Activation of E2F-mediated Transcription by Human T-cell Leukemia Virus Type I Tax Protein in a p16 INK4A -negative T-cell Line*

The human T-cell leukemia virus type I (HTLV-I) is a causative agent of adult T-cell leukemia. Although the exact mechanism by which HTLV-I contributes to leukemogenesis is still unclear, the Tax protein is thought to play a major role in this process. This 40-kDa polypeptide is able to interact with the tumor suppressor p16 INK4A . Consequently, Tax can activate the signaling pathway that lead to the release of E2F that in turn induces expression of factors required for cell cycle progression. In this paper, we demonstrate that Tax can also activate E2F-mediated transcription independently of p16 INK4A . Indeed, when Tax is coexpressed with the E2F-1 transcription factor in CEM T-cells, which lack expression of p16 INK4A , it strongly potentiates the E2F-dependent activation of a reporter construct driven by a promoter containing E2F binding sites. This stimulation is abrogated by mutations affecting the E2F-binding sites. In addition, Tax also stimulates the transcription of the E2F-1 gene itself. Using Tax mutants that fail to activate either ATF- or NF-κB-dependent promoters and different 5′ truncation mutants of the E2F-1 promoter, we show that the Tax-dependent transcriptional control of the E2F1 gene involves, at least in part, the ATF binding site located in the E2F-1 promoter.

Human T-cell leukemia virus type I (HTLV-I) is the etiologic agent of adult T-cell leukemia (ATL). The viral genome codes for regulatory proteins including the 40-kDa Tax protein, which transactivates its own promoter. Tax transactivation involves three 21-base pair regulatory elements containing imperfect cyclic AMP response element (CRE), localized in the U3 region of the long terminal repeat (1)(2)(3)(4). Tax interacts directly with proteins of the activating transcription factor/CRE-binding protein (ATF/CREB) family (5-7) and increases their activ-ity by enhancing their dimerization (8). Moreover, Tax stimulates ATF/CREB binding to CREB-binding protein (9). Tax is highly pleiotropic, as it has been shown to transcriptionally stimulate a wide variety of cellular genes through its ability to activate other transcription factors including the p67 SRF , NF-B/Rel proteins, Ets1, NF-Y, and Sp1 (10 -14), and to repress gene transcription through factors of the basic helix-loop-helix family (15)(16)(17). It has been suggested that Tax plays a key role in the onset of ATL, but it still remains unclear which of the diverse stimulation activities of Tax is essential for immortalization of HTLV-I-infected T-lymphocytes. In studies with Tax mutants (18), transactivation through NF-B was reported to be not required for rat fibroblast transformation, whereas Tax mutants unable to transactivate the ATF/CREB pathway were described to be not oncogenic (19). However, these results are in contrast to an observation of Yamaoka et al. (20), suggesting that constitutive activation of NF-B is essential for transformation of rat fibroblasts by Tax. Nevertheless, transduction in human primary T-lymphocytes of a Tax mutant that is active for ATF/CREB but inactive for NF-B resulted in permanent growth of the cells suggesting that Tax can induce immortalization of T-lymphocytes through a mechanism independent of NF-B pathway (21).
A detailed study of E2F protein complexes in T-cells demonstrated that the prevailing E2F complexes contain E2F-1 and E2F-4 associated with DP-1 (30). Given the crucial role of E2F target genes in cell proliferation, the stimulation of the E2F activity by HTLV-I could be involved in the proliferation and transformation of T-cells in ATL. Several lines of evidence support this model. Indeed, it has been shown that HTLV-Iinfected T-cell lines and leukemic cells obtained from ATL patients contained high levels of the DNA binding activity of E2F and that Tax enhanced this E2F activity in Jurkat T-cells (31). As proposed recently (32), it is believed that Tax mediates this effect via its direct interaction with the cdk inhibitor p16 INK4A (33,34). This interaction would lead to the following successive events (35): (i) an increase in cyclin D/cdk4 kinase activity in infected cells, (ii) an enhanced level of phosphorylation of the retinoblastoma protein family (pRB, p107, and p130), and (iii) an increase in E2F transcriptional activity. However, we demonstrate that Tax also stimulates the E2F activity via other mechanisms. Indeed, using transient transfection analysis, we demonstrate here that Tax remains able to stimulate gene expression through E2F-1 and E2F-4 in the p16 INK4A -negative CEM T-cell line. Furthermore, we show that the stimulation by Tax of a luciferase reporter gene driven by E2F motifs is abrogated by mutations that affect the E2F binding sites. Finally, we demonstrate that Tax regulates E2F-1 gene expression at the level of transcription and that this transcriptional control depends, at least in part, on the ATF binding site of the E2F-1 gene promoter. Together, these results indicate that Tax can likely disregulate the machinery controlling cell cycle progression not only by activating E2Fdependent transcription but also by stimulating the transcription of the E2F-1 gene itself. Since overexpression of E2F-1 in mammalian cells can cause oncogenesis (36 -38), the transactivation of the E2F-1 promoter by Tax could be an important step in the mechanism of HTLV-I transformation.

