An Interferon Regulatory Factor Binding Site in the U5 Region of the Bovine Leukemia Virus Long Terminal Repeat Stimulates Tax-Independent Gene Expression

Abstract

Bovine leukemia virus (BLV) replication is controlled by both cis- and trans-acting elements. The virus-encoded transactivator, Tax, is necessary for efficient transcription from the BLV promoter, although it is not present during the early stages of infection. Therefore, sequences that control Tax-independent transcription must play an important role in the initiation of viral gene expression. This study demonstrates that the R-U5 sequence of BLV stimulates Tax-independent reporter gene expression directed by the BLV promoter. R-U5 was also stimulatory when inserted immediately downstream from the transcription initiation site of a heterologous promoter. Progressive deletion analysis of this region revealed that a 46-bp element corresponding to the 5′ half of U5 is principally responsible for the stimulation. This element exhibited enhancer activity when inserted upstream or downstream from the herpes simplex virus thymidine kinase promoter. This enhancer contains a binding site for the interferon regulatory factors IRF-1 and IRF-2. A 3-bp mutation that destroys the IRF recognition site caused a twofold decrease in Tax-independent BLV long terminal repeat-driven gene expression. These observations suggest that the IRF binding site in the U5 region of BLV plays a role in the initiation of virus replication.

Bovine leukemia virus (BLV) is a naturally occurring B-lymphotropic retrovirus that infects cattle (30). It is the etiologic agent associated with enzootic bovine leukosis, a chronic lymphoproliferative disease complex. The majority of BLV-infected cattle are asymptomatic carriers of the virus. Only about 30% of BLV-infected animals develop a preneoplastic condition termed persistent lymphocytosis, with 2 to 5% developing leukemia and/or lymphoma after a long latency period. Sheep experimentally inoculated with BLV are readily infected, and a high percentage of infected animals develop B-cell lymphoma. Due to structural and biological similarities, BLV is classified in the Retroviridae family along with the human T-lymphotropic virus type 1 (HTLV-1) and type 2 (HTLV-2) (8, 48, 49, 51).

Replication of BLV is transcriptionally and posttranscriptionally regulated by the viral gene products Tax and Rex, which are synthesized from a common doubly spliced mRNA (12, 38, 47). The Rex protein interacts with the Rex-responsive element (RxRE) in the 3′ R region of the viral mRNAs (50) and enhances the cytoplasmic accumulation of singly spliced and unspliced transcripts. This leads to an increase in the production of structural proteins and to a decrease in the level of the doubly spliced tax-rex mRNA (13). The Tax protein transactivates the BLV promoter through a triplicate motif of 21 bp (called the Tax-responsive element, or TxRE) present in the U3 promoter region of the BLV 5′ long terminal repeat (LTR) (12, 27, 59). There is no evidence that Tax binds directly to the TxRE. Rather, Tax is thought to associate with cellular proteins that can bind to the viral DNA. The TxRE sequence contains a cyclic AMP response element (CRE) that has been shown to bind the CRE-binding protein (CREB) and activating transcription factors 1 and 2 (ATF-1 and ATF-2) (1, 2, 60).

Early studies on the transcriptional activity of the BLV LTR concluded that it is a highly restricted promoter which is totally dependent on the presence of the viral transactivator Tax (14, 48). Indeed, transient transfection of various cell lines with plasmids containing the cat (chloramphenicol acetyltransferase) gene under the control of the BLV LTR did not yield detectable Cat activity except in cells expressing Tax (14, 15, 48). However, when cultured cells that do not express Tax are transfected with a plasmid containing a complete proviral genome, viral genes are expressed (56) and sheep can be infected by injection with proviral DNA (61, 62). An internal promoter that can direct expression of the tax gene has not been described so far. Most likely, a low level of LTR-driven Tax-independent transcription occurs and leads to Tax synthesis and accumulation in the early stages of viral infection. Although the elements that control virus transcription in the absence of the viral regulatory proteins likely play an important role in the initiation and maintenance of virus replication, very little is known about the basal transcriptional activity of the BLV LTR.

In uninfected cells, it is possible to induce BLV LTR-driven cat expression in the absence of Tax, by cotransfection of expression vectors for CREB, ATF-1, and ATF-2 in combination with protein kinase A or Ca²⁺/calmodulin-dependent protein kinase IV (1, 60). Furthermore, a functional NF-κB binding site has been identified in the U3 region of the BLV LTR (7). Constitutive expression of NF-κB in B cells could induce low levels of transcriptional activity, which in turn can be upregulated following immunological activation of the cell and thus initiate a positive feedback regulatory loop involving the Tax protein.

The region situated immediately downstream from the transcription start site in the BLV LTR is involved in regulation of viral gene expression. Removal of this region, between position +48 relative to the transcription initiation site and the 3′ end of the LTR (nucleotide [nt] +320), reduces LTR-driven gene expression by 87% in BLV-infected cells (14). This effect was attributed to the R region, since in the absence of viral proteins, a 250-bp element (−22 to +223) containing the R region stimulated gene expression from a simian virus 40 (SV40) minimal promoter. This element is stimulatory independently of its orientation but is effective only when located immediately downstream from the transcription start site (15). Recently, the presence of a 64-bp downstream activator sequence (DAS) at the 3′ end of the R region (+147 to +211) has been reported (31).

