Reciprocal Regulation of C-Maf Tyrosine Phosphorylation by Tec and Ptpn22

C-Maf plays an important role in regulating cytokine production in TH cells. Its transactivation of IL-4 is optimized by phosphorylation at Tyr21, Tyr92, and Tyr131. However, the molecular mechanism regulating its tyrosine phosphorylation remains unknown. In this study, we demonstrate that Tec kinase family member Tec, but not Rlk or Itk, is a tyrosine kinase of c-Maf and that Tec enhances c-Maf-dependent IL-4 promoter activity. This effect of Tec is counteracted by Ptpn22, which physically interacts with and facilitates tyrosine dephosphorylation of c-Maf thereby attenuating its transcriptional activity. We further show that phosphorylation of Tyr21/92/131 of c-Maf is also critical for its recruitment to the IL-21 promoter and optimal production of this cytokine by TH17 cells. Thus, manipulating tyrosine phosphorylation of c-Maf through its kinases and phosphatases can have significant impact on TH cell-mediated immune responses.


Introduction
C-Maf is a leucine-zipper transcription factor and plays an important role in T H cells.It contains an N-terminal transactivation domain, which is connected to the C-terminal DNA binding domain through a hinge domain.It is induced by TCR/CD28 and ICOS signals and preferentially expressed in T H 2, T H 17, T FH (follicular helper T) and Tr1 [1][2][3][4].It directly transactivates IL-4 and is critical for T H 2 differentiation [5].C-Maf also regulates the expansion and maintenance of T H 17 and T FH cells via inducing IL-21 [3].It acts synergistically with Sox5t to induce RORγt to promote T H 17 differentiation but negatively regulates the production of IL-22 [6,7].Moreover, c-Maf is induced by IL-27 and works cooperatively with aryl hydrocarbon receptor to promote the development of Tr1 cells and their expression of IL-10 [8,9].
Tyrosine phosphorylation is a critical regulatory process of signal transduction.It controls various cellular events including cell cycle regulation, cell signaling, and protein trafficking.In addition to cytoplasmic signaling molecules, the activity of a handful of transcription factors is also subject to regulation by tyrosine phosphorylation.We have previously shown that c-Maf can be phosphorylated at Tyr21/92/131 in T H 2 cells.Tyrosine phosphorylation is critical for the recruitment of c-Maf to the IL-4 promoter and the optimal production of IL-4.In addition, T H cells from glycemic NOD mice display attenuated tyrosine phosphorylation of c-Maf compared to those of euglycemic NOD mice [10].Despite these observations, it is unclear how tyrosine phosphorylation of c-Maf is regulated in T H cells and whether this process also plays a role in other T H subsets.
Three Tec kinases, Itk, Rlk and Tec, are highly expressed in T cells and their activity is induced upon antigen engagement [11][12][13].Moreover, Tec kinases are differentially expressed in different T H subsets and play an important role in regulating the function and differentiation of T H cells [14].However, there probably exists functional redundancy among these three Tec kinases.For example, Rlk-deficient mice display a normal T H 1 cytokine profile and marginal defect in T H 1 response against T. gondii infection [15,16]; further, deficiency of Tec has minimal impact on the differentiation of T H 1 and T H 2 cells [17].One of the substrates of Tec kinases is T-bet, a transcription factor that is essential for the differentiation of T H 1 cells.T-bet can be phosphorylated at Tyr525 by Itk [18].The tyrosyl phosphorylated T-bet interacts with GATA-3, preventing GATA-3 from binding to IL-4 promoter [18].It is still unclear whether Tec kinases also act on other transcription factors in T H cells.
Ptpn22, a member of non-transmembrane type protein tyrosine phosphatases (NT-PTPs), is expressed mainly in hematopoietic cells [19].One of its known functions is damping activation signals in lymphocytes via its interaction with various cytoplasmic signaling molecules, including LCK CSK,VAV, and ZAP70 [20][21][22].Accordingly, deficiency of Ptpn22 leads to abnormal expansion of memory/effector T cells and increased antibody production [23].Genome-wide association studies have identified a missense single nucleotide polymorphism of Ptpn22 that is strongly associated with higher risk of several autoimmune diseases, including rheumatoid arthritis and SLE [24][25][26].Although Ptpn22 was originally identified as a cytoplasmic protein, it actually contains a nuclear localization signal and is present in the nucleus of macrophages [27].We have demonstrated that nuclear Ptpn22 is functionally distinct from cytoplasmic Ptpn22 [27].However, it is still unknown as to whether Ptpn22 is also present in the nucleus of T cells and, if it is, what the role of nuclear Ptpn22 is in T cells.
Here, we show that Tec, but not Rlk or Itk, is a tyrosine kinase of c-Maf.We further show that Ptpn22 is also present in the nucleus of T H cells and directly interacts with c-Maf.It counteracts the effect of Tec and dephosphorylates c-Maf.Furthermore, phosphorylation of c-Maf at Try21/92/131 is also critical for optimal expression of IL-21 in T H 17 cells.Our results uncover novel ways to manipulate the status of tyrosine phosphorylation and subsequently the activity of c-Maf.

