Txk, a Member of Nonreceptor Tyrosine Kinase of Tec Family, Acts as a Th1 Cell-Specific Transcription Factor and Regulates IFN-γ Gene Transcription¹

Much interest has focused on Th1 and Th2 subsets that have been characterized on the basis of the discrete cytokine production profiles; Th1 cells secrete IL-2, IFN-γ, and lymphotoxin and are important for the cell-mediated response; Th2 cells produce IL-4, IL-5, IL-10, and IL-13 and provide help for Ig production (1, 2, 3, 4). Accumulating evidence suggests that distinct signaling molecules and transcription factors mediate cytokine expression pattern in Th1 and Th2 cells (5, 6, 7, 8, 9, 10, 11, 12, 13). However, to date, precise mechanisms responsible for the differentiation and development of polarized Th1 responses are not fully clarified in humans. Especially, intracellular signaling pathway specific for Th1 cells remains elucidated.

The Tec family has emerged recently as a subfamily of nonreceptor tyrosine kinases, consisting of Tec, Btk, Itk/Tsk/Emt, Bmx, and Txk/Rlk, all of which are importantly involved in the lymphocyte signaling pathways (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24). Recently, Itk, the T cell-associated Tec family kinase, has been suggested for the involvement of Th2 cell development (5, 25). Txk/Rlk has been shown to be involved in signaling pathways of lymphocyte activation and is presumed to function in vivo as important signaling mediators (26, 27, 28, 29, 30, 31). Schneider et al. (26) suggested that TCR can utilize mouse Rlk (as well as ZAP-70) in the phosphorylation of key sites in the adaptor protein, SLP-76, leading to the up-regulation of Th1-preferred cytokine IL-2. Similarly, Rajagopal et al. (27) identified the T cell-specific adaptor protein, RIBP, which binds to mouse Rlk/Txk and modulates production of IL-2 and IFN-γ.

However, information concerning roles of Txk in human T lymphocyte function is limited. We have recently reported that Txk expression is restricted to Th1/Th0 cells with IFN-γ-producing potential, and that Txk transfection resulted in severalfold increase of IFN-γ mRNA expression and protein production by up-regulating IFN-γ enhancer activity specifically (32). This finding prompted us to study a mechanism of Txk to provoke IFN-γ production in humans.

Human Txk cDNA in λ phage was provided by G. W. Litman (University of South Florida, St. Petersburg, FL) (16). Full-length Txk cDNA was ligated into a mammalian expression vector, pME18S (SR-α promoter; provided by K. Maruyama, Tokyo Medical and Dental University, Tokyo, Japan), as described (32).

IFN-γ promoter plus luciferase plasmids were kindly provided by C. B. Wilson (University of Washington, Seattle, WA) and H. A. Young (National Cancer Institute, Frederick, MD) (33, 34).

The IFN-γ promoter mutant plus luciferase plasmids were created using QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA.). Briefly, pIFN-γ promoter −208 plus luciferase was used as a template. The primers containing the desired mutation were extended during PCR cycling by PfuTurbo DNA polymerase (Stratagene). The amplification cycle consisted of one cycle of denaturation (95°C) for 1 min, followed by 18 cycles of denaturation (95°C) for 30 s, annealing for 1 min (55°C), and polymerization for 10 min (68°C). After PCR cycling, the PCR product was treated with Dpn I that is specific for methylated and hemimethylated DNA, and the synthesized nonmethylated DNA containing the mutation was recovered. The resultant mutant vector was used for transformation of Escherichia coli, DH5α. Their sequences have been verified by DNA sequencing. c-Jun expression vector has been reported previously (35).

Purified plasmids were transfected into Jurkat cells by electroporation, as described (30). In brief, 5 μg of pIFN-γ promoter plus luciferase, 5 μg of pRSV-chloramphenicol acetyltransferase (CAT),⁴ and 10 μg of pME18S-Txk (Txk transfection) or pME18S (empty vector; mock transfection) were cotransfected. Forty-eight hours after transfection, the Jurkat cells were stimulated with 1 μg/ml PHA and cultured for various periods. Thereafter, protein assay, luciferase assay, and CAT-ELISA (Roche Diagnostics, Tokyo, Japan) of the cell lysates were conducted (32). IL-2 promoter plus luciferase vector was also included as a control promoter vector. In some experiments, c-Jun expression vector was used to transfect Jurkat cells to obtain control nuclear proteins (35).

