Abstract
The Polycomb group (PcG) gene products regulate the maintenance of the homeobox gene expression in Drosophila and vertebrates and also the cell cycle progression in thymocytes and Th2 cell differentiation in mature T cells. We herein studied the role of PcG gene bmi-1 product in Th1/Th2 cell differentiation and found that Bmi-1 facilitates Th2 cell differentiation in a Ring finger-dependent manner. Biochemical studies indicate that Bmi-1 interacts with GATA3 in T cells, which is dependent on the Ring finger of Bmi-1. The overexpression of Bmi-1 resulted in a decreased ubiquitination and an increased protein stability of GATA3. In bmi-1-deficient Th cells, the levels of Th2 cell differentiation decreased as the degradation and ubiquitination on GATA3 increased. Therefore, Bmi-1 plays a crucial role in the control of Th2 cell differentiation in a Ring finger-dependent manner by regulating GATA3 protein stability.
The CD4 helper T cell-dependent immune response is controlled by the balance of the generation of Ag-specific Th1 and Th2 cells (1, 2, 3). Th1 cells produce IFN-γ, whereas Th2 cells produce IL-4, IL-5, and IL-13. For Th2 cell differentiation, IL-4R-mediated signal transduction, including STAT6 activation, is required, whereas IL-12-mediated STAT4 activation is important for Th1 cell differentiation (4, 5, 6, 7, 8). Although the role of cytokines in Th1/Th2 cell differentiation have been emphasized, TCR-mediated signaling is also indispensable for Th1/Th2 cell differentiation (4). In particular, the fate of Th1/Th2 cell differentiation appears to be controlled by the TCR-mediated activation of Ras/MAPK cascade (9, 10). We recently found that the Ras-MAPK cascade controls the stability of GATA3 protein through the ubiquitin-proteasome pathway (11). Flavell and colleagues (12, 13, 14) showed that Th1 cell differentiation and Th1 cytokine production are dependent on JNK and the p38 MAPK cascade, respectively. In addition, NF-κB activation has also been suggested to play a role in Th1/Th2 cell differentiation (15, 16, 17).
Several transcription factors that control Th1/Th2 cell differentiation have been revealed. Among them, GATA3 appears to be a key factor for Th2 cell differentiation (18, 19, 20), and T-bet for Th1 (21). Recently, chromatin modification of the IL-4/IL-5/IL-13 locus has been shown to accompany Th2 cell differentiation (22, 23), and it is primarily mediated by GATA3 (17, 24, 25).
The Polycomb group (PcG)3 gene products form heterogeneous and multimeric protein complexes and maintain the early determined gene expression patterns of key developmental regulators such as homeobox genes both in invertebrates and vertebrates (26, 27). Mel-18, Bmi-1, M33, Pc2, Rae-28/Mph1, and Mph2 Ring1A/B constituents of multimeric protein complex similar to the Polycomb-repressive complex (PRC)-1 were identified in Drosophila. Recently, PRC-1 complex, including Ring1B, was shown to possess an activity of histone H2A ubiquitination, which may explain the long-term gene silencing function of the PcG gene products (28, 29). Eed, EzH1, and EzH2 are the members of PRC-2. PRC-2 recruits histone deacetylase (30), while it also possesses an intrinsic histone H3 methyltransferase activity (31, 32, 33, 34), thus suggesting that it likely plays a role in PcG silencing. Mice deficient for individual components of the PRC-1 display severe combined immunodeficiency due to increased apoptosis and the lack of proliferative responses of immature lymphoid cells (35, 36). In mature lymphocytes, PcG gene products play several roles in differentiation and cell fate. Mel-18 controls Th2 cell differentiation through the regulation of GATA3 transcription (37). The EzH2 is involved in the B cell development and controls IgH V(D)J rearrangement (38). More recently, a critical role for Bmi-1 in the maintenance of hemopoietic stem cells (39, 40, 41) and neuronal stem cells (42) has been demonstrated. Among several known PRC-1 PcG genes, a unique role for bmi-1 in hemopoietic stem cells has been indicated (43).
We herein investigated the role of Bmi-1 in the Th1/Th2 cell differentiation and addressed how Bmi-1 regulates Th1/Th2 cell differentiation. Our results indicate that Bmi-1 facilitates Th2 cell differentiation in a Ring finger-dependent manner through the control of GATA3 protein stability.
Materials and Methods
Mice
C57BL/6 mice and BALB/c mice were purchased from Clea. STAT6-deficient mice were provided by Dr. S. Akira (Osaka University, Osaka, Japan) (44). Bmi-1-deficient (bmi-1−/−) mice (35) were backcrossed 12 times with BALB/c. Anti-OVA-specific TCRαβ (DO11.10) transgenic (Tg) mice were provided by D. Loh (Washington University School of Medicine, St. Louis, MO) (45). All mice, including bmi-1−/− × DO11.10 Tg mice, were maintained under specific pathogen-free conditions and then were used at 6–8 wk of age. All animal care was conducted in accordance with the guidelines of Chiba University (Chiba, Japan).
Cell purification
Splenic CD4 T cells were purified using magnetic beads and an AutoMACS Sorter (Miltenyi Biotec), yielding a purity of >98%. Where indicated, CD4 T cells with naive phenotype (CD44low) were isolated from spleens on a FACSVantage cell sorter (BD Biosciences), yielding purity of >98% as described previously (25).
Cell cultures and in vitro T cell differentiation
Freshly prepared CD4 T cells were cultured under Th1, Th2, or neutral conditions as described previously (46). In brief, for Th1 cell differentiation, CD4 T cells were stimulated for 2 days with immobilized anti-TCR mAb (3 μg/ml, H57-597) in the presence of 25 U/ml IL-2 and 100 U/ml IL-12 and anti-IL-4 mAb. For Th2 cell differentiation, immobilized anti-TCR mAb, 25 U/ml IL-2, and 100 U/ml IL-4 were used. Immobilized anti-TCR mAb and 25 U/ml IL-2 were used for neutral conditions. The cells were transferred to new wells and cultured for another 3 days in the presence of only the cytokines present in the initial culture. Sorted DO11.10 Tg CD44lowCD4 T cells were stimulated with antigenic OVA peptide (Loh15, OVA; 323–339, 1 μM) and irradiated BALB/c APCs under Th1 (IL-2 and IL-12 with anti-IL-4 mAb), Th2 (IL-2 and IL-4), or neutral (IL-2) conditions for 6 days in vitro as described previously (47). Where indicated, Th2 culture conditions containing IL-2, IL-4, anti-IL-12, and anti-IFN-γ were used. Th1/Th2 cell differentiation was then assessed by intracellular cytokine staining with anti-IL-4, anti-IFN-γ, and anti-IL-5 or by ELISA as described previously (17). TG40 cells are TCRαβ-negative variant T cell hybridoma cells (48).