Materials and Methods Mice
Ptpn22-deficient (Ptpn22 KO) mice in C57BL/6 background have been previously described [23].C57BL/6 mice were purchased from the Laboratory Animal Center of NTUMC or NAR Labs.All animals were housed under specific pathogen-free conditions.This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.The protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of National Taiwan University College of Medicine and College of Public Health (Permit Number: 20100308).Total of 70 C57BL/6 mice (from NTUMC or NAR Labs) and 20 Ptpn22-deficient (Ptpn22 KO) mice (from Harvard University) were used in this study.Water and food were provided sufficiently daily.All mice were killed by CO 2 .

Retrovirus infection
After 40 hours of stimulation with plate-bound anti-CD3 and soluble anti-CD28 antibodies, CD4 + T cells (1 x 10 6 ) were mixed with retrovirus and polybrene (8 μg/ml; final concentration) in a 24-well plate.Cells were centrifuged at 2200 r.p.m for 1 hour at 30°C and then incubated at 37°C for 30 min.Viral supernatant was replaced by complete RPMI 1640 growth medium.After additional incubation for 48 hours, infected (GFP + ) cells were sorted on FACSAria (BD Biosciences, San Jose, CA).

Immunoprecipitation (IP) and Western blotting (WB)
For detection of tyrosine phosphorylated c-Maf, cells (1x10 7 cells) were pretreated with 100 μM pervanadate [derived from sodium orthovanadate (Sigma-Aldrich)] for 10 min then lysed in 500 μl RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS and 1 mM EDTA) containing 1 mM PMSF (Sigma-Aldrich) and complete protease inhibitor mixture [100 μM pervanadate and plus phosphatase inhibitor (Roche)].Appropriate antibody (2 μg) was added to the lysates, which were shaken for 1-3 hours at 4°C.Protein A/G agarose beads (25 μl per tube) were added into antibody-lysate mixture and gently mixed for an additional 2-3 hours at 4°C.The entire mixture was washed three times with cold PBS and analyzed by Western blotting.Cytoplasmic and nuclear extracts were prepared by washing cells with cold PBS and resuspended in hypotonic lysis buffer (10 mM HEPES [pH 7.9], 1 mM MgCl 2 , 10 mM KCl, 0.1% Triton X-100, 20% glycerol, 0.5 mM PMSF, and protease inhibitors) on ice for 10 min.The supernatant, corresponding to the cytoplasmic fraction, was collected by centrifugation at 13,000 X g for 10 min at 4°C.The nuclear pellet was washed with hypotonic lysis buffer and then resuspended in hypertonic lysis buffer (10 mM HEPES [pH 7.9], 400 mM NaCl, 1 m EDTA, 0.1% Triton X-100, 20% glycerol, 1 mM PMSF, and protease inhibitors) and then incubated on ice for 20 min.Nuclear extract was collected by centrifugation.Western blotting was performed according to a previously described protocol [10].Densitometry readings of Western blots were obtained and analyzed with ImageJ program.Degree of c-Maf tyrosine phosphorylation was calculated by dividing the density of phosphorylated c-Maf with the density of total c-Maf.