The cells were lysed with buffer containing 50 mM Tris, pH 8, 1% Nonidet P-40, 150 mM NaCl, and the protease inhibitors, as described (35). Equivalent amounts of proteins were resolved by SDS-PAGE. Proteins were transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, MA) and blocked with 3.5% BSA. Immunoblotting was performed using goat anti-Txk Ab (Santa Cruz Biotechnology, Santa Cruz, CA) and/or anti-Txk Ab developed by immunizing rabbits with whole Txk protein produced by bacterial cells. Blots were probed with appropriate biotin-conjugated secondary Ab, followed by streptavidin-alkaline phosphatase and detection by chemiluminescence.

Nuclear extracts were prepared from T cells by a modification of the method of Dignam et al. (36). Briefly, cells were homogenized in two cell pellet volumes of 10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂, 1 mM EDTA, 0.5 mM DTT, 10% glycerol, and the protease inhibitors. The resultant nuclear pellet was homogenized in two cell pellet volumes of 20 mM HEPES, pH 7.9, 0.42 M KCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 25% glycerol, and the protease inhibitors. After a 30-min incubation at 4°C, the samples were centrifuged for 20 min and the supernatants were dialyzed against buffer consisting of 20 mM HEPES, 20% glycerol, 0.2 mM EDTA, 0.5 mM PMSF, and 0.5 mM DTT. The protein concentration in the nuclear extracts was determined by Bio-Rad protein assay kit (Bio-Rad, Richmond, CA). In some experiments, human Th1 cell lines were established, as described previously (32), and nuclear proteins of the cells were recovered.

A gel shift assay was performed using digoxigenin gel shift kit (Boehringer Mannheim Biochemica, Mannheim, Germany). In brief, digoxigenin-labeled DNA fragments were incubated at room temperature for 15 min with 5–10 μg of nuclear proteins. Protein-DNA complexes were separated from free probe on a polyacrylamide gel. Thereafter, the gels were electrically transferred to nylon membrane and detected by chemiluminescence. We verified that a 20-fold excess of specific cold oligonucleotide competed the binding of the protein to the digoxigenin-labeled probe, whereas a similar excess from another site would not compete (see Figs. 3 and 4).

The probes were derived from sequences present in the IFN-γ promoter region (33, 34), related promoter regions, and irrelevant promoter regions. Actual DNA sequences synthesized were as follows: IFN-γ gene (designated as IFN −53/−39), −56 to −36 region, ACGTAATCCTCAGGAGACTTC; IFN-γ gene (designated as IFN-irr), −160 to −140 region, AAACTCTAACTACAACACCCA; CCR5 gene (designated as CCR5), −899 to −885 region, CACCAACCGCCAAGAGAGCTT; TNF-α gene (designated as TNF-α), −457 to −437 region, TGGGCCACTGACTGATTTGTG; IL-2 gene (designated as IL-2), −135 to −115 region, AAAGAGTCATCAGAAGAGGAA.

We recently found that Txk expression is restricted to Th1/Th0 cells with IFN-γ-producing potential, and that Txk itself translocates into nuclei and enhances IFN-γ gene transcription in T cells. In this study, we have focused on whether Txk itself or a protein complex including Txk acts as a Th1 cell-specific transcription factor for IFN-γ gene transcription.

It is important to clarify whether Txk protein directly binds to IFN-γ promoter/enhancer region to exert the positive effect for IFN-γ gene transcription. To this end, we labeled IFN-γ promoter −538 with biotin, which was recovered from pIFN-γ promoter −538 plus luciferase vector. Biotin-labeled IFN-γ promoter −538 was reacted with nuclear proteins of Txk-transfected Jurkat cells stimulated with PHA for 1 h. Thereafter, DNA-binding proteins were recovered by streptavidin-Dynabeads (Dynal, Oslo, Norway) and magnet, and analyzed by immunoblotting with anti-Txk Ab. We found that Txk actually binds to IFN-γ promoter −538 region (Fig. 1).

The induction of IFN-γ gene transcription during Th1 cell activation is mediated mainly by a region extending ∼538 bp upstream of the transcription start site (33, 34), and this region contains binding sites for several nuclear proteins (33, 34).

