Transcriptional Regulation of the Bovine Leukemia Virus Promoter by the Cyclic AMP-response Element Modulator τ Isoform*

Bovine leukemia virus (BLV) expression is controlled at the transcriptional level through three TaxBLV-responsive elements (TxREs) responsive to the viral transactivator TaxBLV. The cAMP-responsive element (CRE)-binding protein (CREB) has been shown to interact with CRE-like sequences present in the middle of each of these TxREs and to play critical transcriptional roles in both basal and TaxBLV-transactivated BLV promoter activity. In this study, we have investigated the potential involvement of the cAMP-response element modulator (CREM) in BLV transcriptional regulation, and we have demonstrated that CREM proteins were expressed in BLV-infected cells and bound to the three BLV TxREs in vitro. Chromatin immunoprecipitation assays using BLV-infected cell lines demonstrated in the context of chromatin that CREM proteins were recruited to the BLV promoter TxRE region in vivo. Functional studies, in the absence of TaxBLV, indicated that ectopic CREMτ protein had a CRE-dependent stimulatory effect on BLV promoter transcriptional activity. Cross-link of the B-cell receptor potentiated CREMτ transactivation of the viral promoter. Further experiments supported the notion that this potentiation involved CREMτ Ser-117 phosphorylation and recruitment of CBP/p300 to the BLV promoter. Although CREB and TaxBLV synergistically transactivated the BLV promoter, CREMτ repressed this TaxBLV/CREB synergism, suggesting that a modulation of the level of TaxBLV transactivation through opposite actions of CREB and CREMτ could facilitate immune escape and allow tumor development.

Bovine leukemia virus (BLV) 7 infection is characterized by viral latency in the large majority of infected cells and by the absence of viremia. These features are thought to be due to the transcriptional repression of viral expression in vivo (1,2). BLV transcription initiates at the unique promoter located in the 5Ј-long terminal repeat (5Ј-LTR) of the BLV genome. The 5Ј-LTR is composed of the U3, R, and U5 regions and transcription initiates at the U3-R junction. BLV exhibits two distinct functional transcriptional states as follows: a low basal level of transcription ensured by host cellular transcription factors, and a high level of transcription directed by the virus-encoded transcriptional activator Tax BLV (3,4). In the early stages of BLV transcription, before Tax BLV expression and transactivation, the basal transcriptional promoter activity is ensured by several cis-acting elements located in the 5Ј-LTR. In the U3 region are present the promoter CAAT and TATA boxes (5,6), and three 21-bp enhancers, each containing an imperfectly conserved 8-bp cyclic AMP-responsive element (CRE), which binds at least three proteins: CRE-binding protein (CREB) and activating transcription factors 1 and 2 (ATF-1 and ATF-2) (7)(8)(9)(10). Importantly, the 21-bp enhancers are also called Tax BLV -responsive elements (TxREs) because transactivation of the BLV LTR by Tax BLV requires these enhancers. It has been proposed that Tax BLV activation of transcription could be mediated, as reported for human T-cell leukemia virus-1, through increased binding of the cellular proteins CREB, ATF-1, and ATF-2 (and possibly other factors yet to be identified) to the TxREs (7,9,11,12).
Among the CREB/CREM/ATF family of transcription factors, the cAMP-response element modulator (CREM) has emerged as a unique member because CREM expression is finely regulated, both transcriptionally and post-transcriptionally, leading to the production of various activator and repressor isoforms (13)(14)(15). In contrast to CREB and ATF, the CREM isoforms are expressed in a cell-and tissue-specific manner (13, 14, 16 -19). Production of the different CREM isoforms depends upon RNA processing and upon selection of transcription initiation sites or translation initiation sites. CREM isoforms are highly homologous to CREB, especially in their DNAbinding domain and kinase-inducible domain (KID). However, with the exception of CREM, CREM1, and CREM2, all other CREM isoforms lack a glutamine-rich transcriptional activating domain and are therefore considered to function as repressors of cAMP-regulated genes, either by competing with CREB proteins for binding to CRE motifs or by heterodimerizing with CREB proteins and abolishing their transcription activating potential.
Expression of the CREM isoforms in BLV-infected cells and recruitment of these CREM isoforms to the BLV promoter through the viral CREs have so far not been investigated. In this study, we have examined the potential expression of CREM proteins in BLV-infected cells and the potential specific binding of CREM proteins to the CRE-like motifs of the U3 BLV LTR region, and we have investigated by transient cotransfection experiments the functional role of the CREM isoform in basal and Tax BLV -transactivated BLV promoter activity.

EXPERIMENTAL PROCEDURES
Cell Lines and Cell Culture-The mouse B-lymphoid cell line A20 was obtained from the American Type Culture Collection (Manassas, VA). A20 cells are BALB/c B-lymphoma cells derived from a spontaneous reticulum cell neoplasm found in an old BALB/cAnN mouse. A20 cells were cultured in RPMI 1640 media (Invitrogen) supplemented with 5% fetal bovine serum, 50 units of penicillin/ml, and 50 g of streptomycin/ml. Human HeLa cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 5% fetal bovine serum, 50 units of penicillin/ml, and 50 g of streptomycin/ml. The YR2 B-cell line is derived from leukemic cells of a BLV-infected sheep and contains a single, monoclonally integrated silent provirus with two mutations in Tax BLV that impair the infectious potential of the integrated provirus (20). The YR2 cell line was maintained in OptiMEM medium (Invitrogen) supplemented with 10% fetal bovine serum, 50 units of penicillin/ml, and 50 g of streptomycin/ml. All cells were grown at 37°C in an atmosphere of 5% CO 2 .
Isolation of PBMCs and Cell Culture Conditions-Blood samples were collected from a healthy sheep by jugular venipuncture, mixed with EDTA as an anticoagulant, and centrifuged at 1,750 ϫ g for 25 min at room temperature. The PBMCs were then isolated by Percoll gradient centrifugation (density, 1.129 g/ml; Amersham Biosciences) and washed twice in phos-phate-buffered saline containing 0.075% EDTA. The cells were then washed with phosphate-buffered saline until the supernatant became clear. Cells were resuspended at a concentration of 10 6 cells/ml in RPMI 1640 medium supplemented with 10% fetal calf serum, 50 units of penicillin/ml, and 50 g of streptomycin/ml and cultured at 37°C with 5% CO 2 .
Plasmid Constructs-The pLTRwt-luc, containing the luciferase gene under the control of the complete 5Ј-LTR of the 344 BLV provirus and its derivative pLTR-mutCRE-luc, mutated in the three viral CRE-like motifs, were previously described (9,10,21). The eukaryotic expression vectors pSG-Tax BLV and pSG-CREB2 were gifts from Drs. Luc Willems and Richard Kettmann (12,22). The pSV-CREM-wt expression vector (kindly provided by Dr. Paolo Sassone-Corsi) was described previously (23). Mutation of CREM serine 117 to alanine was performed by the QuikChange site-directed mutagenesis method (Stratagene) using the pSV-CREM-wt as substrate and the following pair of mutagenic oligonucleotides (mutation is highlighted in boldface and the serine 117 codon is underlined on the coding strand primer): CV 867/868 5Ј-CTTTCAC-GAAGACCCGCATATAGAAAAATACTG-3Ј. The mutated construct was fully resequenced after identification by cycle sequencing using the ThermoSequenase DNA sequencing kit (Amersham Biosciences). The resulting plasmid was designated pSV-CREM-S117A. The expression vectors for the cytomegalovirus wild-type and deleted E1A 12S proteins (p12S-E1A-wt and p12S-E1A/2-36del, respectively) have been described previously (24). The pRc/RSV-CBP expression vector (kindly provided by Dr. D. Trouche) was described previously (25).
