SR-Related Protein TAXREB803/SRL300 Is an Important Cellular Factor for the Transactivational Function of Human T-Cell Lymphotropic Virus Type 1 Tax

Abstract

Expression of the human T-cell lymphotropic virus type 1 (HTLV-1) genes is transcriptionally activated by the cognate oncoprotein Tax which enhances the binding of the cyclin AMP-responsive element binding protein (CREB) to the Tax responsive element (TxRE) located in its long terminal repeat (LTR). TxRE is highly homologous to the cyclic AMP-responsive element (CRE) except for the GC-rich sequence flanking the CRE. We cloned the cDNA for a cellular factor, TAXREB803, of which the DNA-binding domain bound to TxRE and the binding was dependent on the 3′ GC-rich sequence in TxRE. TAXREB803 is an SR-related protein composed of 2,752 amino acids including numerous arginine/serine (RS) motifs. TAXREB803 enhanced both the Tax dependent transcription and the CREB binding to TxRE in cooperation with Tax. The interaction of TAXREB803 and Tax was detected by coimmunoprecipitation assays as well as by indirect immunofluorescence assays. Significantly, Tax transactivation for the HTLV-1 LTR decreased dramatically when the expression level of the endogenous TAXREB803 was suppressed by the small interfering RNA. These results suggest that TAXREB803 functions as a transcriptional coactivator for Tax and plays a critical role in the expression of HTLV-1 genes.

Human T-cell lymphotropic virus type 1 (HTLV-1) is a human retrovirus which is the causative agent for adult T-cell leukemia/lymphoma. HTLV-1 is also etiologically related to a neurological disorder called tropical spastic paraparesis/HTLV-1-associated myelopathy as well as other diseases (reviewed in reference 54). The HTLV-1 genome encodes the 40-kDa viral transactivator protein Tax, which acts as a potent activator for gene expression of HTLV-1 (10, 45).

Tax interacts with various cellular proteins, which are transcription-related factors and others, and modulates their transcriptional activities. The cellular genes, including those for interleukin-2 (IL-2) and the IL-2 receptor α (29, 33) and c-fos (13, 14, 48), which possess nuclear factor κB (NF-κB) or serum-responsive factor (SRF) binding sites in their promoters, are transactivated by the interaction of Tax with NF-κB, IκB, IKKγ (22, 50), or SRF (13, 48). Tax also interacts with factors related to cell cycle progression and plays a role in cell growth control. It transsuppresses transcription of genes such as cyclin D3 and CDK inhibitors (2, 21) and enhances the kinase activity of CDK4 resulting from the direct binding of Tax to CDK4 and p16^INK-4A (19, 51). In the others, it is suggested that Tax binding to the human homolog of the Drosophila Disc large tumor suppressor (hDLG) inhibited the tumor-suppressive function of hDLG (52) and network forming by the neuron-specific filament protein α-internexin is inhibited by its binding to Tax (42). Although several cellular proteins have been reported to interact with Tax, the physiological effects of their interactions remain to be clarified.

The expression of the HTLV-1 genes by Tax depends on three imperfectly repeated sequences of the 21-bp element known as the Tax-responsive element (TxRE) localized in the U3 region of the proviral long terminal repeats (LTRs) (5, 15, 41, 47). The transactivational function of Tax with respect to the TxREs is mediated by interaction with the cellular factors such as cyclic AMP-responsive element (CRE) binding protein CREB/ATF family (49, 58), CREB-binding protein (CBP) (17, 25), and p300/CBP-associated factor (PCAF) (23). The TxRE consists of the cellular CRE-like sequence flanked by GC-rich sequences at its 3′ terminus. The cellular CRE-containing genes do not carry the GC-rich flanking sequences in their promoter regions and are not transactivated by Tax (56, 57). Tax interacts directly with CREB through its N terminus and stabilizes the binding of CREB to the TxRE (6, 18). The binding affinity of CREB to CBP is also increased in the presence of Tax. Interestingly, the increase in CREB binding to TxRE or to CBP does not occur in the cellular CRE (26). It is suggested, therefore, that the functional interaction of Tax with TxRE for transactivation depends on the GC-rich flanking sequences in TxRE. However, the mechanism of the specific interactions mediated by the GC-rich sequence remains to be clarified.

We have previously isolated several TxRE-binding proteins by Southwestern blotting, including TAXREB67/CREB-2/ATF-4, TAXREB302, TAXREB107, TAXREB703, and TAXREB803 (38, 39, 53). A piece of cDNA for TAXREB803 was obtained from a peptide that specifically bound to TxRE (53). By searching GenBank for homologous genes, we found a partial cDNA clone with accession number AB002322. We constructed the full cDNA for TAXREB803 with the cDNA of AB002322 and found that TXAREB803 has the capacity to encode a protein with 2,752 amino acid residues. The DNA binding domain (DBD), the peptide initially isolated by us (53), resides between positions 1616 and 1718 of the amino acid sequence of this putative protein.

While investigating the functions of TAXREB803, two proteins, SRL300 and SRm300, were identified by other groups. In amino acid sequence, SRL300 is identical to TAXREB803, but SRm300 has a complicated structure, with many regions identical to the TAXREB803. SRL300 was cloned as an RNA-binding protein bound to the 5′ noncoding region of the AT-rich element binding factor 1 (ATBF1) mRNA (44). SRm300 was isolated as a component of the SRm160/300 complex (SR-related matrix protein complex of 160 kDa and 300 kDa). Although the SRm160/300 complex was found to be capable of associating with the splicing complexes, only SRm160 could activate the splicing reaction, so the function of SRm300 has remained unclear (4). In our study for the functions of TAXREB803, we found that it enhanced Tax-mediated transcription, which was a significantly different feature from the previously reported characteristics of SRL300 and SRm300. Here we report the function of TAXREB803 as a coactivator for Tax-induced transcriptions and discuss the physiological roles of this protein.

Although TAXREB803 is identical to SRL300, we refer to this cellular protein as TAXREB803 because we initially isolated its partial cDNA fragment and named the encoded peptide TAXREB803.

MATERIALS AND METHODS

Cell culture.