To identify to which element Txk binds for up-regulation of IFN-γ gene transcription, we tested a panel of constructs that contain subfragments of the IFN-γ gene linked to the reporter gene luciferase in transient expression system (Fig. 2). The pIFN-γ promoter plus luciferase vectors were transfected into the Jurkat cells. As a control, we used IL-2 promoter plus luciferase vector. The transfected cells were then treated with PHA for 8 h. pME18S-Txk or empty pME18S vector and pRSV-CAT were cotransfected with the luciferase vector. Treatment of the mock (empty pME18S)-transfected Jurkat cells with PHA increased luciferase activity moderately (Fig. 2). Txk-transfected Jurkat cells induced severalfold more luciferase activity than the mock-transfected Jurkat cells. Txk transfection had no detectable effect on the activity of multimers of NF-κB, AP-1, CRE, and glucocorticoid response element (data not shown). A construct containing the IFN-γ promoter −53 had a reproducible increase in response to Txk transfection with mitogenic activation. In contrast, a construct with the IFN-γ promoter −39 did not respond to Txk transfection (Fig. 2). Similarly, IL-2 promoter −568 plus luciferase did not respond to Txk transfection, as has been shown previously (32).

We next performed a gel shift assay to determine whether Txk binds to this region (−53 to −39). Nuclear extracts from Txk-transfected Jurkat cells stimulated with PHA for 30, 60, and 90 min contained binding activity to a double-stranded oligoDNA that corresponded to −53 to −39 (actually −56 to −36: it is generally known that addition of a few nucleotides to both ends of the core recognition sequence is preferable as a DNA probe of a gel shift assay, and actually the Txk protein binds to the −56/−36 oligoDNA much better than the other three oligoDNA, −56/−42, −53/−39, and −50/−36; data not shown) of the human IFN-γ gene. Unstimulated Txk-transfected (Fig. 3,A) or unstimulated and PHA-stimulated mock-transfected Jurkat cells (data not shown) contained marginally detectable binding proteins. Th1 cell lines were established from normal PBLs, as described (32), and tested in the gel shift assay. The DNA-protein complex appeared in the Th1 cell line when stimulated, as Txk-transfected Jurkat cells (Fig. 3 B).

Competition with a 10- and 20-fold molar excess of unlabeled double-stranded IFN-γ −53 to −39 oligoDNA specifically inhibited the binding of the complex; competition with a 10- and 20-fold molar excess of AP-3 binding site had no detectable effect (Fig. 3 C).

It is possible that the protein-DNA complex that appeared in Txk-transfected Jurkat cells contained Txk protein itself. To prove the possibility, we have performed the gel shift assay with anti-Txk Ab to show the complex reacts with the Ab. We found that the complex was depleted by the treatment of the complex with anti-Txk Ab (5 μg/reaction) but not anti-c-Fos Ab (Fig. 3 D). Thus, it is evident that the DNA-protein complex contained Txk.

To more precisely define the recognition sequences of the complex, we performed competition analysis with oligoDNA that contained contiguous five base substitutions dispersed throughout the −53 to −39 region (Fig. 4, A and B). The 20-bp oligoDNA containing −53 to −49 mutant (designated as 53 M), −48 to −44 mutant (48 M), and −43 to −39 mutant (43 M) were synthesized. Ten and 20 times excess of wild-type and the three mutant oligoDNA efficiently inhibited formation of the labeled oligoDNA/Txk complex. These results suggest that promoter region −53 to −39 is important for the Txk binding.

To confirm the above finding, we constructed mutant IFN-γ promoter plus luciferase constructs. We used pIFN-γ promoter −208 as a wild-type vector. We used site-directed mutagenesis to introduce the mutations into the pIFN-γ promoter plus luciferase construct, and obtained −53 to −49 mutant, −48 to −44 mutant, and −43 to −39 mutant vectors. We found that the −53 to −49, −48 to −44, and −43 to −39 mutants did not respond to the Txk transfection (Fig. 4 C), suggesting that the entire sequence (−53 to −39) is critical for the recognition and function of the Txk. Thus, the Txk protein acts on the −53 to −39 region to up-regulate IFN-γ gene transcription.

