Abstract
Binding of IL-4 to its cognate receptor leads to the activation of a number of signaling pathways within the cell. Activation of the transcription factor STAT6 by JAK family protein tyrosine kinases has been shown to be essential for the full response of cells to IL-4. To elucidate the role of STAT6 in IL-4 signaling, we have constructed and expressed in cells a conditionally active form of the protein (STAT6:ER*) by fusing STAT6 to a modified form of the hormone-binding domain of the estrogen receptor. Activation of STAT6:ER* by 4-hydroxytamoxifen leads to specific activation of STAT6-regulated gene expression including the activation of a STAT6 reporter construct and induction of CD23 in B cell lines. Interestingly, in contrast to native STAT6, activation of STAT6:ER* occurs in the absence of detectable tyrosine phosphorylation of the fusion protein. This type of conditional system will be helpful in dissecting the mechanisms and specificity of transcriptional regulation by the STAT family of transcription factors.
Engagement of IL-4 to its receptor leads to heterodimerization of the receptor (1). Receptor dimerization leads to activation of the receptor-associated tyrosine kinases, Janus kinase (JAK)3 1 and JAK3, which then phosphorylate distinct tyrosine residues of the receptor (2). This results in the activation of at least two distinct signaling pathways, including STAT6 activation (3, 4) and the phosphorylation of the insulin receptor substrate 1/2 (5). After phosphorylation of the IL-4 receptor, STAT6 docks onto the receptor through the src homology domain 2 and then is tyrosine phosphorylated on residue Y641. The current model suggests that tyrosine phosphorylation of STAT6 elicits its homodimerization and nuclear translocation. STAT6 homodimers are then able to bind to specific IFN-γ-activated site (GAS)-like sequences found in the promoter region of IL-4-inducible genes. Since the mutation of residue Y641 in STAT6 completely abolishes DNA-binding activity as well as trans-activation (6), it would appear that tyrosine phosphorylation of STAT6 is required for signal transduction.
Gene disruption studies have shown that STAT6 is required for IL-4-mediated biologic functions such as class switching to IgE, induction of IL-4-inducible genes (e.g., CD23, MHC class II, IL-4 receptor), Th2 cell differentiation, and some effects on lymphoid proliferation (7, 8, 9). However, it is unclear whether the activation of STAT6 is sufficient for inducing the IL-4-mediated activities or whether there is some dependence on the activation of alternate signaling molecules as yet undefined.
To study the relative importance of STAT6 in IL-4-induced signal transduction and the mechanisms of activation of STAT6, we developed a conditionally active form of STAT6 by fusing the hormone-binding domain of a modified form of the mouse estrogen receptor (ER*) gene (10, 11, 12) to the murine STAT6. Activation of STAT6:ER* led to induced transcription of an IL-4-responsive promoter and in M12.4.1 B lymphoma cells led to induction of CD23. This system will be useful in dissecting the role of STAT proteins in cytokine and growth factor signaling.
Materials and Methods
Cell lines and reagents
BA/F3 cells were cultured in RPMI 1640 (JRH Biosciences, Lenexa, KS) supplemented with IL-3 (10 ng/ml). M12.4.1 cells, a kind gift from Dr. A. Keegan (American Red Cross, Bethesda, MD), were cultured in RPMI 1640. All tissue culture media were supplemented with 10% FCS, l-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). 4-Hydroxytamoxifen (4-HT) (Research Biochemicals Institute, Natick, MA) was used to activate STAT6 fusion protein.
Plasmid DNA
pMXGSTAT6:ER* (vector pMX (13)) contained STAT6-hbER* fusion (murine STAT6 cDNA (3), hbER* from pBP3:hbER* (12)), and enhanced green fluorescent protein (EGFP) gene that were controlled by SRα (14) and long terminal repeat (LTR) promoter, respectively. Control pMXG plasmid contained the EGFP gene controlled by the LTR promoter.
The reporter construct carrying the luciferase reporter gene (pGLCε) was generated by inserting three copies of human Cε STAT6 binding site oligonucleotide as direct repeats into the BglII site of pGL2-P (Promega, Madison, WI).
Retrovirus infection
Three micrograms of DNA were transfected into the packaging cell line Phoenix (provided by Dr. G. Nolan, Stanford University, Stanford, CA) using lipofectamine (Life Technologies, Gaithersburg, MD). After 2 days of transfection, viral supernatants were used for infection of M12.4.1 and BA/F3 cells.
Luciferase assay
BA/F3 cells were transfected with 10 μg of reporter plasmid pGLCε and 1 μg of reference plasmid pRSV-LacZ by electroporation. After culturing for 24 h in phenol red-free RPMI 1640 (Life Technologies) medium, the cells were stimulated with either 1 μM 4HT or 100 U/ml IL-4 for 24 h at 37°C. The cells were harvested and assayed for luciferase activities using a Luciferase Assay Kit (Promega, Madison, WI).
