The Hepatitis B Virus X Protein Induces HIV-1 Replication and Transcription in Synergy with T-cell Activation Signals

Co-infection with hepatitis B virus (HBV) and human immunodeficiency virus type-1 (HIV-1) is relatively common. However, the impact of this co-infection on the clinical outcome of HIV infection has not been elucidated. We herein demonstrate that the HBV X protein (HBx) superinduces ongoing HIV-1 replication and HIV-1 long terminal repeat (LTR) transcription by synergizing with Tat protein and with T-cell activation signals. Although HBx cooperated with mitogenic stimuli in the induction of reporter plasmids harboring the HIV-1 κB enhancer, in both a NF-κB-dependent manner and a NF-AT-dependent manner, deletion of this element from the LTR did not affect the HBx-mediated up-regulation in the presence of Tat and/or mitogens. In contrast, mutation of the proximal LTR Sp1-binding sites abolished the HBx-mediated synergistic activation, but only when it was accompanied by deletion of the κB enhancer. When HBx was targeted to the nucleus, its ability to synergize with cellular activation stimuli was maintained. Furthermore, mutations of HBx affecting its interaction with the basal transcription machinery abrogated the synergistic activation by HBx, suggesting that this protein exerts its function by acting as a nuclear co-activator. These results indicate that HBx could contribute to a faster progression to AIDS in HBV-HIV co-infected individuals.

pathogenic lentivirus that is associated with the development of AIDS. During the stage of clinical latency, the viral burden in the bloodstream is markedly decreased, and viral replication persists only in the lymphoid organs, whereas in most circulating CD4ϩ T-lymphocytes HIV-1 replication/transcription is not ongoing (1). It is generally accepted that in the cellular environment of resting T-lymphocytes HIV-1 remains in a quiescent state and that HIV-1 reactivation is dependent upon T-cell activation. HIV-1 gene expression and replication are controlled by an interplay of viral and host regulatory proteins that interact with cis-acting sequences located in the HIV-1 long terminal repeat (LTR). Among the multiple regulatory elements of the HIV-1 LTR, the B enhancer, which contains two copies of B elements at nucleotides Ϫ104 to Ϫ81, is considered the main inducible cis-acting element (2). Deletion or site-directed mutations of this regulatory sequence may affect LTR transcriptional activation induced by T-cell activation stimuli and by Tat protein (2)(3)(4). It has been widely demonstrated that the B enhancer element binds and responds to NF-B/Rel family of transcription factors, which are induced by a number of stimuli such as mitogens, cytokines, and specific T-cell activators (5). More recently, it has been reported that members of the NF-AT family of transcription factors also bind the HIV B enhancer (6 -8). The core promoter region of the HIV-1 LTR contains three tandem Sp1-binding sites located upstream of the TATA box. Mutations of these Sp1 sites affect both basal and Tat-induced LTR transcriptional activity (9). In addition, modulation of Sp1 phosphorylation by agents such as phosphatase inhibitors and by Tat-mediated recruitment of Sp1 to DNA-dependent protein kinase complex results in upregulated expression of the HIV-1 LTR (10,11). Other authors have demonstrated that interaction between NF-B and Sp1 mediates HIV-1 LTR activation (12,13). Downstream of the HIV-1 LTR transcription start site is located the TAR element (nucleotides ϩ1 to ϩ59), which forms an RNA stem-loop structure that is recognized by the HIV Tat protein. Two mechanisms have been proposed for Tat-mediated transactivation (2, 14 -16). First, Tat might favor elongation of nascent viral transcripts through its interaction with the Tat-associated kinase complex p-TEFb, leading to phosphorylation and increased processivity of RNA polymerase II. On the other hand, Tat might also enhance transcription initiation rate by its ability to interact with promoter-bound factors such as TATA binding protein, TFIIB, transcription factor-IIH, RNA polymerase II, and Sp1.