Retrovirus vectors and infection
pMx-IRES-GFP, pMxs-IRES-h nerve growth factor receptor (NGFR), and the Plat-E packaging cell lines were provided by Dr. T. Kitamura (University of Tokyo, Tokyo, Japan). cDNA for human GATA3 and GFP-fusion GATA3 were inserted into a multi cloning site of pMx-IRES-GFP and pMxs-IRES-hNGFR, respectively (37). pMx-bmi-1-IRES-GFP was constructed by inserting the full-length bmi-1 cDNA into a multicloning site of pMx-IRES-GFP. The Ring finger-deleted bmi-1 (dRing) and the proline/serine rich region-deleted bmi-1 (dP/S) were generated by PCR-based mutation. The methods for the generation of virus supernatant and CD4 T cell infection on day 2 have been described previously (37). After another 3-day culture, infected cells were harvested, restimulated and stained with PE-conjugated anti-IL-4, allophycocyanin-conjugated anti-IL-5, or allophycocyanin-conjugated anti-IFN-γ. GFP-positive retrovirus-infected cells were sorted with a FACSVantage (BD Biosciences).
ELISA
Sorted GFP-positive retrovirus-infected cells were stimulated immobilized anti-TCR (3 μg/ml) in 96-well flat-bottom plates for 16 h. The production of IL-2, IL-4, IL-5, IL-13, and IFN-γ was measured by ELISA as described previously (25).
Quantitative RT-PCR
Total RNA was isolated using TRIzol reagent (Invitrogen Life Technologies). Reverse transcription was done using Superscript II (Invitrogen Life Technologies). For quantitative real-time PCR, a TaqMan universal PCR Master Mix was used for all reactions (Applied Biosystems), and the ABI Prism 7000 Sequence Detection System was used (47). The primers and TaqMan probes for the detection of mouse GATA3, c-maf, JunB, NF-κB p65, NF-κB p50, NF-AT1, NF-AT2, T-bet, and hypoxanthine phosphoribosyltransferase (HPRT) were purchased from Applied Biosystems. The expression was normalized using the HPRT signal. The data are shown as the relative intensity.
Immunoprecipitation and immunoblotting
Nuclear extracts for the detection of GATA3 and Bmi-1, c-maf, JunB, NF-κB p65, NF-κB p50, NFAT1, and NFAT2 and cytoplasmic extracts for tubulin-α were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagent (78833; Pierce Chemical). Immunoblotting with anti-GATA3 (HG3-31; Santa Cruz Biotechnology), anti-c-maf, anti-NF-AT1, anti-NF-AT2, anti-NF-κB p65, anti-NF-κB p50, and anti-tubulin-α was performed as described previously (25, 47). Anti-JunB (C-11; Santa Cruz Biotechnology), anti-Bmi-1 mouse mAb (229F6; Upstate Biotechnology), anti-GFP mouse mAb (IE4; Medical and Biological Laboratories), anti-Flag mouse mAb (M2; Sigma-Aldrich), and anti-Myc-Tag mouse mAb (PL14; Medical and Biological Laboratories) were also used. Anti-Flag mouse mAb (M2), anti-GATA3 (HG3-31), and anti-Myc-Tag mouse mAb (PL14) were used for immunoprecipitation. For the experiments assessing the physical association between Bmi-1 and GATA3, nuclear extracts of TG40 cells infected with pMx-bmi-1-IRES-GFP and pMxs-IRES-GFP-GATA3-hNGFR were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagent and then were diluted five times with radioimmunoprecipitation assay buffer (1% Nonidet P-40, 0.25% sodium deoxychorate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 μg/ml aprotinin, leupeptin, and pepstain, 1 mM Na3VO4, 1 mM NaF, and 50 mM Tris-HCl (pH 7.4)) before being subjected to immunoprecipitation. Immunoblotting with anti-ubiquitin mAb was described previously (11). In brief, TG40 cells or developing primary Th2 cells were treated with proteasome inhibitor MG132 (20 μM; Sigma-Aldrich) for 2 h, and then, the cells were pelleted, resuspended in radioimmunoprecipitation assay buffer, and lysed on ice for 30 min. The anti-GATA3 immunoprecipitates were subjected to an immunoblot analysis using anti-multiubiquitin mouse mAb (FK2; MBL).
Pulse-chase experiment
TG40 cells or primary developing Th2 cells were washed, preincubated for 30 min in methionine/cysteine-free medium, and pulsed for 30 min with 200 μCi/ml [35S]methionine/cysteine (PerkinElmer). Next, the cells were washed twice with Dulbecco’s modified Eagle’s medium containing nonradioactive 5 mM l-methionine, 3 mM l-cysteine, and 5% FCS and chased in the same medium.
Expression plasmids and transfection
Myc-tagged bmi-1 mutants (pCMV Tag 3B-bmi-1 wt, dRing, and dS/P) were generated by PCR-based mutation. 293T cells were transfected using Fugene reagent (Invitrogen Life Technologies), according to the manufacturer’s protocol.
Luciferase reporter assay
A luciferase reporter assay for IL-5 promoter region was performed using IL-5 promoter (−1200) in the luciferase reporter plasmid pGL3Basic (Promega) (24). Primary developing Th2 cells were used for transfection by electroporation using mouse T cell Nucleofector kit and Nucleofector I (Amaxa). For transfection, 3 μg of promoter reporter construct was used in combination with 2 μg of either a wild-type (wt; pCAGGS-myc-bmi-1 wt) or dRing (pCAGGS-myc-bmi-1 dRing) expression vector or empty pCAGGS vector (49). In addition, 100 ng of a Renilla luciferase reporter vector, pRL-TK (Promega), was added to each transfection as an internal control for transfection efficiency. Thirty-six hours later, cell extracts were prepared and subjected to a luciferase assay using a Dual Luciferase Reporter System (Promega), according to the manufacturer’s instructions.
