BCL3 Acts as a Negative Regulator of Transcription from the Human T-cell Leukemia Virus Type 1 Long Terminal Repeat through Interactions with TORC3*

By associating with cyclic AMP-responsive element-binding protein (CREB), the human T-cell leukemia virus type 1 (HTLV-1) Tax protein activates transcription from the HTLV-1 long terminal repeat (LTR), which contains multiple cyclic AMP-responsive elements. The transducers of regulated CREB activity (TORCs) were a recently identified family of CREB co-activators that bind to CREB to enhance CRE-mediated transcription. TORC3, a TORC family protein, dramatically enhances Tax-mediated transcription from the LTR. In this study, we performed a yeast two-hybrid screen using the N-terminal region of TORC3 as bait and identified B-cell chronic lymphatic leukemia protein 3 (BCL3) as a protein interacting with TORC3. This interaction was confirmed by glutathione S-transferase pulldown assays and co-immunoprecipitation experiments with detection by Western blotting. The ankyrin repeat domain of BCL3 interacted with TORC3. By using a luciferase assay, we determined that BCL3 inhibited transcription from the HTLV-1 LTR in a manner dependent on TORC3. Knockdown of endogenous BCL3 using RNA interference enhanced transcriptional activation of CRE. Treatment with trichostatin A, a potent inhibitor of the transcriptional co-repressor HDAC, partially reversed the inhibitory effect of BCL3. These results suggest that BCL3 functions as a repressor of HTLV-1 LTR-mediated transcription through interactions with TORC3. In addition to stimulating transcription from the HTLV-1 LTR, Tax also enhances BCL3 expression; thus, transcription from the LTR is regulated by both positive and negative feedback mechanisms.

Human T-cell leukemia virus type 1 (HTLV-1) 3 is the causative agent of adult T-cell leukemia and the neurological disor-der HTLV-1-associated myelopathy/tropical spastic paraparesis (1,2). The onset of adult T-cell leukemia is preceded by a long latency period, suggesting that leukemogenesis depends on the accumulation of multiple mutations over many years in conjunction with viral protein expression (3).
The viral oncoprotein Tax, encoded by the env-pX gene (4), plays a principal role in the activation of transcription from the long terminal repeat (LTR) of the virus as well as the proliferation and transformation of HTLV-1-infected T-cells (5)(6)(7). Tax is a pleiotropic factor that interacts with multiple cellular proteins to regulate gene expression (8 -12). Tax also modulates its own transcription by interacting directly with cyclic AMP-responsive element-binding protein (CREB), a cellular transcription factor that activates the LTR (13). The Tax-CREB complex binds to the Tax-responsive element (TxRE) within the HTLV-1 LTR. This binding promotes viral transcription through the recruitment of co-activators, such as CBP/p300 and associated factors, in a process that does not require CREB phosphorylation (14 -16). The HTLV-1 LTR contains three tandem repeats of TxRE, each of which consists of a cyclic AMP-responsive element (CRE) flanked by GC-rich sequences (17).
Recently, a family of CREB co-activators, the transducers of regulated CREB activity (TORCs), was discovered to bind CREB in a phosphorylation-independent manner and enhance CRE-mediated transcription (18,19). This transcriptional activation is facilitated by recruitment of TATA-binding proteinassociated factor 130 (TAF130), a component of the transcription factor IID (TFIID) complex, to the CREB-containing complex. TORC1, TORC2, and TORC3 dramatically enhance Tax-mediated transcription from the LTR (20,21). This enhancement appears to require the formation of a ternary complex composed of Tax, CREB, and TORCs.
To clarify the mechanism by which transcription from the LTR is regulated, we searched for a cellular factor(s) that interacted with TORC3 using a yeast two-hybrid screening system with the N terminus of TORC3 as bait. This approach identified B-cell chronic lymphatic leukemia protein 3 (BCL3) as a protein interacting with TORC3.