SaOS-2, COS-7, HEK 293T, and HeLa cells were grown in Dulbecco's modified Eagle's medium (DMEM; Nissui, Tokyo, Japan) supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, Calif.), 4 mM l-glutamine (Nacalai Tesque, Tokyo, Japan), and 0.1 mg/ml of kanamycin sulfate (Meiji, Tokyo, Japan). MT-2 cells were grown in RPMI 1640 medium (Nissui) supplemented as the same as DMEM except for 2 mM l-glutamine.

cDNA cloning and sequencing.

The 5′ and 3′ coding regions of the mRNA for TAXREB803 were obtained from SaOS-2 by 5′ rapid amplification of cDNA ends (RACE) with the 5′ Full RACE Core Set (TaKaRa, Otsu, Japan) and reverse transcription (RT)-PCR. The cDNAs were subcloned into pGEM-T Easy vector (Promega, Madison, Wis.) and their nucleotide sequences were determined with a DYEnamic ET terminator cycle sequencing premix kit (Amersham, Piscataway, N.J.).

Plasmids.

The cDNA containing the 5′- and the 3′-end regions of TAXREB803 was inserted into AB002322 cDNA (kindly provided by Kazusa DNA Research Inst.) to make the full open reading frame of TAXREB803, and the resulting cDNA was cloned into the vector pBluescript SK⁺ (Stratagene, La Jolla, Calif.). The eukaryotic expression plasmid for TAXREB803 was constructed by inserting the full-length cDNA into vector pcDNA3 (Invitrogen) with KpnI and NotI sites. The Kozak consensus sequence was inserted between the KpnI site and the putative start codon of TAXREB803 to make pcDNA3-TAXREB803. pcDNA3-Flag-TAXREB803 was constructed by inserting the Flag tag between the Kozak sequence and the start codon of TAXREB803. pcDNA3-Tax was used for the expression of Tax (1).

Several luciferase-expressing reporter plasmids were used. pTxRE-luc contains five copies of TxRE in its promoter region (1). pGL3-LTR-luc plasmid was constructed by inserting the HindIII fragment containing the U3 region of the LTR from pHTLV-1 LTR-CAT (47) into the pGL3-basic vector (Promega). pCRE-luc (a tetramer of CRE enhancer), pNF-κB-luc (a pentamer of NF-κB enhancer), pSRE-luc (a pentamer of SRE enhancer), and pFC-MEKK plasmids were purchased from Stratagene, and phRL-TK-luc plasmid was purchased from Promega.

Antibodies.

The rabbit polyclonal antiserum to TAXREB803 was raised by injecting the fusion protein of glutathione S-transferase (GST) and the peptide encoding amino acids 4 to 138 of TAXREB803, GST-TAXREB803(4-138), as an immunogenic antigen. Anti-Tax antibody is the mouse monoclonal antibody against Tax. Rabbit polyclonal antiserum to Flag tag peptide was purchased from Sigma (Saint Louis, Mo.). Goat anti-mouse immunoglobulin lissamine rhodamine B conjugate and goat anti-rabbit immunoglobulin fluorescein isothiocyanate were purchased from Molecular Probes (Eugene, Oreg.). Mouse monoclonal antisera to CREB and SC-35 were purchased from Santa Cruz Biotechnology (Santa Cruz, Calif.). Rabbit normal immunoglobulin G was purchased from Zymed Laboratories (San Francisco, Calif.).

Production of GST fusion proteins.

The cDNA fragments encoding amino acids 4 to 138 and 1616 to 1718 of the TAXREB803 open reading frame were inserted into pGEX-6p vector (Amersham) to produce fusion proteins GST-TAXREB803(4-138) and GST-TAXREB803(1616-1718), respectively. The expression plasmid for GST-CREB was provided by T. Ohshima (Institute for Virus Research, Kyoto University). These GST fusion proteins were produced in Escherichia coli BL21. GST-TAXREB803(4-138) was purified with glutathione-Sepharose resin (Amersham).

Detection of DNA-protein interaction by Southwestern blotting.

The binding of GST fusion proteins to ³²P-labeled TxRE or TxRE mutant probe was analyzed as previously described (55). Crude extracts of GST, GST-TAXREB803(1616-1718), and GST-CREB were separated by SDS-10% PAGE and blotted onto a polyvinylidene difluoride membrane (Millipore, Bedford, Mass.). The membrane was submerged in 50 ml of binding buffer containing 25 mM NaCl, 5 mM MgCl₂, 25 mM HEPES (pH 7.9), and 0.5 mM dithiothreitol supplemented with 6 M guanidine hydrochloride. After gentle shaking for 10 min, the solutions were decanted and replaced with 50 ml of the same buffer. A second buffer containing 6 M guanidine hydrochloride was diluted with a 100% dilution method as previously described (55); after this, the membrane was blocked by incubating with binding buffer containing 5% nonfat dry milk and then washed three times with binding buffer containing 0.25% dry milk. Next, the membrane was exposed to DNA probes by incubating overnight with binding buffer containing 0.25% dry milk, 2 μg of poly(dI-dC) (Amersham) per ml, and 2 × 10⁵ cpm of ³²P-labeled probes per ml. The membrane was washed with binding buffer, air-dried, and subjected to autoradiography.

The probes used for Southwestern blotting were prepared by labeling the annealed DNA of 5′-GACTAAGGCTCTGACGTCTCCCCCC-3′ for the positive strand and 5′-GACTGGGGGGAGACGTCAGAGCCTT-3′ for the negative strand for TxRE and of 5′-GACTAAGGCTCTGACGTCTTAATCG-3′ for the positive strand and 5′-AGCTCGATTAAGACGTCAGAGCCTT-3′ for the negative strand for mutant TxRE. The radiolabeling was performed by the filling-in reaction of Klenow fragment (Toyobo, Osaka, Japan) with [α-³²P]dCTP (Amersham).

Transient transfection and luciferase assays.

Transfection of plasmids was performed with FuGene 6 transfection reagent (Roche Molecular Biochemicals, Indianapolis, Ind.). The reporter assays were performed with the luciferase assay system according to the manufacturer's protocol (Promega). The firefly luciferase activity from pTxRE-luc, pCRE-luc, and pGL3-LTR-luc was measured with the luciferase assay system (Promega). Dual luciferase assay for firefly and Renilla luciferase was performed with the dual luciferase reporter assay system (Promega). pcDNA3 vector plasmid was used as a carrier DNA to equalize the DNA concentration for each transfection. All experiments were performed in triplicate, and results were obtained from at least three separate experiments.