EXPERIMENTAL PROCEDURES
Cells-The lymphoblastoid CEM cell line was obtained from the American Type Culture Collection (Bethesda, MD). The HTLV-I-infected MT4 and C8166 cell lines have been described elsewhere (39,40). The origins of the HSB-2 cell line have been described previously (41). JPX-9 is a clone of Jurkat cells stably transfected by a Tax expression vector in which Tax expression is dependent on heavy metal ions (42). JPX/M cells correspond to cells transfected by a mutant plasmid in which a frameshift mutation is introduced in the coding region of the Tax gene (42); in CdCl 2 -or ZnCl 2 -treated JPX/M cells, the mutant Tax gene is induced at the mRNA level but no functional Tax protein is produced (42). Cells were cultured in RPMI 1640 medium supplemented with 1% penicillin/streptomycin antibiotic mixture, 1% Glutamax (Life Technologies, Eragny, France) and 10% fetal calf serum (Life Technologies), to a density of 5 ϫ 10 5 cells/ml in a 5% CO 2 atmosphere.
Transfections and Luciferase Assays-CEM cells were transiently cotransfected according to the previously published procedure (49). 5 g of a ␤-galactosidase-containing plasmid (pAC␤1) was included in each transfection for controlling of the transfection efficiency. The total amount of DNA in each series of transfection was equal, the balance being made up with empty pSG-5 vector without Tax. Cell extracts equalized for protein content were used for luciferase and ␤-galactosidase assays.
Western Blot Assay-Nuclear extracts were prepared as described previously (49). 20 g of protein from nuclear extracts were electrophoresed onto 10% sodium dodecyl sulfate-polyacrylamide gel and blotted to polyvinylidene difluoride membranes (Millipore). The blot was then incubated for 1 h at room temperature with a blocking solution (phosphate-buffered saline (PBS) containing 10% milk and 0.05% Tween 20) prior to addition of antiserum. After 1 h at 20°C, the blot was washed three times with PBS ϩ 0.05% Tween 20 and incubated for 30 min with goat anti-mouse immunoglobulin-peroxydase conjugate (Immunotech, Marseille, France). After three washes, the membrane was incubated with enhanced chemiluminescence (ECL) reagent (Amersham Pharmacia Biotech). The membrane was then exposed for 0.5 to 5 min to Hyperfilms-ECL (Amersham Pharmacia Biotech). Anti-E2F-1 mAb and anti-actin mAb C4 were purchased, respectively, from Santa Cruz Biotechnology Inc. (Santa Cruz, CA), and ICN Biomedicals Inc. (Costa Mesa, CA); anti-Tax was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH. HTLV-I Tax hybridoma 168A51-42 (Tab176) was from Dr. B. Langton (50).  (33). In agreement with these observations, no p16 INK4A could be detected in the CEM cell line used in our study (data not shown).

