The transcription factor C/EBPβ transactivates the IL-4 gene in murine T lymphocytes and facilitates Th2 cell responses. In this study, we demonstrate that C/EBPβ also acts as a repressor of T cell proliferation. By binding to the c-myc promoter(s), C/EBPβ represses c-Myc expression and, therefore, arrests T cells in the G1 phase of the cell cycle. For C/EBPβ-mediated repression, the integrity of its N-terminal transactivation domain is essential whereas the central regulatory domain is dispensable. This central regulatory domain is sumoylated in vivo which leads to an alteration of the activity of C/EBPβ. Whereas sumoylation does not affect the C/EBPβ-mediated activation of the IL-4 gene, it relieves its repressive effect on c-Myc expression and T cell proliferation. Similar to several other transcription factors, sumoylation redistributes nuclear C/EBPβ and targets it to pericentric heterochromatin. These results suggest an important role of sumoylation in adjusting the finely tuned balance between proliferation and differentiation in peripheral T cells which is controlled by C/EBPβ.
The C/EBPs constitute a family of basic leucine zipper (bZIP)5 transcription factors which play decisive roles in proliferation and differentiation of numerous cell types. C/EBPβ is one prominent member of this family which induces a wide array of genes controlling cell differentiation, innate immunity, inflammatory, and acute-phase responses (1, 2, 3, 4). It also exerts a crucial function in tumorigenesis (5, 6, 7, 8, 9). Due to C/EBPβ’s transactivation of the endogenous murine IL-4 gene, we have concluded that C/EBPβ contributes to murine Th2 cell differentiation (10).
C/EBPβ consists of an N-terminal transactivation, a C-terminal DNA-binding domain, and a central regulatory domain, also known as “bipartite negative regulatory domain” (11). The C/EBPβ gene (also designated as LAP in rats, AGP/EBP in mice, and NF-IL6 in humans) is expressed in three isoforms, termed LAP*, liver-activating protein (LAP), and liver-inactivating protein (LIP). The two activating isoforms, LAP* and LAP, differ by 21 extra amino acids in LAP* to which SWI/SNF chromatin remodeling complexes are recruited (12). The short isoform, LIP, spans only part of the central regulatory domain and the bZIP domain and is considered as a dominant-negative version of C/EBPβ (13). LAP may also act as a transcriptional repressor, since in contrast to the predicted activating role of C/EBPβ, IL-6 levels were found to be strongly increased in C/EBPβ-deficient mice (14). Recently, C/EBPβ was identified as a constitutive repressor of cyclin D1 target genes, where cyclin D1 antagonizes C/EBPβ’s repressor activity (15).
c-Myc contributes to various aspects of tumorigenesis (16) by driving proliferation in response to diverse signals, such as growth factors, cytokines, and mitogens. It can function both as a transcriptional activator and repressor, while activating the expression of a number of cell cycle-promoting genes, but suppressing the expression of cell cycle/growth arrest genes (17, 18). Moreover, a failure to induce c-Myc expression in response to mitogenic signaling inhibits quiescent cells from entering the cell cycle (19, 20). Therefore, precise regulation of c-Myc expression is critical for the adjustment of the finely tuned balance between proliferation and differentiation. Several signaling pathways and transcription factors have been defined to regulate c-Myc expression (21). For example, the translocation and formation of active nuclear NF-κB complexes regulated by opposing signaling pathways lead to differential c-Myc expression in B lymphocytes (22, 23). In contrast, several transcription factors, e.g., myelin basic protein 1, CCCTC-binding factor, B lymphocyte-induced maturation protein 1, lymphocyte enhancer factor, PLZF, GATA-1, and also C/EBPα have been implicated to inhibit c-Myc expression, thereby promoting cellular differentiation (24, 25, 26, 27, 28, 29, 30).
The activity of transcription factors can be regulated by protein modification, such as sumoylation (31). Small ubiquitin-related modifier (SUMO) is ligated to lysine residues within the consensus motif Ψ-Lys-x-Glu (where Ψ is a large hydrophobic amino acid, generally isoleucine or valine, and x any residue) which is most likely preceded or followed by a stretch of prolines (32). Four SUMO family members, SUMO-1/Smt3C, SUMO-2/Smt3A, SUMO-3/Smt3B, and SUMO-4 are known to exist in mammals (31). The sumoylation of transcription factors can lead to enhanced activity, but mostly correlates to repression of the activation potential. SUMO-mediated repression can correlate with relocalization of the targeted proteins to promyelocytic leukemia (PML) nuclear bodies (33, 34, 35, 36). Consistent with a role in repression, several negative regulatory domains of unrelated proteins, including C/EBPs, contain so-called “synergy control motifs” which were subsequently identified as sumoylation motifs (32, 37). For example, sumoylation seems to reduce transactivation of C/EBPα on promoters with multiple C/EBP-binding sites, but has a weak effect on promoters with a single-binding site (38). Contrasting with a negative effect of sumoylation, attachment of SUMO-1 to C/EBPε at the corresponding lysine appears to relieve the inhibitory effect of the regulatory domains (37). The only report on sumoylation of C/EBPβ describes the exclusive modification of C/EBPβ-LAP* by SUMO-2 or SUMO-3 which is needed for the repression of the cyclin D1 promoter in mammary epithelial cells (39).