Transient Transfection and Luciferase Assays-A20 cells were transfected by using the DEAE-dextran procedure as described previously (26). At 20 h post-transfection, transfected cells were mock-treated or treated with a goat antimouse IgG antibody at a concentration of 6.5 g/ml (reference 115-006-006, Jackson ImmunoResearch). At 42 h post-transfection, cells were lysed and assayed for luciferase activity (Promega). Luciferase activities were normalized with respect to protein concentrations using the detergent-compatible protein assay (Bio-Rad). HeLa cells were transfected with jetPEI TM Transfection Reagent (Polyplus Transfection) according to the manufacturer's protocol.
In Vitro Coupled Transcription Translation Reaction-Coupled transcription and translation reactions in wheat germ extracts were performed with the TNT coupled wheat germ extract system (Promega) under the conditions described by the manufacturer. Reactions were performed at 30°C for 90 min in the presence of the T7 RNA polymerase with 1 g of coding DNA (pSG-CREB2, pSV-CREM-wt, or pSG5).
Electrophoretic Mobility Shift Assays and Supershift Assays-Nuclear extracts from YR2 cells or from PBMCs isolated from a healthy sheep were prepared and EMSAs, competition EMSAs, and supershift assays were performed as described previously (10,21). The DNA sequences of the coding strand of the double-stranded TxRE probes are as follows (the viral CRE sites are underlined): TxRE1 (centered at position Ϫ148) 5Ј-CAGACA-GAGACGTCAGCTGCC-3Ј; TxRE2 (centered at position Ϫ123) 5Ј-AAGCTGGTGACGGCAGCTGGT-3Ј; TxRE3 (cen-tered at position Ϫ48) 5Ј-GAGCTGCTGACCTCACCTGCT-3Ј; and TxRE1-mutCRE (5Ј-CAGACAGAGTGGTCAGCT-GCC-3Ј) (mutations are highlighted in boldface). The sequence of the coding strand of the double-stranded Oct-1 consensus oligonucleotide used as competitor was 5Ј-TGTCG-AATGCAAATCACTAGAA-3Ј). For supershift assays, monoclonal antibodies against ATF-1 (Santa Cruz Biotechnology, catalog number sc-243X) and ATF-2 (catalog number sc-242X) and polyclonal antibodies against human CREB (catalog number sc-240X) and human full-length CREM (catalog number sc-440X) were added at a final concentration of 2 g/reaction mixture at the beginning of the binding reaction for 20 min before adding the DNA probe (blocking conditions).
Western Blot Analysis-Nuclear extracts prepared from BLV-infected YR2 cells or in vitro translated CREB or CREM proteins were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The membranes were blocked in Tris-buffered saline (TBS) containing 0.1% Tween 20 and 5% nonfat dry milk and then incubated with anti-CREB (catalog number sc-240) or anti-CREM (catalog number sc-440) antibodies in blocking solution. A second antibody, horseradish peroxidase-conjugated goat anti-rabbit IgG (catalog number sc-2054), was used for enhanced chemiluminescence detection (Cell Signaling).
Preparation of mRNA and cDNA, RT-PCR Amplification, and Cloning of CREM Isoform-One million YR2 cells were used for extracting RNA (RNeasy mini kit, Qiagen). First strand cDNA synthesis was performed using 2 g of YR2 total RNA and (dT) 16 primers (Invitrogen) with the Superscript II RNase H reverse transcriptase reagent, according to the manufacturer's protocol (Invitrogen). RNaseOUT TM RNase inhibitor (Invitrogen) was added to the RT reaction. The resulting cDNA templates were used directly for PCR. After a denaturation period of 5 min at 95°C, PCR was performed as follows: 40 cycles for 1 min at 95°C, 1 min at 50°C, 2 min at 72°C, followed by a final elongation at 72°C for 10 min. Reactions consisted of 10 l of Pfu DNA polymerase 10ϫ buffer, 3 l of 25 mM MgCl 2 , 2 l of dNTP mixture (10 mM each), 5 units of Pfu DNA polymerase (all from Promega), 10 l of sense and antisense primers (20 M each), and 20 l of cDNA in a final volume of 100 l. Primers used for detecting CREM mRNA were designed according to the mouse cDNA sequence (supplemental Fig. S2) and are as follows: 5Ј-CCG GAA TTC ATG ACC ATG GAA ACA GTT GAA TCC CAG-3Ј (this forward primer localized in CREM exon B at the 5Ј end of the translational start codon (underlined) contains an added EcoRI restriction site (in boldface)) and 5Ј-GC TCT AGA CTA ATC TGT TTT GGG AGA GCA AAT GTC-3Ј (this reverse primer localized in CREM exon Ib at the 3Ј end of the DBD II contains an added XbaI restriction site (in boldface)). The forward primer is located at the most 5Ј end of exon B, and the reverse primer is located at the most 3Ј end of exon Ib. To control reproducibility of the results, RNA samples used for RT-PCR were collected from three independent cultures of YR2 cells. PCR products were separated on a 1.5% agarose gel and purified using the QIAEX II gel extraction kit (Qiagen). PCR products were restricted with XbaI/EcoRI and inserted into the XbaI/EcoRI sites of the eukaryotic expression vector pcDNA 3.1 (Invitrogen). The resulting constructs were transformed into TOP10 cells, and several independent clones were sequenced. Homology searches in the nucleotide and genomic DNA data bases were performed with BLAST software accessed through the NCBI web page (ncbi.nlm.nih.gov). The ovine CREM1␣ DNA sequence we cloned is deposited in GenBank TM under accession number EF507802.