Tax Stimulates E2F1-dependent Transcription in the p16 INK4A -negative CEM T-cell Line-To
At first, we tested the effects of Tax on the promoters of the DHFR and cyclin E genes, which are known to be controlled by the E2F factors (29,45). Transient cotransfection assays were carried out using luciferase expression plasmids. The transfection assays were performed in CEM cells in the presence or absence of the Tax expression vector pSG-Tax. Tax synthesis stimulated expression of luciferase gene driven by cyclin E (pCycE-luc) and DHFR (pDHFR-luc) promoters with a 4-and 3.5-fold increase in luciferase activity, respectively (Fig. 1). When the transfection assays were carried out with transcriptionally defective Tax mutants, Tax M9 ( Fig. 1) and Tax M21 (data not shown), such stimulations were not detected. These results indicated that Tax was able to activate the promoters of DHFR and cyclin E genes in a p16 INK4A -negative T-cell line.
To study possible involvement of E2F in Tax transactivation of these cellular promoters, we first tested whether Tax might cooperate with E2F in transactivation. E2F1 and Tax were coexpressed in presence of a reporter construct driven by a minimal promoter with upstream E2F-binding sites (p3xE2F-WT-luc). This construct contains three E2F binding sites cloned immediately upstream of a TATA box controlling the transcription of the luciferase gene. Fig. 2 shows that this reporter was stimulated 8.5-fold in the presence of the expression vector pCMV-E2F-1 alone and 9-fold with pSG-Tax alone. No stimulation was detected with the mutants Tax M9 (Fig. 2) and Tax M21 (data not shown). This Tax stimulation was likely mediated by endogenous E2F factor. Cotransfection of pSG-Tax and pCMV-E2F-1 induced a 44-fold increase in luciferase activity, indicating that Tax was able to stimulate E2F-1 activity in the CEM cell line. Cotransfections with the transcriptionally defective Tax mutants produced no stimulation of E2F-1 activity.
To assess whether transactivation by Tax may be ascribed to a direct effect of the binding of E2F-1 to the tested minimal promoter, Tax stimulation of the luciferase gene driven by three mutated E2F binding sites were analyzed with the plas-mid p3xE2F-MUT-luc. The design of this mutant, which converts the E2F binding site 5Ј-TTTCGCGC-3Ј to 5Ј-TTgCtCGa-3Ј, was based on mutations that are known to abolish binding and activity of E2F (44). As shown in Fig. 3, the mutant promoter was very poorly induced by Tax and E2F-1 compared with the wild type construct. The efficiency of Tax-mediated induction of luciferase activity was reduced by 5.5-fold with Tax alone and 39-fold with Tax and E2F-1 compared with control promoter. The mutant promoter still showed a slight 1.5-fold enhancement by Tax, which corroborates previous published results (see, e.g., Refs. 51-53), and represents low level of Tax induction on a minimal TATA box. Taken together, our results demonstrate that Tax activates E2F-1-dependent transcription.
Tax Also Stimulates Transcription through E2F4 -The E2F family is composed of two subclasses of evolutionarily distinct factors: E2Fs-1, 2 and 3, which interact with pRB; and E2F-4 and E2F-5, which interact with p107 and p130, two other members of the pRB family. Amino acid sequence comparison revealed that E2F-4 and E2F-5 were more closely related to each other than to E2Fs 1, 2 ,and 3. We questioned whether transactivation by Tax of a promoter containing E2F binding sites could be activated in the presence of E2F-4. For this experiment, we used the same approach as described for E2F-1, by testing the Tax transactivation on wild type and mutated E2F sites in cotransfection experiments with or without E2F-4. Although expression of E2F-4 in cells transfected with the vector pCSMYC-E2F-4 induced a 6-fold stimulation of the reporter gene, its coexpression with pSG-Tax gave rise to a 21fold increase in luciferase activity (data not shown). Likewise, as showed for E2F-1, mutation of the E2F site reduced the Tax activity on E2F-4, confirming that stimulation by Tax was due to direct effect of the binding of E2F-4 to the E2F site (data not shown).
Thus, not only does Tax activate E2F-1-dependent transcription, it is also able to stimulate the E2F-4 transcriptional activity.