RNA isolation and RT-PCR.

Total RNA was isolated with Sepasol-RNA I (Nacalai Tesque) according to the manufacturer's protocol. The RNA was further treated with RNase-free DNase I (Stratagene) for 30 min at 37°C, and the DNase I was inactivated by heating for 10 min at 80°C. RT for luciferase mRNA was performed with Superscript II reverse transcriptase (Invitrogen) with the primer 5′-GCTGATGTAGTCTCAGTGAG-3′, which is complementary to the cDNA sequence for the firefly luciferase, and followed by PCR with Taq polymerase (Sigma) with a set of primers, 5′-CAGCCTACCGTGGTGTTC-3′ and 5′-CCTGAAGGCTCCTCAG-3′. Then, the PCR was conducted for 27 cycles, each cycle consisting of denaturation (94°C for 30 s), annealing (57°C for 30 s), and extension (72°C for 1 min). Reverse transcription was carried out in a reaction volume of 20 μl in which 1 μg of total RNA was present. mRNA for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the internal control for the RT-PCR. RT-PCR for TAXREB803 mRNA in HeLa cells was performed with a One Step RNA PCR kit (TaKaRa) with the primers 5′-AAACCTGCAAGCCCCAAGAAG-3′ and 5′-GCTGCTCTGCTGGTC-3′.

RNA transfection.

RNA was transcribed in vitro with a MEGAscript SP6 kit or T7 kit (Ambion, Austin, Tex.) from a linearized plasmid template of pGL3-luc (basic) or pcDNA3-Tax. RNA transfection was performed with DMRIE-C reagent (Invitrogen) according to the manufacturer's protocol. The transfection medium was changed to normal culture medium 4 h after the transfection, and cells were harvested 24 h after the medium change.

DNA affinity precipitations.

Biotin-labeled DNA probe was prepared by the filling-in reaction of Klenow fragment with a template oligonucleotide, 5′-biotin-(TAAGGCTCTGACGTCTCCCCC)×3-3′, containing three repeats of TxRE biotinylated at the 5′ end, and a primer oligonucleotide, 5′-GGGGGAGACGTCAGA-3′. The probe was composed of three repeats of CRE and was prepared with a template oligonucleotide, 5′-biotin-(CCTTGGCTGACGTCAGAGAGA)×3-3′, which is biotinylated at the 5′ end, and a primer oligonucleotide, 5′-TCTCTCTGACGTCAG-3′.

Nuclear extracts were prepared as previously described (38) and diluted with binding buffer consisting of 10 mM HEPES (pH 7.9), 150 mM NaCl, 1 mM EGTA, 5% glycerol, and 1 mM dithiothreitol. The diluted nuclear extract was mixed with 2 μg of biotinylated DNA probe and 15 μg of poly(dI-dC) and then incubated for 30 min on ice. Streptavidin-coated MagneSphere paramagnetic particles (Promega) were added to the mixture. which was rotated for 30 min at 4°C. Streptavidin beads were collected with the magnetic separation stands and washed four times with cold binding buffer. The trapped proteins were dissolved with loading buffer and analyzed by SDS-PAGE followed by immunoblotting.

Immunoprecipitations and immunoblots.

Cell pellets were lysed in 1 ml of immunoprecipitation buffer [50 mM Tris-HCl buffer (pH 8.0) containing 0.5% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 0.25 mM PMSF]. Cell lysates were precleared with protein G-Sepharose beads (Amersham) for 1 h and subjected to immunoprecipitation by incubation with the designated antibodies at 4°C overnight. The immunocomplexes were recovered by adding protein G-Sepharose beads, and the trapped proteins were analyzed by Western blotting as previously described (43).

DNA-protein coimmunoprecipitation assay.

293T cells transfected with the indicated plasmids were harvested and subjected to immunoprecipitation as described above with anti-CREB antibody, anti-TAXREB803 antibody, anti-Tax antibody, anti-Flag antibody, or normal rabbit IgG. The protein G-Sepharose beads containing immunocomplexes were suspended in 50 ml of TE buffer. DNA was extracted from the immunocomplex with TE-saturated phenol and precipitated with ethanol. The DNA was then dissolved in distilled water, and 10% of the recovered DNA was used as a template for the PCR. The primers used to amplify the U3 region of HTLV-1 LTR in pGL3-LTR-luc were 5′-CCGAGAAACAGAAGTCTG-3′ and 5′-GTGATGTCCACCTC-3′. The PCR was conducted for 30 cycles, and each cycle consisted of denaturation (94°C for 30 s), annealing (55°C for 30 s), and extension (72°C for 1 min).

Metabolic labeling of cells with [³⁵S]methionine.

Cells were cultured in methionine-free DMEM or RPMI 1640 (ICN Biomedicals, Aurora, Ohio) supplemented with dialyzed fetal bovine serum (Invitrogen), glutamine, and 100 μCi of [³⁵S]methionine per ml (43.5 TBq/mmole, ICN Biomedicals). At 6 h after the culture, cells were harvested and used for immunoprecipitation as described above. The immunoprecipitants were analyzed by SDS-7% PAGE followed by autoradiography. Protein standards (Bio-Rad Laboratories, Hercules, Calif.) were used as a molecular weight marker.

Immunocytochemistry and confocal microscopic observation.

COS-7 cells were seeded on glass coverslips in six-well plates and transfected with 2 μg of pcDNA3-Flag-TAXREB803 and 0.5 μg of pcDNA3-Tax. At 24 h after the culture, cells were washed twice in phosphate-buffered saline (PBS) and fixed by treatment for 5 min at room temperature with 3.6% formaldehyde in PBS. Then they were permeabilized by treatment for 15 min with 0.1% Nonidet P-40 in PBS. Cells were washed in PBS and incubated with 3% bovine serum albumin in PBS for 1 h at room temperature. Primary antibodies (rabbit anti-Flag antibody for Flag-tagged TAXREB803, mouse anti-Tax antibody, or mouse anti-SC35 antibody) diluted in PBS were added and incubated for 1 h at room temperature. Then, secondary antibody (FITC 488 or rhodamine 568) was added and incubated for 40 min at room temperature in the dark. DNA was counterstained with 4′,6′-diamidino-2-phenylindole (DAPI), and final preparations were mounted in 90% glycerol in PBS containing 0.01% p-phenylenediamine. Observations were performed with a confocal microscope (Axio Vision; Carl Zeiss, Hallbergmoos, Germany).