Chromatin Immunoprecipitation Assays-The chromatin immunoprecipitation (ChIP) assays were performed by using the ChIP assay kit (Upstate Biotechnology, Inc.) according to the manufacturer's recommendations. Formaldehyde crosslinking reactions from 10 7 BLV-infected YR2 cells were performed for 10 min (or 30 min for the detection of CBP recruitment) at room temperature and quenched with 125 mM glycine. Cells were lysed, and chromatin was sonicated to obtain an average DNA length of 600 bp. Following centrifugation, the chromatin was diluted 10-fold and precleared with a protein A-agarose slurry containing salmon sperm DNA and bovine serum albumin (Upstate Biotechnology, Inc.). Precleared chromatin (2 ml) was incubated overnight at 4°C with 5 g of either anti-CREB antibody (catalog number sc-186X), anti-ATF-1 antibody (catalog number sc-243X), anti-ATF-2 antibody (catalog number sc-242X), anti-CREM antibody (catalog number sc-440X), anti-CBP (catalog number sc-369X), or normal rabbit IgG antibody (Upstate Biotechnology, Inc., catalog number 12-370) as control. Incubation with the antibodies was followed by immunoprecipitation with protein A-agarose. Immunoprecipitated complexes were washed and eluted twice with 200 l of elution buffer. The protein-DNA cross-links were reversed by heating at 65°C overnight, and 10% of the recovered DNA was used for radioactive PCR amplification (35 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 1 min) with a primer set amplifying the BLV TxRE promoter region (nt Ϫ172 to ϩ29), 5Ј-nt-172 CCGTAAAC-CAGACAGAGACGTCAG-3Ј/5Ј-ntϩ29 CACGAGGGTCTCAG-GAGGAGAAC-3Ј, or with an unrelated primer set amplifying a gag gene region located ϳ1.8 kb downstream of the LTR region (nt ϩ1736 to ϩ1955): 5Ј-ntϩ1736 CAGCCCTCTCAGAATCAAGC-3Ј/5Ј-ntϩ1955 AATTTGCCATTTATCGAAAG-3Ј. PCR products from all reactions were resolved by PAGE and visualized by autoradiography.

Endogenous CREM Proteins from Nuclear Extracts of BLVinfected YR2 Cells and of Ovine PBMCs Specifically Bind to the BLV LTR CRE Motifs in Vitro-
The potential role of CREM members in regulating BLV gene expression has so far not been examined. CREM is a bZIP protein belonging to the CREB/ CREM/ATF family of transcription factors that bind to DNAregulatory regions via CRE target sequences. As CREM expression is known to be finely regulated in a cell-and tissue-specific manner (16,19,27,28), we decided to investigate whether CREM proteins were expressed in BLV-infected cells, and whether these CREM proteins were able to bind to the BLV LTR CRE motifs. To this end, we performed supershift EMSA experiments using the three 21-bp TxREs as probes. These probes were incubated with nuclear extracts prepared from YR2 cells derived from leukemic cells of a BLV-infected sheep (20), either mock-treated or treated for 1 h with phorbol ester phorbol 12-myristate 13-acetate plus Ca 2ϩ ionophore ionomycin (P ϩ I), a combination known to activate BLV expression (29,30). Incubations were carried out in the presence of either a polyclonal anti-CREM antibody (Fig. 1A, lanes 3, 4, 7, 8, 11, and 12) or purified rabbit IgG as control (Fig. 1A, lanes 1, 2, 5, 6, 9, and 10). The anti-CREM antibody is raised against the fulllength protein sequence, including the CREM DNA-binding domain present in all CREM isoforms. Therefore, this antibody is predicted to be active against all CREM isoforms. Results presented in Fig. 1 showed that the mock-treated YR2 nuclear extracts formed three major retarded complexes (C1, C2, and C3) when incubated with the TxRE1, TxRE2, or TxRE3 probe (Fig. 1A, lanes 1, 5, or 9, respectively). Treatment of BLV-infected YR2 cells with P ϩ I caused an increased binding activity to the TxREs (Fig. 1A, compare lanes 1 and 2, lanes 5 and 6, and lanes 9 and 10). Because P ϩ I treatment is known to activate BLV expression, these data suggest that initiation of BLV transcription could occur, in response to P ϩ I, through increased recruitment of CREB/CREM/ATF cellular transcription factors at the viral LTR TxREs. To evaluate the specificity of these interactions, unlabeled double-stranded oligonucleotides were prepared and used as competitors in EMSAs (Fig. 1B). Binding of proteins in complexes C1, C2, and C3 was shown to be sequence-specific to the CRE motif. Indeed, formation of these complexes was inhibited by a 80-fold molar excess of the unlabeled homologous TxRE1 oligonucleotide (Fig. 1B, compare  lanes 1 and 4) but not by the same molar excess of an unlabeled TxRE1 oligonucleotide mutated at the CRE site or an unlabeled oligonucleotide with an unrelated sequence containing an Oct-1 consensus binding site (Fig. 1B, compare lanes 2 and 3 with lane 4). Addition of the anti-CREM antibody in the binding reactions resulted in the complete (mocktreated cells) or partial (P ϩ I-treated cells) disappearance of the C2 complex concomitant with the appearance of a supershifted complex of lower mobility (S-CREM) but did not modify the migration profile of complexes C1 and C3 (Fig. 1A, lanes 3, 4, 7, 8, 11, and 12). In contrast, the presence of the IgG control did not affect formation of complexes C1, C2, and C3 (data not shown), indicating the specificity of the protein-antibody interaction. These results indicated that the C2 complex contained CREM isoform(s).
To further evaluate the binding of CREM to the BLV TxREs and its potential heterodimerization with other members of the CREB/ CREM/ATF family of transcription factors, we performed additional supershift experiments with antibodies directed against CREB, ATF-1, or ATF-2 alone or in combination with the anti-CREM antibody ( Fig. 2A). Addition of the anti-CREB antibody to the binding reaction interfered with formation of complex C2 and resulted in the appearance of a supershifted complex ( Fig. 2A, lane 2, 1 S-CREB), indicating that CREB is part of complex C2. As shown in Fig. 1, addition of the anti-CREM antibody impaired complex C2 formation concomitant with a supershifted complex with lower mobility ( Fig 5), indicating that formation of this complex results from ATF-2 binding to the TxRE1 in vitro. As negative control, we used the same antibodies in supershift experiments performed with a probe corresponding to the CRE-mutated TxRE1 site ( Fig. 2A, lanes 12-16), and we observed that mutation of the CRE motif abolished formation of complexes C1 and C2, indicating that binding of CREB, CREM, ATF-1, and ATF-2 to the BLV TxRE1 was specific to the CRE motif.