Downstream regulatory sequences have also been identified in the HTLV-1 LTR. A 45-bp element that is located at the boundary of R-U5 and binds the YB-1 transcription factor is required for Tax-independent transcription (25, 26). On the other hand, binding of the Sp1 and Sp3 transcription factors to the HTLV-1 U5 region has been associated with transcriptional repression of the LTR (40, 41). Furthermore, it has been suggested that the interaction of CREB and ATF-2 with the R region of the HTLV-1 LTR is associated with viral latency (63, 64).

This study further characterizes the regulatory activity exerted on Tax-independent BLV promoter-driven gene expression by the LTR regions located downstream from the transcription start site. We have identified a transcriptional enhancer in the 5′ portion of the BLV U5 region that acts independently of any viral regulatory proteins. This element contains a binding site for interferon (IFN) regulatory factor 1 and 2 (IRF-1 and IRF-2), as demonstrated by gel retardation assay. A 3-bp mutation that abolishes protein binding to this motif caused a twofold decrease in LTR Tax-independent promoter activity.

MATERIALS AND METHODS

Plasmid constructs.

The BLV LTR used in this study is the LTR of the T15 provirus described by Couez et al. (9). Our sequencing data indicated three errors in the published sequence: we found a 1-nt insertion (G) between positions −185 and −186, a T in position −115 instead of a C, and a C in position −116 instead of a T. The nucleotide numbering of the LTR refers to the RNA initiation site as defined in reference 9 but takes into account the additional nucleotide at position −186. The first nucleotide of R and the last nucleotide of U3 are considered +1 and −1, respectively.

The pTK-cat vector was derived from pBLCAT6 (a gift from Günther Schütz) (4) by insertion of a 157-bp SphI/XbaI fragment containing the minimal promoter of the thymidine kinase gene of the herpes simplex virus (the HSV TK promoter). The HSV TK promoter fragment extends from −106 to +51 relative to the CAP site and was obtained by PCR amplification using pBLCAT2 (37) as a template and primers introducing an SphI site at the 5′ extremity and an XbaI site at the 3′ extremity. All LTR fragments were obtained by PCR amplification with primers containing the restriction sites used for subcloning into the pTK-cat or the pGL2-basic vector (Promega). The U5-I fragment was a synthetic double-stranded oligonucleotide. The LTR fragments cloned into pTK-cat were inserted either into the XbaI-BglII sites, downstream from the promoter, or into the HindIII site, upstream from the promoter. Cloning into pGL2-basic was performed by using the KpnI-BglII sites of the vector, upstream from the firefly luc gene. Mutation of the IRF-binding site within the LTR was generated by a two-step PCR process (22). All constructs were verified by cycle sequencing using the Thermosequenase DNA sequencing kit (Amersham). The pHHcat plasmid containing the human Mx promoter was a gift from Jean Content (24).

Cell culture.

All media, sera, and supplements were from GIBCO-BRL. Raji cells were grown in RPMI 1640-Glutamax I medium supplemented with 10% fetal bovine serum, 50 U of penicillin/ml, and 50 μg of streptomycin/ml. Daudi cells were maintained in RPMI 1640-Glutamax I medium with 10% Myoclone Superplus fetal bovine serum, 10 mM HEPES buffer, 1 mM sodium pyruvate, nonessential amino-acids, 50 U of penicillin/ml, and 50 μg of streptomycin/ml. OVK and MDBK cells were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, 1 mM sodium pyruvate, 2 mM glutamine, nonessential amino acids, and 100 μg of kanamycin/ml. All cells were grown at 37°C in an atmosphere of 5% CO₂.

Transient transfection.

Daudi cells were transfected by electroporation. Cells were harvested in exponential-growth phase and resuspended in supplemented RPMI 1640 at a concentration of 10⁷ cells per 400 μl. Then 400 μl of cells were mixed with 5 μg of plasmid DNA, incubated for 15 min at room temperature, transferred to electroporation vials, and electroporated at 250 V with a capacitance of 960 μF (by using a Bio-Rad gene pulser). Transfected cells were collected, plated out immediately in 15 ml of preheated medium (a 1:1 mixture of fresh culture medium and supernatant of a 24-h culture), and grown for 48 h at 37°C.

MDBK cells were transfected by using DEAE-dextran (16). Cells were seeded at 5 × 10⁵ cells per 10-cm-diameter dish 24 h prior to transfection. The plasmid DNA (20 μg) was mixed with 2.5 ml of transfection solution (50 mM Tris-Cl, 20 mM Na-HEPES, 750 μg of DEAE-dextran/ml in serum-free medium [pH 7.2]) and added to cells that had been washed twice with serum-free medium. Cells were incubated for 4 h at 37°C in the presence of the transfection complex, washed twice with serum-free medium, and grown for 48 h in fresh supplemented medium.