RNA interference.

A duplex of 21-nucleotide small interfering RNA (siRNA) corresponding to the sequence of mRNA for TAXREB803 (nucleotides 649 to 669) with the addition of 2 nucleotides of 2-deoxythymidine attached to the 3′ end was synthesized and annealed (B-Bridge International, San Jose, Calif.). The sequence of the duplex is 5′-GCACAGGUCAGAAUCUGAGdTdT in the sense orientation and 5′-CUCAGAUGACCUGUGCdTdT in the antisense orientation. HeLa cells were transfected with 25 μl of 20 μM siRNA duplex per 6-cm dish with Oligofectamine (Invitrogen) as previously described (9). Four hours after the siRNA transfection, the medium was replaced with the normal culture medium. For the RNA analysis, 48 h after the siRNA transfection, total RNA was isolated from the cells. For the protein analysis, cells were incubated for 66 h after the siRNA transfection and subjected to metabolic labeling with [³⁵S]methionine for 6 h. For the reporter assay, 24 h after the siRNA transfection, cells were reseeded into 12-well dishes and incubated for another 24 h, and then reporter plasmids were transfected with pcDNA3-Tax. At 24 h after the transfection of the reporter plasmids, the expressions of firefly luciferase from pGL3-LTR-luc and Renilla luciferase activity from phRL-TK-luc were evaluated with the dual luciferase reporter assay system (Promega).

RESULTS

Cloning of full-length cDNA for TxRE-binding protein TAXREB803.

We previously cloned a partial cDNA for a TxRE-binding protein by Southwestern blotting with a probe containing the 21-bp enhancer sequence (TxRE) residing in the HTLV-1 LTR and named the product of the gene TAXREB803 (53). The cloned cDNA has an open reading frame coding a peptide with 103 amino acid residues. To analyze the region of TxRE which associates with this peptide, we performed a Southwestern blotting analysis with ³²P-labeled oligonucleotides for the wild-type TxRE and mutant TxRE in which the 3′ GC-rich sequences were mutated. The GST-fused protein containing the 103 amino acids of TAXREB803 (actually defined as amino acids 1616 to 1718 of the full sequence) bound to the probe of wild-type TxRE but failed to interact with the mutant TxRE (Fig. 1A). GST-fused CREB protein bound to both of the probes independently of the 3′ GC-rich sequences.

FIG. 1.

TxRE-binding activity of the peptide derived from TAXREB803 and schematic structure of TAXREB803. (A) Southwestern blotting. Binding specificity of a peptide derived from TAXREB803 for TxRE. The 21-nucleotide sequences of the TxRE and mutant TxRE probes are described. CRE-like sequences are underlined. GC-rich and mutated sequences in TxRE and mutant TxRE, respectively, are double underlined. The GST-fused peptide containing TAXREB803 residues 1616 to 1718 and GST-CREB were expressed in E. coli, and the crude lysates were separated by SDS-10% PAGE. The gel was blotted onto a polyvinylidene difluoride membrane, denatured with 6 M guanidine hydrochloride, then renatured with the 100% dilution method, and incubated with ³²P-labeled probe. The protein bound to the probe was examined by autoradiography. (B) The schematic structure of TAXREB803. The SRm300 sequences different from TAXREB803/SRL300 are depicted as black boxes. Characteristic domains are indicated as follows: RNA binding domain (RBD), checkered; DNA binding domain (DBD), hatched; and RS (arginine/serine-rich) domains (RSD-1, RSD-2), shaded. Serine clusters are shown by dark shading.

Previous analysis showed that TAXREB803 interacted with the pyrimidine-rich region of the TxRE (data not shown), and thus, it was suggested that the sites for the interaction with TAXREB803 and CREB did not overlap at the TxRE. In order to clone the full-length cDNA for TAXREB803, we searched GenBank and found one homologous sequence with accession number AB002322. The cDNA of AB002322 is a partial clone lacking the 5′ coding region of TAXREB803. We conducted an analysis on the entire cDNA corresponding to AB002322 in cDNA from human SaOS-2 cells with RT-PCR and found a cDNA that was different from the 3′ region of AB002322. We obtained a cDNA for the 5′ upstream region containing the putative start codon by 5′-RACE. We constructed the full-length cDNA for TAXREB803 with cDNAs of the 5′ and 3′ regions cloned from SaOS-2 cells and of AB002322,

The protein deduced from the sequence of the cDNA consisted of 2,752 amino acid residues, with several unique features in its sequence. It contains many arginine/serine-rich sequences (RS motifs) which are conserved in the SR family proteins (12), two serine stretches (repeats of 25 and 42 residues), and a proline-rich domain in the C terminus. TAXREB803 lacked the typical DNA binding motif as well as the RNA recognition motif which is often seen in other SR family proteins. The DNA binding domain (DBD) is situated between positions 1616 and 1718 of the amino acid sequence of this putative protein. The schematic structure of this protein, together with those of SRL300 and SRm300, is shown in Fig. 1B.

TAXREB803 strongly enhanced Tax-induced transcription from HTLV-1 LTR.

It is thought that Tax forms a ternary complex with CREB/ATF at TxRE sites and lead to transcriptional activation (56, 57). To investigate the effect of TAXREB803 on Tax-mediated LTR transactivation, we analyzed the transactivational function of Tax with the pTxRE-luc and pCRE-luc reporter genes. These contain five repeats of TxRE and four repeats of cellular CRE in their promoter regions, respectively (Fig. 2A). No significant transcriptional activation of pTxRE-luc and pCRE-luc was observed when TAXREB803 was overexpressed alone in COS-7 cells. However, cotransfection of plasmids expressing TAXREB803 and Tax resulted in the marked enhancement of Tax-induced expression from pTxRE-luc (Fig. 2B, lanes 5 and 7) but not from pCRE-luc (Fig. 2B, lanes 6 and 8). This enhanced transactivation of the luciferase gene with the TxRE promoter was dependent on the dose of TAXREB803 and accompanied an increase in the mRNA for luciferase (Fig. 2C). These observations clearly show that TAXREB803 enhances Tax transactivation but leave it unclear at which step it contributes.