BCL3 was originally identified through molecular cloning of the chromosomal breakpoint in a subset of human B-cell chronic lymphatic leukemia containing the t(14;19) chromosome translocation (22). BCL3 is unique among the members of the ankyrin repeat-containing IB family of NF-B inhibitors for its ability to regulate gene expression in multiple ways (23). In addition to binding NF-B p50 and p52 homodimers, BCL3 also interacts with co-activators, such as CBP/p300, SRC-1, and Tip60, and co-repressors, such as HDAC1 (24 -29). Differential associations also target genes to be either activated or repressed by BCL3. During our investigation of the role of BCL3 in the regulation of transcription from the HTLV-1 LTR, we discovered that BCL3 suppressed transcription from its promoter through interaction with TORC3.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screening Assay-The DNA fragment encodes amino acids 1-298 of TORC3 into the pGBT9 vector (Clontech). This plasmid was used as the bait to screen a human spleen cDNA library (Clontech) that had been cloned into the yeast cell line Y190 according to the manufacturer's instructions (Clontech). Transformants were selected in medium lacking tryptophan, leucine, and histidine containing 50 mM 3-amino-1H-1,2,4-triazole. After incubation for 7 days at 30°C, we screened a total of 2.0 ϫ 10 6 transformants by ␤-galactosidase activity.
Cell Culture and Treatment-HEK-293T and HeLa cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Jurkat, JPX-9, MT2, C91-PL, SLB1, and Raji cells were cultured in RPMI 1640 medium (Nissui) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Cells were maintained at 37°C in a 5% CO 2 humidified chamber. Raji cells were purchased from American Type Culture Collection (ATCC). The histone deacetylase inhibitor trichostatin A (TSA; Sigma) was used at a final concentration of 100 nM. JPX-9 is a Jurkat cell derivative with inducible expression of Tax cDNA under the control of the metallothionein promoter. In this cell, Tax expression is induced by the presence of cadmium (20 M CdCl 2 ) (30). HeLa-LEC1 cells were transfected with pLEC1 using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions into HeLa cells. Selection of cells was performed in complete Dulbecco's modified Eagle's medium with 1 mg/ml G418 (Invitrogen), resulting in isolation of the stable transformant, HeLa-LEC1, carrying a gene consisting of HTLV-1 LTR followed by the coding region of GFP.
Plasmid Transfection-HEK-293T cells were grown to 50% confluence in 6-cm dishes or 24-well plates in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells were transfected with expression vectors or the corresponding empty vector constructs using FuGENE 6 reagent (Roche Applied Science) according to the manufacturer's instructions.
GST Pulldown Assay-GST fusion proteins expressed in Escherichia coli were isolated using glutathione-Sepharose 4B beads (Amersham Biosciences). Resin-bound GST fusion proteins were then incubated with [ 35 S]methionine-labeled BCL3, which had been synthesized by in vitro translation using a TNTcoupled transcription-translation system (Promega). Binding reactions were performed in 600 l of GST binding buffer (50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 5% glycerol, 0.5% Nonidet P-40, 1 mM EDTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, supplemented with a Complete protease inhibitor mixture tablet (Roche Applied Science)) for 3 h at 4°C. Beads were washed three times in 1 ml of GST binding buffer. Bound proteins were eluted in 20 l of sample buffer and resolved by SDS-PAGE. Gels were visualized by autoradiography using a BAS-5000 image analyzer (Fujix).
Co-immunoprecipitation Assay-Co-immunoprecipitation assays utilized HEK-293T cells. Cells were lysed in radioimmune precipitation buffer (50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, supplemented with a Complete protease inhibitor mixture tablet). Cell debris was removed by centrifugation for 5 min. Lysates were first cleared with protein G-Sepharose beads (Amersham Biosciences) for 15 min and then incubated with 2 g of anti-FLAG antibody for 3 h at 4°C. Antibody-protein mixtures were then incubated for another 30 min with 20 l of protein G-Sepharose beads. The beads were then washed three times in radioimmune precipitation buffer; immunoprecipitated proteins were resolved by SDS-PAGE and analyzed by Western blotting.