Results
Th2 cell differentiation is enhanced by the overexpression of Bmi-1
We previously reported that the PcG mel-18 gene product regulates Th2 cell differentiation by controlling GATA3 transcription (37). The aim of this study is to clarify how another PcG gene, bmi-1, controls Th1/Th2 cell differentiation. The mRNA expression levels of bmi-1 and mel-18 were higher in developing Th2 cells than those of developing Th1 cells (data not shown). First, we induced an overexpression of Bmi-1 in developing Th1/Th2 cells and assessed the efficiency of Th1/Th2 cell differentiation. Freshly prepared CD4 T cells from normal C57BL/6 or BALB/c mice were stimulated in vitro with immobilized anti-TCRβ mAb under Th1, Th2, or neutral conditions for 2 days and then they were infected with a retrovirus containing a control (pMX-IRES-GFP) or the bmi-1 gene (pMx-myc-tag-bmi-1-IRES-GFP). Three days after infection, bmi-1-infected cells expressed ∼15-fold increased levels of bmi-1 mRNA and ∼8-fold Bmi-1 protein in comparison with those in mock-infected control cells (data not shown). Fig. 1,A shows the IFN-γ/IL-4 profiles of mock-infected control, bmi-1- or gata3-infected CD4 T cells cultured under Th1, Th2, or neutral conditions. The generation of IL-4-producing cells cultured under Th2 and neutral conditions was enhanced in bmi-1-infected cells (C57BL/6 Th2-conditions: 37.5 vs 55.3%; C57BL/6 neutral conditions: 33.1 vs 48.7%; and BALB/c Th2 conditions: 51.4 vs 66.4%). Under Th1 conditions, the generation of IFN-γ-producing cells was not increased in bmi-1-infected cells, while the generation of IFN-γ/IL-4 double-producing cells was enhanced (C57BL/6: 5.6 vs 9.1%; and BALB/c: 3.4 vs 6.5%). The generation of IL-5/IL-4 double-producing Th2 cells was also increased in bmi-1-infected cells cultured under Th2 conditions (2.9 vs 8.5%) (Fig. 1 B). To confirm the effect of overexpression of Bmi-1, we compared IFN-γ/IL-4 profiles of electronically gated GFP-negative, GFP-low expressing and GFP-high expressing cells and found that the increase in the IL-4-producing Th2 cells appeared to be more prominent in GFP-high expressing cell populations (data not shown), suggesting a dose-dependent control by Bmi-1.
Introduction of bmi-1 enhances Th2 cell differentiation under Th2 conditions. A, Freshly prepared splenic CD4 T cells from C57BL/6 and BALB/c mice were stimulated under the indicated conditions and infected on day 2 with retrovirus encoding bmi-1 or gata3 bicistronically with enhanced GFP. Three days after infection, the cells were restimulated, and intracellular IFN-γ/IL-4 profiles of electronically gated GFP+ populations were determined. Representative results of eight independent experiments are shown with percentages of cells in each area. The control represents infection with an enhanced GFP-containing retrovirus vector. B, Intracellular IL-5/IL-4 profiles of the same cell preparation (C57BL/6) shown in A. Three experiments were performed with similar results. C and D, GFP-expressing infected CD4 T cells prepared as in A were isolated by cell sorting. The cell lysates with a 3-fold serial dilution (0.3 × 106 and 1.0 × 106) were subjected to immunoblotting with anti-GATA3 or anti-tubulin-α Ab. Arbitrary densitometric units are shown under each band. Five independent experiments were performed with similar results. E, GFP-expressing infected CD4 T cells were stimulated with immobilized anti-TCR mAb for 16 h. The amounts of IL-4, IL-5, IL-13, IL-2, and IFN-γ in the culture supernatant were assessed by ELISA. The mean values with SDs of triplicate cultures are shown. Eight experiments were performed. ∗, Not detectable. F, Freshly prepared splenic CD4 T cells from STAT6-deficient (STAT6 KO) mice were stimulated under Th2 conditions, and then retrovirus infection and intracellular staining were performed as described in A. Two independent experiments were performed with similar results.
Introduction of bmi-1 enhances Th2 cell differentiation under Th2 conditions. A, Freshly prepared splenic CD4 T cells from C57BL/6 and BALB/c mice were stimulated under the indicated conditions and infected on day 2 with retrovirus encoding bmi-1 or gata3 bicistronically with enhanced GFP. Three days after infection, the cells were restimulated, and intracellular IFN-γ/IL-4 profiles of electronically gated GFP+ populations were determined. Representative results of eight independent experiments are shown with percentages of cells in each area. The control represents infection with an enhanced GFP-containing retrovirus vector. B, Intracellular IL-5/IL-4 profiles of the same cell preparation (C57BL/6) shown in A. Three experiments were performed with similar results. C and D, GFP-expressing infected CD4 T cells prepared as in A were isolated by cell sorting. The cell lysates with a 3-fold serial dilution (0.3 × 106 and 1.0 × 106) were subjected to immunoblotting with anti-GATA3 or anti-tubulin-α Ab. Arbitrary densitometric units are shown under each band. Five independent experiments were performed with similar results. E, GFP-expressing infected CD4 T cells were stimulated with immobilized anti-TCR mAb for 16 h. The amounts of IL-4, IL-5, IL-13, IL-2, and IFN-γ in the culture supernatant were assessed by ELISA. The mean values with SDs of triplicate cultures are shown. Eight experiments were performed. ∗, Not detectable. F, Freshly prepared splenic CD4 T cells from STAT6-deficient (STAT6 KO) mice were stimulated under Th2 conditions, and then retrovirus infection and intracellular staining were performed as described in A. Two independent experiments were performed with similar results.
The GFP-expressing bmi-1 or gata3-infected CD4T cells prepared as shown in Fig. 1,A were purified by cell sorting, and the levels of GATA3 protein were assessed by immunoblotting with specific Abs. Two- to 3-fold increases in the level of GATA3 protein were detected in the Bmi-1-overexpressing Th2 cells as compared with the expression in the control mock-infected developing Th2 cells (Fig. 1,C). The level of GATA3 in the Bmi-1-overexpressing Th2 cells was lower than that in the GATA3-overexpressing Th2 cells (Fig. 1,C). Under Th1 conditions, the level of GATA3 protein did not increase in the Bmi-1-overexpressing cells in comparison to the expression in the control mock-infected developing Th2 cells (Fig. 1 D).
The GFP-expressing infected CD4T cells were restimulated with immobilized anti-TCRβ mAb, and the levels of cytokine production were assessed by ELISA (Fig. 1 E). The production of Th2 cytokines (IL-4, IL-5, and IL-13) dramatically increased in bmi-1-infected Th2 cells, whereas the production of IL-2 and IFN-γ did not increase in bmi-1-infected Th1 cells. Thus, Bmi-1 appeared to control Th2 cell differentiation, but no obvious effect was observed in Th1 cell differentiation.
We next examined the requirement for STAT6 in the Bmi-1-mediated enhancement of Th2 cell differentiation. The number of Th2 cells generated in bmi-1-infected STAT6-deficient T cell cultures was negligible in comparison with those in the gata3-infected STAT6-deficient T cells (Fig. 1 F), thus indicating that STAT6 is required for the Bmi-1-mediated enhancement of Th2 cell differentiation.