Luciferase assay
HEK 293T cells (1x10 5 ) were co-transfected with pGL3-Il4-Luc reporter, Renilla reporter plasmid (pTK-RL) and various expression vectors.Twenty-four hours after incubation at 37°C, cells were lysed and analyzed by using the Dual-Glo Luciferase Assay System (Promega, Madison, WI) according to the manufacturer's instructions.Firefly luciferase activities were normalized against Renilla luciferase levels.
Measurement of IL-4, IL-17 and IL-21 production in primary T cells by ELISA CD4 + T cells (1 x 10 6 /ml) were cultured with complete RPMI 1640 growth medium and stimulated with plated-bound anti-CD3 (1 μg/ml) for 24 hours.The supernatant of the culture medium was collected and the concentration of IL-4, IL-17 and IL-21 was measured using Ready-Set-Go kit (eBioscience), according to the manufacturer's instruction.

Confocal microscopy analysis
HEK 293T Cells (4x10 5 cells/well) were cultured on glass cover slips in six-well culture plates containing 10% FBS DMEM for 24 hours and were transfected with vectors expressing DsRed2-c-Maf or DsRed and EYFP-Ptpn22 or EYFP.Twenty-four hours after transfection, cells were fixed, stained with DAPI, and then analyzed with a Leica confocal microscope (Leica TCS SP5, 63X (oil) objective lens).

Chromatin immunoprecipitation (ChIP)
ChIP was performed according to a published protocol [10].Briefly, T H cells (1x10 7 ) were fixed with 1% formaldehyde.Chromatin was sheared by sonication to 200 to 600 base pairs in size.Anti-FLAG antibody (M2; Sigma-Aldrich) and normal mouse IgG (Santa Cruz Biotechnology) were used for immunoprecipitation.The precipitated DNA was analyzed by quantitative PCR using primers corresponding to mouse Il21 promoter and the data are presented as percentage of input DNA.The following primers were used: Il21 promoter forward primer: 5'-TGGTGAATGCTGAAAACTGGA-3'; Il21 promoter reverse primer: 5'-CTAGGTG-TACGTGTGCGTGT-3'

Statistical analysis
Statistical analyses were performed with unpaired, two-tailed Student's t-tests unless stated otherwise.A value of P<0.05 was considered statistically significant.Tyrosine phosphorylation of c-Maf is required for optimal expression of IL-4 [10].Therefore, we were interested in determining whether c-Maf-dependent IL-4 expression was enhanced by Tec.We found that co-expression of Tec increased c-Maf-induced IL-4 promoterluciferase activity.Y3F c-Maf was less active than WT c-Maf and its activity was not further enhanced by Tec (Fig 2A).However, we were not able to detect a reproducible effect of Tec overexpression on endogenous IL-4 production by T H 2 cells partly due to unexpected cytotoxic effects of exogenous Tec.Tyrosine phosphorylation of c-Maf facilitates its recruitment to the IL-4 promoter [10].We found that overexpression of Tec enhanced the binding of c-Maf to the MARE site derived from the IL-4 promoter (Fig 2B).Thus, Tec is a tyrosine kinase of c-Maf and promotes IL-4 expression by facilitating the binding of c-Maf to the IL-4 promoter.

Ptpn22 interacts with and dephosphorylates c-Maf
In a parallel experiment, we used the hinge domain of c-Maf as bait to search for its interacting proteins.We obtained 42 positive colonies encoding fifteen candidates (Table 1).Three of the colonies encoded overlapping peptide fragments matching the PTP domain of Ptpn22 (S2 Fig) .We subsequently expressed HA-tagged c-Maf and FLAG-tagged Ptpn22 in HEK 293T cells and confirmed their physical interaction with co-immunoprecipitation (Fig 3A).To further visualize the subcellular localization of c-Maf and Ptpn22, we co-expressed DsRed2-c-Maf and EYFP-Ptpn22 in HEK 293T cells.Expectedly, c-Maf was present exclusively in the nucleus.Ptpn22 was not only detected in the cytoplasmic membrane but also co-localized with c-Maf in the nucleus (Fig 3B).As c-Maf is located mainly in the nucleus of T cells, our data prompted us to examine whether Ptpn22 was also present in the nucleus of primary T cells.Indeed, a substantial fraction of Ptpn22 of primary T H cells was located in the nucleus and the nuclear