OVK cells were transfected by the calcium phosphate precipitate method (20). Briefly, the cells were seeded at 5 × 10⁵ cells per 10-cm-diameter dish 24 h prior to transfection. Calcium phosphate precipitate was allowed to form for 20 min, at room temperature, in the presence of 20 μg of plasmid DNA, and then was added to the cells. Cells were incubated for 16 h in the presence of the transfection complex, washed with serum-free medium, and grown for an additional 48 h in fresh supplemented medium.

Raji cells were transfected by using either DEAE-dextran or the Superfect transfection reagent (Qiagen). The Superfect transfection reagent was used to transfect the cat reporter plasmids, according to the manufacturer’s protocol, with 5 × 10⁶ cells and 5 μg of DNA per transfection. The DEAE-dextran procedure was used to transfect luc reporter plasmids. Cells were harvested at a density of 10⁶/ml, washed with STBS (25 mM Tris-Cl [pH 7.5], 137 mM NaCl, 5 mM KCl, 700 μM CaCl₂, 500 μM MgCl₂, 600 μM Na₂HPO₄), and resuspended at a concentration of 3 × 10⁶ cells in 300 μl of a mixture containing 260 ng of plasmid DNA (250 ng of the pGL2-basic-derived construct and 10 ng of the pRL-SV40 control plasmid [Promega]) and 450 μg of DEAE-dextran (Pharmacia)/ml in STBS. Cells were incubated for 1 h at 37°C, washed twice with 1 ml of STBS and once with culture medium, and grown in 3 ml of supplemented medium for 24 h.

Cat assay.

Cat activity was assayed as described in reference 19 with some slight modifications. Cells were harvested 48 h after transfection, washed once with TNE (40 mM Tris-Cl [pH 7.8], 150 mM NaCl, 1 mM EDTA), and resuspended in 100 μl of 250 mM Tris-Cl (pH 7.8). After three freeze-thaw cycles and a 5-min incubation at 65°C, the lysate was centrifuged for 10 min at 13,000 × g, and the supernatant was recovered. The protein content of cellular extracts was determined by the Bradford protein assay (Bio-Rad) (6), and Cat assays were performed in 100-μl reaction mixtures, in the presence of 1 mM acetyl coenzyme A and 0.125 μCi of deoxychloramphenicol (Amersham). The substrate and product were separated by thin-layer chromatography, and the results were quantified with a Molecular Dynamics PhosphorImager.

Luciferase assay.

Luciferase assays were performed as previously described (57).

EMSA.

Nuclear extracts were prepared by a rapid method previously described (42, 57). For preparation of nuclear extracts from IFN-α-treated cells, Daudi cells were cultured for 6 h in the presence of 500 U of IFN-α (a gift from Samira Majjaj)/ml prior to extract preparation. Protein concentration was determined by the method of Bradford (6). The sequences of the oligonucleotides used for this study are listed in Fig. 4A. Electrophoretic mobility shift assays (EMSAs) were performed as previously described (57). The concentration of poly(dI)-poly(dC) (Pharmacia) used as a nonspecific competitor in the binding reaction was optimized for each probe and is mentioned in the Fig. 4 legend. For supershift assays, polyclonal antibodies against Stat-1 and Stat-2 (gifts from Chris Schindler) or rabbit preimmune serum was added to the binding reaction mixture as described elsewhere (57). Rabbit polyclonal antibodies against human IRF-1 and IRF-2 (Santa Cruz Biochemical) were used according to the manufacturer’s recommendations, with purified rabbit immunoglobulin G (IgG) as the control (a gift from Christine Metz).

FIG. 4.

Characterization of factors that bind to the U5-I element. (A) Oligonucleotides used in the EMSA. The ISG15 ISRE and the ISRE-like motif in U5 are underlined. Boldface characters represent mutated residues. The sequences presented are the positive strand of the BLV LTR and the antisense strand of the ISG15 promoter. (B) Competition EMSAs. Oligonucleotides containing the well-characterized ISRE from the ISG15 gene (ISRE_ISG15wt) or the BLV ISRE-like motif (IRF_BLVwt) were tested in competition experiments. The labelled U5-I oligonucleotide was incubated with Raji cell nuclear extracts and 5 μg of poly(dI)-poly(dC), either in the absence of a specific competitor (control lanes, indicated by minus signs) or in the presence of increasing concentrations (4-, 20-, 100-, and 500-fold molar excesses) of the homologous oligonucleotide (U5-I), IRF_BLVwt, an oligonucleotide containing three substitutions in the ISRE-like motif (IRF_BLVmut), ISRE_ISG15wt, or a mutated oligonucleotide (ISRE_ISG15mut) in which the ISRE was destroyed. (C) Competition EMSAs. The labelled IRF_BLVwt oligonucleotide was incubated with nuclear extracts of Raji cells (lane 1) or Daudi cells (lanes 2 to 17) in the absence of specific competitor (controls, lanes 1 and 2) or in the presence of increasing concentrations (4-, 20-, and 100-fold molar excesses) of the homologous competitor (IRF_BLVwt), IRF_BLVmut, ISRE_ISG15wt, or a nonspecific competitor (Sp1 consensus). Poly(dI)-poly(dC) was 1.5 μg per binding reaction. (D) Supershift assay. The ISRE_ISG15wt (lanes 1 to 14) and IRF_BLVwt (lanes 15 to 28) oligonucleotide probes were incubated with nuclear extracts from untreated (−IFNα) or treated (+IFNα; 500 U/ml for 6 h) Daudi cells. Supershift assays were performed in the absence of antibodies (lanes 1, 8, 15, and 22) or in the presence of anti-IRF-1 antibody (lanes 3, 10, 17, and 24), anti-IRF-2 antibody (lanes 4, 11, 18, and 25), purified rabbit IgG as a control (lanes 2, 9, 16, and 23), anti-Stat-1 antibody (lanes 6, 13, 20, and 27), anti-Stat-2 antibody (lanes 7, 14, 21, and 28) or preimmune serum for the control (lanes 5, 12, 19, and 26). Poly(dI)-poly(dC) was 0.5 μg per binding reaction. A 48-h exposure of the autoradiograms is shown for nuclear extracts of untreated cells (lanes 1 to 7 and 15 to 21), and a 16-h exposure is shown for the extracts of IFN-treated cells (lanes 8 to 14 and 22 to 28). Lanes 22′ to 25′ are a longer exposure (48 h) of the gel portion from lanes 22 to 25. (E) Competition EMSAs. The ISRE_ISG15wt probe was incubated with nuclear extract from IFN-α-treated Daudi cells and 0.5 μg of poly(dI)-poly(dC), either in the absence of a specific competitor or in the presence of a 50-fold molar excess of the ISRE_ISG15wt, ISRE_ISG15mut, IRF_BLVwt, or IRF_BLVmut competitor.