Ring finger domain of Bmi-1 is required for the enhancement of Th2 cell differentiation
To identify the functional domains of the Bmi-1 molecule, we made truncated Bmi-1 molecules with Myc-tag, a dRing and a deleted mutant missing the dP/S (Fig. 2,A), and tested their effect on Th2 cell differentiation. The expression levels of wt, dRing, and dP/S Bmi-1 protein determined by immunoblotting with anti-Myc Ab were comparable (data not shown). The enhancement of Th2 cell differentiation was detected by the overexpression of wt and dP/S mutant but not by the overexpression of dRing (Fig. 2,B). Fig. 2,C shows a summary of the five independent experiments of Fig. 2,B, and the enhancement by dRing was <10% of that by the wt Bmi-1. Concurrently, the production of IL-4 and IL-5 in developing Th2 cells infected with wt and dRing was assessed by ELISA (Fig. 2 D). A substantial enhancement was detected in wt but not dRing-overexpressing Th2 cells. These results indicate that the Ring finger domain of Bmi-1 plays a crucial role in the regulation of Th2 cell differentiation.
Ring Finger domain of Bmi-1 is critical for the enhancement of Th2 cell differentiation. A, Schematic representation of Myc-tagged bmi-1 mutants. wt and two mutants (dRing and dP/S) are shown with the location of the Myc-tag (Myc), Ring finger domain (Ring), Helix-turn-helix like domains (HTH like), and proline/serine-rich region (P/S rich). B, Freshly prepared splenic CD4 T cells were stimulated under Th2 conditions and infected with retrovirus encoding wt, dRing, or dP/S bicistronically with enhanced GFP on day 2. Representative IFN-γ/IL-4 profiles determined as in Fig. 1 are shown. C, The results of five independent experiments as shown in B are summarized. The extent of the enhancement with wt bmi-1 is set as 100%, and the mean relative enhancement (%) in each group is shown with SDs. D, The amounts of IL-4 and IL-5 produced in the culture supernatant of Th2 cells infected with wt or dRing as in B were assessed by ELISA. The mean values with SDs of triplicate cultures are shown. Three experiments were performed with similar results.
Ring Finger domain of Bmi-1 is critical for the enhancement of Th2 cell differentiation. A, Schematic representation of Myc-tagged bmi-1 mutants. wt and two mutants (dRing and dP/S) are shown with the location of the Myc-tag (Myc), Ring finger domain (Ring), Helix-turn-helix like domains (HTH like), and proline/serine-rich region (P/S rich). B, Freshly prepared splenic CD4 T cells were stimulated under Th2 conditions and infected with retrovirus encoding wt, dRing, or dP/S bicistronically with enhanced GFP on day 2. Representative IFN-γ/IL-4 profiles determined as in Fig. 1 are shown. C, The results of five independent experiments as shown in B are summarized. The extent of the enhancement with wt bmi-1 is set as 100%, and the mean relative enhancement (%) in each group is shown with SDs. D, The amounts of IL-4 and IL-5 produced in the culture supernatant of Th2 cells infected with wt or dRing as in B were assessed by ELISA. The mean values with SDs of triplicate cultures are shown. Three experiments were performed with similar results.
The enhanced protein expression of GATA3 in Bmi-1-overexpressing developing Th2 cells
We next assessed the mRNA and protein expression levels of various transcription factors that are known to be involved in Th1/Th2 cell differentiation in the Bmi-1-overexpressing Th2 cells. No significant difference in the expression levels of mRNA for GATA3, c-maf, JunB, NF-κB p65, NF-κB p50, NF-AT1, NF-AT2, or T-bet was observed (Fig. 3,A). In contrast, an increased level of GATA3 protein was detected in the Bmi-1-overexpressing Th2 cells (Fig. 3 B). The increase was ∼2-fold, but it was reproducibly detected. No obvious difference was detected in the protein expression of other transcription factors tested.
Regulation of the stability of GATA3 protein by Bmi-1. A, Freshly prepared splenic CD4 T cells were stimulated under Th2 conditions and infected with retrovirus encoding bmi-1 bicistronically with enhanced GFP on day 2. Three days after infection, the expression levels of GATA3 mRNA were assessed by quantitative PCR assay. The expression was normalized with HPRT expression. Three independent experiments were performed with similar results. B, Control and bmi-1-infected Th2 cells prepared as in A were lysed, and cell lysates with a 3-fold serial dilution (0.3 × 106 and 1.0 × 106) were subjected to immunoblotting using the indicated Abs. Arbitrary densitometric units are shown under each band. Three experiments were performed with similar results. C, A mouse T cell line, TG40 cells, was infected with retrovirus encoding bmi-1, and 2 days later, the GFP-expressing infected cells were sorted and then were cultured for the indicated times with CHX (100 μM). The amount of GATA3 protein was assessed by immunoblotting with anti-GATA3. Arbitrary densitometric units are shown under each band, and the percentages of each point are shown in a graph. Three experiments were performed with similar results. D, Degradation of GATA3 determined by pulse-chase analysis. TG40 cells infected with retrovirus encoding Myc-tagged bmi-1 were labeled with [35S]methionine and [35S]cysteine and chased in medium containing 5% FCS and nonradioactive methionine and cysteine. 35S-Labeled GATA3 protein was visualized by autoradiography. Arbitrary densitometric units are shown under each band, and the percentages of each point are shown in a graph. Three experiments were performed with similar results. E, The introduction of wt bmi-1 inhibited multiubiquitination of GATA3. TG40 cells infected with retrovirus encoding Myc-tagged wt or dRing (dR) bmi-1 were prepared as in C and cultured at 37°C for 2 h in the presence of MG132 (20 μM). Then, the ubiquitination of GATA3 was assessed. GATA3 were immunoprecipitated (IP) with anti-GATA3 Ab, and the levels of ubiquitination was assessed by immunoblotting (IB) with anti-ubiquitin (Ub) Abs (left panel). Arbitrary densitometric units of the major Ub-GATA3 band are shown under each lane. The levels of expression of wt and dRing Bmi-1 and GATA3 were also assessed by anti-Myc and anti-GATA3 Abs, respectively (right panel). Three experiments were performed with similar results. F and G, Luciferase assay for IL-5 promoter. Developing Th1/Th2 cells were transfected with IL-5 reporter construct and wt or dRing bmi-1 on day 4 of the culture. Thirty-six hours after infection, a luciferase assay was performed. Two (F) and three (G) experiments were performed with similar results.