CD4+ Th cells expand and differentiate into large populations of highly active effector T cells and small populations of memory cells. Based on the synthesis of lymphokine subsets, effector Th cells have been subdivided into Th1 and Th2 cells that produce predominantly either IFN-γ, IL-2, and lymphotoxin, or IL-4, IL-5, IL-10, and IL-13, respectively (40). In the case of Th1 cells, the size of T effector cell population is regulated by FasL-Fas-mediated activation-induced cell death, while Th2 effectors exhibit a resistance to this pathway (41).
We demonstrate here that in T cells, C/EBPβ is predominately expressed and activated in Th2 cells in which both transactivating isoforms of C/EBPβ, LAP and LAP*, exert at least two functions. Their activity leads 1) to transactivation of the IL-4 promoter and endogenous IL-4 gene (10, 42) and 2), as we show here, to repression of the c-myc gene, thereby arresting T cells in the G1 phase of the cell cycle and facilitating differentiation along the Th2 lineage. Most importantly, C/EBPβ is modified by SUMO-1 in vivo and our results indicate that sumoylation within the central regulatory domain counteracts C/EBPβ’s transactivation domain (TAD)-mediated repression of c-Myc, whereas transactivation of the IL-4 gene remains unaffected.
Materials and Methods
Preparation and in vitro differentiation of primary murine T cells
DO11.10 TCR transgenic (tg) BALB/c mice (43) were bred under barrier-contained conditions and were used at an age of 6–8 wk. CD4+ T cells were prepared from lymph node cells, passing them over CD4 T cell recovery columns (Cedarlane Laboratories) according to the manufacturer’s instructions. They were cultured at 5 × 105/ml in X-Vivo (BioWhittaker) supplemented with 5% FCS, glutamine, nonessential amino acids, pyruvate (all 2 mM), antibiotics (penicillin, streptomycin), and 5 × 10−5 M 2-ME. Dendritic cells were enriched from spleen cells, prepared as previously described (44) and used as APCs. Th cell polarization was induced as described previously (45). Briefly, purified CD4+ T cells were cultured at 5 × 105/ml with the cognate OVA peptide (0.5 mg/ml) and dendritic cells (5 × 104/ml) under the following conditions: IL-2 (50 U/ml) for Th0 cells, IL-12 (0.3 mg/ml), anti-IL-4 (10 mg/ml) and IL-2 (50 U/ml) for Th1 cells, IL-4 (1000 U/ml) and anti-IFN-γ (5 mg/ml) for Th2 cells. On day 3 of stimulation, cells were transferred to new plates and fresh medium was added. For restimulation, cells were harvested, washed twice with balanced salt solution, resuspended in fresh medium, and stimulated for 6 h on plate-bound anti-CD3 (10 μg/ml) in combination with soluble anti-CD28 (5 μg/ml).
Cell culture and stimulations
Murine CD4+O-tetradecanoylphorbol-13-acetate (TPA; 10 ng/ml; Sigma-Aldrich) and ionomycin (0.5 μM; Calbiochem), indicated as TPA/ionomycin (T/I).
Transient transfection and viral infection
To detect SUMO modification in transient transfections, 293T cells were grown to 50% confluency in plates of 90 mm in diameter and transfected with 10 μg of C/EBPβ-ER constructs and 10 μg of pcDNA, Flag-SUMO, or Flag-ubiquitin constructs (46) by conventional calcium phosphate precipitation method. Cells were induced by Tm 24 h posttransfection and harvested after a further 20 h. Transfections/infections were conducted as described previously (10, 47).
Cultured cells were suspended and stained at 4°C with the various Abs for 20 min in 100 ml of PBS/0.1% (w/v) BSA/0.02% (w/v) sodium azide and analyzed on a FACScan (BD Pharmingen) using CellQuest software. FcγRII were blocked by preincubation with 2.4G2 mAb. Cytokine production was determined by intracellular staining as previously described (48).
LIP = β152n (5′-AATTCGAATTCGCGCCACCATGGCGGCCGGTTTCCCGTTCGC 3′) was cloned in parallel. Mouse C/EBPδ derived from pMEXCRP3 (49) and c-Myc (50) were fused to the modified hormone-binding domain of the estrogen receptor as well and cloned into the EYZ vector. Deletion constructs were generated by exchanging the 355-bp EcoR1-BstB1 fragment of C/EBPβ-ER. For PCR, the following forward primers were used which kept unchanged the Kozak sequence in all clones: β38n (5′-AATTCGAATTCGCGCCACCATGGCCAAGGCGGCCCGCGCCG-3′), β53n (5′-AATTCGAATTCGCGCCACCATGGCCATCGGCGAGCACGAGCG-3′), β83n (5′-AATTCGAATTCGCGCCACCATGGCGCACCACGACTTCCTTTCCG-3′), β105n (5′-AATTCGAATTCGCGCCACCATGGCCTCCGACTACGGTTACGTGA GCC-3′), and the common reverse primer BstB1 (5′-GCGGGTTCGAAGCCCGGCTC-3′).
β131n was cloned as a double-stranded oligo into EcoRI/BstB1 sites of C/EBP-ER (5′-AATTCGCGCCACCATGGCCGCACTCAAGGCCGAGCCGGGCTT-3′) and (5′-CGAACGCCGGCTCGGCCTTGAGTGCGGCCATGGTGGCGCG-3′).
Internal deletion and ΔSUMO clones used LAP-Eco (5′-CTAGAGAATTCGCGCCACCATGGAAG-3′) and βΔ53i-BstB1 (5′-GGGTTCGAACGGCTCGGCGGCGGGGG-3′), βΔ105i-BstB1 (5′-GGGTTCGAACGGCTTCTTGCTCGGCTTGG-3′), or (5′-GGTTCGAAGCCCGGCTCGGCCCGGGGTG-3′).