FIG. 2.

TAXREB803 enhanced transcription of TxRE-containing gene. (A) Schematic structures of the reporter genes. pTxRE-luc contains five repeats of TxRE, and pCRE-luc contains four repeats of CRE in the promoter region. The underlined letters indicate the CRE and CRE-like sequences. (B) COS-7 cells were transfected with 0.1 μg of pTxRE-luc or pCRE-luc together with 0.1 μg of pcDNA3-Tax, with 1 μg of TAXREB803, or with 0.1 μg of pcDNA3-Tax plus 1 μg of pcDNA3-TAXREB803. At 24 h after the transfection, luciferase activity was measured. Values are given in terms of activation above the basal levels of the lysates prepared from mock-transfected cells, and the average of three independent experiments is shown. The error bars are standard deviations. (C) The increase in luciferase activity caused by TAXREB803 was accompanied by an elevation of the RNA level of luciferase, which paralleled the dose of TAXREB803. COS-7 cells seeded in six-well plates were transfected with 0.1 μg of pTxRE-luc and 0.1 μg of pcDNA3-Tax together with an increasing amount (0, 1, and 2 μg) of pcDNA3-TAXREB803. At 24 h after the transfection, total RNA was isolated and RT-PCR was performed. GAPDH was used as the internal standard for RT-PCR. (D)Enhanced transcription of HTLV-1 LTR by TAXREB803 is not due to an increase in Tax protein level. COS-7 cells were transfected with 0.1 μg of pTxRE-luc together with empty vector or 2 μg of pcDNA3-TAXREB803. Then 2 μg of in vitro-transcribed Tax mRNA was transfected 6 h after the DNA transfection, and luciferase activity was measured after another 24 h. The Tax protein levels were evaluated by immunoprecipitation and immunoblotting with the anti-Tax antibody. (E) TAXREB803 did not affect luciferase expression at the posttranscriptional level. Two sets of COS-7 cells were transfected with 0.1 μg of pcDNA3-Tax with or without 1 μg of pcDNA3-TAXREB803 as indicated. Six hours after the transfection, one set of cells was transfected with 0.1 μg of pTxRE-luc (DNA-luc) and the other with 2 μg of in vitro-transcribed luciferase mRNA (RNA-luc). Luciferase activity was measured 24 h after the second transfection. Error bars show standard deviations.

We observed slightly enhanced expression of Tax when pcDNA3-Tax was cotransfected with pcDNA3-TAXREB803, and it was also observed that the luciferase expression from pCMV-luc was slightly enhanced by TAXREB803 (data not shown), so there was a possibility that TAXREB803 enhanced the expression of Tax through cytomegalovirus promoter of pcDNA3 vector and that the increased luciferase activity by Tax and TAXREB803 resulted from the increased production of Tax. To analyze the context of this increase in luciferase activity more precisely, we prepared in vitro-transcribed RNA for Tax and introduced it into COS-7 cells with the expression plasmid of TAXREB803. At 24 h after the transfection, the expression of Tax was examined by immunoprecipitation and immunoblotting with an anti-Tax antibody. While the luciferase expression was enhanced by the addition of TAXREB803, the Tax expression levels were not affected (Fig. 2D, lanes 3 and 4). This result shows that TAXREB803 has a function of enhancing the Tax-induced transcription independently of the expression level of Tax.

Next, we examined the effect of TAXREB803 on the stability of luciferase RNA. In vitro-transcribed RNA for luciferase (RNA-luc) was transfected into the cells that had been transfected with the expression plasmids for TAXREB803 and/or Tax at 6 h before the RNA transfection. The production of luciferase from RNA-luc was not changed by the presence of Tax and/or TAXREB803 (Fig. 2E, lanes 4, 6, and 8). These results indicate that the enhancement of transactivation of Tax by TAXREB803 occurs at the transcriptional level, not the posttranscriptional level.

TAXREB803 enhanced association of CREB with TxRE.

To investigate whether TAXREB803 plays a role in the binding of CREB to TxRE, we used an in vitro DNA affinity precipitation assay. Nuclear extracts were prepared from 293T cells transfected with an empty vector, with TAXREB803 expression plasmid, or with Tax expression plasmid. Each nuclear extract was mixed as designated in Fig. 3A (mock, Tax, or Tax plus TAXREB803) and then added with the 5′-biotinylated probe composed of three repeats of the TxRE or CRE. In the input of nuclear extracts, the amounts of CREB or Tax were adjusted to the same level. The amount of CREB bound to TxRE increased in the presence of Tax (Fig. 3A, lanes 1 and 2), which is consistent with previous reports (6, 18, 56, 57). This enhanced binding of CREB to TxRE was further enhanced by TAXREB803 (Fig. 3A, lanes 2 and 3). But CREB binding to CRE was not changed (Fig. 3A, lanes 5 and 6).

FIG. 3.

TAXREB803 enhanced the association of CREB with TxRE in vitro and in vivo. (A) TAXREB803 enhanced CREB binding to TxRE in vitro. A biotin-labeled DNA probe containing three repeats of TxRE or CRE was incubated with a mixture of nuclear extracts prepared from 293T cells transfected with empty vector, with 1 μg of pcDNA3-Tax, or with 5 μg of pcDNA3-TAXREB803. The probe DNA was recovered with streptavidin-coated beads, and the probe-bound proteins were eluted with SDS loading buffer and analyzed by Western blotting with the anti-CREB antibody. The levels of CREB and Tax in the lysates used for the binding assay were equal, as shown by immunoblot analysis with a 10% input of lysate. (B) Schematic structure of the promoter and luciferase gene of pGL3-LTR-luc, which consists of the U3 region containing three TxREs of the HTLV-1 LTR. The primer sites used for LTR detection are shown as thick arrows. (C) TAXREB803 enhanced the association of CREB with TxRE in vivo; 2 μg of pGL3-LTR-luc was cotransfected with empty vector, with 1 μg of pcDNA3-Tax, or with 1 μg of pcDNA3-Tax plus 5 μg of pcDNA3-Flag-TAXREB803 into 293T cells, which were then seeded in 10-cm dishes. At 48 h after transfection, cell lysates were prepared and immunoprecipitated with anti-CREB, anti-Tax, or anti-Flag antibody or IgG. Immunocomplexed DNA was amplified by PCR with the indicated primers.