Regulation of the stability of GATA3 protein by Bmi-1. A, Freshly prepared splenic CD4 T cells were stimulated under Th2 conditions and infected with retrovirus encoding bmi-1 bicistronically with enhanced GFP on day 2. Three days after infection, the expression levels of GATA3 mRNA were assessed by quantitative PCR assay. The expression was normalized with HPRT expression. Three independent experiments were performed with similar results. B, Control and bmi-1-infected Th2 cells prepared as in A were lysed, and cell lysates with a 3-fold serial dilution (0.3 × 106 and 1.0 × 106) were subjected to immunoblotting using the indicated Abs. Arbitrary densitometric units are shown under each band. Three experiments were performed with similar results. C, A mouse T cell line, TG40 cells, was infected with retrovirus encoding bmi-1, and 2 days later, the GFP-expressing infected cells were sorted and then were cultured for the indicated times with CHX (100 μM). The amount of GATA3 protein was assessed by immunoblotting with anti-GATA3. Arbitrary densitometric units are shown under each band, and the percentages of each point are shown in a graph. Three experiments were performed with similar results. D, Degradation of GATA3 determined by pulse-chase analysis. TG40 cells infected with retrovirus encoding Myc-tagged bmi-1 were labeled with [35S]methionine and [35S]cysteine and chased in medium containing 5% FCS and nonradioactive methionine and cysteine. 35S-Labeled GATA3 protein was visualized by autoradiography. Arbitrary densitometric units are shown under each band, and the percentages of each point are shown in a graph. Three experiments were performed with similar results. E, The introduction of wt bmi-1 inhibited multiubiquitination of GATA3. TG40 cells infected with retrovirus encoding Myc-tagged wt or dRing (dR) bmi-1 were prepared as in C and cultured at 37°C for 2 h in the presence of MG132 (20 μM). Then, the ubiquitination of GATA3 was assessed. GATA3 were immunoprecipitated (IP) with anti-GATA3 Ab, and the levels of ubiquitination was assessed by immunoblotting (IB) with anti-ubiquitin (Ub) Abs (left panel). Arbitrary densitometric units of the major Ub-GATA3 band are shown under each lane. The levels of expression of wt and dRing Bmi-1 and GATA3 were also assessed by anti-Myc and anti-GATA3 Abs, respectively (right panel). Three experiments were performed with similar results. F and G, Luciferase assay for IL-5 promoter. Developing Th1/Th2 cells were transfected with IL-5 reporter construct and wt or dRing bmi-1 on day 4 of the culture. Thirty-six hours after infection, a luciferase assay was performed. Two (F) and three (G) experiments were performed with similar results.
Regulation of the stability of GATA3 protein by Bmi-1
Consequently, we studied the effect of an overexpression of Bmi-1 on the degradation of GATA3 protein. A GATA3-expressing T cell line, TG40 cells were infected with bmi-1, and the expression levels of GATA3 were assessed in the cells cultured for 1, 2 and 4 h in the presence of cycloheximide (CHX). Without protein synthesis, the expression of GATA3 protein decreased very rapidly in the control TG40 cells, whereas the degradation of GATA3 was inhibited in bmi-1 introduced TG40 cells, particularly at 2- and 4-h time points (Fig. 3,C). We next performed a pulse-chase experiment with 35S-labeling to follow the degradation of GATA3 and found that the degradation of 35S-labeled nascent GATA3 protein was inhibited in bmi-1-introduced TG40 cells (Fig. 3 D). These results suggest that Bmi-1 regulates the stability of GATA3 protein.
We previously reported the expression levels of GATA3 protein to be regulated by the ubiquitination and proteasome-dependent degradation system (11). We examined whether the overexpression of Bmi-1 affects the ubiquitination of GATA3 (Fig. 3,E). TG40 cells were introduced with myc-tagged wt or dRing and then were treated with a proteasome inhibitor MG132 for 2 h. Cell lysates were subjected to immunoprecipitation with anti-GATA3 mAb and immunoblotting with anti-ubiquitin mAb. Total lysates were also used for immunoblotting with anti-Myc, anti-GATA3, and tubulin-α mAb. In wt bmi-1-introduced TG40 cells, the levels of ubiquitinated form of GATA3 dramatically decreased, while the expression levels of GATA3 were increased ∼2-fold (Fig. 3 E). In contrast, the levels of GATA3 ubiquitination did not decrease in dRing-introduced TG40 cells. The increase in the levels in GATA3 protein was also marginal. These results indicate that Bmi-1 regulates the ubiquitination of GATA3 in a Ring finger-dependent manner.
Increased transcriptional activity of GATA3 by Bmi-1 overexpression
To assess the possible increased transcriptional activity of GATA3 by Bmi-1 overexpression, a luciferase assay for IL-5 promoter was performed with bmi-1-overexpressed Th2 cells. Developing Th1/Th2 cells were transfected with IL-5 reporter construct and wt or dRing bmi-1 on day 4 of the culture as described in Materials and Methods. Thirty-six hours after infection, a luciferase assay was performed. The IL-5 promoter activity was increased after the introduction of wt bmi-1 in developing Th2 cells but not in Th1 cells (Fig. 3,F). Moreover, the IL-5 promoter activity significantly increased after the introduction of wt but not dRing bmi-1 in developing Th2 cells (Fig. 3 G). These results indicate that the GATA3 activity in the transactivation of IL-5 promoter is enhanced by the overexpression of Bmi-1 in a Ring finger-dependent manner.
Physical association of Bmi-1 with GATA3 in T cells
To gain insight into the molecular basis for Bmi-1-mediated stabilization and inhibition of GATA3 ubiquitination, we assessed the potential physical interaction between GATA3 and Bmi-1. 293T cells were cotransfected with Flag-tagged gata3 and Myc-tagged bmi-1, and their total lysates were subjected to anti-Flag immunoprecipitation and subsequent anti-Myc immunoblotting. As shown in Fig. 4,A, a substantial amount of Bmi-1 was easily coprecipitated with GATA3, thus indicating the interaction between Bmi-1 and GATA3 in 293T cells. We next wanted to know whether the Ring finger-deleted mutant dRing is able to interact with GATA3. 293T cells were cotransfected with Flag-tagged gata3 and Myc-tagged wt or dRing bmi-1. The association with GATA3 was almost completely abrogated by the deletion of the Ring finger domain of Bmi-1 (Fig. 4,B). Concurrently, TG40 T cells were infected with pMx-Myc-tagged-bmi-1-IRES-GFP and pMxs-GFP fusion-GATA3-IRES-hNGFR. Nuclear extracts of GFP and hNGFR-expressing infected cells were subjected to anti-Myc immunoprecipitation and subsequent anti-GFP immunoblotting. As shown in Fig. 4,C, Bmi-1 was coprecipitated with GATA3, suggesting the interaction between Bmi-1 and GATA3 in TG40 T cells. To identify the Ring finger of Bmi-1 is crucial for the association between GATA3 and Bmi-1 in TG40 T cells, wt or dRing Bmi-1 were coinfected with GFP fusion-GATA3. The association of Bmi-1 with GATA3 was dramatically reduced by the deletion of Ring Finder (Fig. 4 D). Taken together, these results indicate that Bmi-1 is physically associated with GATA3 in 293 T cells and TG40 T lineage cells, and the association is dependent on the Ring finger of Bmi-1.