DNA Affinity Precipitation Assay (DNAP)-HEK-293T cells were lysed with DNAP binding buffer (25 mM Tris-HCl (pH 8.0), 100 mM NaCl, 1 mM EDTA, 0.25% Nonidet P-40, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, supplemented with a Complete protease inhibitor mixture tablet). Cell debris was removed by centrifugation for 5 min. Lysates were first incubated with streptavidin-Sepharose beads (Amersham Biosciences) for 15 min to eliminate nonspecific binding and then incubated with 15 g of poly(dI-dC) and 2 g of biotinylated DNA probe for 1 h at 4°C. Streptavidin-Sepharose beads were then incubated with these mixtures for an additional 30 min at 4°C. After washing the beads three times in DNAP binding buffer, precipitated proteins were eluted in loading buffer. Samples were analyzed by SDS-PAGE followed by immunoblotting. The biotin-labeled DNA probes were annealed to complementary oligonucleotides. We used DNA probe sequences representative of CRE, 5Ј-biotin-(AGCCTG-ACGTCAGAGA)x3-3Ј, and TxRE, 5Ј-biotin-(GAAGGCTCTGACGTCTCCCCCA)x3-3Ј.
Luciferase Reporter Assay-After seeding onto 24-well plates for 24 h, HEK-293T cells were transfected with expression plasmids using FuGENE 6. After an additional 24-h incubation, we performed luciferase assays using a dual luciferase assay system (Promega) on a Berthold Lumat LB 9507 luminometer according to the manufacturer's instructions. We normalized luciferase activities in cell lysates to the measured Renilla luciferase activity derived from co-transfected pRL-RSV-luc. All reporter assays were performed in triplicate. Standard errors are denoted by the bars in the figures.
RNA Interference-A 21-nucleotide small interfering RNA (siRNA) duplex specific for BCL3 was purchased from Ambion (predesigned siRNA). A control nucleotide, si-control, was purchased from Dharmacon (nonspecific control duplex IX). HeLa cells were transfected with 25 l of 20 M duplex RNA per 6-cm dish using Oligofectamine (Invitrogen) according to the manufacturer's protocol. At 4 h after transfection, medium was replaced with normal culture medium. Total RNA isolated from cells 48 h after siRNA transfection was used as a template for RT-PCR using the following primers: 5Ј-CTCATC-CACGCCGTGGAAAAC-3Ј (forward) and 5Ј-TCAGCTGC-CTCCTGGAGCTGG-3Ј (reverse).
Chromatin Immunoprecipitation Assay (ChIP)-This assay was performed essentially as described (protocol from Upstate Biotechnology, Inc.). Formaldehyde cross-linked chromatin from 1 ϫ 10 6 cells for each reaction with the antibody was used for analyzing histone modifications or association with transcriptional regulators. Cross-linking reactions were quenched with 125 mM glycine; cell were lysed, and chromatin was sonicated to obtain DNA fragments with an average 500 bp. Following centrifugation, the chromatin was diluted 10-fold with ChIP dilution buffer, and precleared with protein A-agarose beads containing salmon sperm DNA and bovine serum albumin. Aliquots of the precleared chromatin fraction were incubated with different antibodies. After overnight rotation at 4°C, the immune complexes were collected by addition of protein A-agarose beads. DNA was purified by proteinase K digestion, phenol extraction, and followed by ethanol precipitation. The recovered DNA was amplified by PCR using primers specific for HTLV-1 LTR 5Ј-TTCCGAGAAACAGAAGTCTG-3Ј and 5Ј-CTCCTGTTAGTTTATTGAGC-3Ј. The PCR products were then analyzed by electrophoresis using 2% agarose gel and visualized with ethidium bromide staining.

Identification of BCL3 as a Factor
Interacting with TORC3-TORC3, a TORC family protein, dramatically enhances Taxmediated transcription from the HTLV-1 LTR (20). This activation requires the formation of a ternary complex containing Tax, CREB, and TORC3. To clarify the molecular mechanisms governing the regulation of transcription from the LTR by these proteins, we searched for a cellular factor(s) that interacts with TORC3 using a yeast two-hybrid screening system with the TORC3 N terminus as bait. The N-terminal 298 amino acids of TORC3 were subcloned into pGBT9 (GAL4 system; Clontech); this plasmid was co-transformed with human spleen cDNA library (Clontech) into yeast Y190 cells. We obtained 29 clones from 2.0 ϫ 10 6 yeast transformants that exhibited a positive phenotype. One of these clones encoded a portion of BCL3.