Tax Stimulates E2F-1 Gene Expression-As it has been proposed that the E2F1 gene might be regulated by its own product, we tested whether Tax would also stimulate the E2F1 promoter, thereby enhancing its effect on cellular E2F activity. FIG. 1. Tax stimulation of the luciferase reporter gene driven by the DHFR or cyclin E promoters. CEM cells (5 ϫ 10 6 ) were transfected with 2 g of luciferase gene driven by cyclin E promoter (pCycE-luc) or by DHFR promoter (pDHFR-luc) ϩ 5 g of the Tax expression vectors (producing either the wild type Tax or the transcriptionally defective mutant Tax M9) or empty pSG-5 vector. 5 g of a ␤-galactosidase containing plasmid (pAC␤1) was included in each transfection for controlling of the transfection efficiency. After 48 h, cell extracts prepared and equalized for protein content as described previously (49) were used for luciferase and ␤-galactosidase assays. Luciferase activities were assayed using the Promega's luciferase assay system and normalized for ␤-galactosidase activities. The activities of luciferase gene driven by the cyclin E and DHFR promoters in the absence of Tax were arbitrarily given a value of 1 and the activities of the other transfections were adjusted relative to these activities. Values represent the mean Ϯ S.D. (n ϭ 3).

FIG. 2. Tax stimulation of gene expression mediated by E2F-1.
Assays were performed in CEM cells using for each transient cotransfection 10 g of p3xE2F-WT-luc bearing three copies of E2F binding site (mentioned E2F-Box in the figure) and 5 g of pAC␤1. The effects of Tax and E2F-1 were analyzed by using either 0.1 g of pCMV-E2F-1 ϩ 0.5 g of pDP1 (E2F-Box ϩ E2F-1), or 5 g Tax expression vectors (E2F-Box ϩ Tax or E2F-Box ϩ Tax M9), or 0.1 g of pCMV-E2F-1 ϩ 0.5 g of pDP1 ϩ 5 g of Tax expression vectors (E2F-Box ϩ E2F-1 ϩ Tax or E2F-Box ϩ E2F-1 ϩ Tax M9), the balance of total amount of transfected DNA being made up with pSG-5. Luciferase values were normalized for ␤-galactosidase activity. The activity of luciferase gene driven by E2F-Box in the absence of E2F-1 or/and Tax was arbitrarily given a value of 1, and the activities of the other transfections were adjusted relative to this activity. Values represent the mean Ϯ S.D. (n ϭ 3). To address this question, we first checked the level of endogenous E2F-1 protein in two human T-cell lines infected with HTLV-I, MT4, and C8166 cells. As shown in Fig. 4, immunoblotting analysis using an anti-E2F-1 mAb indicated that expression of E2F-1 was increased in nuclear extracts of Taxpositive cells compared with the level seen in uninfected Jurkat and HSB-2 cells. This result suggests that Tax expression could stimulate E2F-1 gene expression.
In order to test whether this effect was due to a direct transcriptional effect of Tax on E2F-1 gene expression, we then compared E2F-1 mRNA levels in presence or absence of Tax protein. This experiment was carried out with the JPX-9 clone of Jurkat cells containing the Tax gene under a promoter whose expression is stimulated by heavy metal ions (42). As shown in Fig. 5A, the level of E2F-1 mRNA was clearly increased in JPX-9 cells treated with either ZnCl 2 or CdCl 2 , whereas the expression of control GAPDH mRNA was unchanged. JPX/M cells that expressed nonfunctional Tax were treated in the same way but did not show any increase in the level of E2F-1 mRNA (Fig. 5B). These results indicate that stimulation of E2F-1 mRNA synthesis correlates with the presence of Tax.
We finally tested whether Tax can stimulate E2F-1 promoter activity. Transfection assays were performed with the plasmid pGL2-AN that contains the E2F-1 promoter cloned upstream of the luciferase gene. Cotransfection of pSG-Tax and pGL2-AN induced a 5.5-fold stimulation in luciferase activity (Fig. 6), indicating that Tax is effectively able to activate the E2F-1 promoter.
Tax Transactivation of the E2F-1 Promoter Is Dependent, at Least in Part, on the ATF Binding Site-To further explore the mechanisms underlying this transcriptional activation the same cotransfection assay was carried out in presence of Tax mutants that fail to activate ATF (M47)-or NF-B (M22)-dependent promoters (18). Tax M22 activated the promoter to levels similar to wild type Tax whereas Tax M47 was able to stimulate only about 2-fold (Fig. 6). From these results, we conclude that Tax transactivation of the E2F-1 promoter is independent on NF-B but involves the ATF pathway.