SUMO PCR products for gene fusion also included EcoRI and NcoI (5′-AGCTGAATTCGCGCCACGGCCTCTGACCAGGAGGCAAAACCTTC-3′) and (5′-AGCTCCATGGCAACTGTTGAATGACCCCCCG-3′).
Underlined portion represents start codon for translation followed in all cases by an alanine-coding triplet and the gene-specific sequence.
Cell cycle and proliferation analyses
For proliferation assays, 2 × 104 EL-4 cells were cultured in 96-well plates in triplicate in complete RPMI 1640 medium. Cells were left untreated or stimulated with Tm for the time points indicated. On days 1, 2, and 3, cells were pulsed with 1 μCi [3H]thymidine/well (ICN Pharmaceuticals) and harvested after 10–16 h. Alternatively, proliferation was measured by trypan blue exclusion and direct cell counts over 3 days. In this case, proliferation was expressed as the percentage of Tm-treated from untreated cells. Cell cycle analyses were performed by using the CycleTEST PLUS DNA Reagent kit (BD Biosciences) according to the manufacturer’s protocol in which cell nuclei were stained with propidium iodide. The percentage in G1, S, and G2-M phases was determined by FACS analysis and evaluation by the program ModFit LT (Verity Soft Ware House).
RNA purification and RNase protection assay (RPA)
Protein purification and detection
For analyses of construct expression after viral infection, total cellular protein was extracted by radioimmunoprecipitation assay buffer (1× PBS without Mg2+ and Ca2+; 1% Nonidet P-40; 0.5% sodium deoxycholate; 0.1% SDS; protease inhibitor mix) and cleared from cellular debris by centrifugation. For detection of SUMO modification, cells were harvested in 200 μl of a 1/3 dilution of high detergent lysis buffer I (5% SDS; 0.15 M Tris (pH 6.7); 30% glycerol) and buffer II (25 mM Tris (pH 8.2); 50 mM NaCl; 0.5% Nonidet P-40 (NP40); 0.5% deoxycholate; 0.1% SDS; 0.1% azide; 0.1 mM PMSF; protease inhibitor mix, Complete; Roche) (46). DNA was sheared by sonification. A total of 10 μl of cell lysates were directly separated by gel electrophoresis and 190 μl were diluted with 1 ml PBS + 0.5% NP40. Cellular debris was precipitated by high-speed centrifugation, and supernatants were subjected to immunoprecipitation (IP) using anti-Flag mAb (M2; Sigma-Aldrich; F-3165) and protein G Agarose (Upstate Biotechnology). The IPs were washed with PBS + 0.5% NP40 three times. All proteins were separated on 10% SDS-PAGE, and immunodetected by using rabbit anti-estrogen receptor (ER) polyclonal Ab (MC-20; Santa Cruz Biotechnology; sc-542X) for exogenous and mouse monoclonal IgG2a anti-C/EBPβ (H-7; sc-7962) for endogenous C/EBPβ protein. Nuclear protein preparation and EMSA were performed as described (10). C/EBP-specific AGATTGCGCAATCTGCA oligo was used as a probe and competitor (wild type (WT)) along with C/EBP-mutant ATTGCAGAGACTAGTCTCTGC and Oct-specific CTAGAGCAGAAATGCAAATTATACCCG oligos.
Chromatin IP (ChIP) assays
Proteins were cross-linked to DNA by adding formaldehyde directly to the culture medium to a final concentration of 1% for 5 min and quenching with glycine to a final concentration of 0.125 M for 5 min at room temperature and harvested. To collect nuclei, cells were washed with 0.1% BSA/PBS, resuspended in swelling buffer (25 mM HEPES (pH 7.8); 10 mM KCl; 1.5 mM MgCl2; 0.1% NP40; protease inhibitor mix), kept on ice for 40 min, vortexed in the presence of 0.5% NP40, and centrifuged at 600 × g. Nuclei were suspended in sonification buffer (50 mM HEPES (pH 7.9); 140 mM NaCl; 1 mM EDTA; 1% Trition X-100; 0.1% sodium deoxycholate, 0.1% SDS; protease inhibitor mix) and chromatin was sheared by sonification. After preclearing by using 3 μg of unrelated Ab, IPs were performed using anti-ER rabbit polyclonal (MC-20; Santa Cruz Biotechnology; sc-542X) and salmon sperm DNA/protein A agarose (Upstate Biotechnology). IPs were washed twice with the following buffers: sonification, high salt (sonification buffer containing 500 mM NaCl), LiCl (20 mM Tris (pH 8.0); 250 mM LiCl; 1 mM EDTA; 0.5% NP40; 0.5% sodium deoxycholate), and TE (20 mM Tris (pH 8); 1 mM EDTA). Chromatin was eluted with elution buffer (50 mM Tris (pH 8.0); 1 mM EDTA; 1% SDS) by heating at 65°C for 15 min. Cross-links were removed by adding NaCl to a final concentration of 210 mM in rotating samples at 65°C overnight. Proteins were removed by proteinase K treatment and extensive phenol/chloroform extraction. After EtOH precipitation, DNA was subjected to PCR (25–30 cycles) including 2 μCi [32P]dCTP per reaction. PCR products were separated by native 6% PAGE followed by exposure to x-ray films. The PCR primer pairs used for mIL-4 were (5′-AAAGGCCGATTATGGTGTAATTTC-3′) and mIL-4 reverse (5′-CAATAGCTCTGTGCCGTCAGTG-3′) giving raise to 320 bp of the proximal IL-4 promoter and murine cMyc for (5′-ACCGTACAGAAAGGGAAAGGA-3′) and murine cMyc reverse (5′-GTCAGAAAAAAACGCCCGAA-3′) giving raise to 464 bp of the proximal c-myc promoter.