Immunoprecipitation and Western blotting
Cell lysates were immunoprecipitated with anti-mouse STAT6 Ab (Santa Cruz Biotechnology, Santa Cruz, CA) as previously described (15). Western blotting was performed with either anti-STAT6 Ab, anti-ER Ab (Santa Cruz), or monoclonal anti-phosphotyrosine Ab (Upstate Biotechnology, Lake Placid, NY) using the ECL detection system (Amersham, Amersham, Buckinghamshire, U.K.).
Electrophoretic mobility shift assay for STAT6
Nuclear extracts and electrophoretic mobility shift assay (EMSA) were prepared as previously described (16), except that binding buffer included 1 μM 4HT. In the supershift assay, nuclear extracts were incubated with 1 μg of each Ab at room temperature for 10 min before adding the probe.
Results
Construction and expression of conditionally active STAT6:ER*
To derive a conditionally active form of STAT6, we made a retrovirus construct, pMXGSTAT6:ER*, encoding the EGFP marker and, a chimeric protein of mouse STAT6 and a modified form of the hormone-binding domain of ER* that fails to respond to β-estradiol but retains responsiveness to estrogen analogues such as 4-HT and ICI 182,780 (11). Retrovirus infection was followed by several rounds of sorting for EGFP-positive cells in a mouse pre-B cell line (BA/F3) and a mouse B lymphoma line (M12.4.1) to yield stable cell lines, BS6ER* and MS6ER*, respectively. As shown in Figure 1, both BS6ER* and MS6ER* expressed comparable amounts of STAT6:ER* fusion protein (160 kDa) and endogenous STAT6 (100 kDa). The control pMXG infecting BA/F3 and M12.4.1 cell lines (named BGFP and MGFP, respectively) expressed only endogenous STAT6.
The fusion protein STAT6:ER* binds to the STAT6-specific GAS-like element and activates transcription in response to 4-hydroxytamoxifen
To examine 4HT-dependent activation of the STAT6:ER* protein, we first performed EMSA on nuclear extracts from MS6ER*, using a probe containing GAS-like sequence specific for STAT6 from the human Cε promoter (17). As shown in Figure 2 A, treatment of cells with murine IL-4 induced formation of complex (I), which appeared in nuclear extracts from both MS6ER* and MGFP cells. By contrast, activation of STAT6:ER* by 4HT-induced complex (II) which had reduced mobility compared with complex (I). Complex (II) appeared only in nuclear extracts from MS6ER* cells but not in those from MGFP cells. Furthermore, complex (II) was supershifted by both an anti-STAT6 Ab and an anti-ER Ab, indicating that complex (II) contained STAT6:ER*. In accordance with the above observations, IL-4-induced complex (I) was supershifted only by the anti-STAT6 Ab in MS6ER* cells. A minor part of complex (I) was also supershifted by adding the anti-ER Ab in MS6ER* cells, suggesting that complex (I) consists of either a STAT6-STAT6 homodimer or a small amount of STAT6-STAT6:ER* heterodimer. These data indicate that STAT6:ER* DNA-binding activity is strongly and rapidly (within 1 h, data not shown) induced after addition of 4HT to these cells. They also indicate that STAT6:ER* does not respond to IL-4 in the absence of 4HT.
Next, to test the trans-activation of the STAT6:ER* protein, pGLCε was transfected into BS6ER* and BGFP by electroporation. In BS6ER* cells, the reporter gene was efficiently transcribed in the presence of 4HT. Importantly, the luciferase activity induced by 4HT was similar to that induced by IL-4. As expected, 4HT did not induce luciferase activity in control BGFP cells, whereas IL-4 did (Fig. 2 B). These data indicate that STAT6:ER* activates Cε-driven transcription in a 4HT-dependent fashion.
Tyrosine phosphorylation is not required for activation of STAT6:ER*
It has been shown that tyrosine phosphorylation of STAT proteins is essential for their activation (18). Thus, we investigated whether STAT6:ER* activation by 4-HT was accompanied by tyrosine phosphorylation by Western blot (Fig. 3). Treatment of cells with IL-4 led to tyrosine phosphorylation of both STAT6 and STAT6:ER*, which was readily detected after 15 min of cytokine treatment and was sustained for up to 24 h after the addition of IL-4. By contrast, activation of STAT6:ER* by the addition of 4HT did not result in any detectable phosphorylation in spite of the expression of STAT6:ER* protein. High levels of STAT6:ER* protein accumulation were observed after 6 to 24 h of 4HT addition, possibly due to the selective stabilization of the protein as observed in ΔRaf:ER fusion proteins after prolonged addition of 4HT (19).