RESULTS

The R-U5 region of the BLV LTR stimulates gene expression.

In order to determine the contribution of the BLV R-U5 region to basal (Tax-independent) LTR-driven gene expression, U3- and LTR-driven luciferase gene (luc) expression was compared in transient transfection experiments. A fragment of the BLV LTR containing the U3 region (fragment from nt −211 to +47) or corresponding to the complete LTR (−211 to +320) was inserted into the pGL2-basic vector, upstream from the firefly luc gene. The two resulting plasmids, pLTR(−211/+47)-luc and pLTR-luc (Fig. 1A), were transfected into Raji cells (a human B lymphoblastoid cell line) together with the pRL-SV40 vector. The latter vector contains the Renilla luc gene under the transcriptional control of the SV40 promoter and is used as an internal control for transfection efficiency. Luciferase (Luc) activities (firefly and Renilla) in cell lysates were assayed 24 h after transfection. Basal luc expression increased 5- to 10-fold when the R-U5 downstream region was present in the BLV promoter (Fig. 1B).

FIG. 1.

Stimulation of gene expression by the R-U5 region of BLV. (A) Schematic representation of the plasmid constructs. The numbering of the LTR sequence refers to the transcription initiation site (first nucleotide of R), considered +1. Solid and open arrows indicate transcription initiation sites in the BLV and HSV TK promoters, respectively. (B) Transient transfection of Raji cells with pLTR(−211/+47)-luc and pLTR-luc. These plasmids were cotransfected with the pRL-SV40 control plasmid, in which the SV40 early promoter is driving Renilla luc gene expression. Luc activities in the cellular extracts were measured 24 h after transfection. For each sample, the firefly Luc activity was corrected relative to the Renilla Luc value. Averages of triplicate samples from one of four experiments are shown. (C) Transient transfection of pTK-cat and pTK-(−22/+320)-cat in different cell lines. These two plasmids were transfected into Raji, Daudi, OVK, and MDBK cells. The Cat activities in the cell lysates were measured 48 h after transfection and are expressed as percentages of deoxychloramphenicol converted to the acetylated product. The average of three independent transfection experiments is shown for each cell line. (D) Transient transfection of pTK-cat, pTK-(−22/+320)-cat, pTK-(+26/+320)-cat, and pTK-(−22/+47)-cat in Raji cells. Cat assays were performed as described for panel C. Stimulation of HSV TK promoter-driven cat expression is defined as the ratio of the Cat activity obtained with an individual plasmid relative to the Cat activity yielded by the pTK-cat vector. The average values from four independent transfection experiments are shown.

The ability of the R-U5 region to stimulate gene expression from a heterologous promoter was determined by inserting the −22-to-+320 fragment of the BLV LTR into the pTK-cat vector, downstream from the HSV TK promoter (see Materials and Methods). The resulting plasmid, pTK-(−22/+320)-cat (Fig. 1A), and pTK-cat were each transiently transfected into human B cells (Raji and Daudi), ovine fibroblasts (OVK), and bovine epithelial cells (MDBK). In all cell lines tested, the R-U5 sequence increased HSV TK promoter-driven cat expression 15- to 20-fold (Fig. 1C).