The prevalence of hepatitis B virus (HBV) infection in patients infected with HIV-1 is very common (17,18). In addition, although HBV has a marked hepatic tropism, it has been shown that this virus is also able to infect T-lymphocytes (19,20), suggesting that HIV-1 and HBV may encounter each other at the cellular level in co-infected patients. However, the impact of this co-infection on the clinical outcome of HIV-1infected patients has not been clearly established so far, probably because many other factors may also influence HIV-1 outcome (21)(22)(23)(24)(25).
The HBV genome encodes a 17-kDa protein, termed HBx, that has been shown to function as a transcriptional transactivator of a variety of viral and cellular promoter/enhancer elements. HBx does not bind directly to DNA, but it is able to transactivate transcription through multiple cis-acting elements. The exact mechanism of HBx-mediated transactivation still remains unresolved. It has been shown that HBx interacts in the nucleus with components of the basal transcription machinery and with transcription factors, mimicking the cellular co-activator functions (26 -34). Another proposed mechanism for HBx function involves the activation of cytoplasmic signal transduction pathways, leading to functional activation of a variety of transcription factors (35)(36)(37)(38)(39)(40)(41).
In this report, we demonstrate that HBx synergizes with cellular activation to induce HIV-1 replication and with Tat and cellular activation to transactivate the HIV-1 LTR in Jurkat cells. We show that isolated HIV-1 B enhancer is a target for HBx, in both a NF-B-dependent manner and a NF-ATdependent manner, but its deletion from the HIV-1 LTR does not affect the synergistic activation by HBx. In addition, we identify the two most proximal Sp1-binding sites of HIV-1 LTR as new HBx-responsive elements. Finally, we demonstrate that HBx induces the HIV-1 LTR transcriptional activity by acting as a nuclear co-activatior.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-The human lymphoblastoid T-cell line Jurkat was grown at 37°C with a 5% CO 2 atmosphere in RPMI 1640, supplemented with 10% fetal calf serum, 2 mM glutamine, and 50 g/ml gentamycin. PMA and the calcium ionophore A23187 were obtained from Sigma.
Plasmid Constructs-The expression vectors pSV-HBx and pSVhygro, harboring the HBx open reading frame from HBV and the bacterial hygromycin phosphotransferase gene, respectively, under the control of the simian virus 40 promoter, have been described previously (40). The expression vectors pNLSHBx-FLAG and pSLNHBx-FLAG, harboring a FLAG-tagged HBx open reading frame either with a nuclear localization signal (NLS) or with a nonfunctional mutated nuclear localization signal (SLN), respectively, were kindly provided by Dr. R. J. Schneider (Kaplan Cancer Center, New York, NY) and have been described elsewhere (37). Substitution HBx mutants m93 and m138 were generated by site-directed mutagenesis (Stratagene, La Jolla, CA) using pNLSHBx-FLAG as template. In HBxm93 amino acid residues 93-96 (Leu-His-Lys-Arg) were substituted by Gly-Ala-Gly-Ala, which has been described to interfere with the binding of HBx to the RNA polymerase subunit RPB5 (30). In HBxm138 amino acid residues 138 -140 (Arg-His-Lys), located within the TFIIB-interaction region (29 -31), were changed to Ala-Ala-Ala. The vectors pSG5UTPL-HBx wt, 5D1, 5D2, 5D3, 5D4, 3D5, 3D39, and 3D99, expressing either full-length or truncated HBx proteins, were kindly provided by Dr. S. Murakami (Cancer Research Institute, Kanazawa, Japan) and have been described elsewhere (42). The expression plasmid pGFP-VIVIT, harboring the highly specific NF-AT inhibitor VIVIT peptide fused to the green fluorescence protein, was kindly provided by Dr. Anjana Rao (Harvard Medical School, Boston, MA) (43). The plasmid pCMV-Tat contains the two exons of HIV Tat gene (subtype LAI) and the simian virus 40 poly(A) cloned between XbaI and HindIII sites of pcDNA3 (Invitrogen Corp., Carlsbad, CA). The vector pNL-Luc was generated by cloning the luciferase gene in HIV-1 proviral clone NL4 -3 (44). Luciferase gene was excised from the plasmid pNL4 -3.Luc.R-E (National Institutes of Health AIDS Research and Reference Reagent Program, catalog number 3418) with XhoI and BamHI and inserted in HIV-1 Nef gene. This proviral construct is fully competent for replication and express luciferase activity as a marker of viral gene expression. The vectors pXP1LTR wt and pXP1LTR ⌬B, which has been deleted in the B enhancer element, were obtained by subcloning the fragment Ϫ644 to ϩ77 of the HIV-1 LTR from LAI strain in the pXP1-Luciferase vector. pXP1LTR ⌬TAR was obtained from pXP1LTR wt; briefly, the fragment Ϫ644 (BamHI) to ϩ38 (SacI) was cloned in BamHI/SacI sites of pXP1. Then the vector was digested with SacI, blunt-ended with T4 DNApolimerase, and ligated to generate pXP1LTR ⌬TAR, which contains a disrupted TAR element and a mutated downstream NFB element (45). The same approach was followed to generate pXP1LTR ⌬B ⌬TAR from pXP1LTR ⌬B. The pXP1B-Sp and pXP1Sp set of vectors were made by subcloning in pXP1 the Ϫ107 to ϩ 77 and Ϫ81 to ϩ 77 HIV-1 LTR fragments, respectively, obtained from the described previously NF-Sp CAT set of vectors (9). The pB-TKLuc and pmB-TKLuc vectors contain three copies of the B enhancer sequence wild type (ACAAGG-GACTTTCCGCTGGGGACTTTCCAGGGA) or four copies of the B enhancer sequence mutated in both NF-B-binding sites (ACAACTCA-CTTTCCGCTGCTCACTTTTCCAGGGA), respectively, upstream of the TK promoter and have been described previously as ENH-TK and mENH-TK luciferase vectors (46). This last construct still contains the described NF-AT-responsive element of the HIV-1 LTR (6).
In Vitro HIV-1 Replication Assays-Jurkat cells were transfected by electroporation. Briefly, 5 ϫ 10 6 cells were washed and resuspended in 0.35 ml of RPMI 10% serum with 7.5 g of plasmid DNA. After incubation on ice, the cells were subjected to an electrical pulse (0.32 kV, 1500 microfarads, R ϱ) using an Equibio (UK) apparatus. Then the cells were diluted in fresh medium and were either left untreated or stimulated with PMA (10 ng/ml) plus calcium ionophore (0.5 M). At day 1 after transfection, half of the cells were harvested, and the cell extracts were assayed for luciferase activity using a Lumat LB9501 luminometer. The remaining cells were maintained in culture, and the supernatants were collected at days 1, 3, and 7 after transfection for CA-p24 detection using a p24 enzyme-linked immunosorbent assay kit from Innogenetics (Ghent, Belgium).
Cell Transfection and Luciferase Assays-5 ϫ 10 6 Jurkat cells were transfected with 4.5-6 g of the indicated plasmid DNAs using Lipofectin (Life Technologies, Inc.) according to the manufacturer's recommendations. 36 h post-transfection cells were left untreated or stimulated for 16 h with PMA (10 ng/ml) plus calcium ionophore (0.5 M). The cells were harvested, and the cell extracts were assayed for luciferase activity using a Lumat LB9501 luminometer.