To confirm the interaction of BCL3 with TORC3, we performed a GST pulldown assay using bacterially expressed GST-TORC3 (1-298 amino acids). Under conditions in which in vitro translated BCL3 did not bind GST-bound Sepharose beads alone (Fig. 1B, lanes 2, 5, and 8), it could be detected in the fraction bound to Sepharose beads conjugated to recombinant GST-TORC3 (1-298 amino acids) (Fig. 1B, lane 3). This result suggests that BCL3 interacts specifically with TORC3. To identify the region of BCL3 responsible for interacting with TORC3, we repeated this GST pulldown assay using several BCL3 deletion mutants (Fig. 1A). The in vitro synthesized BCL3⌬NC fragment (125-359 amino acids), but not ⌬ANK (which lacks the ankyrin repeats), co-precipitated with GST-TORC3 (Fig. 1B,  lanes 6 and 9).
To analyze the in vivo interaction of BCL3 with TORC3, we performed co-immunoprecipitation experiments. HEK-293T cells were co-transfected with expression plasmids encoding FLAG-tagged TORC3 and HA-tagged BCL3. Immunocomplexes precipitated with the anti-FLAG antibody were resolved by SDS-PAGE. Western blotting using the anti-HA antibody identified a band corresponding to the expected size of BCL3 (Fig. 1C, lane 2). To define the domains within BCL3 and TORC3 responsible for the interaction, we performed immunoprecipitation assays using the BCL3 deletion mutants (Fig.   1A). HA-tagged BCL3⌬NC (Fig. 1C, lane 4), but not HA-tagged BCL3⌬ANK (Fig. 1C, lane 6), could be detected in the immunocomplexes isolated from cell lysates with an anti-FLAG antibody. Taken together, these results suggested that TORC3 interacts with the ankyrin repeats of BCL3.
BCL3, TORC3, and CREB Form the Ternary Complex on CRE and TxRE-As TORCs are involved in CREB complex (18,19), we used an in vitro DNA affinity precipitation assay to determine whether BCL3 is also recruited to CRE and TxRE. HEK-293T cells were transfected with empty vector, HA-tagged BCL3, or FLAG-tagged TORC3 expression plasmids. Cell lysates were incubated with a 5Ј-biotin-labeled DNA probe consisting of three repeats of either CRE or TxRE. Probe DNA was recovered with streptavidin-coated beads; probe-bound proteins were eluted in SDS loading buffer and detected by Western blotting with anti-FLAG, anti-HA, and anti-CREB antibodies. The amount of BCL3 bound to either CRE or TxRE increased in the presence of TORC3 (Fig. 2, A, lane 4, and B,  lane 4); the presence or absence of BCL3 did not change the amounts of TORC3 and CREB (Fig. 2, A and B). This result suggests that BCL3 is recruited to CRE and TxRE in a TORC3dependent manner.
BCL3 Inhibits Transcription from the LTR, TxRE, and CRE by Tax and/or TORC3-As TORC3 is a co-activator of the transcription factor CREB, which itself enhances CRE-mediated target gene transcription (18,19), we explored the effect of BCL3 on TORC3-and Tax-dependent transcription from the CRE-containing HTLV-1 LTR. A luciferase assay using the pLTR-luc reporter plasmid revealed that luciferase activity was enhanced by ectopic expression of either Tax or TORC3. BCL3 expression, however, suppressed this activation of transcription from the HTLV-1 LTR (Fig. 3A).