To evaluate bZIP complex formation on the BLV TxREs, we next performed additional supershift experiments by including combinations of the antibodies in the binding reactions. We observed a dramatic decrease of complex C2 formation con-FIGURE 1. Binding of CREM proteins from nuclear extracts of BLV-infected YR2 cells to the three BLV TxREs. A, oligonucleotides corresponding to the BLV TxRE1, TxRE2, and TxRE3 were 5Ј-end-labeled and used as probes in EMSAs. Prior to the addition of the TxRE1, -2, or -3 probe, 20 g of nuclear extracts from BLV-infected YR2 cells either mock-treated (lanes 1, 3, 5, 7, 9, and 11) or treated with phorbol 12-myristate 13-acetate (P) plus ionomycin (I) (P ϩ I: lanes 2, 4, 6, 8, 10, and 12) were incubated with either an anti-CREM antibody (lanes 3, 4, 7, 8, 11, and 12) or purified rabbit IgG as negative control (lanes 1, 2, 5, 6, 9, and 10). Binding reactions were analyzed by PAGE, and the retarded complexes were visualized by autoradiography. The three major DNAprotein complexes C1, C2, and C3 are indicated by arrows. The free probe was run out of the gel for better separation of the complexes. The supershifted (S-CREM) complexes are indicated by an arrow. B, specificity of the C1, C2, and C3 complexes was tested by competition experiments. The TxRE1 oligonucleotide was 5Ј-endlabeled and incubated with 10 g of nuclear extracts from BLV-infected YR2 cells in the absence of competitor comitant with the appearance of a super-supershifted complex ( Fig. 2A, lane 6, 5 SS-CREB ϩ CREM) when both the anti-CREB and the anti-CREM antibodies were added to the same binding reaction mixture, indicating that CREM⅐CREB heterocomplexes bind to the BLV TxRE1. As a control for cross-reactivity between the anti-CREM antibody and the CREB protein, we performed Western blot analysis on CREB (TNT-CREB) and CREM (TNT-CREM) proteins produced in vitro in a transcription-translation-coupled wheat germ extract reaction (Fig. 2B). This control experiment shows that the anti-CREM antibody had no cross-reactivity with the TNT-CREB protein (Fig. 2B, top panel, lane 3), thereby demonstrating the relevance of the anti-CREM antibody to distinguish between CREM-and CREB-related proteins. In parallel, the same amounts of the TNT-control, TNT-CREM, and TNT-CREB proteins were used in a similar Western blot experiment with the anti-CREB antibody, and we observed the specific recognition of the TNT-CREB protein (Fig. 2B, bottom panel, lane 3). Therefore, both the anti-CREM and the anti-CREB antibodies appear to be highly specific for recognition of their respective target proteins. Moreover, we performed additional supershift experiments further confirming the specificity of the anti-CREM antibody against CREM isoforms (supplemental Fig.  S1). The addition of both the anti-CREM and anti-ATF-1 antibodies to the same binding reaction resulted in a decreased complex C2 intensity and in the appearance of the 2 S-CREM supershifted complex, but also in the appearance of a super-supershifted complex ( Fig.  2A, lane 9, 7 SS-CREM ϩ ATF-1), suggesting the presence of CREM⅐ ATF-1 heterocomplexes in complex C2. When both the anti-CREB and the anti-ATF-1 antibodies were added to the same binding reaction, we observed an important decrease of complex C2 formation concomitant with the appearance of the 1 S-CREB supershifted complex but also with the appearance of a supersupershifted complex (Fig. 2A, lane 7  anti-ATF-2 antibody combined with either the anti-CREB antibody or the anti-CREM antibody or the anti-ATF-1 antibody did not result in the appearance of any super-supershifted complex, suggesting that ATF-2 did not bind to the BLV TxRE1 as heterocomplexes with those bZIP proteins ( Fig. 2A, lanes 8, 10, and 11).
In addition, we also performed EMSA and supershift experiments using nuclear extracts from PBMCs isolated from a healthy sheep (Fig. 2C) or from BLV-infected ovine PBMCs (data not shown). Results presented in Fig. 2C showed that incubation of the TxRE1 probe with the uninfected PBMC nuclear extracts caused the formation of two major retarded complexes (called N1 and N2, respectively) (Fig. 2C, lane 1). Addition of the anti-CREM antibody to the binding reaction resulted in a drastic reduction of the N1 complex concomitant with the appearance of a supershifted complex (S-CREM) but did not modify the migration profile of the complex N2 (Fig. 2C,  lane 2).
Taken together, our in vitro binding studies demonstrate that CREM proteins are expressed in a BLV-infected B lymphoma cell line as well as in ovine-infected and uninfected PBMCs and that these CREM proteins bind to the TxREs of the BLV promoter in vitro in a CRE-dependent manner. Moreover, our results suggest that complex C1 contains ATF-2 homodimers and that formation of complex C2 results from the binding of the CREB, CREM, and ATF-1 bZIP proteins to the BLV TxREs as homo-or heterodimers or as heterocomplexes.
CREB, CREM, and ATF Transcription Factors Are Recruited to the TxRE Region of the BLV Promoter in Vivo-To demonstrate in vivo, in the context of chromatin, the relevance of our in vitro binding studies, we performed chromatin immunoprecipitation (ChIP) assays using BLV-transformed YR2 cells. Formaldehyde cross-linked chromatin from these cells was used for immunoprecipitation with antibodies directed against ATF-1, ATF-2, CREB, or CREM or with a purified rabbit IgG as negative control. Following reverse of the cross-links, the purified DNA was subjected to radioactive PCR analysis using a set of primers flanking the TxRE region (nt Ϫ172 to ϩ29) of the BLV promoter. Fig. 3 shows PCR amplification of the DNA after immunoprecipitation (Fig. 3, IP antibody, lanes 3-7), PCR amplification of the input DNA (Fig. 3, lane 2), and PCR amplification in the absence of DNA as negative control (Fig. 3, PCR control, lane 1). Analysis of PCR products from immunoprecipitated DNA showed significant enrichment of the TxRE region when immunoprecipitation was carried out with the anti-CREB, anti-ATF-1, anti-ATF-2, or anti-CREM antibody (Fig. 3, lanes 4 -7, respectively). In contrast, no such enrichment was observed following immunoprecipitation of the cross-linked chromatin with purified rabbit IgG (Fig. 3, lane 3). These data confirmed the binding of CREB to the BLV TxREs in vivo, as demonstrated previously by our laboratory (9), and demonstrated in vivo the binding of other CREB/CREM/ATF members, including CREM protein(s), to this LTR region. As a control, we used another set of primers amplifying a region of the gag gene of the BLV provirus located ϳ1.8 kb downstream of the TxRE region, which has not been reported so far as binding CREB/ATF family members. Immunoprecipitations with all the antibodies did not enrich eluates with DNA from this control region in YR2 cell chromatin, demonstrating the spec-ificity of the TxRE interactions (data not shown). Thus, these data demonstrate the occupancy in vivo of the BLV TxRE region by CREB, ATF-1, ATF-2 and CREM proteins in the context of an integrated, chromatin-assembled BLV provirus.
Ectopic Expression of the CREM Activator Isoform Up-regulates BLV Promoter Activity through the Three BLV LTR CRElike Motifs-To assess the potential functional role of CREM factors in transcriptional regulation of the BLV promoter, we studied the effect of overexpression of a CREM isoform on BLV LTR-driven luciferase reporter gene expression. To this end, mouse B-lymphoid A20 cells were transiently transfected with a pLTRwt-luc reporter construct and with increasing amounts of an expression vector for the murine CREM activator isoform (pSV-CREM-wt) (23) and then assayed for luciferase activity (Fig. 4). Consistent with its role as a transcriptional activator, cotransfection of the CREM expression vector with the pLTRwt-luc construct resulted in a dose-dependent stimulation of the luciferase activity (up to 3.3-fold) (Fig. 4). This effect required intact CRE-like motifs located in the middle of each TxRE, because mutations in these motifs (pLTR-mutCREluc) almost completely abrogated CREM-mediated transactivation of the BLV LTR (Fig. 4). Because transiently transfected DNA does not always form proper chromatin structure, we next confirmed the above transient transfection results in the nucleosomal context of an episomally replicating pREP-based BLV LTR-luc plasmid (10), and we obtained CREM transactivations of the BLV-promoter similar to those observed in Fig. 4 with the nonepisomal pLTRwt-luc plasmid (data not shown).