HBx Synergizes with Cellular Activation to Induce HIV-1
Replication-To analyze whether HBx may influence HIV-1 replication, Jurkat cells were transiently co-transfected with the HIV-1 infectious recombinant clone pNL-Luc, in which the nef gene has been replaced by luciferase gene, along with the HBx expression vector pSV-HBx or the control plasmid pSVhygro. The transfected cells were either left untreated or stimulated with PMA plus Ca 2ϩ ionophore (PMA ϩ Io), which mimic the T-cell activation signals (49), and the HIV-1 replication rate was analyzed by measuring CA-p24 antigen production in the culture supernatants at different time points (1, 3 and 7 days). As shown in Fig. 1A, both HBx expression and PMA ϩ Io treatment were able to significantly enhance HIV-1 replication. Interestingly, when these two stimuli, HBx and PMA ϩ Io, were combined, a strong synergistic effect was observed, reaching CA-p24 levels 4 -5-fold higher than those obtained with PMA ϩ Io alone at day 3 after transfection. Similar results were obtained with a noninfectious HIV-1 recombinant clone lacking the env gene that allows only a single cycle of viral replication (data not shown), thus ruling out that the up-regulation of CA-p24 production by HBx and/or PMA ϩ Io was due to increased re-infection rather than enhanced HIV-1 production.
Because enzyme-linked immunosorbent assay detection of CA-p24 antigen was not sensitive enough to detect differences of CA-p24 production at day 1 post-transfection, the effects of HBx and/or PMA ϩ Io on HIV-1 replication at this time were also analyzed using the highly sensitive luciferase activity assay. In agreement with CA-p24 production data, HBx expression cooperated with PMA ϩ Io treatment in the induction of HIV-1 replication at early times after transfection and cell activation (Fig. 1B).
HBx Further Increases Tat-and T-cell Activation-induced HIV-1 LTR Transcriptional Activity-To analyze whether the above effects of HBx on HIV-1 replication were regulated at the level of HIV-1 LTR transcriptional activity, Jurkat cells were transiently co-transfected with the plasmid pXP1LTR wt, containing the complete HIV-1 LTR (nucleotides Ϫ644 to ϩ77) upstream of the luciferase reporter gene (Fig. 2), along with the HBx expression vector pSV-HBx or the control plasmid pSVhygro. In addition, to evaluate the possible functional interplay between HBx and the HIV Tat protein, the construct pCMV-Tat or its control vector pcDNA3 was also included in the co-transfection experiments. As expected, the expression of Tat protein strongly enhanced the transcriptional activity of the HIV-1 LTR (Fig. 3A). In contrast, the expression of HBx induced only about 4 -6-fold the HIV-1 LTR activity. However, a multiplicative induction was observed when HBx and Tat proteins were co-expressed (Fig. 3A).
To study the influence of HBx and/or Tat on HIV-1 LTR function in an activated cellular background, the transfected cells were treated with PMA ϩ Io. Inductions of ϳ35and 97-fold over the effect of PMA ϩ Io were observed by HBx and Tat expression, respectively, which were greater than multiplicative (Fig. 3B). Thus, the expression of HBx or Tat proteins exerted a marked synergism with the mitogenic stimulation of the HIV-1 LTR. Interestingly, when both viral proteins were co-expressed a strong superinduction of 2800-fold over the effect of PMA ϩ Io was obtained (Fig. 3B). Taken together, these FIG. 1. HBx synergizes with cellular activation to induce HIV-1 replication. Jurkat cells were co-transfected with 2.5 g of the infectious clone pNL-Luc along with 5 g of either pSV-HBx or the control plasmid pSV-hygro. The cells were then stimulated with PMA (10 ng/ ml) plus calcium ionophore (0.5 M) or left untreated. A, viral replication was assessed by detection of HIV CA-p24 antigen in culture supernatants collected at days 1, 3, and 7 after transfection. Three independent experiments were performed, and the data of a representative one are shown. B, at day 1 after transfection, the cells for each condition were harvested, and the cell extracts were assayed for luciferase activity. The values represent the mean relative luciferase activities Ϯ S.E. of three independent experiments. results indicate that HBx and Tat act at different levels and add their effects, whereas these viral proteins strongly synergize cellular activation, which implies cooperativity between HBx and/or Tat proteins and the signals elicited by PMA ϩ Io.