Association of Tax as well as TORC3 with CREB on the TxRE within the HTLV-1 5Ј-LTR activates viral gene expression. To  determine the effect of these proteins on viral gene transcription, we performed a reporter assay using pTxRE-luc, a reporter plasmid with five copies of the Tax-responsive element in its promoter region. Luciferase activity was enhanced by the expression of either Tax or TORC3. BCL3, however, decreased activation 3-fold from the TxRE by Tax or TORC3 (Fig. 3B). Next, we explored the effect of BCL3 on TORC3-dependent transcription from the CRE. HEK-293T cells were transfected with expression plasmids encoding TORC3 and BCL3 together with the luciferase reporter plas- , or pCRE-luc (C) and 1 ng of pRL-RSV-luc as an internal control. The total DNA concentration in each well was adjusted to 411 ng with empty vector. Twenty four hours after transfection, we evaluated the expression of firefly luciferase from pLTR-luc, pTxRE-luc, or pCRE-luc and Renilla luciferase with a dual luciferase reporter assay system. D, suppression of endogenous BCL3 by siRNA treatment. 21-Nucleotide siRNA-targeting mRNA duplexes specific for BCL3 and a control siRNA (random 21-nucleotide sequences) were transfected into HeLa cells plated in 6-cm dishes. mRNAs encoding BCL3 and GAPDH (control) were examined by RT-PCR 48 h after transfection. E, HeLa cells were seeded onto 24-well plates. Cells were transfected with BCL3 (solid bar) or control (open bar) siRNAs. Twenty four hours after transfection, 10 ng of pCRE-luc and 1 ng of pRL-RSV-luc were transfected into HeLa cells in the presence (ϩ) or absence (Ϫ) of 200 ng of pcDNA3-FLAG-TORC3. Twenty four hours after transfection, we measured luciferase activities in the cell lysates. F, HeLa-LEC1 cells were seeded onto 6-well plates. Cells were transfected with 2 g of pcDNA3-FLAG-TORC3 or 2 g of pcDNA3-HA-BCL3. The total amount of DNA in each well was adjusted to 4 g with empty vector. Twenty four hours after transfection, we evaluated the expression of GFP mRNA by semiquantitative RT-PCR with the ABI 7500 real time PCR system. G, HeLa-LEC1 cells were seeded onto 6-cm dishes. Cells were transfected with BCL3 (solid bar) or control (open bar) siRNAs. Twenty four hours after transfection, 4 g of pcDNA3-FLAG-TORC3 was transfected (ϩ) or mock-transfected (Ϫ). Twenty four hours after the transfection, we evaluated the expression of GFP mRNA by semiquantitative RT-PCR with the ABI 7500 real time PCR system. The average values of three independent experiments are shown throughout the figures; standard error is denoted by error bars. mid, pCRE-luc, which contains four repeats of the CRE consensus sequence. TORC3-activated pCRE-luc expression, generating luciferase levels ϳ80-fold greater than untransfected control cells. BCL3, however, decreased TORC3-dependent activation of luciferase expression from pCRE-luc in a dose-dependent manner, whereas Tax was unable to activate transcription from CRE (Fig. 3C). To examine the function of BCL3 under physiological conditions, we employed siRNA technology to suppress endogenous BCL3. Treatment with BCL3 siRNA efficiently reduced the levels of BCL3 mRNA in HeLa cells, whereas GAPDH mRNA levels remained unchanged (Fig. 3D). Knockdown of endogenous BCL3 by RNA interference enhanced the transcription activation of CRE by TORC3 (Fig. 3E).
Next, to examine the effect of BCL3 under physiological condition, we used stable transformant of HeLa cell clone, HeLa-LEC, transduced with pLEC1. These cells are appropriated to assess the physiological relevance of the HTLV-1 LTR and superior to HTLV-1-infected lymphoid cell lines because of higher efficiency of transfection. A semiquantitative RT-PCR assay revealed that GFP mRNA was enhanced by ectopic expression of TORC3. BCL3 expression, however, suppressed this activation of transcription from the HTLV-1 LTR (Fig. 3F). Moreover, we observed that knockdown of endogenous BCL3 by RNA interference enhanced the transcription activation of GFP mRNA by TORC3 (Fig. 3G). These results suggest that BCL3 represses transcription from the HTLV-1 LTR.
TORC3-dependent Transcription, but Not Tax-dependent Transcription, from a GAL4-driven Promoter Is Suppressed by BCL3-To examine the effect of BCL3 on TORC3-or Tax-dependent transcription, we analyzed GAL4-driven luciferase activity. Luciferase expression from this promoter is activated by GAL4 or GAL4 fusions with other transcription factor(s), allowing detection of direct interactions between factors. This system is more appropriate to determine the direct effect of BLC3 on Tax-TORC3-dependent transcription than the complex HTLV-1 LTR. We introduced GAL4-TORC3 and GAL4-Tax into HEK-293T cells with pGAL4-luc, whose promoter contains five GAL4-binding sites following an SV40 minimal promoter, and then measured luciferase expression in the presence or absence of a co-transfected BCL3 expression plasmid. Transcriptional activation from pGAL4-luc by GAL4-TORC3 was suppressed by the co-production of BCL3 in dose-dependent manner (Fig. 4A). Luciferase activity following GAL4-Tax expression, however, was unaffected by the presence or absence of BCL3 (Fig. 4B).