To further map the regulatory elements of the E2F1 promoter required for activation by Tax, a series of 5Ј truncation mutants of pGL2-AN (48) were cotransfected with pSG-Tax (Fig. 7). The deletion that encompasses the 5Ј region of the promoter, from position Ϫ211 to Ϫ131, increased the stimulation by Tax (from 6-fold for the wild type to 10-fold for the deleted promoter), suggesting that Sp1 binding sites could have an inhibitory effect on the E2F-1 promoter activation by Tax. Although the deletion of the CCAAT box had no effect, the deletion of ATF binding motif significantly reduced the activation by Tax (from 10-fold to 5.5-fold). Finally, the deleted pGL2-AN vector containing the E2F binding sites was still stimulated by Tax, confirming the results already described in this paper with the plasmid p3xE2F-WT-luc that contained three copies of the E2F binding site.
Taken together, these results confirm the involvement of an ATF-dependent pathway, at least in part, in the transcriptional control of E2F-1 gene by Tax. ATF Pathway Is Necessary but Not Sufficient to Explain the Transactivation by Tax-In this paper, we show that Tax is able to activate E2F-1-dependent transcription in CEM cells (Figs. 2 and 3) and to stimulate E2F-1 gene transcription through the ATF-responsive element localized in the promoter (Figs. 6 and 7). Stimulation of the E2F-1 gene by Tax could explain how Tax transactivates E2F-1-dependent transcription. To determine effects of this stimulation in our tests, transfection assays already described in Fig. 2 were performed again, but in these assays cells were cotransfected with the plasmid p3xE2F-WT-luc containing E2F-binding sites, the E2F-1 expression vector pCMV-E2F-1, together with the mutant Tax M47. As shown in Fig. 8, the wild type Tax and the mutant Tax M22 induced about a 40-fold increase in luciferase activity, whereas Tax M47 stimulated luciferase activity by 14-fold. Since Tax M47 fails to activate ATF-dependent promoters, this stimulation cannot be explained by the effect of Tax on the stimulation of the E2F-1 gene through the ATF pathway.
This result confirms the importance of the ATF pathway but also indicates that other activation mechanisms could contribute to E2F-1 stimulation by Tax. DISCUSSION The regulation of E2F is a key target for oncoviruses. Binding of pRB and the other pocket proteins to ligands such as, adenovirus E1A, simian virus large T antigen, and papillomavirus E7, leads to a stimulation of E2F-dependent transcription and cellular transformation (54 -56). The human cytomegalovirus IE72 protein is able to phosphorylate E2F-1-3 and the pocket proteins p107 and p130, and this phosphorylation step would play an essential role in the mechanism of cell proliferation (57). Induction of E2F DNA binding activity in HTLV-Iinfected T-cell lines and leukemic cells obtained from ATL patients (31) suggests that the activation of E2F-dependent transcription by HTLV-I could also be involved in the proliferative response during HTLV-I infection. Suzuki et al. (33) demonstrated that Tax interacted with p16 INK4A and suggested that the inactivation of p16 INK4A by Tax would contribute to cellular immortalization and transformation induced by HTLV-I infection. The observations that Tax released T-lymphocytes from cell cycle arrest induced by p16 INK4A but also that p16 INK4A overexpression blocked Tax-dependent stimulation of DNA synthesis (34) are effectively consistent with a deregulation of cell cycle progression by Tax in a p16 INK4A -dependent manner. However, these results did not exclude the possibility that Tax could also activate E2F-dependent transcription in a p16 INK4A -independent manner.