EL-4 cells stably expressing C/EBPβ proteins upon Tm treatment were cytospun (Cytospin 2 centrifuge; Shandon) on siliconized slides for 3 min at 300 rpm. Cells were fixed in acetone for 10 min at −20°C, rehydrated in PBS, and incubated with 5% donkey serum to avoid unspecific staining. They were stained with a rabbit polyclonal anti-ERα (MC-20) Ab (1:200 in PBS; Santa Cruz Biotechnology; sc-542X) and/or a rabbit polyclonal anti-trimethyl-Histone H3 (Lys9) (Upstate Biotechnology no. 07-442) as well as 4′,6′-diamidino-2-phenylindole (DAPI). The Abs were prelabeled with Alexa Fluor 555 (anti-ER) or Alexa Fluor 488 (anti-H3K9) by Zenon Tricolor, rabbit IgG labeling kit no. 1 (Molecular Probes). The slides were mounted with Fluoromount-G (Southern Biotechnology Associates) and images were collected with confocal microscope (Leica TCS SP2 equipment).
C/EBPβ decreases the number of CD4+ cells
Previously, we showed that the transcription factor C/EBPβ promotes IL-4 expression, the leading lymphokine in Th2 cells (10). When we monitored C/EBPβ-mRNA expression in primary CD4+ cells, differentiating in vitro, we detected its expression in all three subtypes, namely Th0, Th1, and Th2 cells upon 6 h restimulation. However, in Th2 cells, C/EBPβ mRNA expression was already observed before restimulation; additionally, it prolonged with extended restimulation. (Fig. 1⇓A). Furthermore, at the peak of expression, i.e., 6 h of restimulation, the DNA-binding activity of the C/EBPβ protein was much stronger in Th2 than in Th1 cells (Fig. 1⇓B). Accordingly, C/EBPβ increased IL-4 expression in primary Th2 cells, and it slightly reduced IFN-γ in Th1 when exogenously expressed by retroviral infection. In addition, C/EBPβ decreased the number of both Th1 and Th2 cells to ∼25%, measured by the ratio of C/EBPβ- to vector-transduced cells (Fig. 1⇓C).
C/EBPβ arrests T cells in the G1 phase of the cell cycle
To elucidate the cause of T cell loss upon C/EBPβ expression and activation, C/EBPβ was expressed as chimeric LAP (or dominant negative/LIP) protein linked to a modified estrogen receptor in T lymphoma cells. Therefore, chimeric C/EBPβ isoforms were constitutively expressed, but silenced by dimerization with heat shock protein 70 in cytoplasm which gave us the opportunity to establish a stable pool of infected cells. Upon Tm treatment, chimeric C/EBPβ was translocated into the nucleus, as shown for EL-4 T cells (Fig. 2⇓A).
According to the data with primary CD4+ cells (Fig. 1⇑C), C/EBPβ repressed proliferation of all tested T and B cell lines, such as human A3.01, murine EL-4, or A20J cells. Monitoring stably transduced EL-4 cells for proliferation over 3 days of Tm activation, thymidine incorporation diminished. The repressive effect of C/EBPβ on T cell proliferation occurred only when C/EBPβ-LAP (or LAP*, data not shown) was expressed, but not in cells transduced with C/EBPβ-LIP or EGZ vectors (Fig. 2⇑B). Inhibition of proliferation was due to an arrest in G1 phase since a 2-fold increase was detected in the number of cells in the G1 phase of the cell cycle after 3 days of Tm treatment. Moreover, a minor population of cells was found to be in sub-G1 phase, indicating that they had undergone apoptosis (Fig. 2⇑C).
C/EBPβ binds to the c-Myc promoter and impairs its activity
Because c-Myc expression is strongly linked to proliferation, and G1 arrest was shown to be induced by factors which down-regulate c-Myc, RPAs were performed for c-Myc and further E-box-binding factors. Indeed, C/EBPβ-transduced EL-4 cells (and all other T and B cell lines tested, data not shown) lost c-Myc RNA expression upon Tm-mediated LAP induction (Fig. 3⇓A, lane 6), compared with uninduced cells (lane 5) and Tm-treated, but vector-transduced, cells (lane 2) which expressed c-Myc RNA at normal levels. In contrast, RNA expression of max dimerization protein 4 which is negatively controlled by c-Myc (51) was found to be increased upon Tm treatment of C/EBPβ-transduced cells (lane 6). Furthermore, we tested expression of cell cycle inhibitors, revealing p27kip1, a negatively regulated c-Myc target (52), along with the inhibitors p18INK4C and p19INK4D, to be increased in both human A3.01 and murine EL-4 T cells (data not shown).
To verify the down-regulation of c-myc as a consequence of C/EBPβ activation and not as an artifact of the ER-containing construct, C/EBPβ without the ER-fusion partner was used for infection experiments. As expected, we were unable to establish stable cells, indicating that C/EBPβ itself has a negative effect on proliferation. If the limited amount of cells enriched for transgene expression over 3 days after infection was analyzed for c-Myc expression, the population containing C/EBPβ without ER (Fig. 3⇑B, lane 3) showed repressed c-Myc expression like the Tm-treated C/EBPβ-ER-infected pool (lane 2). Also max dimerization protein 4 expression was clearly up-regulated.