We analyzed the effect of BCL3 on pGAL4-luc transcription in the presence of both GAL4-TORC3 and Tax or GAL4-Tax Twenty four hours after transfection, luciferase activity was measured. C and D, HEK-293T cells were seeded onto 24-well plates. Cells were transfected with 100 ng of GAL4-TORC3 (C) or GAL4-Tax (D), 100 ng of pcDNA3-HA-BCL3, 10 ng of GAL4-luciferase reporter plasmid (pGAL4-luc), 1 ng of pRL-RSV-luc, and either 500 ng of pcDNA3-FLAG-TORC3 or 500 ng of pH 2 R-Tax. The total amount of DNA in each well was adjusted to 711 ng. Twenty four hours after transfection, we measured luciferase activity. and TORC3. BCL3 induced a significant reduction in the luciferase activity promoted by GAL4-TORC3 plus Tax (Fig. 4C,  lane 3 versus 4 and lane 5 versus 6). The percentage reduction, however, was similar to that facilitated by BCL3 in the presence of GAL4-TORC3 alone. Luciferase activity activated by GAL4-Tax was not affected by BCL3 (Fig. 4D, lane 3 versus 4). On the other hand, the augmented luciferase activity by TORC3 was again suppressed by BCL3 (Fig. 4D, lane 5 versus 6). These results suggest that BCL3 inhibits TORC3dependent, but not Tax-dependent, transcriptional activity.
BCL3 Represses Transcription from the HTLV-1 LTR by Recruiting HDAC1-We hypothesized that this inhibition may be mediated by HDAC recruitment to BCL3. To test this possibility, we analyzed the effect of TSA, a potent inhibitor of HDACs, on CRE-mediated gene transcription. HEK-293T cells transfected with effector plasmids expressing TORC3 and BCL3 and the reporter plasmid pCRE-luc were incubated in the presence or absence of TSA (100 nM) 8 h after transfection. Cells were harvested 16 h after treatment, and luciferase activity was measured. Treatment with TSA partially reversed the inhibitory effect of BCL3 (Fig. 5A).
To analyze the interaction of BCL3 with HDAC1 in vivo, we performed co-immunoprecipitation experiments. HEK-293T cells were cotransfected with expression plasmids encoding FLAG-tagged HDAC1 and HA-tagged BCL3. Immunocomplexes precipitated with the anti-FLAG antibody were resolved by SDS-PAGE. Western blotting using the anti-HA antibody identified a band corresponding to the expected size of BCL3 (Fig. 5B, lane 2). Next, we used an in vitro DNA affinity precipitation assay to determine whether HDAC1 is also recruited to CRE containing DNA. HEK-293T cells were transfected with empty vector, HA-tagged BCL3, or FLAG-tagged HDAC1 expression plasmids. Cell lysates were incubated with a 5Ј-biotin-labeled DNA probe consisting of three repeats of CRE. The probe DNA was recovered with streptavidincoated beads; proteins bound to the probe were eluted in SDS loading buffer and resolved by SDS-PAGE. The protein bands were detected by Western blotting with anti-FLAG and anti-HA antibodies. The amount of HDAC1 bound to CRE increased in the presence of BCL3 (Fig. 5C, lane 2). We observed that HDAC3, a member of the HDAC family, also interacted with BCL3 and was recruited on CRE containing DNA in BCL3-dependent manner (data not shown).