In this report, we demonstrate that Tax stimulates the activity of E2F-1 and E2F-4 in CEM T-cells, which lack expression of p16 INK4A . This stimulation is abrogated by mutations affecting the E2F binding sites, thereby confirming that stimulation by Tax is due to direct effect of the binding of E2F-1 and E2F-4 to E2F boxes. In addition, we demonstrate that (i) Tax stimulates E2F-1 mRNA synthesis in Jurkat cells, (ii) Tax transactivates the E2F-1 promoter, and (iii) this transactivation is dependent, at least in part, on the ATF pathway. Tax is well known to interact directly with proteins of the ATF/CREB family and to stimulate their activity (5)(6)(7). Tax interaction with the ATF pathway is important in the development of the neoplastic phenotype in the case of adult rat fibroblasts (19), and mutants of Tax, which failed to activate the endogenous  (n ϭ 3). The E2F-1 transcription start site is indicated by an arrow, with the locations of two CCAAT boxes and potential binding sites for MBF-1, Sp-1, ATF, E4F, and E2F. ATF pathway, had a reduced ability to induce tumors in transgenic mice (58). Stimulation of E2F-1 mRNA at the transcriptional level may be a critical step in cell cycle control. Overproduction of E2F-1 can induce cell proliferation, presumably by titrating pRB from the relevant promoter and thus leading resting cells to inappropriate entry into S phase (59,60), but can also induce p53-dependent apoptosis (61,62). However, it has been shown that p53 is fully inactivated by Tax in the HTLV-I-transformed T-cells (63,64), confirming that stimulation of E2F-1 gene by Tax could effectively be involved in T-cell transformation.
The activation by ATF pathway of the E2F-1 promoter is not sufficient to completely explain the effect of Tax on E2F-mediated transcription in p16 INK4A -negative T-cells. Indeed, the mutant Tax M47 that fails to activate ATF pathway remains able to transactivate exogenous E2F-1, suggesting that a mechanism, not yet characterized, could be also involved in a p16 INK4A -independent activation of E2F-mediated transcription by Tax. Growth factors such as IL-2 are necessary to carry T-cells from the G1 to the S phase of the cell cycle. The ability of Tax to stimulate the transcription of IL-2 and to trigger the constitutive expression of the ␣ subunit of its receptor, IL-2R␣, has led to the hypothesis that constitutive signaling by the IL-2 receptor may be a key factor in developing ATL (65). However, a progression from IL-2 dependence to independence has been observed in HTLV-I immortalized cells (66 -68), and expression of the IL-2R␣ was recently reported as being unnecessary to the in vitro growth of HTLV-I transformed cell lines (69). Signaling cascades initiated by IL-2 include protein kinase-dependent pathways involving the Lck and Fyn kinases (70), two Janus family kinases, Jaks 1 and 3 (71,72), and the phosphatidylinositol 3-kinase (PI3K) (73). It is well established that neither the Lck nor the Fyn kinases are expressed in HTLV-Iinfected human T-cells (17,74), whereas the Jak/STAT pathway is constitutively activated in HTLV-I immortalized cells (75,76) suggesting that this pathway could participate in HTLV-I-mediated T-cell transformation. Although the mecha-nism by which HTLV-I transformation induces activation of the Jak pathway is not well understood, Tax is probably not involved in this activation (75). On the other hand, a recent report suggests that PI3K is both necessary and sufficient to couple the IL-2R to the transcriptional activation of E2F in T-cells (73), E2F transcriptional activity being regulated by the action of PI3K on the activation of protein kinase B. Inhibition of PI3K inhibits phosphorylation of pRB, induction of cyclin D3, and degradation of p27 kip1 (73). Altogether, these results establish a crucial PI3K/protein kinase B-mediated link between the IL-2 receptor and the cell cycle machinery. Thus, the direct activation of this pathway by Tax could contribute to explain the p16 INK4A -independent activation of E2F-mediated transcription by Tax and the switch to IL-2 independence in HTLV-I immortalized cells. Yet other mechanisms could explain the effects of Tax including, for example, the direct interaction of Tax with E2F-1, as described for several cellular transcriptional factors. Experiments are under way to further elucidate the role of Tax in E2F-mediated transcription independently on p16 INK4A . The identification of this novel mechanism will provide a useful tool to understand how HTLV-I could induce T-cell transformation.
Addendum-During the submission of this paper, Neuveut et al. (77) published results indicating that Tax can increase cyclin D-cdk activity and induce a phosphorylation of cyclin D3 in T-cells null for p16 INK4A . They suggested that this Tax-associated phosphorylation of cyclin D3 might stabilize the cyclin-cdk complexes, and thus enhance the pRB phosphorylation and increase the E2F transcriptional activity.