Next, we analyzed the in vivo association of CREB and TxRE with a DNA-protein coprecipitation assay. We constructed pGL3-LTR-luc, where the promoter included the unique 3′ sequence (U3) containing three TxRE repeats of the HTLV-1 provirus (Fig. 3B). 293T cells were cotransfected with pGL3-LTR-luc together with the expression plasmids, as shown in Fig. 3C. After preparing the cell lysate, immunoprecipitation was performed with the anti-CREB, anti-Flag, and anti-Tax antibodies and IgG. Then, we analyzed the immunocomplex-associated LTR. DNA was recovered from immunocomplexws and amplified by PCR with the primers indicated by a thick arrow in the diagram of pGL3-LTR-luc (Fig. 3B). The expected size of the PCR product was about 900 bases. LTR was detected in the immunoprecipitant by anti-Tax antibody as well as by anti-Flag antibody (Fig. 3B, lanes 1 and 2). The level of CREB-associated LTR was increased to the higher level in the cells overproducing both Tax and TAXREB803 than Tax alone (lanes 3 and 4). These results suggest that TAXREB803 enhances the association of CREB with TxRE in cooperation with Tax.

Association of TAXREB803 with HTLV-1 LTR in vivo.

To investigate whether TAXREB803 interacts with the HTLV-1 LTR, we used a DNA-protein coimmunoprecipitation assay. An anti-TAXREB803 antibody was raised in a rabbit with a GST-fused peptide containing TAXREB803 residues 4 to 138. This antibody recognized [³⁵S]methionine-labeled TAXREB803 (both endogenously and transiently produced) in 293T cells (Fig. 4A). TAXREB803 migrated at a slower rate than expected from the molecular weight on SDS-PAGE, and the band was detected over a broad range. This suggests that TAXREB803 is highly modified, possibly by phosphorylation in cells, as shown by another group (44).

FIG. 4.

Association of HTLV-1 LTR with TAXREB803 in vivo. (A) Detection of endogenous as well as ectopically expressed TAXREB803 with rabbit antiserum to TAXREB803. The titer of the antiserum was examined by immunoprecipitation of the [³⁵S]methionine-labeled 293T lysate transfected with empty vector or with 5 μg of pcDNA3-TAXREB803. The immunoprecipitant was analyzed by SDS-7% PAGE followed by autoradiography. A molecular size marker of 250 kDa is indicated. (B) Coimmunoprecipitation of HTLV-1 LTR with TAXREB803; 2 μg of pGL3-LTR-luc was transfected into 293T cells in 10-cm dishes together with 1 μg of pcDNA3-Tax or with 1 μg of pcDNA3-Tax plus 5 μg of pcDNA3-TAXREB803. Lysates were prepared 48 h after the transfection and immunoprecipitated with either rabbit preimmune serum or the antiserum to TAXREB803. DNA in the immunocomplex was extracted and amplified by PCR with the pGL3-LTR-luc-specific primers shown in Fig. 3B. (C) The specific association of the LTR with TAXREB803. 293T cells transfected with 2 μg of pGL3-LTR-luc, 1 μg of pcDNA3-Tax, and 5 μg of pcDNA3-TAXREB803 in 10-cm dishes. Depleted anti-TAXREB803 serum with GST or GST-TAXREB803(4-138) was used for the immunoprecipitation. Anti-CREB antibody and normal rabbit IgG were used as the positive and negative control antibodies, respectively, in identifying the associated LTR.

For the DNA-protein coimmunoprecipitation assay, 293T cells were transfected with pGL3-LTR-luc together with the TAXREB803 expression plasmid, as shown in Fig. 4B, and immunocomplexes were prepared by precipitation with either preimmune or anti-TAXREB803 serum. DNA was extracted from these immunocomplexes and analyzed by PCR. The LTR was present in the immunocomplexes precipitated by anti-TAXREB803 serum, but not detected in those precipitated by preimmune serum (Fig. 4B). The LTR was also detectable in the immunocomplexes derived from cells in which TAXREB803 was not ectopically expressed, presumably due to the presence of endogenous TAXREB803 (Fig. 4B, lane 2). Thus, coprecipitation of the LTR with TAXREB803 suggests that TAXREB803 associates with the HTLV-1 LTR in vivo. Furthermore, when antiserum to TAXREB803 was depleted beforehand with the same peptide epitope used in its preparation and then used for immunoprecipitation, the LTR was not detected in the immunocomplex, but antiserum treated with GST could still precipitate the LTR (Fig. 4C, lanes 1 and 2).

Association of TAXREB803 with Tax.

In the promoter of the HTLV-1 gene, many factors, including Tax, CREB, p300, CBP, and AP-1 family members (17, 23, 25, 45, 49, 58), cooperate and promote transcription. We postulated that TAXREB803 could be involved in their transcriptional complex, through interaction with certain factors. To verify our hypothesis, we examined the association of TAXREB803 with Tax in cells. In an immunoprecipitation assay, Tax was coimmunoprecipitated with TAXREB803 in MT-2 cells, HTLV-1-infected T cells, as well as in 293T cells in which Tax and TAXREB803 were transiently overproduced by expression plasmids (Fig. 5A, lanes 3 and 7), and in a reciprocal experiment, ³⁵S-labeled TAXREB803 was detected in an immunocomplex precipitated with the anti-Tax antibody (Fig. 5B, lanes 3 and 7). These results show that Tax and TAXREB803 interact in the cells, either directly or indirectly.

FIG. 5.