Next, we assessed in vivo protein complex formation on the HTLV-1 LTR by ChIP assay (Fig. 5D). HeLa-LEC1 cells were FIGURE 5. BCL3 represses transcription from the HTLV-1 LTR by recruiting HDAC1. A, HEK-293T cells were seeded onto 24-well plates. Cells were transfected with 200 ng of pcDNA3-FLAG-TORC3 and 50, 100, or 200 ng of pcDNA3-HA-BCL3 together with 10 ng of pCRE-luc and 1 ng of pRL-RSV-luc as an internal control. The total amount of DNA in each well was adjusted to 411 ng with empty vector. Eight hours after transfection, cells were incubated in the presence or absence of TSA (100 nM). Sixteen hours after TSA treatment, we evaluated the expression of firefly and Renilla luciferases from pCRE-luc and pRL-RSV-luc, respectively, using a dual luciferase reporter assay system. B, HEK-293T cells were seeded onto 6-cm dishes. Cells were transfected with 3 g of pcDNA3-HA-BCL3 in the presence or absence of 3 g of pcDNA3-FLAG-HDAC1. Twenty four hours after transfection, proteins were immunoprecipitated (IP) from cell lysates with an anti-FLAG antibody; co-immunoprecipitated proteins were resolved by SDS-PAGE and analyzed by Western blot with an anti-HA antibody. IB, immunoblot. C, HEK-293T cells were seeded onto 6-cm dishes. Cells were transfected with 3 g of pcDNA3-HA-BCL3 and 3 g of pcDNA3-FLAG-HDAC1. Twenty four hours after transfection, cell lysates were incubated with a biotinylated DNA probe containing three repeats of CRE. The DNA probe was recovered with streptavidin-coated beads. Probe-bound proteins were eluted in SDS loading buffer and analyzed by Western blotting with anti-FLAG and anti-HA antibodies. D, HeLa-LEC1 cells were seeded onto 10-cm dishes. Cells were transfected with 12 g of pcDNA3-FLAG-BCL3 and 12 g of pcDNA3-Myc-TORC3. After formaldehyde fixation, the cross-linked DNA-protein complexes were immunoprecipitated with anti-FLAG, anti-HA, anti-HDAC1, and anti-AcH3 antibodies or normal mouse IgG. 1% input of total cell lysate was used as positive control. The DNA extracted from each immunoprecipitate was used as the template for PCR to detect the HTLV-1 LTR fragment containing CRE.
transfected with empty vector, FLAG-tagged BCL3, or Myctagged TORC3 expression plasmids. We observed BCL3 recruited on the HTLV-1 LTR in TORC3-dependent manner. Moreover, BCL3 increased the amount of HDAC1 recruited on the HTLV-1 LTR with TORC3. Additionally, we also detected histone H3 acetylation level was increased in a TORC3dependent manner, and it decreased its level in a BCL3-dependent manner. These results suggest that BCL3 functions to repress transcription from the HTLV-1 LTR by recruiting HDAC1.
Induction of BCL3 mRNA Expression by Tax-To explore the role of BCL3 in HTLV-1-infected cells, we examined the expression levels of BCL3 mRNA in HTLV-1-infected cell lines, MT2, C91-PL, and SLB1, and an uninfected cell line, such as Jurkat and JPX-9. JPX-9 is a Jurkat cell derivative with inducible expression of Tax cDNA under the control of the metallothionein promoter. RT-PCR analysis demonstrated higher levels of BCL3 mRNA in Tax-expressing cell lines than in Jurkat and JPX-9 cells (Fig. 6A). We next investigated if BCL3 mRNA expression is induced by Tax. In JPX-9 cells, Tax expression is induced upon treatment with CdCl 2 , and this treatment also enhanced expression of BCL3 but did not show any effect in Jurkat cells (Fig. 6B). These results suggest that Tax up-regulates BCL3 mRNA expression.

DISCUSSION
By using a yeast two-hybrid system, we determined that BCL3 interacts with TORC3; this interaction of both proteins can be confirmed in cells exogenously expressing these proteins. This interaction occurs between the N terminus of TORC3 and the ankyrin repeat-containing domain of BCL3 (Fig. 1). BCL3, which interacts with the CREB-containing complex on the TxRE, suppressed TORC3-mediated transcription (Figs. 2-4). The up-regulation by TORC3 of transcription from CRE-containing promoters was also abrogated by BCL3, indicating that BCL3 is involved in the regulation of CRE-mediated transcription through its association with TORC3. It was not clear, however, if the suppression of transcription from the LTR or TxRE-containing promoters by BCL3 results from influencing both TORC3 and Tax or only one of these factors. To obtain insight into the role of BCL3 on LTR-mediated and CRE-dependent transcription, we employed a GAL4-driven reporter assay. Luciferase activity from the pGAL4-luc construct was enhanced by both GAL4-TORC3 and GAL4-Tax. Expression of BCL3 abrogated the enhancement of transcription mediated by TORC3, whereas that induced by Tax was unaffected (Fig. 4). This observation suggests that the suppression of transcription by BCL3 is a consequence of the inhibition of TORC3 function but not inhibition of Tax.