We conclude from these experiments that ectopic CREM protein has a CRE-dependent stimulatory effect on the BLV promoter. These results thus establish the functional significance of CREM through the CRE sites present in the BLV 5Ј-LTR. Our data suggest that transcriptional activity of this promoter is positively regulated by CREM and that CREM could initiate BLV transcription in the absence of Tax BLV in newly infected cells.
BCR Cross-link Potentiates CREM Transactivation of the BLV Promoter-CREM is known to be phosphorylated on a particular key serine residue, Ser-117, located in its KID domain. Phosphorylation of the KID allows recruitment of two transcriptional coactivators CBP and p300 and leads to an important increase in CREM activating potential (14,15,31). It has been demonstrated that, ex vivo, BLV gene expression can be induced by cross-link of the B-cell receptor (BCR) with anti-IgG antibody (32). In B-cells, signaling events induced following BCR cross-link can lead to activation of several different pathways, i.e. the protein kinase C pathway, the Ca 2ϩ /calmodulin pathway, the Ras/extracellular signal-regulated kinase (ERK) pathway, and the phosphatidylinositol 3-kinase/Akt pathway (31,(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44). Although initially described as specific cAMP-responsive factors, CREB and CREM seem to be phosphorylated by a growing number of kinases, in response to different signaling cascades (45). Importantly, the phosphoacceptor sites (Ser-133 in CREB and Ser-117 in CREM) are the same as those targeted by protein kinase A (31,45). All the kinases activated in response to BCR cross-link would therefore be susceptible to phosphorylate Ser-117 of CREM.
To assess the role of BCR cross-link in CREM transactivation of the BLV promoter, we transiently cotransfected A20 cells with pLTRwt-luc and with increasing amounts of pSV-CREM-wt. Transfected cells were mock-treated or treated with anti-IgG antibody and then lysed and assayed for luciferase activity. Results presented in Fig. 5, A and B, showed that, in the absence of anti-IgG treatment, CREM transactivated BLV LTR-directed gene expression in a dose-dependent manner (up to 3.7-fold), whereas in the presence of anti-IgG treatment, we observed a dose-dependent increase in luciferase activity by ectopically expressed CREM (up to 5.7-fold) (Fig. 5B). These results indicate that cross-link of BCR with anti-IgG potentiates CREM transactivation of the BLV promoter and suggest that phosphorylation of CREM could modulate its ability to transactivate BLV gene expression.
CREM has been extensively studied in spermatogenesis (14, 15, 17, 46 -48). Whether this isoform of CREM is also involved in regulation of B-cell gene expression has been poorly documented. We therefore performed Western blot experiments using an anti-CREM antibody and nuclear extracts from YR2 cells treated or not with the combination P ϩ I for different times (Fig. 6A). These Western blot analyses showed that YR2 cells expressed a CREM isoform with an apparent molecular mass of 43 kDa, which corresponds to the expected molecular mass of the CREM isoform (49). However, the expression of this CREM isoform did not seem to be influenced by P ϩ I treatment (Fig. 6A, compare lane 1 with lanes 2-4). To further address the relevance of CREM involvement in BLV regulation in the YR2 BLV-infected cells, we have cloned a CREM1␣ cDNA by RT-PCR from YR2 cell mRNA. Cloned sequences were translated in amino acid sequences, and homology searches in protein and cDNA data bases led to the identification of a clone (found in three independent experiments) with a translated cDNA sequence harboring 92% of identity with the murine CREM1␣ isoform (supplemental Fig. S2). The ovine CREM1␣ cDNA sequence was then cloned into the eukaryotic expression vector pcDNA 3.1 (pCREM1␣) and tested in transient transfection experiments to confirm its transactivation potential on BLV promoter activity. A20 cells were transiently transfected with the pLTRwt-luc reporter construct and with increasing amounts of the ovine pCREM1␣ expression vector (Fig. 6B). Transfected cells were mock-treated or treated with anti-IgG antibody and then lysed and assayed for luciferase activity. Our results showed that, in the presence of the anti-IgG treatment, the ovine CREM1␣ isoform isolated from YR2 cells was able to activate BLV LTR-driven expression up to 4.8-fold (Fig. 6B). Moreover, similarly to what we observed in Figs. 4 and 5 with the murine pSV-CREM-wt, this effect required intact CRE-like motifs located in the middle of each TxRE, because mutations in these motifs (pLTR-mutCRE-luc) abrogated CREM1␣-mediated transactivation of the BLV LTR (Fig. 6B). Together, our data demonstrated that YR2 BLV-infected B-cells express a CREM1␣ isoform and that this isoform is able to transactivate BLV promoter-driven transcription following anti-IgG activation through the three CRE elements present at the middle of each TxRE.

CREM Ser-117 Mutation Reduces the BCR Cross-link Potentiation of CREM Transactivation of the BLV Promoter-To
address the role of CREM phosphorylation at Ser-117 in CREM-mediated anti-IgG stimulation of BLV LTR gene regulation, we substituted the Ser-117 residue for an alanine residue in the context of the pSV-CREM-wt vector, thereby generating pSV-CREM-S117A. We next transiently cotransfected A20 B-cells with pLTRwt-luc and with either the pSV-CREM-wt or the pSV-CREM-S117A expression vector. Cells were mock-treated or treated with anti-IgG antibody, and luciferase activities were measured in cell lysates (Fig. 7). We observed that mutation of CREM Ser-117 resulted in a decrease in CREM transactivation of the BLV promoter in the presence of anti-IgG treatment (Fig. 7, anti-IgG, compare pSV-CREM and pSV-CREM-S117A). These data indicate that CREM Ser-117 mutation reduced anti-IgG potentiation of CREM transactivation of the BLV 5Ј-LTR, suggesting that this potentiation was mediated, at least in part, through CREM Ser-117 phosphorylation.
CBP/p300 Coactivators Are Involved in CREM Transactivation of the BLV 5Ј-LTR-Ser-117 phosphorylation is known to activate CREM-directed transcription through the recruitment of two coactivators, CBP and p300. Therefore, we decided  to examine the potential functional effect of CBP/p300 on CREM-mediated transactivation of the BLV promoter. To this end, A20 cells were cotransfected with pLTRwt-luc, with pSV-CREM-wt, and with an expression vector coding for the adenovirus wild-type E1A 12S protein (pE1A-12S-wt) or an expression vector coding for a mutant form of the E1A 12S protein (pE1A-12S/2-36del) as negative control. The adenovirus E1A oncoprotein is known to bind to CBP/p300 and to inhibit their capacity to coactivate transcription (50 -52). The pE1A-12S/2-36del expression vector codes for an E1A mutant protein that lacks the ability to interact with CBP/p300 and fails to inhibit CBP/p300 activity. Results presented in Fig. 8A showed that coexpression of the wild-type E1A 12S protein decreased (by ϳ2-fold) CREM-mediated transactivation of the BLV promoter, both in the presence and in the absence of anti-IgG treatment, when compared with coexpression of the mutant E1A/2-36del protein (Fig. 8A). These results suggest that CBP/p300 are functionally involved in CREM transactivation of the BLV LTR.