To determine whether the loss of c-Myc is causally related to the inhibition of proliferation upon C/EBPβ induction, double infection experiments were performed. c-Myc (and C/EBPδ) was cloned into the EYZ vector that expresses the “yellow” (instead of the “green”) fluorescence protein. Cells doubly infected and expressing both green and yellow fluorescent proteins were sorted by FACS and monitored for proliferation over 3 days. As shown in Fig. 3⇑C, c-Myc-expressing virus was able to cure the negative effect on proliferation by C/EBPβ. This indicates that the suppression of c-Myc by C/EBPβ is sufficient to achieve G1 arrest of the cell cycle. In contrast to C/EBPβ, retroviral expression of C/EBPδ was without any effect on proliferation, and C/EBPβ dominated the phenotype over C/EBPδ.
Inhibition of c-Myc expression could be due to a direct or an indirect effect of C/EBPβ on the c-myc promoter. To test the in vivo situation, ChIP assays were performed. The IL-4 promoter which is targeted by C/EBPβ via defined sequence motifs (10), was included as control. The results of these assays show that upon Tm induction of EL-4 cells for 2 h, C/EBPβ binds to both the c-myc and the IL-4 promoter in vivo (Fig. 3⇑D). This identifies the c-myc promoter as a direct target for C/EBPβ. However, while double treatment of cells by Tm and T/I did not affect C/EBPβ binding to the IL-4 promoter, T/I suppressed C/EBPβ recruitment to the c-myc promoter. This suggests that complex formation of C/EBPβ with the c-myc promoter DNA which inhibits its activity differs from C/EBPβ complex formation with the IL-4 promoter which leads to promoter activation (10).
An intact TAD of C/EBPβ is necessary to repress c-myc, but not to support IL-4 promoter activity
Analyses of proliferation and c-Myc expression revealed the equivalence of LAP and LAP* (for which the construct carried the WT 21 aa added in front of LAP) since either transactivating isoform exerted a strong inhibitory effect on proliferation and c-Myc expression (data not shown). The inability of LIP, the short dominant-negative version of C/EBPβ, to inhibit proliferation (Fig. 2⇑B) indicates that DNA binding alone is insufficient for suppression of c-Myc and proliferation by C/EBPβ. Now we wanted to know whether C/EBPβ’s N-terminal TAD or its central regulatory domain are responsible for its repressive activity. Therefore, successive N-terminal deletion mutants of C/EBPβ were expressed as ER-fusion proteins. The deletions were mainly constructed according to defined subdomains or modules in C/EBPβ (53). All constructs were retrovirally transduced into EL-4 cells and found to be properly expressed (Fig. 4⇓A). Measurements of proliferation in the presence of Tm indicated that inhibition of proliferation was strongly reduced as soon as part of C/EBPβ’s TAD was deleted (Fig. 4⇓B). To determine whether the decrease in proliferation was reflected in c-Myc suppression, each transduced population was left untreated or stimulated by Tm followed by RNA extraction and RNase protection. Fig. 4⇓C shows that already deletion clone β-38n which lacks only 16 aa of TAD was unable to repress c-Myc RNA levels. This was in noted contrast to C/EBPβ-mediated IL-4 induction in parallel EL-4 cell cultures which were induced by Tm and T/I. As long as the C-terminal part of C/EBPβ’s TAD was present, IL-4 expression was strongly enhanced by C/EBPβ (Fig. 4⇓C).
To define the contributions of C/EBPβ’s TAD and regulatory domains in repression of c-Myc, two internal deletion clones were also constructed targeting either the C terminus of TAD and the regulatory domain (clone Δ53i), or the regulatory domain alone (Δ105i). Although in both clones ∼50 N-terminal aa residues of the regulatory domain were deleted, the TAD remained intact in clone Δ105i, but was shortened by 50 aa at the C terminus in Δ53i. Both constructs were properly expressed (Fig. 4⇑A, β-Δ53i and β-Δ105i), and clone Δ105i with a complete TAD, but not Δ53i, suppressed proliferation and c-Myc expression ((Fig. 4⇑, B and C). On the same hand, clone Δ53i was a poor inducer of IL-4 RNA synthesis in EL-4 cells stimulated with T/I plus Tm, compared with clone Δ105i which stimulated IL-4 expression as well as C/EBPβ-LAP (Fig. 4⇑C). These data demonstrate again that an intact TAD, but not the central regulatory domain is necessary for inhibition of c-Myc expression and proliferation, whereas IL-4 transactivation is controlled mainly by the C-terminal amino acid residues of C/EBPβ’s TAD.
C/EBPβ is SUMOylated in vivo
Computer-assisted searches defined a consensus site for sumoylation within the central parts of all C/EBP proteins, including C/EBPβ (32). In accordance with experimental data published for other proteins, sumoylation was shown to interfere with the activity of C/EBPα and C/EBPε (37, 38). The putative sumoylation site of C/EBPβ is located within its regulatory domain around lysine 134 (Fig. 5⇓A) and, therefore, in a region which we found to be dispensable for repression.