To further characterize binding of Txk to IFN-γ gene, biotin-labeled double-stranded IFN-γ −53 to −39 region was synthesized. Nuclear proteins were prepared from c-Jun-transfected Jurkat cells (35) and Txk-transfected Jurkat cells. The nuclear proteins were incubated with the oligoDNA. The binding proteins were recovered by streptavidin-Dynabeads and magnet, and analyzed by immunoblotting with anti-Txk Ab. As shown in Fig. 3 E, c-Jun-transfected Jurkat cells contained undetectable levels of Txk protein bound to the −53 to −39 region. In contrast, Txk-transfected Jurkat cells contained full-length Txk (64 kDa). Thus, the region −53 to −39 is specifically involved in the binding of Txk.

Similar sequences to this DNA-binding motif were found within the 5′ flanking regions of IFN-γ promoter of several mammals and several human Th1 cell-associated protein genes, including CCR5 and TNF-α (Fig. 5). Thus, we have conducted gel shift assays using the double-stranded oligoDNA corresponding to several human Th1 cell-associated protein gene promoters. As shown in Fig. 3 F, the same nuclear protein included binding protein to the IFN-irr region −160 to −140, but the DNA-protein complex did not contain Txk protein in the complex, confirming the specificity of the Txk binding to the region −53 to −39. Similarly, the same nuclear protein contained binding protein to human IL-2 promoter region −135 to −115, but the binding complex did not contain Txk. In contrast, the same nuclear protein bound to the CCR5 and TNF-α promoters, and the complexes were disappeared by the anti-Txk Ab, suggesting that Txk bound to the Th1 cell-associated gene promoters.

In this study, we found that Txk specifically binds to the IFN-γ promoter −53/−39 region to exert a positive effect on IFN-γ gene transcription. Similar sequences of this element are found within the 5′ flanking regions of several Th1 cell-associated protein genes, and Txk also bound to the regions in the gel shift assay. Thus, it is possible that Txk is expressed on Th1/Th0 cells with the IFN-γ production and acts as a Th1 cell-specific transcription factor.

The importance of the proximal element (−71 to −43) for the IFN-γ production has already been reported by C. B. Wilson et al. (37); they suggested that activating transcription factor-2 and c-Jun play an important role in the induction of transcription by this proximal element. The Txk binding region we identified (−53 to −39) partly overlaps with the proximal element (−71 to −43), supporting their finding that the proximal region is important for IFN-γ gene transcription.

Recently, SLP-76 and RIBP were shown to be adaptor proteins of mouse Rlk (26, 27). We found that the −53 to −39 binding protein containing Txk did not contain activating transcription factor-2 (70 kDa) and c-Jun (39 kDa) by the gel shift assay (data not shown and Fig. 3 E). It thus is important to identify the adaptor protein of human Txk.

In addition, they suggested that CpG dinucleotide in the proximal element is selectively methylated in Th2 cells that do not express IFN-γ, and is demethylated in Th1 cells, and that methylation of this region correlates strongly and inversely with the capacity of T cells to express IFN-γ (37, 38). These findings partly support our present result that the region −53 to −39 is the Txk binding site and is important for Th1 cell-specific IFN-γ gene transcription.

Because Txk is a nonreceptor tyrosine kinase, it is important to clarify whether phosphorylation of Txk is involved in the function of Txk. With regard to the kinase activity of Txk, Chamorro et al. (39) demonstrated that Rlk/Txk can be phosphorylated and activated by Src kinases. They suggested that Rlk/Txk is phosphorylated by Src family kinases in response to TCR engagement. We found that phosphorylation of Txk is necessary to exert its positive effect on IFN-γ production.⁵

It has been shown that Rlk/Txk has two isoforms generated by alternative translation start sites in mice (21). As shown in Fig. 3 E, we have recovered the IFN-γ promoter (region −53/−39) binding protein and conducted immunoblotting analysis employing anti-Txk Ab (this Ab was developed against whole Txk protein). We found that the binding protein contained a longer type of Txk preferentially. However, it is possible that small amount of a shorter form of Txk is involved in the binding to IFN-γ promoter.

Our study indicates that Txk can greatly increase IFN-γ enhancer activity as a Th1 cell-specific transcription factor. The region of the IFN-γ enhancer that responds to Txk is absolutely conserved between the human and other mammalian IFN-γ genes, and similar sequences are present in the 5′ flanking regions of several Th1 cell-associated genes. This suggests an important function of this signal transduction pathway and DNA-binding complex involving Txk for the Th1 cell development.