To determine whether CBP is recruited at the BLV promoter region in BLV-infected YR2 cells, we performed ChIP assays. For these assays, YR2 cells were treated with P ϩ I for 30 min and then formaldehyde cross-linked for another 30 min. Chromatin from these cells was used for immunoprecipitation with anti-CBP antibody or with a purified rabbit IgG as negative control. Following reverse of the cross-links, the purified DNA was subjected to radioactive PCR analysis with primers flanking the TxRE region of the BLV promoter. Fig. 8B shows PCR amplification of the DNA after immunoprecipitation. Enrichment of the TxRE region when immunoprecipitation was carried out with the anti-CBP antibody demonstrated that CBP is recruited at the BLV promoter in vivo (Fig. 8B, lane 4). As a control, we used another set of primers amplifying a region of the gag gene of the BLV provirus located ϳ1.8 kb downstream of the TxRE region, which has not been reported so far as binding CBP. Immunoprecipitations with the CBP antibody did not enrich eluates with DNA from this control region in YR2 cell chromatin, demonstrating the specificity of the TxRE interactions (data not shown). Moreover, we performed functional studies to further illustrate the involvement of CBP/p300 in FIGURE 8. Functional involvement of CBP in transactivation of the BLV LTR by CREM. A, A20 cells were transiently cotransfected using the DEAEdextran procedure with 500 ng of pLTRwt-luc and 500 ng of either pSV-CREM-wt or the empty vector pSG5 in the presence of either pE1A-12S-wt or pE1A-12S/2-36del. At 20 h post-transfection, cells were mock-treated or treated with anti-IgG antibody (6.5 g/ml). Luciferase activities were measured in cell lysates 42 h after transfection and were normalized with respect to protein concentrations of the lysates. Results are presented as histograms indicating relative light units (RLU) with respect to the basal activity of pLTRwt-luc in the absence of CREM and in the absence of anti-IgG treatment, which was assigned a value of 1. B, ChIP assays were performed in order to demonstrate CBP recruitment to the BLV promoter. Cells were treated with P ϩ I for 1 h; DNA and protein were cross-linked with formaldehyde for 30 min, and DNA was sheared. The cross-linked DNA-protein complexes were immunoprecipitated (IP) with the anti-CBP antibody (lane 4) or with purified rabbit IgG as negative control (lane 3). Cross-links were reversed, and purified DNA was amplified by semi-quantitative radioactive PCR using a primer set amplifying the BLV promoter TxRE region (nt Ϫ172 to ϩ29). PCR of the input (sample representing PCR amplification from a 1:25 dilution of total input chromatin from the ChIP experiment) is shown in lane 2. The PCR control represents the PCR amplification in the absence of DNA (lane 1). C, HeLa cells were transiently cotransfected using the jetPEI TM procedure with 500 ng of pLTRwt-luc and increasing amounts (from 0 to 200 ng of plasmid DNA) of pRc-RSV-CBP. To maintain the same amount of transfected DNA and to avoid squelching artifacts, the different amounts of CBP expression vector cotransfected were complemented to 200 ng of DNA by using the parental empty vector. Luciferase activities were measured in cell lysates 42 h after transfection and were normalized with respect to protein concentrations of the lysates. Results are presented as histograms indicating relative light units (RLU). D, HeLa cells were transiently cotransfected using the jetPEI TM procedure with 500 ng of pLTRwt-luc and increasing amounts (from 0 to 60 ng of plasmid DNA) of pSV-CREM-wt, in the presence or the absence of 50 ng of pRc-RSV-CBP. To maintain the same amount of transfected DNA and to avoid squelching artifacts, the different amounts of CREM expression vector cotransfected were complemented to 60 ng of DNA by using the empty vector pSG5. Luciferase activities were measured in cell lysates 42 h after transfection and were normalized with respect to protein concentrations of the lysates. Results are presented as histograms indicating relative light units (RLU). Means Ϯ S.E. from a representative experiment performed in triplicate of three independent transfections are shown.
CREM transactivation of the BLV promoter. HeLa cells were cotransfected with the pLTRwt-luc reporter construct together with increasing amounts of an expression vector for the CBP coactivator (pSV-CBP), and assayed for luciferase activity (Fig.  8C). Cotransfection of the CBP expression vector resulted in a dose-dependent stimulation of the BLV LTR-driven luciferase activity (up to 7-fold) (Fig. 8C). Additional cotransfection experiments with the CBP expression vector along with increasing amounts of the pSV-CREM-wt expression vector demonstrated that CBP overexpression potentiated CREM transactivation of the BLV promoter (Fig. 8D). Taken together, these data demonstrate for the first time the role of CBP in BLV transcriptional regulation and suggest that CBP recruitment to the BLV promoter is involved in the mechanism of CREM transactivation.

CREM Has No Effect on Transactivation of the BLV Promoter by the Viral Transactivator Tax BLV , whereas CREB and Tax BLV Synergistically Activate BLV Promoter Transcriptional
Activity-Several studies have suggested that Tax BLV transactivation of the BLV promoter is mediated through interactions of Tax BLV with cellular factors that bind to the CRE motifs of the viral TxREs. Our laboratory has reported previously that a dominant negative inhibitor of CREB (A-CREB) efficiently inhibits Tax BLV transactivation of the BLV promoter, suggesting that recruitment of CREB to the BLV TxREs plays a critical role in the mechanism of Tax BLV transactivation (9). However, the effect of CREM expression on Tax BLV -mediated transactivation of the BLV 5Ј-LTR has not been examined so far. Therefore, we transiently cotransfected the B-lymphoid A20 cell line with pLTRwt-luc, the CREM expression vector, and a Tax BLV expression vector (pSG-Tax BLV ). Because Ser-117 phosphorylation of CREM could also intervene in Tax BLV -mediated transactivation, transfected cells were mock-treated or treated with anti-IgG antibody. As control, we compared in the same transfection experiment the effect of CREB expression on Tax BLV -mediated transactivation of the BLV LTR, by including cotransfections with a CREB expression vector (pSG-CREB) instead of the CREM expression vector. Results presented in Fig. 9 showed that, in the absence of anti-IgG treatment, cotransfection of 20 ng of the pSG-Tax BLV alone resulted in a 40.7-fold stimulation of the luciferase activity (Fig. 9) and that cotransfection of pSV-CREM-wt or pSG-CREB-wt alone resulted in a 1.6-or 10.9-fold activation of transcription, respectively (Fig. 9). Remarkably, when cells were cotransfected with both the pSG-Tax BLV and pSG-CREB-wt, a strong synergism was observed between Tax BLV and CREB, resulting in a 197-fold activation of the BLV promoter activity (Fig. 9). Transcriptional activators synergize when their combination produces a transcriptional rate that is greater than the sum of the effects produced by the individual activators (53). In the latter case, the amount of transcription in the presence of both Tax BLV and CREB (197-fold) is 4.9-fold greater (fold synergism) than the sum of the effect produced by each activator separately (10.9 ϩ 40.7). These results demonstrate that CREB synergistically enhances transcriptional activation of the BLV promoter by Tax BLV , and we confirm that CREB plays an important role in the mechanism of Tax BLV transactivation. On the contrary, when cells were cotransfected with both pSG-Tax BLV and pSG-CREM-wt, no synergism between Tax BLV and CREM was observed. Indeed cotransfection of both expression vectors led to a 38-fold stimulation of the luciferase activity, which is similar to the sum of the effect produced by Tax BLV and CREM individually (40.7 ϩ 1.6) (Fig. 9), suggesting that CREM is not involved in Tax BLV transactivation of the BLV LTR.
