ϩ can overcome the effect of verapamil; and 3) whether IL-3 signaling leading to the activation of c-fos is dependent on intracellular Ca 2ϩ . c-fos is induced early upon activation and is known to play a role in cell growth and proliferation. We have shown previously that c-fos is one of the IEGs activated upon IL-3 stimulation of Baf3 cells (1). We evaluated the effects of manipulating the IL-3 signals on a c-fos promotor-driven luciferase gene construct. The Cn A subunit was always co-transfected with the Cn B subunit for optimal Cn function. Baf3 cells transfected with vector alone, wild type (WT) Cn A, or a constitutively activated form of Cn-Cn A(␦) together with a c-fos luciferase construct were IL-3-starved for 16 -18 h and then left unstimulated or stimulated with IL-3 for 6 h. As reported earlier, there is a robust activation of the c-fos promoter upon IL-3 stimulation of Baf3 cells (Fig. 1A). This IL-3-induced c-fos-driven luciferase gene activity was inhibited upon pretreatment with verapamil, an ion channel blocker, and this inhibition was overcome by the expression of Cn A(␦) but not by WT Cn (Fig. 1A). Starving Baf3 cells resulted in complete loss of [Ca 2ϩ ] i stores that could be released by thapsigargin (data not shown). Upon starvation and restimulation with IL-3, the Baf3 cells showed a robust activation of the c-fos promoter (Fig. 1B). Chelation of intracellular Ca 2ϩ by BAPTA led to a nearly complete inhibition of c-fos promoter activity even in the presence of ectopically expressed WT Cn, whereas there was no effect of chelation when Cn A(␦) was expressed (Fig. 1B).
Gab2, a scaffolding adaptor protein, is robustly phosphorylated upon IL-3 stimulation (1). To investigate the possible role for both Cn and Gab2 in IL-3-mediated c-fos activation, we sought to further understand the nature of the complex that is assembled upon IL-3 stimulation. To determine whether there is a direct interaction between Cn and Gab2 upon IL-3 stimulation, we conducted co-immunoprecipita-tion experiments. We expressed either the WT or Cn A(␦) in the presence of HA-tagged Gab2 in Baf3 cells. A Gab2-Cn association could be detected by co-immunoprecipitation with an anti-Cn B antibody and blotting for Gab2. (Fig. 1C). Although IL-3 stimulation leads to an increased binding of Gab2 to WT Cn, the constitutively activated form showed no dependence on IL-3 for binding. Cn has been studied extensively as an upstream phosphatase regulating the transcription factor nuclear factor of activated T-cells (NFAT); however, the activation of c-fos via Ca 2ϩ and Cn has not been demonstrated. We, therefore, assessed whether the IL-3-mediated Gab2-Cn interaction could lead to a more robust activation of c-fos. Gab2 and Cn constructs were expressed alone or together along with a c-fos reporter as described earlier.
Although expression of either Gab2 or Cn individually increased c-fos promoter activity (5-and 14-fold, respectively), there was remarkable synergy when the two proteins were expressed together, leading to a 45-fold increase in c-fos activity as compared with the unstimulated vector control (Fig. 1D). The expression of the constitutively activated form of Cn resulted in c-fos promoter activity being independent of IL-3 stimulation, whereas this result was not seen when WT Cn was co-expressed. Taken together, these data show that Cn and Gab2 can directly interact, and co-expression of Cn with Gab2 greatly augments c-fos promoter activity. We also expressed the phosphatase dead point mutant Cn A␦ (H151Q) in Baf3 cells along with the c-fos luciferase construct. As shown in Fig. 1D, the expression of the phosphatase dead Cn A␦ point mutant resulted in almost complete inhibition of IL-3-induced c-fos promoter activity even in the presence of Gab2 overexpression. Knock down of Gab2 by small interfering RNA also inhibited c-fos activity to background levels (data not shown). These results indicate an important role for both Gab2 and Cn in IL-3-mediated c-fos activation.