Interaction of TAXREB803 with Tax. (A) Coimmunoprecipitation of Tax with TAXREB803. 293T cells seeded in 10-cm dishes were transfected with 5 μg of pcDNA3-TAXREB803 plus 2 μg of pcDNA3-Tax or with 5 μg of pcDNA3-TAXREB803 alone. Cell lysates of 293T cells were prepared at 48 h posttransfection and lysates of MT-2 cells were immunoprecipitated with the preimmune serum, anti-TAXREB803 serum, the anti-Tax antibody, or rabbit normal IgG. The immunocomplexes were subjected to SDS-PAGE and analyzed by Western blotting with anti-Tax antibody. (B) Coimmunoprecipitation of TAXREB803 with Tax. 293T cells were transfected with 2 μg of pcDNA3-Tax plus 5 μg of pcDNA3-TAXREB803, or with 5 μg of pcDNA3-TAXREB803 alone. At 24 h posttransfection, cells were metabolically labeled with [³⁵S]methionine for6 h. MT-2 cells were also labeled for 6 h. The lysates prepared from the cells were immunoprecipitated with the anti-Tax antibody. The resulting immunocomplex was analyzed by SDS-7% PAGE followed by autoradiography. The position of a protein of 250 kDa is shown as a molecular size marker. (C) The subcellular colocalization of Tax and TAXREB803. COS-7 cells were seeded on glass coverslips. After incubation for 24 h, 0.5 μg of pcDNA3-Tax and 2 μg of pcDNA3-Flag-TAXREB803 were transfected. The cells were fixed at 24 h posttransfection, permeabilized, and stained by dual immunofluorescence with primary antibodies (rabbit polyclonal antibody to Flag and mouse monoclonal antibody to Tax and SC-35) and fluoroconjugated secondary antibodies. DNA was counterstained with DAPI, while Tax and SC-35 were stained with rhodamine conjugate (red), and TAXREB803 was stained with FITC conjugate (green). The overlay of TXAREB803 with SC-35 (upper panel) or with Tax (lower panel) is shown as yellow.

We also examined the cellular localization of these two proteins by expressing Tax and Flag tagged TAXREB803 in COS-7 cells. TAXREB803 localized to the nuclear speckles of interchromatin granule clusters (IGCs)/RNA splicing bodies. We could identify these speckles as IGCs because SC-35, a splicing body component, colocalized with TAXREB803 (Fig. 5C, upper panels). It has been reported that Tax also localizes to IGCs (46), and indeed we observed a tight colocalization of Tax with TAXREB803 in IGCs (Fig. 5C, lower panels). This is consistent with our earlier observation of the in vivo association of Tax and TAXREB803 (Fig. 5A and B). Interestingly, IGC/RNA splicing bodies were enlarged in those cells in which TAXREB803 was exogenously overexpressed (Fig. 5C).

TAXREB803 is an important factor in Tax transactivation.

The TAXREB803 gene is ubiquitously expressed in many different cell lines and tissues (data not shown). The fact that TAXREB803 is endogenously expressed in 293T cells (Fig. 4B) allowed us to examine the capacity for endogenous TAXREB803-mediated enhancement of Tax-induced transactivation. To make cells in which the expression of endogenous TAXREB803 was repressed, we treated HeLa cells with small interfering RNA (siRNA), the duplex RNA encoding nucleotides 651 to 669 of TAXREB803. TAXREB803 expression was significantly suppressed at both the RNA and protein levels, while the level of other cellular protein such as GAPDH was not changed (Fig. 6A). TAXREB803 expression was not affected by the control siRNA with randomized sequences. Interestingly, when expression of endogenous TAXREB803 was suppressed, Tax-induced luciferase expression of pTxRE-luc decreased by almost 60% (Fig. 6B, left panel, black bars). However, Renilla luciferase expression from phRL-TK-luc, which contains the herpes simplex virus thymidine kinase promoter, showed no significant change (Fig. 6B, right panel). Taken together, these results show that TAXREB803 plays a key role in Tax-induced HTLV-1 LTR transactivation.

FIG. 6.

Suppression of endogenous TAXREB803 expression caused a substantial decrease in Tax-induced LTR transactivation. (A) Suppression of endogenous TAXREB803 production by siRNA. The 21-nucleotide RNA duplex (siRNA) directed against the TAXREB803 sequence was transfected into HeLa cells in 6-cm dishes. The level of TAXREB803 mRNA was evaluated by RT-PCR at 48 h posttransfection, and the protein level was examined by the immunoprecipitation of [³⁵S]methionine-labeled cell lysate at 72 h posttransfection. The 21-nucleotide RNA duplex with randomized sequences was used as a control. (B) The reporter assay for pTxRE-luc in cells in which the endogenous TAXREB803 was depleted. The siRNA was transfected into HeLa cells as described for A. Cells were reseeded into 12-well dishes at 24 h after the siRNA transfection and incubated for another 24 h. Then, 0.1 μg of pcDNA3-Tax and 0.1 μg of pTxRE-luc were transfected into the cells, together with 0.05 μg of phRL-TK-luc as a control reporter plasmid. Luciferase activities from pTxRE-luc (left) and phRL-TK-luc (right) were measured with a dual luciferase assay system at 24 h after transfection of reporter genes.

TAXREB803 enhanced Tax-mediated activation of NF-κB and SRE enhancer.

In the above results, it was suggested that TAXREB803 was an important cellular factor for the transcriptional function of Tax, so we examined whether TAXREB803 affected Tax-dependent activation of genes containing an NF-κB or SRF enhancer. In the reporter gene assay, Tax-induced transactivation of pNF-κB-luc and pSRE-luc was strongly enhanced by TAXREB803 by approximately 10-fold (Fig. 7A and B, lanes 3 and 4). However, MEKK-derived activation of these genes was only slightly enhanced by TAXREB803 (about 1.5-fold; Fig. 7A and B, lanes 5 and 6). From this result, we suggest that TAXREB803 is an important factor for Tax transactivation for NF-κB and the SRF enhancer as well as the HTLV-1 LTR.

FIG. 7.

TAXREB803 enhanced Tax transactivation for NF-κB and SRF enhancers. 293T cells were transfected with 0.1 μg of pNF-κB-luc (A) or pSRE-luc (B) with various combinations of 0.1 μg of pcDNA3-Tax, 1 μg of TAXREB803, and 0.1 μg of pFC-MEKK. At 24 h after the transfection, luciferase activity was measured. Values are given in terms of activation above the basal levels of the lysates prepared from mock-transfected cells, and the average of three independent experiments is shown. The error bars are standard deviations.