The Tax BLV /CREB synergism was also observed following anti-IgG treatment. Indeed, when cells were cotransfected with both the Tax BLV expression vector and the CREB expression vector, anti-IgG treatment resulted in a 1242-fold stimulation of the luciferase activity, which was 6.5-fold (fold synergism) greater than the sum of the activations by Tax BLV alone and CREB alone (19.5 ϩ 192) in the presence of anti-IgG treatment (Fig. 9). On the contrary, no synergism was observed between Tax BLV and CREM in the presence of anti-IgG treatment. Indeed, in the presence of anti-IgG treatment, cotransfection of Tax BLV and CREM expression vectors resulted in a 210-fold increase of the luciferase activity, which was approximately equal to the sum of the activations by Tax BLV alone and CREM alone (192 ϩ 3.8) (Fig. 9). This suggests that, in the presence of IgG-treatment, CREM is not involved in Tax BLV transactivation of the 5Ј-LTR .
Thus, our results demonstrate a strong synergism between Tax BLV and CREB in transcriptional transactivation of the BLV promoter, both in the absence and presence of BCR cross-link, and confirm previous studies from our laboratory indicating that CREB recruitment at the BLV promoter plays an important role in the mechanism of transactivation by Tax BLV (9). In contrast, our results show no synergism between Tax BLV and FIGURE 9. CREB and Tax BLV synergistically activate BLV promoter transcriptional activity, whereas CREM has no effect on Tax BLV -mediated transactivation of the BLV promoter. A20 cells were transiently cotransfected using the DEAE-dextran procedure with 500 ng of pLTRwt-luc and with 20 ng of a Tax BLV expression vector (pSG-Tax BLV ) in the presence or the absence of 500 ng of pSV-CREM-wt or pSG-CREB-wt. To maintain the same amount of transfected DNA and to avoid squelching artifacts, the different amounts of expression vectors cotransfected were complemented to 520 ng of DNA by using the empty vector pSG5. At 20 h post-transfection, cells were mock-treated or treated with anti-IgG antibody (6.5 g/ml). Luciferase activities were measured in cell lysates 42 h after transfection and were normalized with respect to protein concentrations of the lysates. Results are presented as histograms indicating relative light units (RLU) with respect to the basal activity of pLTRwt-luc in the absence of CREM, CREB, Tax BLV , and anti-IgG treatment, which was assigned a value of 1. Means Ϯ S.E. from a representative experiment performed in triplicate of three independent transfections are shown.
CREM, suggesting that CREM is not involved in the mechanism of Tax BLV -mediated transactivation.
CREM Counteracts the Functional Action of CREB and Affects the Tax BLV /CREB Synergistic Activation of the BLV Promoter-Because both CREB and CREM are recruited to the BLV promoter through the CRE-like motifs of the TxREs, we next wanted to determine their mutual influence on Tax BLVmediated transactivation. We therefore tested whether CREM could counteract the action of CREB and affect the Tax BLV /CREB synergistic activation of BLV promoter transcriptional activity and/or whether CREB could counteract the action of CREM and synergize with Tax BLV in the presence of CREM. To this end, we performed cotransfection experiments in A20 B-cells with pLTRwt-luc, with pSG-Tax BLV , and with a fixed amount of pSG-CREB in the presence of increasing amounts of pSV-CREM (Fig. 10A), or with pLTRwt-luc, with pSG-Tax BLV , and with a fixed amount of pSV-CREM in the presence of increasing amounts of pSG-CREB (Fig. 10B). Results showed that, in both the absence and the presence of anti-IgG treatment, the Tax BLV /CREB synergism was repressed by increasing amounts of ectopically expressed CREM (Fig.  10A). Moreover, cotransfection of increasing amounts of CREB in the presence of a fixed amount of CREM activated synergistically with Tax BLV LTR-directed reporter gene expression (Fig.  10B). Taken together, our results suggest that CREB and CREM exert opposite roles in Tax BLV -mediated transactivation of the BLV promoter and that a modulation of the level of Tax BLV transactivation takes place through a competition between the actions of CREB and CREM. This competition would depend on the relative spatio-temporal expression levels of those proteins in BLV-infected cells.

DISCUSSION
In this study, we have shown that CREM proteins were expressed in the YR2 cell line derived from leukemic cells of a BLV-infected sheep as well as in infected and uninfected ovine PBMCs and that these CREM proteins bound to the BLV promoter in vitro, through the three CRE-like motifs present in the middle of each viral TxRE. Supershift analysis suggested that in vitro CREM formed heterocomplexes with CREB and ATF-1 to bind to the BLV TxREs. Moreover, chromatin immunoprecipitation analysis using the BLV-infected YR2 B-cell line demonstrated the recruitment of CREM proteins to the BLV promoter in vivo, in the chromatin context of a BLV-integrated provirus. Functionally, we showed that in the absence of the viral Tax BLV protein, ectopic expression of the CREM activator isoform up-regulated BLV promoter activity through the three BLV LTR CRE-like motifs. Cross-link of the B-cell receptor potentiated CREM transactivation of the viral promoter. Further experiments supported the notion that this potentiation could involve CREM Ser-117 phosphorylation and recruitment of CBP/p300 by CREM. Finally, CREB and CREM seem to play opposite roles in Tax BLV -mediated transactivation of the BLV promoter. Although CREB and Tax BLV synergistically activated the LTR, CREM repressed this Tax BLV /CREB synergism.
In conclusion, these results suggested that activation of Tax BLVindependent BLV transcription could occur in vivo, in response to B-cell activation. CREM could thus initiate a low level of BLV transcription and lead to the synthesis of small amounts of the Tax BLV transactivator, which would then amplify transcription from the 5Ј-LTR. In this regard, we have demonstrated that CREM proteins are expressed in PBMCs isolated from a healthy sheep and are able to bind to the BLV promoter in vitro, supporting our hypothesis that CREM could initiate BLV transcription in the absence of Tax BLV in newly infected cells. Moreover, we have shown that CREM did not potentiate Tax BLV -mediated transactivation of BLV LTR-directed gene expression. Indeed, our cotransfection experi- FIGURE 10. Opposite roles of CREB and CREM during Tax BLV -mediated transactivation of the BLV promoter. A, A20 cells were transiently cotransfected using the DEAE-dextran procedure with 500 ng of pLTRwt-luc, 250 ng of pSG-CREB, 20 ng of pSG-Tax BLV , and increasing amounts of pSV-CREM (from 0 to 750 ng of plasmid DNA). To maintain the same amount of transfected DNA and to avoid squelching artifacts, the different amounts of CREM expression vector cotransfected were complemented to 750 ng of DNA by using the empty vector pSG5. B, A20 cells were transiently cotransfected using the DEAE-dextran procedure with 500 ng of pLTRwt-luc, 250 ng of pSV-CREM, 20 ng of pSG-Tax BLV , and increasing amounts of pSG-CREB (from 0 to 750 ng of plasmid DNA). To maintain the same amount of transfected DNA and to avoid squelching artifacts, the different amounts of CREB expression vector cotransfected were complemented to 750 ng of DNA by using the parental empty vector pSG5. A and B, at 20 h post-transfection, cells were mock-treated or treated with anti-IgG antibody (6.5 g/ml). Luciferase activities were measured in cell lysates 42 h after transfection and were normalized with respect to protein concentrations of the lysates. Results are presented as histograms indicating relative light units (RLU) with respect to the basal activity of pLTRwt-luc in the absence of CREB, CREM, Tax BLV , and anti-IgG treatment, which was assigned a value of 1. Means Ϯ S.E. from a representative experiment performed in triplicate of three independent transfections are shown.