C-Maf undergoes tyrosine phosphorylation in T H 17 cells
C-Maf is also expressed in T H 17 cells [10].To further examine whether c-Maf was also tyrosine phosphorylated in T H 17 cells, primary CD4 + T cells were polarized into T H 17 cells for 72 hours.Cell lysate of the polarized cells was then subjected to immunoprecipitation with either anti-c-Maf or control IgG and immunoblotted with anti-p-Tyr.As shown in Fig 6A , anti-c-Maf but not control IgG was able to precipitate a tyrosyl phospho-protein corresponding to the size of c-Maf.These data indicate that c-Maf is tyrosine phosphorylated in both T H 2 and T H 17 cells.Tyr21/92/131 are the dominant tyrosine phosphorylation sites of c-Maf in T H 2 cells [10].
We then asked whether these three tyrosine residues were also the dominant phosphorylation sites of c-Maf in T H 17 cells.We expressed FLAG-tagged WT or the Y3F mutant of c-Maf in normalized according to Materials and Methods and then against the value obtained with empty expression vector, which was arbitrarily set as 1.Each experiment was done in triplicate.The data shown are mean ± SEM from three independent experiments.** P<0.01.NS stands for not significant.(B) Nuclear extracts were prepared from HEK 293T cells expressing c-Maf and Tec, and subjected to EMSA using a 33 bp biotinylated probe derived from the MARE element (-31 to -64) within the IL-4 promoter (biotin-IL-4 probe).Unlabeled IL-4 probe (cold-IL-4 probe) and anti-c-Maf were added to the indicated lanes.A fraction of the nuclear extract used in lanes 2 and 3 was analyzed with Western blotting using anti-c-Maf (the lower panel).doi:10.1371/journal.pone.0127617.g002Table 1.List of fifteen c-Maf-interacting candidates obtained from yeast two-hybrid experiment.The observation that c-Maf tyrosine phosphorylation and the expression of IL-4 and IL-21 are intact in Ptpn22KO T H cells strongly suggests that there are additional PTPs of c-Maf.Although the ternary structure of PTP domain is conserved among all PTPs, its surface is quite diverse and displays a wide spectrum of electrostatic potential, allowing the grouping of PTPs [31].Phylogenetically, Ptpn22 is closely related to Ptpn12 and Ptpn18, both of which are expressed in T cells according to the Immgen database.It is possible that these two PTPs can also use c-Maf as a substrate and compensate for the loss of Ptpn22.Deficiency of Ptpn12 affected secondary T cell response presumably by acting on Pyk2 but it did not affect the development and primary response of T cells [32].The role of Ptpn18 in regulating IL-4 expression is still unknown.It will be of great interest to determine whether Ptpn12, Ptpn18, and other PTPs also physically interact with c-Maf and regulate its tyrosine phosphorylation.An alternative explanation for the intact IL-4/IL-21 production in Ptpn22KO T H cells is that the basal c-Maf activity and IL-4/IL-21 production in T H 2/T H 17 cells may be relatively independent of Ptpn22, which becomes influential only when Tec kinase expression/activity is enhanced.This alternative scenario raises the possibility that Tec may directly or indirectly recruit Ptpn22.This possibility is being investigated.