The BLV region present in the pTK-(−22/+320)-cat plasmid contains a small fragment that corresponds to the last 22 nt of U3, in addition to the R and U5 regions. The 5′ boundary of this fragment is therefore the same as the 250-bp stimulatory element previously described by Derse and Casey (15). Thus, the pTK-(−22/+320)-cat construct contains two potential transcription initiation sites (in the HSV TK promoter and in the BLV LTR). In order to determine whether the BLV transcription initiation site was involved in stimulating gene expression, we constructed a deletion mutant, pTK-(+26/+320)-cat, by removal of the BLV transcription initiation site (Fig. 1A). As shown in Fig. 1D, this plasmid yielded Cat activity similar to that of pTK-(−22/+320)-cat. Furthermore, if a small fragment corresponding to the BLV initiation site (fragment −22 to +47) was cloned into pTK-cat, no stimulation of cat expression was observed (Fig. 1A and D). Taken together, these results demonstrate that stimulation by the R-U5 region of HSV TK promoter-driven gene expression is independent of the BLV transcription initiation site.

Both the R and U5 regions of the BLV LTR contain elements that stimulate gene expression.

A series of 3′ deletion mutants of R-U5 were obtained by PCR amplification. These fragments were cloned into the pTK-cat vector, downstream from the HSV TK promoter. The resulting plasmids were transfected into Raji cells, and Cat activity in cell lysates was measured 48 h after transfection. The stimulatory activity of the R-U5 region was decreased five- to sixfold when the U5 region (nt +230 to +320) was deleted, demonstrating that U5 plays a major role in this stimulation (Fig. 2). Further deletion of the R sequence caused a progressive decrease in cat expression, suggesting that additional regulatory elements are located in the R region.

FIG. 2.

Progressive deletion analysis of the R-U5 stimulatory activity. A set of 3′ deletion mutants of the R-U5 region were obtained by PCR and cloned into pTK-cat, downstream from the HSV TK promoter. Cat activity was assayed after transient transfection of these plasmids in Raji cells. Stimulation of HSV TK promoter-driven cat expression was defined as for Fig. 1D. The stimulation by each deletion mutant is plotted against the boundary of the 3′ deletion in the R-U5 region. The designation of the corresponding plasmid is given below the x axis. Each point is the average from four independent transfection experiments. The R-U5 fragment is diagrammed at the top in scale with the x axis.

The 5′ half of U5 contains a transcriptional enhancer.

Further investigation into the role of the U5 region in LTR-driven gene expression was accomplished by the construction of two LTR deletion mutants. A 45-bp deletion from the 3′ end of the LTR was created to produce the −211-to-+275 fragment, and a 91-bp 3′ deletion removed the entire U5 region, leading to the −211-to-+229 fragment. These fragments were subcloned into the pGL2-basic vector, upstream from the firefly luc gene (Fig. 3A). Luc expression obtained after transient transfection of the two plasmids, pLTR(−211/+275)-luc and pLTR(−211/+229)-luc, into Raji cells was compared with that observed for the pLTR(−211/+47)-luc and pLTR-luc plasmids. The pLTR(−211/+275)-luc construct yielded higher luc expression than pLTR-luc (Fig. 3B). A further 46-bp 3′ deletion (nt +275 to +230) strongly decreased Luc activity. These results demonstrate that U5 contains a positive regulatory element in its 5′ region (nt +230 to +275) that is partially counteracted by a negative regulatory element in the 3′ region (nt +276 to +320). The 5′ half of U5 containing the positive element (+230 to +275) is referred to as U5-I.

FIG. 3.

U5 contains a transcriptional enhancer. (A) Schematic representation of the plasmids pLTR(−211/+275)-luc and pLTR(−211/+229)-luc. (B) Transient transfection of the pLTR-luc, pLTR(−211/+275)-luc, pLTR(−211/+229)-luc, and pLTR(−211/+47)-luc constructs in Raji cells. pRL-SV40 was cotransfected as an internal control for transfection efficiency. Luc activities (firefly and Renilla) were determined as described for Fig. 1B. Normalized Luc values from triplicate samples were averaged, and results from one representative experiment are shown. These results were reproduced three times by using two different plasmid preparations. (C) Transient transfection assay in Raji cells. The pTK-U5-I(s)cat plasmid contains the +230-to-+275 U5-I fragment downstream from the promoter, in the natural orientation. The pU5-I(s)-TK-cat and pU5-I(as)-TK-cat constructs contain the U5-I fragment upstream from the promoter, in the sense and antisense orientations, respectively. Cat assays were performed as described for Fig. 1C. Averages obtained from triplicate samples from one of four independent experiments are shown.

The stimulation exerted by U5-I can occur at either the transcriptional or posttranscriptional level, since it is located downstream from the transcription initiation site. In order to determine whether this element can activate heterologous promoter-driven gene expression when located outside of the transcription unit, the U5-I segment was cloned into the pTK-cat vector either downstream from the promoter in its natural orientation [pTK-U5-I(s)cat] or upstream from the promoter, in both the natural and the inverted orientation [pU5-I(s)-TK-cat and pU5-I(as)-TK-cat, respectively]. The resulting constructs were transfected into Raji cells, and Cat activity in cell lysates was assayed 48 h after transfection. The U5-I DNA segment stimulated cat expression three- to fivefold when cloned downstream from the promoter (Fig. 3C). Similar results were obtained in the Daudi and OVK cell lines (data not shown). Moreover, stimulation of HSV TK promoter-driven cat expression by U5-I was independent of its position and orientation relative to the promoter (Fig. 3C). Thus, the U5-I element has all the characteristics of a classical transcriptional enhancer.