ments indicated that overexpression of CREM repressed the synergistic activation of the BLV promoter by Tax BLV and CREB. The absence of viral proteins at the surface of BLVinfected cells is necessary to escape from an efficient host immune attack and to allow BLV-induced tumor development. However, because the tax BLV gene product is considered to account for viral leukemogenicity (22, 54 -56), a low level of basal BLV expression is likely necessary at the early stage of BLV infection to induce transformation of the infected cells. Based on our results we suggest a bimodal role of CREM in BLV LTR-directed transcription: in the absence of Tax BLV , CREM would initiate BLV transcription in response to activation of the BLV-infected B-cells, and on the contrary, in the presence of Tax BLV , CREM would decrease Tax BLV -mediated transactivation of the BLV promoter by competing with CREB for binding to the viral CREs. CREM would thus be able to initiate a low level of BLV expression necessary for Tax BLVinduced transformation of the infected cells, but in the presence of Tax BLV , CREM would reduce BLV expression and thus facilitate immune escape and allow tumor development.
BLV infection is characterized by viral latency, rendering the infected animals asymptomatic seropositive carriers, for whom only a small fraction will develop persistent lymphocytosis or B-cell lymphosarcoma several years later (32,56). During this latency period, expression of the virus is thought to be blocked at the transcriptional level (32,57,58). However, little is known about specific molecular events responsible for this viral latency. Some studies have suggested that a blocking factor could be implicated in the inhibition of viral synthesis in vivo (59 -62), but the nature of this factor and its mechanism of action have never been established. To date, all the regulatory proteins known to be recruited to the BLV promoter and to influence BLV transcriptional activity are activating transcription factors (CREB, ATF-1, PU.1/Spi-B, USF1, IRF, etc.). In this study, we propose a role for CREM proteins in transcriptional repression of the BLV LTR in the presence of Tax BLV . Moreover, even though the CREM isoform transactivates the BLV promoter, we did not investigate the potential functional role of other CREM isoforms. The CREM gene consists of 14 exons (63,64). Alternative exon splicing and utilization of alternative promoters and translation initiation codons result in the production of functionally different CREM proteins with either activating or repressing potential on target gene expression (23,46). However, as all CREM isoforms contain at least the bZIP domain, they potentially could be recruited to the BLV promoter through the three CRE-like motifs and could then up-or down-regulate BLV provirus transcription. Preliminary results from our laboratory suggest that a CREM repressor isoform is expressed in latently BLV-infected cells. 8 Recruitment of such repressor to the BLV promoter could contribute to the silent status of the BLV promoter observed in the majority of BLVinfected cells in vivo.
Regulation of the expression of the different activator and repressor isoforms of the CREM gene has been extensively studied in spermatogenesis (14, 15, 17, 46 -48). During sper-matogenesis, only CREM repressors are expressed in pre-meiotic germ cells, but a switch to the expression of the CREM and CREM2 activators occurs in post-meiotic germ cells (15,17,46,65) and allows expression of testis-specific genes such as calspermine, TP1, and RT7 (66 -68). However, little is known about regulation of CREM expression in B-cells. In this study we have shown that BLV-infected YR2 B-cells express a CREM1␣ activator isoform and demonstrate the functional role of this isoform in BLV expression. Two different studies have studied specific expression of the CREM gene in B-lymphoid cells. The first study, published by Pongubala and Atchison (69), has suggested through RNA analysis that the CREM gene is expressed differentially in the B-cell lineage. More specifically, their hypothesis is that repressor isoforms of CREM could predominate at the pre-B-cell stage and that an activator form of CREM could replace these repressive forms of CREM at later stages of B-cell development. The second study, published by van der Stoep et al. (70)., has demonstrated that ICER downregulates the constitutive transcriptional activity of the CIITA-PIII promoter in B-cells, indicating the implication of another CREM isoform (ICER) during B-cell specific gene expression. Together, these two studies suggest that, like in germ cells, expression of the different CREM isoforms would be finely regulated during differentiation and activation of B-cells. At the early stage of BLV infection, predominant expression of CREM repressor isoforms in the infected B-cells would explain the absence of BLV expression. However, a switch in CREM gene expression following B-cell activation or differentiation would allow initiation of BLV transcription, thereby generating the first Tax BLV molecules. In this regard, it has been demonstrated that, in BLV-infected but asymptomatic sheep, BLV integrated both in CD5Ϫ and CD5ϩ B-cells. In lymphoma, however, BLV provirus was detected only in CD5Ϫ-B-cells but not in CD5ϩ B-cells (71). Furthermore, van den Broeke et al. (72) have demonstrated that a variety of ovine B-cell populations can support a productive BLV infection, suggesting that BLV can infect and replicate in both immature and mature B-cells.
In the case of the closely related HTLV-I retrovirus, several studies have investigated the potential role of different CREM isoforms during basal and Tax HTLV-I -activated transcription of the HTLV-I promoter. Bodor et al. (73) have demonstrated that CREM binds to all three HTLV-I CRE-like motifs and attenuates the activation of a HTLV-I LTR-CAT reporter construct by Tax HTLV-I and protein kinase A. Further studies by the same group have shown that elevated cAMP levels in T cells correlate with expression of the potent transcriptional repressor ICER (74). Moreover, it has been shown that quiescent PBMCs respond to cell activation by producing sustained ICER RNA levels through 18 h post-stimulation and that ICER is able to inhibit Tax HTLV-I -mediated transactivation of the HTLV-I promoter (75), suggesting a role for ICER in the establishment and maintenance of a persistent HTLV-I infection. In contrast, Laurance et al. (76) have shown that binding of the CREM␣ isoform to the HTLV-I CREs allows Tax HTLV-I to stimulate viral LTR expression, but with concomitant inhibition of Tax HTLV-I effects on other CRE-containing genes. Together, these studies indicate that the level of HTLV-I transcription is influenced by the expression pattern of the different CREM isoforms accord-  JULY 20, 2007 • VOLUME 282 • NUMBER 29

Regulation of the BLV Promoter by CREM
ing to the state of T-cell activation and/or differentiation. Both BLV and HTLV-I oncoretroviruses could therefore use similar strategies, depending on the CREB and CREM gene expression levels, to inhibit and/or activate viral transcription according to the activation state of the infected cell.