Initially, we asked whether sumoylation occurs on endogenous C/EBPβ in vivo. And indeed, CD4+ cells, but only when extracted with high detergent buffer, clearly gave rise to an additional higher molecular mass band of the expected size (Fig. 5⇑B). To test whether this additional band corresponds to sumoylated C/EBPβ, 293T cells were cotransfected with vectors expressing chimeric C/EBPβ-ER protein and pcDNA, Flag-tagged ubiquitin or Flag-tagged SUMO, respectively. The transfected cells were stimulated by Tm after 24 h and lysates were prepared with high detergent buffer after another 24 h. Besides the unmodified (65 kDa) C/EBPβ-ER band, again a slower migrating band was detected in lysates from cells cotransfected with the control vector (Fig. 5⇑C, lane 0) as well as the Flag-ubiquitin vector (lane 1). When Flag-SUMO vector was cotransfected, the upper band occurred as a double band (lane 2, left panel), the most upper of which could be immunoprecipitated by an anti-Flag Ab (lane 2′, right panel), indicating the other band as C/EBPβ-ER protein modified by endogenous SUMO. No modified C/EBPβ could be immunoprecipitated by anti-Flag Ab from lysate of cells transfected with the Flag-ubiquitin vector (lane 1′, right).
To verify the predicted sumoylation site within C/EBPβ, the motif around position 134 was mutated. To erase the acceptor site without altering its charge, lysine 134 was changed to arginine (Fig. 5⇑A). This SUMO-site mutant (βΔSUMO) and both the N-terminal deletion construct β131n (which just includes this site) and LIP (starting at aa 152) were tested in cotransfection assays with Flag-ubiquitin and Flag-SUMO. Clearly, only transfection of β131n (lanes 5 and 6) and not the LIP construct (lanes 7 and 8) resulted in one and, in cotransfections with Flag-SUMO, in two slower migrating bands. The upper of the two bands could be immunoprecipitated by anti-Flag Ab (lane 6′). Even more convincingly, in cotransfections with a βΔSUMO construct, no slower migrating bands appeared or could be precipitated after cotransfection with a Flag-SUMO construct (lanes 3 + 4 and 3′ + 4′). These results demonstrate that C/EBPβ harbors an unique sumoylation site at position 133-136 which can be sumoylated in vivo.
SUMOylation suppresses the C/EBPβ-mediated inhibition of proliferation
Next, we examined the effect of sumoylation on C/EBPβ function. To test loss of function, the construct βΔSUMO (Figs. 5⇑A and 6⇓A) was stably transduced into EL-4 cells. To test gain of function, EL-4 cells were created which expressed chimeric proteins in which SUMO protein was either covalently attached to the N terminus of WT C/EBPβ (giving rise to SUMO-β) or to βDSUMO (SUMO-βΔS) (Fig. 6⇓A). The cell lines expressed equal levels of the chimeric C/EBPβ proteins (data not shown). To verify sumoylation of stably transduced C/EBPβ, EL-4 cells were extracted with high detergent buffer. Upon Tm treatment, sumoylation could be detected for the WT-C/EBPβ, but not for the βΔSUMO version (Fig. 6⇓B).
A strong inhibition of proliferation was detected for βΔSUMO-transduced cells, whereas the SUMO-fusion proteins SUMO-β and SUMO-βΔS were unable to arrest proliferation (Fig. 6⇑C). Similarly, βΔSUMO also resulted in a pronounced c-Myc repression, while neither of the two SUMO-C/EBPβ-fusion proteins was able to repress c-Myc. The SUMO-C/EBPβ-fusion proteins did not affect IL-4 expression, whereas βΔSUMO had a moderately increased stimulatory potential on IL-4 expression compared with WT-C/EBPβ (Fig. 6⇑D). These data indicate that sumoylation of C/EBPβ relieves the negative effect on c-Myc expression and proliferation, but exerts only a weak, if any, positive effect on IL-4 expression.
SUMOylation targets C/EBPβ to heterochromatin
Sumoylation of transcription factors can alter their localization, and several sumoylated transcription factors have been shown to accumulate in specific subnuclear structures, described as PML nuclear bodies (33, 34, 35, 36). To investigate whether in T cells sumoylation leads to the accumulation of C/EBPβ in nuclear bodies, EL-4 cells expressing WT-C/EBPβ, βΔSUMO, and chimeric SUMO-C/EBPβ proteins were analyzed by immunofluorescent confocal microscopy. Prior stimulation, all four proteins were found in cytoplasm of EL-4 cells (Fig. 2⇑A), whereas after activation by Tm C/EBPβ proteins accumulated within the nucleus. The main part of WT-C/EBPβ was found to be evenly distributed, while a minor portion concentrated in speckles (Fig. 7⇓). Remarkably, a clear-cut difference in subnuclear distribution was observed for SUMO-modified C/EBPβ proteins on the one hand and βΔSUMO, the C/EBPβ protein bearing a mutated SUMO-site, on the other hand. Whereas the nonsumoylated βΔSUMO protein performed a totally diffuse pattern, both chimeric SUMO-C/EBPβ proteins (only SUMO-WT shown) appeared to be localized exclusively in speckles. Therefore, the nuclear distribution pattern of WT-C/EBPβ must be due to a mixture of sumoylated and nonsumoylated protein. This is in line with the observation that, in EL-4 cells, a minor portion of the C/EBPβ protein is sumoylated after Tm treatment (Fig. 6⇑B). Because sumoylation can target proteins to PML nuclear bodies, we counterstained with an Ab directed against PML. However, we did not detect any colocalization of C/EBPβ with PML and therefore PML nuclear bodies (data not shown). On the contrary, when cells were counterstained with DAPI, whose bright fluorescent condensations correspond to pericentric heterochromatic DNA, it became evident that sumoylated C/EBPβ localizes to heterochromatin (Fig. 7⇓). Because pericentric heterochromatin is selectively enriched for H3-K27 monomethylation and H3-K9 trimethylation (54), we verified the DAPI-bright staining by the use of an Ab specific for trimethyl H3-K9. As demonstrated in Fig. 7⇓, DAPI-bright staining and the punctuated pattern achieved by trimethyl H3-K9 are the same. More importantly, the speckles formed by sumoylation of C/EBPβ, either endogenously of the wt-C/EBPβ or exogenously by the fusion construct SUMO-β, overlap with pericentric heterochromatin indicated by trimethyl H3-K9. We conclude that at least in T cells endogenous C/EBPβ is in part sumoylated, drawn to pericentric heterochromatin thereby allowing proliferation.