The U5-I DNA element specifically binds IRF-1 and IRF-2 in vitro.

In order to identify cellular factors that bind to the positive regulatory region in U5, EMSAs were performed by using U5-I as the probe and nuclear extracts from the Raji cell line. In the presence of the nonspecific competitor poly(dI)-poly(dC), a single, specific band was shifted (Fig. 4B, control lanes). The U5-I region contains a sequence highly similar to an IFN-stimulated response element (ISRE). The ISRE motif was originally described in the promoters of several IFN-inducible genes (34, 35, 45) and is the recognition site for members of the IRF family (reviewed in references 23 and 32). This includes IRF-1 and IRF-2, two antagonistic transcription factors that are constitutively expressed and can be further induced in response to IFN (21). The ISRE motif also binds the multiprotein complex ISGF3 (IFN-stimulated gene factor 3), composed of p48 and the phosphorylated Stat-1 and Stat-2 proteins (signal transducers and activators of transcription 1 and 2 [reviewed in references 11 and 53]). The phosphorylation of Stat-1 and Stat-2 in response to type 1 IFN rapidly induces the formation of the ISGF3 complex and its translocation to the nucleus (29, 33). The region of the ISRE important for IRF-1 and IRF-2 recognition (the core ISRE) is contained within the broader ISGF3 binding site. The residues flanking the core ISRE do not play a role in IRF binding but are important for ISGF3 recognition (28). In order to determine whether the ISRE-like motif identified in U5 was actually involved in the formation of the DNA-protein complex, gel shift competition experiments were performed by using an excess of several unlabelled double-stranded oligonucleotides (Fig. 4A). A 24-bp oligonucleotide, IRF_BLVwt, composed of the BLV sequence centered on the ISRE-like motif, was used to determine whether this region was responsible for protein binding to the 46-bp U5-I probe. A well-characterized ISRE motif from the ISG15 gene was also used as a competitor (ISRE_ISG15wt) to compare the binding specificity of the U5-I sequence with that of a classical ISRE (5). The formation of the low-mobility complex after incubation of the U5-I probe with a Raji nuclear extract could be competed for by the IRF_BLVwt and the ISRE_ISG15wt oligonucleotides as efficiently as by the homologous U5-I oligonucleotide (Fig. 4B). In contrast, the ISRE_ISG15mut oligonucleotide, containing four point mutations known to abolish protein binding to the ISRE (28, 57), did not have any inhibitory effect on complex formation with the U5-I probe. Similarly, a 3-bp mutation modifying crucial residues in the BLV ISRE-like motif (10, 28) (oligonucleotide IRF_BLVmut) abolished competition for the retarded band (Fig. 4B). Similar results were obtained with nuclear extracts from the OVK ovine cell line (data not shown). When the IRF_BLVwt oligonucleotide was used as a probe, three major retarded bands appeared upon incubation with nuclear extracts from Raji and Daudi cells (Fig. 4C, lanes 1 and 2). The two lower bands, appearing as a doublet, were reproducibly observed, in contrast to the upper band and other retarded bands which were not consistently observed. Competition experiments performed with Daudi nuclear extracts revealed that the doublet is specific to the ISRE-like motif, since both bands were competed for by the IRF_BLVwt and ISRE_ISG15wt oligonucleotides and not competed for by either the IRF_BLVmut competitor or an unrelated Sp1 consensus oligonucleotide (Fig. 4C). In contrast, the slowest-migrating band resulted from nonspecific binding since it was competed by all oligonucleotides tested, including those of unrelated sequence. Similar results were obtained with Raji nuclear extracts (data not shown). No specific retarded complex was formed with the IRF_BLVmut probe, even at low poly(dI)-poly(dC) concentrations (data not shown). Taken together, these results indicate that the same proteins are binding to the U5-I, IRF_BLVwt, and ISRE_ISG15wt oligonucleotides.

Supershift assays using antibodies specific for individual members of the IRF family were performed to identify the proteins present in the retarded complexes. Daudi cells were used for these experiments because they are highly responsive to IFN treatment. Labelled IRF_BLVwt oligonucleotide was incubated with nuclear extracts from untreated or IFN-α-treated Daudi cells. Addition of anti-IRF-1 and anti-IRF-2 antibodies interfered with the formation of the higher-mobility and lower-mobility shifted bands, respectively (Fig. 4D, lanes 17, 18, 24, and 25). The anti-IRF-2 antibody seemed also to decrease the intensity of the lower IRF-1-containing band. A similar observation was previously reported with an antibody directed against the same region of the IRF-2 protein (5). In IFN-treated cells, it was possible to detect the supershifted complexes generated with these antibodies (Fig. 4D, lanes 24′ and 25′), although a control purified rabbit IgG did not affect the binding pattern (lane 23′). IFN-α treatment of cells induces ISGF3; however, we observed that the binding patterns obtained with extracts from IFN-α-treated and those obtained with extracts from untreated cells were identical and could not be affected by the addition of anti-Stat-1 or anti-Stat-2 antibodies (Fig. 4D, lanes 19 to 21 and 26 to 28), showing that the retarded complexes did not involve ISGF3. When the ISRE_ISG15wt oligonucleotide (5) was used as a probe, an additional shifted complex of low mobility was observed upon treatment of the cells with IFN-α. This complex could be supershifted by the addition of either anti-Stat-1 or anti-Stat-2 antibodies to the binding reaction (Fig. 4D, lanes 13 and 14), while preimmune serum did not affect the complex mobility (lane 12). Thus, as expected (5), supershift analysis showed that the complex formed with the ISRE_ISG15wt probe after IFN-α treatment of the cells corresponds to the binding of ISGF3, confirming that this multiprotein factor was present in the extracts of IFN-α-treated cells. In competition experiments with the ISRE_ISG15wt probe and nuclear extracts from IFN-α-treated Daudi cells (Fig. 4E), the bands that correspond to the binding of IRF-1 and IRF-2 could be competed for by an excess of unlabelled ISRE_ISG15wt or IRF_BLVwt oligonucleotides, while the band corresponding to ISGF3 was competed for by unlabelled ISRE_ISG15wt but remained unaffected by the IRF_BLVwt competitor.