Group Name of Gene Frequency
While we were able to co-immunoprecipitate c-Maf and Ptpn22 when both proteins were overexpressed in HEK 293T cells (Fig 3A ), we were unable to demonstrate physical interaction between endogenous c-Maf and Ptpn22 in T H cells.There are several possible explanations for this negative result.Both c-Maf and Ptpn22 are not abundant proteins and their interaction may be very transient.In addition, as discussed above, the interaction between c-Maf and Ptpn22 in primary T H cells may occur only when Tec activity is induced.Thus, detailed analysis of the kinetics of c-Maf phosphorylation and expression of Tec and Ptpn22 in T H cells under different stimulation conditions may eventually establish a definite role of Ptpn22 in regulating the phosphorylation and activity of c-Maf.
Ptpn22 is known to be present in the nucleus of macrophages.However, our data is the first to demonstrate that it is also expressed in the nucleus of primary T cells.Although deficiency of Ptpn22 has little impact on the expression of IL-4, the observation that Ptpn22 interacts with c-Maf in the nucleus indicates that Ptpn22 can modulate the function of T cells by acting on nuclear proteins in addition to regulating the strength of activation signals.Given the strong association of Ptpn22 in several human autoimmune diseases [24][25][26], identifying nuclear substrates of Ptpn22 may shed light on the pathogenesis of those diseases.by the Cell Sorting Core Facility of the First Core Laboratory, National Taiwan University College of Medicine and Sorting Core Facility at NTUH.
Given the high expression of Tec kinases in T cells, we first examined if any of the Tec kinases was a tyrosine kinase of c-Maf.We co-expressed c-Maf with one of the Tec members in HEK 293T cells.All three Tec kinases were capable of auto-phosphorylation (S1 Fig).We detected a high level of tyrosine phosphorylated c-Maf only when Tec, but not Rlk or Itk, was co-expressed (Fig1A).This effect was dependent on the kinase activity of Tec because co-expression of a kinase-dead Tec (Tec KD) did not lead to tyrosine phosphorylation of c-Maf.Moreover, both Tec and Tec KD, but not Rlk or Itk, co-immunoprecipitated with c-Maf (Fig1A).In addition, the dominant Tec-induced phosphorylation sites were located at the N-terminal activation domain (residues 1-118) (Fig1B).Trace phosphorylation was also detected in the hinge domain (residues 119-254).In contrast, no phosphorylation was detected in the C-terminal DNA binding domain (residues 255-370).We subsequently investigated whether Tec phosphorylated c-Maf at Tyr21/92/131.Indeed, mutation of Tyr21 and Tyr92 attenuated Tec-induced tyrosine phosphorylation and mutation of all three residues (Y3F) nearly completely ablated Tec-induced tyrosine phosphorylation, an effect very comparable to that seen with mutation of all 8 tyrosine residues of c-Maf (Y8F in Fig 1C).To determine whether Tec was able to phosphorylate endogenous c-Maf in primary T cells, T H 2 cells were transduced with retrovirus expressing murine Tec or lentivirus expressing shRNA specific for Tec (shTec) (Fig 1D).We then examined the intensity of tyrosine phosphorylation of c-Maf 48 hours after transduction.Similar to the result obtained from HEK 293T cells, tyrosine phosphorylation of c-Maf was enhanced in activated T H 2 cells by ectopically expressed Tec, whereas c-Maf tyrosine phosphorylation was reduced by knocking down Tec (Fig 1D).

Fig 1 .Fig 2 .
Fig 1. Tec induces tyrosine phosphorylation of c-Maf.(A) HA-c-Maf was co-expressed with various C-terminal 3XFLAG-tagged Tec kinase members (Tec, Rlk or Itk) or a kinase dead Tec (Tec KD) in HEK 293T cells.Whole cell extract was harvested after 24 hours and immuno-precipitated (IP) with anti-c-Maf antibody.The immunoprecipitant was then immuno-blotted (IB) with anti-p-Tyr, anti-c-Maf and anti-FLAG antibodies.(B) Tec was co-expressed with EGFP-fused full-length, activation domain (AD), HINGE domain (HINGE), or DNA binding domain (DBD) of c-Maf in EL4 cells.The cells were lysed 24 hours after transfection and cell extract was immunoprecipitated with anti-EGFP antibody.The immunoprecipitant was then probed with 4G10 or anti-EGFP antibody.The various forms of c-Maf were marked with arrowheads.(C) Wild type c-Maf or various c-Maf mutants, in which tyrosine residues were converted to phenylalanine, were co-expressed with Tec in EL4 cells.The transfected cells were lysed 24 hours later and then immunoprecipitated with anti-c-Maf antibody.The immunoprecipitant was then probed with 4G10 or anti-c-Maf antibody.Y8F carries Y-to-F mutation at Tyr21/91/92/97/131/181/341/345.Phosphorylation ratios of c-Maf are also shown.(D) Naïve T H cells were stimulated under T H 2 skewing conditions for 48 hours and then transduced with retrovirus expressing GFP alone (GFPRV mock) or along with murine Tec (GFPRV Tec) or lentivirus expressing shRNA specific for Tec (LV shTec).Fortyeight hours later, the transduced cells were re-stimulated with PMA/ionomycin (P+I) for 5 hr, lysed and immunoprecipitated with anti-c-Maf antibody.The immunoprecipitant was then probed with 4G10 or anti-c-Maf antibody.A fraction of un-precipitated extract was probed with anti-Tec and anti-Tubulin (the bottom two panels).Phosphorylation ratios of c-Maf are also shown.doi:10.1371/journal.pone.0127617.g001