C/EBPβ can function as a transcriptional repressor
We show here that the transactivating C/EBPβ isoforms, LAP (and LAP*), act as repressors for c-Myc expression. In contrast to the repressive influence of LIP which lacks a TAD and acts as a dominant-negative version for LAP/LAP*-mediated transactivation (13, 55), we identified an intact TAD to be essential for LAP-mediated repression. Although the central regulatory domain is dispensable for repression, it interferes with TAD-mediated inhibition when sumoylated. In contrast, transactivation of the IL-4 promoter remained unaffected by sumoylation. This suggests TAD-mediated repression and its SUMO-mediated relief as distinct and individual processes in controlling C/EBPβ activity.
Recently, a MAPK phosphorylation site located within C/EBPβ’s central regulatory domain has been identified to control the interaction of C/EBPβ with transcriptional mediator complexes (56). This site is located at threonine (189) within LIP/β152n which is unable to repress c-Myc and proliferation. This indicates that MAPK phosphorylation is insufficient and the regulatory domain dispensable for suppression by C/EBPβ. However, it cannot be ruled out that MAPK phosphorylation can function in conjunction with TAD.
The increase of IL-6 expression in C/EBPβ-deficient mice suggested that C/EBPβ-LAP, in addition to its transactivating capacity, might also function as a repressor. However, it remained unclear whether C/EBPβ-LAP or, rather, C/EBPβ-LIP binds to the IL-6 promoter in vivo, thereby repressing its activity, or whether C/EBPβ-LAP activates a transcriptional repressor of the IL-6 promoter. For the C/EBPβ-mediated c-Myc repression in T cells, we demonstrated that C/EBPβ-LAP targets the c-myc gene itself, albeit the ChIP technique applied could not distinguish between direct and indirect binding to DNA. A recent study revealed C/EBPβ as a binding site-dependent and direct repressor of cyclin D1 target genes (15). However, the c-myc promoter lacks any C/EBP consensus-binding site. Studies on c-Myc repression by C/EBPα in granulocytes defined the E2F-binding site located between the two promoters of c-myc as essential, but failed to detect any binding of C/EBPα to this site (30). To test this E2F site directly, we performed transient transfection assays with luciferase reporter constructs controlled by c-myc promoter/-Exon 1, but were unable to detect any inhibition of luciferase activity upon C/EBPβ activation (data not shown). Therefore, we suggest that C/EBPβ-mediated suppression of c-Myc in T cells is mechanistically distinct from C/EBPα-mediated repression of c-Myc in granulocytes.
Possible mechanism of repression
There are several transcription factors which can function both as activators or repressors. For example, the c-Myc protein activates expression of cell cycle promoting genes, but suppresses expression of cell cycle/growth arrest genes (17, 18). These opposing functions are achieved by different complex formation (57). For C/EBPβ-mediated regulation in T cells, the direct binding to its consensus sequences within the IL-4 promoter leads to transcriptional activation (10). As a mechanism for repression, one could assume that C/EBPβ binds one of its described interaction partners thereby contacting the c-Myc promoter and, more importantly, repressing the activity of this interaction partner.
Transcriptional activation and repression depends on the recruitment of coactivators and corepressors. C/EBPβ has been shown to bind to the coactivator p300 via its N-terminal part which we defined as necessary for repression (58). Among other functions, p300 and the closely related CREB-binding protein (CBP) covalently modify chromatin by their histone acetylation activity thereby opening chromatin locally for transcription (59). Thus, p300 could be withdrawn by C/EBPβ from NF-κB and c-Myb which transactivate c-Myc and also interact with p300 (58, 60). CBP is also known as a corepressor when it acetylates lymphocyte enhancer factor 1/T cell factor (61). Ectopic expression of p300, with or without histone acetylation activity, can suppress c-Myc and arrest rat cell lines in the G1 phase of the cell cycle (62). These data suggest that p300/CBP might also act as corepressors for C/EBPβ-mediated suppression of c-Myc. Another such corepressor could be histone-deacetylase 1 which interacts with C/EBPβ (63) and suppresses the c-myc promoter when recruited by the repressor Blimp-1 (64).