Taken together, these results demonstrate that IRF-1 and IRF-2 bind to the ISRE-like motif in the U5 region of the BLV LTR. However, this sequence is not capable of binding ISGF3. Consequently, the motif extending from position +251 to +261 in U5 can be considered an IRF binding site rather than a classical ISRE.

The IRF binding site in U5 is not sufficient to confer IFN-α responsiveness on the BLV LTR.

We tested the effect of IFN-α treatment on BLV LTR-driven luc expression. Sixteen hours after transfection by the pLTR-luc plasmid, Daudi cells were treated with 500 U of IFN-α/ml for different periods of time. The BLV IRF binding site did not confer IFN-α inducibility on the LTR, whereas a control Mx promoter (24) was strongly activated (data not shown). This failure of the BLV ISRE-like motif to confer IFN responsiveness is most probably due to its inability to bind ISGF3.

The IRF binding site in U5 is required for optimal basal gene expression from the BLV LTR.

In order to examine the importance of the IRF binding site in the basal activity of the BLV promoter, site-directed mutagenesis was used to destroy this motif within the BLV LTR. The same 3-bp mutation shown to abolish protein binding to the BLV IRF binding site in gel retardation assays was introduced into the LTR. This mutated LTR was subcloned into the pGL2-basic vector, upstream from the firefly luc gene, resulting in the pLTR(IRF*)-luc plasmid. The constructs pLTR(−211/+47)-luc and pLTR(−211/+229)-luc (in which luc expression is driven by the U3 and the U3-R region, respectively), pLTR-luc, and pLTR(IRF*)-luc were transfected into Raji cells, and Luc activity was measured in cell extracts 24 h after transfection. Mutation of the IRF binding site decreased LTR basal activity two- to threefold (Fig. 5). However, the Luc activity obtained with the mutated LTR was slightly higher than that obtained after deletion of the complete U5 region. Similarly, introduction of this mutation into the U5-I region reduced but did not abolish the capacity of this region to stimulate a heterologous promoter (data not shown), suggesting that other uncharacterized elements may contribute to the U5 stimulatory activity.

FIG. 5.

Mutation of the IRF binding site in U5 decreases Tax-independent LTR-driven gene expression. Plasmids containing the firefly luc gene under the transcriptional control of either the BLV U3 region (−211 to +47), U3-R region (−211 to +229), wild-type LTR, or mutated LTR (where the IRF binding site was destroyed) were cotransfected in Raji cells with the pRL-SV40 control plasmid. Luc activity was measured and corrected as described for Fig. 1B. Averages obtained from triplicate samples in a representative experiment are shown. Similar results have been obtained in five independent transfection experiments using two different plasmid preparations.

DISCUSSION

Because of the low level of BLV LTR promoter activity in the absence of Tax, regulation of LTR basal activity has not been well characterized previously, and most studies have focused on Tax responsiveness. We have succeeded in measuring Tax-independent BLV LTR promoter activity in B-cell transient transfection experiments, using the sensitive luc reporter system. Our work provides evidence that the R-U5 region of BLV is important for Tax-independent LTR-driven gene expression. We have shown that this stimulation by R-U5 is, in part, due to the presence of a transcriptional enhancer in U5, which is capable of binding members of the IRF family of proteins, IRF-1 and IRF-2. Moreover, mutation of the IRF binding site can alter Tax-independent LTR-driven gene expression.

On the other hand, we found a negative regulatory element at the 3′ end of U5, and experiments are in progress to characterize this repressor. The existence of such a repressor has been suggested in a previous report (14). Similarly, a transcriptional repressor has been described in the U5 region of HTLV-1 and coincides with an Sp1- and Sp3-binding site (40, 41).