Bmi-1 is associated physically with GATA3. A, 293 T cells were transfected with expression plasmids encoding Flag-tagged gata3 and Myc-tagged bmi-1. Two days later, immunoprecipitates (IP) with anti-Flag mAb were subjected to immunoblotting (IB) with anti-Myc mAb (left panel). Total lysates were also run in parallel (right panel). Two experiments were performed with similar results. B, 293 T cells were transfected with expression plasmids encoding Flag-tagged gata3 and Myc-tagged wt or dRing bmi-1. Two days later, the amounts of Myc-tagged Bmi-1 (wt and dRing) associated with Flag-tagged GATA3 were assessed as in A. Three experiments were performed with similar results. C, A mouse GATA3 expressing T cell line TG40 was infected with retrovirus encoding bmi-1 bicistronically with enhanced GFP and encoding GFP-fusion GATA3 bicistronically with hNGFR. And then, immunoprecipitates with anti-Myc mAb were subjected to immunoblotting with anti-GFP mAb (left panel). Total lysates were also run in parallel for a control (right panel). Two experiments were performed with similar results. D, The amounts of GFP-fusion GATA3 associated with Myc-tagged Bmi-1 (wt and dRing) were assessed as in C. Three experiments were performed with similar results.
Bmi-1 is associated physically with GATA3. A, 293 T cells were transfected with expression plasmids encoding Flag-tagged gata3 and Myc-tagged bmi-1. Two days later, immunoprecipitates (IP) with anti-Flag mAb were subjected to immunoblotting (IB) with anti-Myc mAb (left panel). Total lysates were also run in parallel (right panel). Two experiments were performed with similar results. B, 293 T cells were transfected with expression plasmids encoding Flag-tagged gata3 and Myc-tagged wt or dRing bmi-1. Two days later, the amounts of Myc-tagged Bmi-1 (wt and dRing) associated with Flag-tagged GATA3 were assessed as in A. Three experiments were performed with similar results. C, A mouse GATA3 expressing T cell line TG40 was infected with retrovirus encoding bmi-1 bicistronically with enhanced GFP and encoding GFP-fusion GATA3 bicistronically with hNGFR. And then, immunoprecipitates with anti-Myc mAb were subjected to immunoblotting with anti-GFP mAb (left panel). Total lysates were also run in parallel for a control (right panel). Two experiments were performed with similar results. D, The amounts of GFP-fusion GATA3 associated with Myc-tagged Bmi-1 (wt and dRing) were assessed as in C. Three experiments were performed with similar results.
Th2 cell differentiation and GATA3 degradation in bmi-1−/− CD4 T cells
Since the results shown thus far were all obtained from so-called overexpression experimental systems, we next investigated the role of Bmi-1 in nonoverexpression systems using primary bmi-1−/− mouse T cells. Bmi-1−/− mice were previously shown to have a defect in hemopoietic cell generation (35). However, we found that a moderate number of CD4 and CD8 T cells were present in the spleen of bmi-1−/− mice with BALB/c background (Fig. 5,A). The cell surface expression of TCRβ, CD3ε, CD69, CD25, CD62L, CD44, IL-7R, IL-4Rα, and common γ-chain (Cγ) on splenic CD4 T cells from bmi-1−/− mice was comparable to those of bmi-1+/− or bmi-1+/+ controls (data not shown). We assessed the capability of bmi-1+/− and bmi-1−/− CD4 T cells to differentiate into Th1/Th2 cells. Naive CD4 T cells (CD4+CD44low) from bmi-1−/− × OVA-specific TCRαβ Tg (DO11.10 Tg) mice were purified by cell sorting and stimulated with 1 μM antigenic OVA peptide and irradiated BALB/c splenocytes under Th1, Th2, or neutral conditions. No significant effect was observed in the generation of Th1 cells under Th1 culture conditions (Fig. 5,B, upper panels). In addition, the generation of IL-4 producing Th2 cells under Th2 culture conditions was not apparently inhibited in bmi-1+/− or bmi-1−/− CD4 T cells (Fig. 5,B, middle panels). However, under neutral conditions, where no exogenous IL-4 was added, the generation of Th2 cells was decreased in a bmi-1 gene dosage-dependent manner (bmi-1+/+: 20.7%; bmi-1+/−: 12.0%; and bmi-1−/−: 1.1%) (Fig. 5 B, lower panels). The number of IFN-γ-producing Th1 cells increased in a dose-dependent manner (bmi-1+/+: 6.1%; bmi-1+/−: 11.7%; and bmi-1−/−: 21.4%).
Th2 cell differentiation and GATA3 degradation in bmi-1−/− CD4 T cells. A, Representative CD4/CD8 profiles of splenocytes of bmi-1+/+, bmi-1+/−, and bmi-1−/− mice with BALB/c background are shown. The yields of splenocytes are shown in boxed numbers. Eight individual mice in each group showed similar results. B, Naive CD4 T cells (CD4+CD44low) from bmi-1−/− × DO11.10 TCRαβ Tg mice were stimulated under the indicated conditions with OVA peptide and irradiated BALB/c APC for 6 days. Th1 (IL-2 and IL-12 with anti-IL-4 mAb), Th2 (IL-2 and IL-4 with anti-IL-12 and anti-IFN-γ), and neutral (IL-2) conditions were used. Intracellular staining was performed with allophycocyanin-conjugated anti-IFN-γ mAb and PE-conjugated anti-IL-4 mAb. The percentages of cells in the each quadrant are shown. Three experiments were performed with similar results. C, The levels of GATA3 protein in developing bmi-1−/− × DO11.10 TCRαβ Tg cells cultured under neutral or Th2 conditions. Cell lysates were subjected to immunoblotting with anti-GATA3 or anti-tubulin-α Abs. The ratios (GATA3/tubulin-α) of the band intensity are shown. Three experiments were performed with similar results. D, Freshly prepared splenic CD4 T cells from BALB/c background bmi-1−/− mice were stimulated with immobilized anti-TCR mAb under Th2 or neutral conditions for 5 days. Representative intracellular IFN-γ/IL-4 profiles are shown. Two independent experiments were performed with similar results. E, The stability of GATA3 protein in developing bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells cultured under Th2 conditions for 6 days shown in A was assessed by the method used in Fig. 3,C. MG132 (100 μM) was added during the chase culture. Three experiments were performed with similar results. F, Degradation of GATA3 determined by a pulse-chase analysis. Developing bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells was used. A pulse-chase experiment was performed as in Fig. 3,D. Three experiments were performed with similar results. G, The levels of multiubiquitination of GATA3 in bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells was assessed as in Fig. 3 E. The levels of GATA3 and tubulin-α protein in total lysates were also assessed by immunoblotting with anti-GATA3 and anti-tubulin-α Abs, respectively (right panel). Two experiments were performed with similar results.