DISCUSSION

In this paper, we report a new coactivator, TAXREB803, for Tax-mediated transactivation of HTLV-1 gene expression. Transactivation of the HTLV-1 LTR by Tax requires the participations of various cellular transcription factors and coactivators (17, 23, 25, 49, 58). We had previously cloned a partial cDNA of TAXREB803 as a TxRE-binding protein by Southwestern blotting. This partial cDNA encoded 103 amino acid residues, and the binding of the peptide to the TxRE required the 3′ GC-rich sequence in TxRE (Fig. 1A). In spite of its TxRE-binding ability, a characteristic DNA-binding motif was not found in the peptide. On the other hand, TAXREB803 has distinctive amino acid sequences which include arginine/serine (RS)-rich domains scattered throughout. In addition to these RS motifs, there are two clusters with an extraordinarily high number of repeated serine residues in the 3′-terminal region.

Exogenously expressed TAXREB803 markedly enhanced Tax-induced transcription of the gene containing the HTLV-1 LTR. We observed several characteristic features of TAXREB803 which might be closely related to its ability to enhance transcription: TAXREB803 enhanced CREB binding to TxRE both in vitro and in vivo, TAXREB803 associated with the TxRE-containing region of the LTR in vivo, TAXREB803 associated with Tax in vivo, and Tax and TAXREB803 very tightly colocalized to ICG/RNA splicing bodies. In order to analyze the contribution of endogenous TAXREB803 to enhancing Tax-mediated LTR transactivation, we suppressed endogenous production of TAXREB803 by RNA interference. The transfection of siRNA directed against TAXREB803 substantially reduced the expression of endogenous TAXREB803 and suppressed Tax-mediated transactivation of the HTLV-1 LTR to a remarkable extent. These results suggest that Tax requires TAXREB803 as a coactivator in transactivating the HTLV-1 LTR. Furthermore, TAXREB803 showed a strong enhancing effect on the transcription of NF-κB and SRF enhancer by coexpression with Tax.

We think that TAXREB803 enhances TxRE-dependent transcription through association with the GC-rich sequence of the promoter and simultaneous interaction with Tax, while it enhances NF-κB and SRF-dependent transcriptions probably through the interaction with Tax. For the NF-κB enhancer, it is suggested that Tax has at least two distinguished functions. One is to enhance the nuclear transport of NF-κB through the interaction with IκB and IKKγ. Another is the association of Tax with NF-κB in the nucleus, but the nuclear function of Tax remains unclear. We think that TAXREB803 is related to the function of Tax for transactivation of NF-κB. On the basis of the above consideration, we suggest that Tax utilizes TAXREB803 to transactivate its target genes.

Previously, Tax was shown to interact with the target DNA (24, 27, 28, 31). However, there are no data in those papers to show that Tax directly binds to the target DNA without any cellular factor. It was reported that the Tax recruited CREB to the CRE site, and through this interaction Tax was able to contact the GC-rich region of the TxRE (24). Possibly, the interaction of Tax with the 3′ GC-rich region of the TxRE in the presence of CREB is important for the Tax-dependent transactivation, because the TxRE with a mutation in its 3′ GC-rich region did not function as an enhancer of Tax (Fig. 2). Since TAXREB803 interacted with the region and with Tax, we think that TAXREB803 and CREB are necessary to recruit Tax properly in that region for the full activity of transcription for Tax. However, the precise mechanism of CREB and TAXREB803 cooperation, if any, to recruit Tax on the target DNA and to activate the transcription remains to be clarified.

TAXREB803 has numerous consensus sites for phosphorylation by various kinases. In fact, we detected TAXREB803 as a broad band running at the location for a larger molecular size than expected on SDS-PAGE (Fig. 4A). Therefore we speculate that multiple phosphorylation may be one factor to retard its mobility. In our studies, Tax associated TAXREB803 was detected as a narrower band than the whole TAXREB803 (Fig. 5B). This observation suggests that a definite modification of TAXREB803 is required for interaction with Tax.

SRL300, which has an amino acid sequence identical to that of TAXREB803, was isolated as an RNA-binding protein bound to the 5′ noncoding region of ATBF-1 (41). SRm300, which is closely related to TAXREB803 genetically, is identical to TAXREB803/SRL300 in most of its sequences (Fig. 1C). SRm300 is a nuclear matrix protein and a component of the SRm160/SRm300 splicing complex. SRm160 is known to enhance the splicing of specific pre-mRNAs. However, the role of SRm300 in the splicing complex remains unclear, since depletion of SRm300 from the splicing reaction mixture does not prevent SRm160/SRm300-dependent splicing (4). Furthermore, the depletion of the Caenorhabditis elegans homologue of human SRm300, CeSRm300, by RNA interference led to early larval arrest and lethality, but that of CeSRm160 yielded no change in the C. elegans phenotype. Moreover, the phenotypic change caused by CeSRm300 depletion was unaffected by the simultaneous suppression of any other C. elegans SR family protein, including CeSRm160. Thus, it has been proposed that the function of CeSRm300 is different from those of other SR family proteins involved in splicing activation (30).

Exogenously expressed TAXREB803 mainly localizes to the IGCs/RNA splicing bodies. Although the function of IGCs remains unknown, they are believed to supply splicing factors to active genes, and phosphorylation of SR proteins, a prominent component of IGCs, is thought to play a key role in this process (35, 36). Recent studies have shown that SR family proteins are closely related to the coupling of transcription and splicing (32, 41). Furthermore, SR family members and other proteins have recently been identified as bifunctional proteins, which act as both transcription factors and RNA processing factors. These proteins, including heterogeneous nuclear ribonucleoprotein K (34), p54^nrb (3), nucleolin, (20), WT1 (7), p52 (16), PGC-1 (37), and small nuclear ribonucleoproteins (11), are involved in transcription as well as in RNA processing. One of them, pre-mRNA-binding protein hnRNP K, which is known to facilitate mRNA biogenesis, was found to act as a transcription factor (34). WT1 possesses both DNA binding and transcriptional activity; moreover, it is incorporated into spliceosomes and interacts with the essential splicing factor (7). PGC-1, an SR-like protein containing both RNA binding and RS domains, functions as a transcriptional coactivator through association with promoters (37).

These bifunctional proteins are expected to participate in both transcription and RNA processing and thereby to promote efficient gene expression. TAXREB803 reported in this report may have the features of such bifunctional protein. Further characterization of TAXREB803 will elucidate the physiological roles of this protein in gene expression.