It has been previously postulated that the serine/threonine kinase Akt could be a regulator of Gab2 activity by phosphorylating it at Ser-159 (human)/Ser-160 (mouse) (4). We did an extensive internal deletion analysis to determine the binding domain of Gab2 that interacts with Cn (data not shown). This analysis revealed a direct Gab2-Cn interaction, and the serinerich region C-terminal to the pleckstrin homology domain of Gab2 was found to be important for the interaction. It was interesting that the serine-rich domain that we identified as critical for Cn binding contains the Ser-160. We, therefore, investigated the possible IL-3-dependent regulation of Gab2 S160 by Akt. Akt has been previously implicated in cell survival and is activated upon IL-3 stimulation. When Baf3 cells were stimulated with IL-3, we could detect serine phosphorylation of Gab2 as indicated by the anti-phosphoGab2 S160 antibody (Cell Signaling) blot ( Fig. 2A). We then determined whether the serine phosphorylated Gab2 could be immunoblotted with an anti-phospho-Akt substrate antibody. For this experiment, Gab2 was expressed with either wild type Akt or kinase dead Akt. We found that Gab2 could be detected by the anti-phospho-Akt substrate antibody only when the WT Akt was coexpressed and not when the kinase dead Akt was expressed (Fig.  2B). We then determined whether Akt had any effect on IL-3- induced activation of c-fos mediated by Gab2-Cn. When we co-expressed Akt (WT) along with Gab2 and Cn, c-fos activation was greatly diminished, indicating negative regulation of the Gab2-Cn complex (Fig. 2C). These results indicate that Akt could phosphorylate Gab2 and regulate the Gab2-Cn complex and thereby c-fos activation by IL-3. To test whether serine phosphorylated Gab2 could be a substrate for Cn, we transfected in Gab2 and Akt along with either Cn A(␦) or Cn A␦ (H151Q). As shown in Fig. 2D, Gab2 could be robustly detected by phospho-Akt substrate antibody only in the presence of phosphatase dead Cn and not activated Cn. These results indicate that serine phosphorylated Gab2 can be a substrate for Cn.
We and others have previously shown that calcium, Cn, and Gab2 could all be individually involved in promoting cell growth and proliferation (3,(22)(23)(24). In this report, we show that upon IL-3 stimulation, Cn and Gab2 cooperate to activate c-fos. We, therefore, assessed whether IL-3-induced activation of c-fos promoter could regulate cell growth and proliferation. We performed cell cycle analysis of Baf3 cells expressing Gab2 and Cn A(␦) either alone or together. Positive transfectants were sorted by fluorescence-activated cell sorter utilizing the enhanced green fluorescent protein plasmid that was co-transfected at 1/1000. Positive transfectants were analyzed for DNA content by propidium iodide staining on fluorescence-acti-vated cell sorter. The percentages of cells in the S-phase of cell cycle for each transfection are shown in Table 1. As shown in Table 1, about 5% of Baf3 cells transfected with vector control were in the S-phase of cell division when grown for 24 h without IL-3 as compared with ϳ32% with IL-3. Expression of Gab2 or Cn A(␦) did not significantly alter the percentage of cells in the S-phase under IL-3 starvation or IL-3 stimulation conditions. If Gab2 were a negative regulator of proliferation, it would be expected that the expression of Gab2 should reduce the percentage of cells in the S-phase of cell division. No such inhibition is observed, and this could be due to the requirement of Gab2 s160 to be phosphorylated by Akt to be a negative regulator; this phosphorylation could be a finely tuned event that regulates the overall cell survival and proliferation. Upon coexpression of Cn A(␦) along with Gab2, a large number of cells were rescued, and the percentage of cells in the S-phase of the cell cycle increased from 5 to ϳ26% under conditions of IL-3 starvation (Wilcoxon rank sum test, n ϭ 4, p ϭ 0.02). There was also a significant increase in the percentage of cells in S-phase with IL-3 stimulation from ϳ32 to ϳ38% (p ϭ 0.02). Interestingly, when Cn A␦ (H151Q) was expressed along with Gab2, almost no rescue of cells from apoptosis upon IL-3 withdrawal was observed. When co-expressed with Gab2, the phosphatase activity of Cn is required as indicated by the reduction in the percentage of cells from 25 to 6% (p value 0.02) under conditions of IL-3 withdrawal and from 38 to 23% (p ϭ 0.05) under IL-3 stimulation. Although this reduction is only modest, this could be due to the phosphatase dead Cn acting as a dominant negative and sequestering Gab2 and/or other proteins from the IL-2-induced complex. Alternatively, the phosphatase activity of Cn could be required for down modulating and releasing Gab2 to return to the G 0 /G 1 state as demonstrated for CDK4 (25).