Considering that the recruitment of transcriptional regulatory factors to promoter-bound complexes can be affected by the promoter sequence itself, Tax may recruit BCL3 as a transcriptional repressor through associations with specific promoters. In this manner, Tax-mediated up-regulation of BCL3 may mediate trans-suppression of several genes, such as those encoding DNA polymerase ␤, c-Myb, or cyclin A (31)(32)(33).
The coiled-coil domain within the N-terminal 1-435 amino acids of TORC1 interacts with CREB to facilitate its co-activating function. This interaction enhances association of the CREB complex with transcription factors, such as TAF130, to augment CREB-dependent transcription (18). We confirmed the enhancement of CREB activity by TORC by assessing the interaction of CREB with TAF130 (data not shown). CREB also interacts with other transcription co-activators, such as CBP, p300, and P/CAF (34,35), prompting us to examine the possibility that the levels of those proteins in the CREB-containing complex were altered by BCL3 expression. We did not observe any such alteration in the levels of those factors (data not shown). Furthermore, the levels of CREB recruited to TxREand CRE-containing DNA sequences were not altered by ectopic BCL3 expression (Fig. 2).
Suppressive effect of BCL3 on gene expression occurs through a variety of mechanisms as follows. 1) Depletion of p50 by its association with BCL3 suppresses transcription of the TNF␣ gene potentially by reducing the amount of RelA/p50, a complex needed to activate TNF␣ gene expression (36). 2) Recruitment of FIGURE 6. Induction of BCL3 mRNA expression by Tax. A, expression level of BCL3 mRNA was examined by RT-PCR in HTLV-1 infected (MT2, C91-PL, and SLB1) and uninfected (Jurkat and JPX-9) cell lines. B, Tax and BCL3 mRNA expression was induced in CdCl 2 -treated JPX-9 cells. JPX-9 and Jurkat cells were treated with CdCl 2 for the indicated periods. Tax (top panels), BCL3 (middle panels), and GAPDH (bottom panels) mRNA levels were examined by RT-PCR. HDAC1, -3, and -6 through the association with BCL3 to promoters decreases transcriptional activity (37). 3) Enhanced expression of BCL3, as well as HDAC1, by lipopolysaccharide results in suppression of TNF␣ gene expression (28). 4) Association of BCL3, whose expression was enhanced by Rous sarcoma virus infection, with the interleukin 8 promoter, and HDAC1, resulted in suppression of interleukin 8 expression (29). Although the molecular mechanism underlying the suppression of TORC3-mediated transcription from TxRE-or CRE-containing promoters by BCL3 remains elusive, preliminary studies proposed that BCL3 recruits HDACs to promoters through similar interactions as those seen for N-coR/SMRT and mSin3. The presence of HDAC1, -2, and -3 on the HTLV-1 LTR, association of Tax with transcriptional repressors, or negative regulation of transcription by HBZ suggests the existence of positive and negative regulatory mechanisms governing transcription of HTLV-1 gene expression (38 -45). The up-regulation of BCL3 by Tax suggests that transcription from the LTR as well as from cellular genes containing CRE in their promoters is down-regulated by infection with HTLV-1 (Fig. 6). In this study, we provide evidence suggesting that BCL3 functions as a repressor of transcription from the HTLV-1 LTR through interactions with TORC3. Because we observed that Tax enhanced BCL3 expression, transcription from the LTR is regulated by positive and negative feedback mechanisms.
BCL3 transforms NIH3T3 cells and enhances proliferation of other cell types (37). Furthermore, BCL3 expression is elevated in certain cancers, suggesting oncogenic properties. Tax plays a pivotal role in the proliferation and immortalization of HTLV-1-infected cells and is likely important in the development of adult T-cell leukemia (1-3, 6, 46, 47). Therefore, the up-regulation of BCL3 by Tax may help to mediate these effects of Tax to promote cell proliferation.