Th2 cell differentiation and GATA3 degradation in bmi-1−/− CD4 T cells. A, Representative CD4/CD8 profiles of splenocytes of bmi-1+/+, bmi-1+/−, and bmi-1−/− mice with BALB/c background are shown. The yields of splenocytes are shown in boxed numbers. Eight individual mice in each group showed similar results. B, Naive CD4 T cells (CD4+CD44low) from bmi-1−/− × DO11.10 TCRαβ Tg mice were stimulated under the indicated conditions with OVA peptide and irradiated BALB/c APC for 6 days. Th1 (IL-2 and IL-12 with anti-IL-4 mAb), Th2 (IL-2 and IL-4 with anti-IL-12 and anti-IFN-γ), and neutral (IL-2) conditions were used. Intracellular staining was performed with allophycocyanin-conjugated anti-IFN-γ mAb and PE-conjugated anti-IL-4 mAb. The percentages of cells in the each quadrant are shown. Three experiments were performed with similar results. C, The levels of GATA3 protein in developing bmi-1−/− × DO11.10 TCRαβ Tg cells cultured under neutral or Th2 conditions. Cell lysates were subjected to immunoblotting with anti-GATA3 or anti-tubulin-α Abs. The ratios (GATA3/tubulin-α) of the band intensity are shown. Three experiments were performed with similar results. D, Freshly prepared splenic CD4 T cells from BALB/c background bmi-1−/− mice were stimulated with immobilized anti-TCR mAb under Th2 or neutral conditions for 5 days. Representative intracellular IFN-γ/IL-4 profiles are shown. Two independent experiments were performed with similar results. E, The stability of GATA3 protein in developing bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells cultured under Th2 conditions for 6 days shown in A was assessed by the method used in Fig. 3,C. MG132 (100 μM) was added during the chase culture. Three experiments were performed with similar results. F, Degradation of GATA3 determined by a pulse-chase analysis. Developing bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells was used. A pulse-chase experiment was performed as in Fig. 3,D. Three experiments were performed with similar results. G, The levels of multiubiquitination of GATA3 in bmi-1−/− × DO11.10 TCRαβ Tg Th2 cells was assessed as in Fig. 3 E. The levels of GATA3 and tubulin-α protein in total lysates were also assessed by immunoblotting with anti-GATA3 and anti-tubulin-α Abs, respectively (right panel). Two experiments were performed with similar results.
We next assessed the levels of GATA3 protein expression in bmi-1+/+, bmi-1+/−, and bmi-1−/− cells cultured under neutral and Th2 conditions (Fig. 5,C). As we expected, the expression levels of GATA3 in the cells cultured under neutral conditions were decreased in a bmi-1 gene dosage-dependent manner (bmi-1+/+: 1.0; bmi-1+/−: 0.6; and bmi-1−/−: 0.2) (Fig. 5,C, left panel). In addition, substantial levels of GATA3 were expressed under Th2 conditions regardless of the bmi-1 deficiency (Fig. 5 C, right panel). It is also noted that the levels of GATA3 protein in bmi-1+/+ cells cultured under neutral condition were lower than those in the cells cultured under Th2 conditions (1.0 vs 2.6). These results indicate that the levels of Th2 cell generation are well associated with the expression levels of GATA3, particularly in the cells cultured under neutral conditions.
To confirm the results of DO11.10 Tg T cell cultures, we isolated CD4 T cells from non TCR Tg bmi-1+/+, bmi-1+/− and bmi-1−/− mice, and stimulated them with immobilized anti-TCRβ mAb under Th2- or neutral conditions (Fig. 5 D). Under neutral conditions, a substantial decrease in the generation of IL-4-producing Th2 cells was observed in bmi-1+/− T cell cultures, and a more prominent decrease was detected in bmi-1−/− T cell cultures. No effect on the generation of Th2 cells was detected under Th2 culture conditions.
We next assessed the levels of degradation of GATA3 protein using Th2 cell populations developed by cultivation for 6 days under Th2 conditions as shown in Fig. 5,B, middle panels. As shown in Fig. 5,E, the degradation of GATA3 protein was enhanced in bmi-1−/− Th2 cells, and the degradation was inhibited by the presence of MG132. Moreover, a pulse-chase experiment in which no CHX or MG132 was added revealed that the degradation of 35S-labeled nascent GATA3 protein was enhanced in bmi-1−/− Th2 cells, particularly at 2.5- to 3-h time points (Fig. 5,F). Finally, the levels of ubiquitination on GATA3 protein in bmi-1−/− Th2 cells were determined and were found to be twice as much as those in bmi-1+/+ Th2 cells (Fig. 5 G). These results indicate that Bmi-1 controls the stability of GATA3 protein and GATA3 ubiquitination in primary Th2 cells.
Discussion
In this study, we demonstrate that the Ring finger of Bmi-1plays a crucial role in the regulation of Th2 cell differentiation. Bmi-1 appears to control the stability of GATA3 protein involving the regulation of ubiquitination and proteasome-dependent degradation of GATA3. This is confirmed in bmi-1−/− Th2 cells, where the degradation of GATA3 protein is enhanced with the increased ubiquitination on GATA3.
The physical association between Bmi-1 and GATA3 is detected in a T cell line, TG40, as well as 293 T cells, and the association is highly dependent on the Ring finger of Bmi-1 (Fig. 4). The ubiquitination of GATA3 is inhibited by the overexpression of wt Bmi-1 but not dRing (Fig. 3,E). The transcriptional activity for the IL-5 gene, one of the most well-established functions of GATA3, is enhanced by the overexpression of wt but not dRing Bmi-1 (Fig. 3, F and G). Bmi-1-mediated enhancement of Th2 cell differentiation is dependent on the Ring finger domain of Bmi-1 (Fig. 2). These findings all point to the importance of the Ring finger of Bmi-1. It is thus likely that the stabilization of GATA3 and the resultant increased GATA3 activity are mediated through the Ring finger-dependent physical association of Bmi-1 with GATA3.