HBx Synergizes with T-cell Activation Signals to Induce HIV-1 B Enhancer through Both NF-B and NF-AT-binding
Sites-Among the regulatory elements of the HIV-1 LTR, the B enhancer is considered the main inducible cis-acting element. Recently, it has been shown that in addition to NF-B proteins, members of the NF-AT family also bind to the HIV-1 B enhancer (6 -8).
The stimuli employed in this study to trigger T-cell activation (PMA ϩ Io) may lead to activation and nuclear translocation of both NF-B and NF-AT proteins. Therefore, to characterize the inducible protein complexes that interact with the B enhancer, electrophoretic mobility shift assays were performed using nuclear extracts from unstimulated or PMA ϩ Io-treated Jurkat cells. Four inducible DNA-protein complexes (complexes 1-4) were observed when the B oligonucleotide probe was incubated with nuclear extracts from PMA ϩ Io-treated cells (Fig. 4A). To identify the nature of these complexes, the binding reactions were preincubated with antibodies specific to p50, p65, NF-ATp and NF-ATc. The anti-p50 antiserum abolished the formation of complexes 2-4, and anti-p65 antiserum prevented the formation of complex 2. On the other hand, the anti-NF-ATc monoclonal antibody and the anti-NF-ATp antiserum induced the disappearance or reduced the intensity of complex 1. The addition of a preimmune antiserum did not affect any of the inducible complexes (Fig. 4A). These results demonstrate that T-cell activation with PMA ϩ Io induces the formation of NF-B and NF-AT-containing complexes in the HIV-1 B enhancer.
To analyze whether HBx was able to synergize with cellular activation to induce the HIV-1 B enhancer, Jurkat cells were transiently co-transfected with the plasmids pBTK-Luc or pmBTK-Luc (46), containing tandem copies of wt or mutated B enhancer, along with pSV-HBx or pSV-hygro. The transfected cells were either left untreated or stimulated with PMA ϩ Io. In addition, to determine the functional contribution of NF-AT in the activation of the B enhancer, the plasmid pGFP-VIVIT (43), encoding a specific NF-AT inhibitor, or the control vector pGFP were also included in the co-transfection experiments. As shown in Fig. 4B, the transcriptional activity of pBTK-Luc was induced about 72-fold by PMA ϩ Io treatment. In addition, HBx expression in combination with PMA ϩ Io further enhanced up to 438-fold the B-mediated transcriptional activity. On the other hand, the plasmid pmBTK-Luc, harboring tandem copies of the B enhancer mutated in sequences involved in NF-B binding but not in NF-AT binding, displayed a marked decrease in the response to PMA ϩ Io treatment (10-fold). The expression of HBx in the presence of PMA ϩ Io further increased up to 71-fold the transcriptional activity of this plasmid. In terms of fold induction over the effect of PMA ϩ Io, HBx activated both constructs, pBTK-Luc and pmBTK-Luc, to a similar extent (6 -7-fold). In the absence of PMA ϩ Io, HBx expression only slightly induced the transcriptional activity of those plasmids (data not shown). Interestingly, the plasmid encoding the NF-AT inhibitor, pGFP-VIVIT, did not significantly affect the up-regulated transcriptional activity of the plasmid pBTK-Luc. In contrast, the induction of pmBTK-Luc by PMA ϩ Io, either in the presence or the absence of HBx, was clearly prevented by the NF-AT inhibitor. The plasmids pKBF-Luc and pNF-AT-Luc, containing tandem copies of NF-B-and NF-AT-responsive elements, respectively, were used as controls to test the specificity of pGFP-VIVIT (data not shown). Taken together these results indicate that wt B enhancer functions mainly as a NF-B-responsive element when both transcription factors, NF-B and NF-AT, are activated by mitogenic stimuli. However, the B enhancer may also function as a NF-AT-responsive element in instances when T-lymphocytes receive stimuli that preferentially activate NF-AT translocation. Independent of which transcription factor is targeted to the B enhancer, the expression of HBx is able to further enhance the transcriptional activity of this cis-acting element.