DISCUSSION
Cytokine receptors trigger multiple signaling cascades that regulate many cellular functions including survival, growth, and proliferation. IL-3 is known to play an essential role in apoptosis, survival, and proliferation of hematopoietic cells both in vivo and ex vivo. IL-3 triggers many pathways that lead to distinct cellular outcomes. Although IL-3 signals via the PI3-kinase/Akt pathway for cell survival, it also utilizes the SHP2/MAPK (mitogen-activated protein kinase)/Elk1 pathway for cell growth and the STAT5 pathway for cell differentiation. The mechanism by which activation of a cytokine receptor leads to multiple signals that are chan-

TABLE 1 Cell cycle analysis showing percentages of cells in the S-phase of cell division
Mean percentages of cells Ϯ S.D. in the S-phase of cell division are presented. Comparison among the transfections were done using the nonparametric Kruskal-Wallis test (n ϭ 4, p ϭ 0.02 (for -IL-3 groups) and p ϭ 0.01 (for ϩIL-3 groups)). neled within the cell to achieve different outcomes remains a puzzle that is yet to be solved. The scaffolding adaptor Gab2 is one of the proteins that is robustly phosphorylated upon IL-3 signaling (1, 3). More recent reports implicate Gab2 in numerous proliferation signaling cascades including receptor activator for nuclear factor B ligand (RANKL)/NFB, PI3-kinase/Akt, and c-Kit/IgE (7,24,26). We provide evidence here that Gab2 modulates IL-3 signaling cascades by integrating or coordinating with other proliferation signals, resulting in cell growth and proliferation. We show that IL-3-induced Cn activation and that its association with Gab2 modulates the transcriptional regulation of c-fos. Although calcium and Cn have been previously implicated in cell growth and proliferation, the modulation of its function by association with adaptor proteins probably provides the cellular context or niche for the signals to be regulated in an organized manner. It could be envisioned that these contextspecific interactions can then help cells modulate their cell cycle depending on the cues received from both external and internal stimuli. A recent report of E2F regulation of Gab2 transcription and Akt activation lends credence to this idea (12). Cell cycle-regulated activation of Akt by Gab2 might provide the feed-forward regulation of IEGs and proliferation via the calcium/Cn pathway in the context of cell survival signals mediated via Akt. Cn is a ubiquitously expressed protein that also plays an essential role in many cellular functions such as cell proliferation, differentiation, and death (27). The role of Cn in nuclear factor of activated T-cells signaling is best studied, but there is a growing list of Cn-binding proteins that play a role in regulating or modulating signaling pathways. Some Cn binding interactions, for example with adaptor proteins such as AKAP-79, brings Cn together with protein kinase C and cAMP-dependent protein kinase (28) within cellular domains such as plasma membrane. Other examples include FKBP12, which brings Cn to the inositol 1,4,5-trisphosphate complex (29) and calsarcin that links Cn to ␣-actinin in muscles (30). All of these adaptor proteins seem to provide Cn a context for its activity. Likewise, Gab2, an adaptor protein, also possibly provides the cellular context for calcium and Cn function in cell growth and differentiation. Another possibility is that Gab2 and Cn acting together is akin to FKBP12/FK506, where the two together result in a gain of function in activating c-fos, which in turn is part of the cell growth and proliferation cascade.
In this report, we provide evidence for Gab2 and Cn directly interacting and providing the context and convergence of IL-3-stimulated cell survival signals via Akt and cell proliferation signals downstream of c-fos. The modulation of Gab2 phosphorylation provides the framework for assembling signaling complexes in a given context. How these signals of cell survival and proliferation are coordinately regulated and amplified in different cellular contexts remains to be elucidated.