We previously reported that TCR-mediated activation of the Ras-ERK MAPK cascade controls GATA3 protein stability through the ubiquitin-proteasome pathway in developing Th2 cells (11). Mdm2 acts as an E3 ligase for GATA3 ubiquitination. There is no direct evidence indicating that Bmi-1 inhibits the Mdm2-mediated ubiquitination of GATA3 at this time. However, the expression levels of Mdm2 mRNA and Mdm2 protein were not changed in a bmi-1-overexpressing Th2 cells or bmi-1-deficient developing Th2 cells, suggesting that Bmi-1 does not control the Mdm2 expression in Th2 cells (H. Hosokawa and T. Nakayama, unpublished observation). Ring1B has been shown to possess an activity of histone H2A ubiquitination (28, 29), while there is no direct evidence indicating that Bmi-1 molecules have an intrinsic activity of ubiquitination. As a result, although both Ring1B and Bmi-1 have the Ring finger domain and are the members of the PRC-1 complex, these two molecules appear to play different roles in different processes. In any event, it appears to be clear that Bmi-1 is an important molecule that controls the stability of GATA3 protein through the ubiquitin-proteasome pathway in developing Th2 cells.
Enhanced Th2 cell differentiation by the overexpression of Bmi-1 is observed in Th2 cultures where large amounts of GATA3 protein are expressed (Fig. 1). This contrasts with the results of gata3 introduction, i.e., Th2 cells are generated even under Th1 conditions. STAT6 induces GATA3 transcription and is required for the Bmi-1-mediated enhancement of Th2 cell differentiation (Fig. 1,F). Bmi-1 thus appears to be involved in the higher-level regulation of STAT6/GATA3-mediated Th2 cell generation. We detected the enhanced transcriptional activity of IL-5 in primary Th2 cells when Bmi-1 is overexpressed (Fig. 3, F and G). This result also implicates that Bmi-1 regulates a function of GATA3 in primary Th2 cells. As for the role of STAT6 in the Bmi-1-mediated enhancement of Th2 cell differentiation, it seems most likely that STAT6 plays a role indirectly, probably through the transcriptional regulation of the gata3 gene. However, a direct regulation of the Th2 cell differentiation is also likely because STAT6 was reported to bind to a locus control region within the Th2 cytokine gene cluster (50).
Another interesting observation is that the Bmi-1-mediated support of the Th2 cell differentiation was easily demonstrated under neutral conditions, but not Th2 culture conditions, in bmi-1−/− CD4 T cell cultures (Fig. 5, B and D). The protein expression levels of GATA3 were limiting under neutral culture conditions in comparison to the Th2 conditions (Fig. 5 C). Therefore, the contribution of Bmi-1-mediated stabilization of GATA3 appear to be more important in the regulation of Th2 cell differentiation under the conditions where GATA3 expression is not robust (e.g., neutral conditions), which are probably a more physiological setting for Th cell differentiation in vivo.
Our results obtained from Bmi-1 overexpression systems indicate that Bmi-1 facilitates Th2 cell differentiation through a mechanism distinct from the activity to control the expression levels of GATA3 protein. Two- to 3-fold increases in the expression of GATA3 protein were detected in bmi-1-infected cells, but an ∼6-fold increase was detected in gata3-infected Th2 cells (Fig. 1,C). However, the levels of the generation of IL-4-producing cells in the bmi-1-infected Th2 cultures were higher than those in the gata3-infected Th2 cultures. In addition, the production of Th2 cytokines (IL-4, IL-5, and IL-13) was apparently higher in bmi-1-infected cells (Fig. 1,E). We need to await a more precise investigation addressing the molecular mechanisms operating in the Bmi-1-overexpressed developing Th2 cells. However, because Bmi-1 is able to associate with GATA3 in TG40 T cells (Fig. 4), it is very likely that Bmi-1 controls the functional quality of the GATA3 complex through recruiting various molecules, including a component of HAT complex. Another possibility is that the multiubiquitinated GATA3 protein is functionally different from the deubiquitinated GATA3.
We previously reported that a PcG gene product Mel-18 controls the Th2 cell differentiation through the regulation of GATA3 transcription (37). In contrast, we have reproducibly observed that Bmi-1 controls protein stability of GATA3 (Figs. 3, C–E, and 5, E–G). Mel-18 and Bmi-1 thus appear to regulate Th2 cell differentiation in a distinct fashion, i.e., transcription and posttranscriptional protein stabilization of GATA3, respectively. In fact, although both Mel-18 and Bmi-1 are members of the same PRC-1 complex, possess structural similarities, including N-terminal Ring finger and Helix-turn-Helix domains (51), and show similar phenotypes in the posterior transformation of the axial skeleton and severe combined immunodeficiency in deficient mice (35, 36), it has also been noted that some differences exist in the phenotype such as neurological defects in bmi-1−/− mice (35) and smooth muscle defects in mel-18−/− mice (52), and more recently the differential role in the self-renewal of hemopoietic stem cells (43).
Both p16INK4a and p19ARF have been reported to be a target gene of Bmi-1 (53), and the proliferation of immature lymphoid cells and neuronal cells is regulated by the expression levels of Bmi-1 (53, 54). However, this regulation does not appear to play an important role in peripheral CD4 T cells. We detected an increased mRNA expression of p16INK4a in bmi-1−/− T cells (H. Hosokawa and T. Nakayama, unpublished observation), but no significant effect was observed regarding the extent of TCR-mediated early proliferative responses in bmi-1−/− mice (M. Yamashita and T. Nakayama, unpublished observation).
In summary, the results of this study indicate that Bmi-1 plays a crucial role in the control of Th2 cell differentiation by regulating the stability of GATA3 protein in a Ring finger-dependent manner.
Acknowledgments
We are grateful to Dr. Atsushi Iwama for helpful comments and constructive criticisms in the preparation of the manuscript. We thank Hikari Asou, Satoko Norikane, Toshihiro Ito, and Kaoru Sugaya for their excellent technical assistance.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology (Japan) (Grants-in-Aid for Scientific Research in Priority Areas 17016010 and 17047007; Scientific Research B 17390139 and Scientific Research C 18590466; Grant-in-Aid for Young Scientists 17790318; and Special Coordination Funds for Promoting Science and Technology), the Ministry of Health, Labor and Welfare (Japan), the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Immovation (Japan), the Japan Health Science Foundation, the Kanae Foundation, the Uehara Memorial Foundation, and the Mochida Foundation.
Abbreviations used in this paper: PcG, Polycomb group; bmi-1 deficient, bmi-1−/−; CHX, cycloheximide; dP/S, proline/serine rich region-deleted bmi-1; dRing, Ring finger-deleted bmi-1; HPRT, hypoxanthine phosphoribosyltransferase; PRC, Polycomb-repressive complex; Tg, transgenic; wt, wild type; NGFR, nerve growth factor receptor.