A 5 B 3 C 3 D 3 Emusculus golgi phosphoprotein 3-like 2 F 1 G 1 H 1 I 1 J 1 K 1 L 1 M 1 N 1 O 2 Fig 3 .
Fig 3. Ptpn22 is located in the nucleus of T cells and interacts with c-Maf.(A) HEK 293T cells were transfected with expression vectors encoding HA-c-Maf and/or FLAG-Ptpn22.Whole cell extract was harvested from the transfected cells 24 hours later and immunoprecipitated (IP) with anti-HA or anti-FLAG antibodies.The immunoprecipitant was then Immuno-blotted (IB) with anti-FLAG or anti-c-Maf antibodies.A fraction of the whole cell lysate was also analyzed directly with Western blot analysis using anti-c-Maf and anti-FLAG antibodies (Input).(B) HEK 293T cells were transfected with vectors expressing DsRed2-c-Maf or DsRed and EYFP-Ptpn22 or EYFP.Twenty-four hours later, the transfected cells were stained with DAPI and examined with confocal microscopy.DIC stands for differential interference contrast.(C) In vitro differentiated primary mouse T H 2 cells were left unstimulated or re-stimulated with anti-CD3 antibody for 16 hours or with PMA and ionomycin (P + I) for 6 hours.Cytoplasmic and nuclear extract was separately prepared and analyzed with Western blotting using indicated antibodies.Hsp90 is a cytoplasmic protein whereas Oct1 represents nuclear protein.doi:10.1371/journal.pone.0127617.g003

Fig 4 .Fig 5 .Fig 6 .
Fig 4. Ptpn22 attenuates tyrosine phosphorylation of c-Maf and reduces c-Maf-dependent transactivation of the IL-4 reporter.(A) HEK 293T cells were transfected with plasmids expressing HA-c-Maf, Tec and/or Ptpn22.The transfected cells were lysed and then immunoprecipitated with anti-c-Maf antibody.The immunoprecipitant was then examined with Western blotting using 4G10 or c-Maf antibody.(B) HEK 293T cells were transfected with pGL3-Il4-Luc, pRL-TK, along with vectors expressing c-Maf and escalating levels Ptpn22.Luciferase activity was quantified 24 hours later and normalized according to Materials and Methods.The normalized luciferase activity obtained from cells transfected with empty expression vector was arbitrarily set as 1. (C) HEK 293T cells were transfected with pGL3-Il4-Luc, pRL-TK together with vectors expressing WT c-Maf, Tec, and/or Ptpn22.The normalized luciferase activity was calculated as in (B).(B) and (C) were performed in triplicate.The data shown are mean ± SEM from three independent experiments.* P<0.05, ** P<0.01 and *** P<0.001.doi:10.1371/journal.pone.0127617.g004

Fig 7 .
Fig 7. Normal protein level and degree of tyrosine phosphorylation of c-Maf in Ptpn22-deficient T H 17 cells.(A) HEK 293T cells were transfected with pGL3-Il21-Luc, pRL-TK together with vectors expressing WT c-Maf, Tec, and/or Ptpn22.Luciferase activity was quantified 24 hours later and normalized according to Materials and Methods.The normalized luciferase activity obtained from cells transfected with empty expression vector was arbitrarily set as 1.The data shown are means ± SEM from three independent experiments.(B & C) Wild type (Ptpn22 WT) and Ptpn22 KO T H cells were differentiated into T H 17 cells.The differentiated cells were re-stimulated with anti-CD3 antibody for 6 hr or 24 hr and the expression of IL-21 gene was measured by qPCR analysis (B).Mean and SEM values were obtained from three independent experiments.NS stands for not significant.Whole cell extract was harvested after restimulation with anti-CD3 for 24 hr, and subjected to immunoprecipitation with anti-c-Maf antibody or control antibody.The immunoprecipitant was then probed with 4G10 or anti-c-Maf antibody (C).Phosphorylation ratios of c-Maf are also shown.doi:10.1371/journal.pone.0127617.g007