SUMO modification of C/EBPβ relieves suppression of c-myc and proliferation
Sumoylation of C/EBPβ leads to relief of its TAD-mediated repression and to the concurrent accumulation in subnuclear speckles. Upon sumoylation, several transcription factors are targeted to PML nuclear bodies which were also designated as PML oncogenic domains or ND10. Therefore, it was unexpected that the punctuated pattern of sumoylated C/EBPβ does not coincide with PML, but with bright DAPI and most importantly, with trimethyl histone 3-lysine 9 staining of pericentric heterochromatin. A reversible punctuated pattern of C/EBPβ has long been known for macrophages (65) and a transient localization to heterochromatin, where it binds to specific C/EBP consensus-binding sites, has been demonstrated for adipocytes (66). The exclusive relocalization of C/EBPβ upon sumoylation from a widely distributed nuclear appearance to heterochromatin and the striking loss of its repressor function makes it very likely that the tightly packed chromatin structure inactivates the TAD-mediated repressor function of C/EBPβ. Recently, localization of C/EBPβ to heterochromatin was shown to be dependent on MAPK activity in preadipocyte fibroblasts (67). It is tempting to speculate and preliminary data indicate that phosphorylation has to precede sumoylation which then leads to relocalization.
Furthermore, genes are thought to shuttle toward or from pericentromeric heterochromatin, in relation to their differentiation stage-dependent activity, which might be an important strategy to regulate gene expression during cell differentiation (68). Conformational changes induced by sumoylation might influence the ability of C/EBPβ to interact with pericentromeric C/EBP-binding sites and heterochromatic proteins, thereby withdrawing the c-myc gene to heterochromatic regions.
Because transactivation of IL-4 gene by C/EBPβ is unaffected by sumoylation and its accumulation to heterochromatin, we consider the transactivation potential of C/EBPβ on the IL-4 promoter as very strong and comparable to other genes whose transcription can be achieved when they are localized to or close to pericentric heterochromatin (69, 70). Consistently, in C/EBP factors, the sumoylation sites map to so-called synergy control motifs (32). These small domains possess the capacity to regulate transcriptional synergy of factors that occupy multiple binding sites in promoters, the activity of which depends on cooperative DNA binding (38). Therefore, it was concluded that sumoylation facilitates interaction with a protein that inhibits synergistic activation. Studies on the IL-4 promoter revealed two binding sites for C/EBPβ, exhibiting very different affinity (10). The predominance of a single binding site might explain why IL-4 expression remained unaffected by sumoylation. Furthermore, while transactivation of the IL-4 promoter depends only on the 3′ part of C/EBPβ’s TAD, a complete TAD is necessary for c-Myc repression. This makes it likely that multiple factors need to interact with C/EBPβ’s TAD to suppress c-Myc. Thus, c-Myc repression as well as its relief upon C/EBPβ sumoylation might also be mediated through multiple protein binding sites within the c-myc promoter region.
Sumoylation is a dynamic process which is under tight control of isopeptidases. This allows a reversible regulation of c-Myc and proliferation necessary for normal cell growth. The central role of c-Myc for proliferation/differentiation control by C/EBPβ is in good agreement with many data on c-Myc. The importance of c-Myc regulation in differentiation is underlined by the elucidation of specific repressors in different cell types (24, 25, 26, 27, 28, 29, 30). Our data elucidate C/EBPβ as a specific repressor of c-Myc in T cells, especially Th2 cells and sumoylation of C/EBPβ as a silencer of its repressive activity. Thus, C/EBPβ appears to control the size of Th2 cell population (Fig. 1⇑) which is inefficiently controlled by the Fas/FasL pathway (41). This is in line with the elevated level of Th2 cells in C/EBPβ-deficient mice (14). The confusing fact that C/EBPβ-KO mice also show increased IL-4 levels might be explained by the high expression of IL-6 in those mice, because IL-6 represses Th1 and induces Th2 differentiation by mechanisms independent of C/EBPβ (71).
We thank Ursula Sauer for her excellent technical help throughout this work. For support in DNA transfections, we thank Doris Michel and Ilona Pietrowski. We are indebted to our colleagues Andris Avots, Alois Palmetshofer, and Devapriya Samanta for critical reading the manuscript and helpful discussions. For the kind gift of DNA constructs and helpful discussions, we thank Heike Hofmann (Erlangen). We also thank Achim Leutz (Berlin-Buch) for helpful suggestions.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported from the Deutsche Forschungsgemeinschaft (through SFB 466, Erlangen and FOR 330, Wuerzburg), the Wilhelm Sander Foundation (to E.S. and I.B.) and the Fritz Thyssen Foundation (to F.B.-S. and E.S.).
↵2 Address correspondence and reprint requests to Dr. Friederike Berberich-Siebelt, Department of Molecular Pathology, Institute of Pathology, Josef-Schneider-Strasse 2, 97080 Wuerzburg, Germany. E-mail address:
↵3 Current address: University of Sidney, Sidney, Australia.
↵4 Current address: Yale University School of Medicine, New Haven, CT 06520.
↵5 Abbreviations used in this paper: bZIP, basic leucine zipper; LAP, liver-activating protein; LIP, liver-inactivating protein; SUMO, small ubiquitin-related modifier; PML, promyelocytic leukemia; TAD, transactivation domain; tg, transgenic; PSG, penicillin-streptomycin-glutamine; Tm, 4-hydroxytamoxifen; TPA, 12-O-tetradecanoylphorbol-13-acetate; T/I, TPA/ionomycin; RPA, RNase protection assay; m, murine; NP40, Nonidet P-40; IP, immunoprecipitation; ER, estrogen receptor; WT, wild type; ChIP, chromatin IP; DAPI, 4′,6′-diamidino-2-phenylindole; CBP, CREB-binding protein.
- Received May 18, 2005.
- Accepted January 25, 2006.
- Copyright © 2006 by The American Association of Immunologists