R-U5 is a stimulatory element which is capable of inducing a 10- to 30-fold increase in gene expression either from the BLV promoter or a heterologous promoter; however, basal LTR transcriptional activity in the absence of Tax is still extremely low. This paradox could be explained by the presence of a potent inhibitor in the U3 region, which could counteract the enhancing effect of R-U5. However, we failed to detect a repressive element in U3 by progressive deletion from the LTR 5′ end or by cloning of U3 subfragments upstream from an active promoter (30a). This discrepancy between the strong enhancer activity of R-U5 and the low level of LTR-driven gene expression could also be explained by the fact that the potential BLV TATA box matches the consensus poorly. It is conceivable that in the absence of Tax, transcription initiation is inefficient because of weak interactions of basal transcription factors with the BLV promoter. Low levels of transcriptional activity could occur after stimulation of the promoter by upstream or downstream elements such as the CRE, NF-κB, or IRF binding sites. This would lead to the synthesis and accumulation of Tax, which in turn might stabilize the transcription initiation complex.

Our results showed that the major contribution to the stimulatory effect of R-U5 comes from the U5 region. However, when U5 is deleted, a residual fivefold stimulation by R is still observed. A 250-bp element corresponding to the R region was previously described as a stimulatory element that is orientation independent but active only when located between the transcription and the translation initiation sites (15). This Tax-independent stimulatory activity was attributed to a 64-bp DAS (nt +147 to +211) by Kiss-Toth and Unk (31). Our results from progressive deletion analysis are not in agreement with those of this study, since removal of the sequence containing this DAS element does not significantly decrease stimulation by R (Fig. 2). This discrepancy might be explained by the differences in experimental conditions. Kiss-Toth and Unk (31) reported measurements of LTR-driven gene expression upon transfection of epithelial HeLa cells in the presence of Tax, while we used a heterologous promoter to assess stimulation by R, in the absence of any viral proteins. Furthermore, our transfection experiments were performed with the Raji cell line, which, like the BLV target cells, is of B lymphoid origin. Thus, the presence of Tax or the availability of some B-cell-specific transcription factor(s) might explain these contradictory results.

Mutation of the IRF binding site in U5 decreased Tax-independent activity but did not abolish stimulation by U5. This observation demonstrates that this sequence is important for promoter activity and suggests that another cis-acting element(s) is present within this region. The potential to chose among a variety of regulatory elements in order to increase the strength of the promoter/enhancer unit could be a mechanism that is designed to broaden the variety of cellular conditions under which the virus can be active.

The weak transcriptional activator, IRF-1, is a possible candidate for mediating activation by U5 (21, 43; reviewed in reference 23). Although IRF-1 synthesis is induced in response to IFN-α treatment, its binding site in U5 does not confer IFN-α responsiveness to the BLV promoter. This observation is consistent with previous reports demonstrating that IRF-1 binding to the ISRE is not sufficient to mediate IFN inducibility (10, 18, 28, 44). Other IRF binding sites unable to bind ISGF3 have been described (23, 28, 57). Although such elements are not sufficient to confer IFN responsiveness, they have been shown to influence the activity of the promoters where they reside. Additionally, IRF-1-deficient mice have two- to threefold-lower constitutive levels of major histocompatibility complex (MHC) class I, the expression of which is controlled by an ISRE. However, upon IFN treatment, the IRF-1-deficient mice show normal activation of MHC class I and other IFN-stimulated genes (46). Therefore, despite being called IFN regulatory factor, IRF-1 is dispensable for IFN induction of at least some of the IFN-stimulated genes. Since mutations that impair IRF-1 binding to the ISRE diminish the transcriptional activity of target genes even in the absence of IFN treatment (28), IRF-1 should be considered a factor confering a basal transcriptional activity.

The IFN-inducible transcriptional repressor IRF-2 (21) is also capable of binding to the U5 element. While its functional role has not been fully characterized, it is thought to be a transcriptional repressor, or at least an inhibitor of the IRF-1 stimulatory function (21, 54). However, IRF-2 has also been associated with IFN-dependent or -independent transcriptional activation of viral and cellular genes (39, 52, 58). Furthermore, it is also a target for inducible processing, which can produce a truncated version of IRF-2, with modified regulatory properties (reviewed in reference 23).

Other IRF family members have also been shown to bind to the IRF binding site or related sequences (3, 17, 39, 65, 66), and we cannot exclude the possibility that one or more can bind to U5 in infected cells.

The IRF binding site in the BLV U5 region is susceptible to recognition by both stimulatory and inhibitory transcription factors, depending on the cellular conditions. We have shown that a stimulatory role is associated with the IRF binding site, although a repressive function during the course of BLV infection cannot yet be ruled out. Members of the IRF family of transcription factors have been shown to positively and negatively regulate other viruses, including Epstein-Barr virus (39, 52, 66) and human immunodeficiency virus (HIV-1). Mutation of the IRF binding site in the HIV-1 leader sequence was shown to decrease LTR-driven transcription, and in concert with at least one additional mutation was deleterious for viral replication (36, 57). Moreover, HIV-1 infection is strongly inhibited in monocytic cells expressing a dominant-negative factor of the IRF family (55). Interestingly, we found that there is a potential IRF recognition sequence in the U5 region of HTLV-1. Therefore, an IRF binding site could be part of a common strategy of HIV, HTLV, and BLV to regulate their genome expression in the early stages of infection or in response to extracellular signals. The detection of Tax-independent BLV LTR-driven gene expression provides us with a new means of gaining insight into this process.