4HT induced up-regulation of CD23 expression in M12.4.1 cells infected by pMXGST6:ER*
It has previously been shown that treatment of the M12.4.1 cell lines with IL-4 leads to an elevation of CD23 expression (20). We examined whether STAT6:ER* could mimic the effects of IL-4 in the regulation of CD23 expression. Cells expressing STAT6:ER* as well as the control obtained by infecting pMXGST6:ER* and pMXG, respectively, were treated with either IL-4 or 4HT for 16 h. The activation of STAT6:ER* by 4HT or IL-4 was evaluated by the level of cell surface CD23 expression assessed by flow cytometry (Fig. 4). It was apparent that CD23 induction was observed only in STAT6:ER expressing populations (Fig. 4,f) but not in cells infected with the control virus (Fig. 4,c) after addition of 4HT, although IL-4 caused increased CD23 expression in all cell populations (Fig. 4, b, c, e, and f). These data indicate that STAT6:ER* can mimic IL-4 function to induce CD23 expression.
Discussion
We describe the construction and utility of a conditionally active form of STAT6 derived by fusing mouse STAT6 to a modified form of the hormone-binding domain of the mouse ER. Activation of STAT6:ER* with 4HT leads to the rapid induction of specific DNA-binding activity and activation of a STAT6-specific reporter gene. When expressed stably in cells, activation of STAT6:ER* mimics some of the effects of IL-4 in that it induces the cell surface expression of CD23. We observed that the rate and kinetics of CD23 induction by STAT6:ER* were similar to those observed in response to IL-4. In addition, we confirmed that STAT6:ER* and native STAT6 share the same specificity for DNA binding and promoter activation (data not shown). Importantly, however, activation of STAT6:ER* occurred in the absence of tyrosine phosphorylation of the fusion protein.
These data indicate that we can activate STAT6 protein in cells in the absence of cytokine receptor engagement, allowing the analysis of STAT6-specific events in the absence of other “parallel” signal transduction pathways. This conditional system can be readily exploited to identify STAT6-regulated genes in a variety of cell systems.
The hormone-binding domain of steroid receptors has proven remarkably versatile in the establishment of conditional forms of transcription factors (E1A, c-Myc, c-Fos, CEBP, v-Rel) (21, 22, 23, 24, 25) and signaling molecules (Raf, c-Abl) (26, 27). It has been clearly demonstrated that the isolated C-terminal hormone-binding domain of ER dimerizes and undergoes conformational changes as observed in the whole molecule after ligand binding (28, 29). Although the exact mechanism of the observed conditionality remains unclear, it is likely that hormone-regulated dimerization may be important. In this regard, hormone-induced dimerization may mimic the functional consequences of STAT6 tyrosine phosphorylation believed to be essential for dimerization, which is a prerequisite for transcriptional regulation (6). Moreover, since STAT6:ER* activation occurred in the absence of tyrosine phosphorylation, it seems likely that tyrosine phosphorylation is required only for dimerization and not for activation of transcription per se. Furthermore, the low level of reporter activity in the absence of 4HT indicates that the STAT6:ER* fusion protein is not leaky.
Both tyrosine phosphorylation and DNA binding of STAT6:ER* resisted the effects of IL-4. However, in the presence of 4HT, both tyrosine phosphorylation and DNA binding of STAT6:ER* were strongly potentiated in response to costimulation with IL-4. These data suggest that the ER* domain renders the STAT6:ER* fusion resistant to the activation of IL-4 receptor by ligand engagement. Neither the inactive form of STAT6:ER* nor the phosphorylated form of STAT6:ER* displayed a dominant-negative effect, since activation of cells in combination of IL-4 and 4HT caused additive effects on endogenous CD23 expression as well as trans-activation of the reporter gene (data not shown).
Taken together, these data indicate the feasibility of using the hormone-binding domain of the ER to regulate the function of STAT family transcription factors. Activation of STAT6:ER* clearly mimicked one of the effects of IL-4 treatment of B lymphoma cells. We are presently identifying additional STAT6-regulated genes and attempting to identify their role in the effects of IL-4. Finally, given the degree of structural similarity of the members of the STAT family of transcription factors, it seems likely that this approach will have broad utility in the analysis of STAT-mediated regulation of gene expression.
Acknowledgements
We thank Dr. J. Ihle for the gift of the STAT6 cDNA, Dr. G. Nolan for retrovirus packaging cell line, Dr. T. Kitamura for a retroviral vector pMX, Dr. M. Howard for the BA/F3 cell line, and Dr. S. Menon for mouse IL-3 and IL-4. We thank J. Cupp, E. Callas, D. Polakoff, and E. Murphy for cell sorting and D. Liggett for for oligonucleotide synthesis and Drs. T. Migone, E. Masuda, A. Mui, D. Robinson, Y. Amasaki, R. Imamura, D. Wylie, and K. Arai for critical comments and reading of the manuscript.
Footnotes
DNAX Research Institute of Molecular and Cellular Biology is supported by the Schering-Plough Corporation.
Abbreviations used in this paper: JAK, Janus kinase; 4HT, 4-hydroxytamoxifen; ER, estrogen receptor; GAS, IFN-γ-activated site; LTR, long terminal repeat; EGFP, enhanced green fluorescent protein; EMSA, electrophoretic mobility shift assay.