Synergistic Activation of the HIV-1 LTR by HBx Is Maintained in the Absence of the B Enhancer-Given that HIV-1 replication takes place in activated T-cells and that the stimulatory effects of HBx on HIV-1 replication/transcription are more evident in synergy with cellular activation, the next experiments were focused mainly on the effect of HBx in a cellular activation background.
To analyze the functional role of the B enhancer in the synergistic activation of the HIV-1 LTR by HBx, the plasmid pXP1LTR ⌬B, in which the B enhancer had been deleted (Fig. 2), was employed in transient co-transfection assays in Jurkat cells. The response to PMA ϩ Io of this construct was significantly diminished in comparison with pXP1LTR wt (Fig.  5A). However, the fold induction by HBx expression over the effect of PMA ϩ Io was not affected by deletion of the B

FIG. 3. HBx further increases Tatand T-cell activation-mediated HIV-1 LTR transcriptional up-regulation.
Jurkat cells were co-transfected with 1 g of pXP1LTR wt along with 2 g of either pSV-HBx or the control plasmid pSV-hygro and 0.5 g of either pCMV-Tat or the control plasmid pcDNA3. 36 h after transfection cells were either left untreated (A) or stimulated with PMA (10 ng/ml) plus calcium ionophore (0.5 M) for 16 h prior to luciferase assay (B). The luciferase activities are represented as fold induction over the expression of pXP1LTR wt in the absence of any stimuli. The values shown represent the mean fold inductions Ϯ S.E. of at least four independent experiments. enhancer (Fig. 5A). In addition, the response of pXP1LTR ⌬B to Tat expression plus PMA ϩ Io treatment was also diminished in comparison with pXP1LTR wt (Fig. 5B). In contrast, the fold induction by HBx expression over the effect of Tat plus PMA ϩ Io was not altered in the absence of the B enhancer (Fig. 5B). Therefore, the absence of the B enhancer in the HIV-1 LTR reduces the Tat-and PMA ϩ Io-mediated transactivation but not the synergistic activation obtained by HBx expression.
To explore the functional relevance of a second NF-B-binding site located within the HIV-1 TAR element (45), the plasmids pXP1LTR ⌬TAR and pXP1LTR ⌬B ⌬TAR, in which the TAR element and the downstream NF-B sequence had been disrupted (Fig. 2), were also included in the co-transfection assays in Jurkat cells. Removal of the downstream NF-B-binding site alone (pXP1LTR ⌬TAR) had no effect on the response to PMA ϩ Io and on the synergistic activation by HBx expression, indicating that this putative NF-B-binding site is nonfunctional in our system (Fig. 5A). As happened with pXP1LTR ⌬B, disruption of both NF-B elements (pXP1LTR ⌬B ⌬TAR) affected the re-sponse to PMA ϩ Io but not the fold induction by HBx expression over the effect of PMA ϩ Io (Fig. 5A). These data reinforced the idea that the B enhancer is not necessary to mediate the synergistic activation of HIV-1 LTR by HBx. As expected, the expression of Tat alone did not significantly induce the activity of these ⌬TAR constructs (Fig. 5B).
The Sp1-binding Sites of HIV-1 LTR Are Sufficient to Mediate the Synergistic Activation by HBx but They Are Not Necessary in the Presence of the B Enhancer-It has been reported that the HIV-1 LTR Sp1-binding sites are not only mere components of the basal transcription machinery but that they also mediate the up-regulation of the transcriptional activity by Tat and other stimuli (10,11). In addition, it has been shown that HBx induces the expression of insulin-like growth factor II (IGF-II) through Sp1-binding sites located in the proximal promoter region of the IGF-II-encoding gene (50). Therefore, to analyze the relevance of the Sp1-binding sequences of HIV-1 LTR, a set of reporter plasmids containing the proximal region of HIV-1 LTR (nucleotides Ϫ81 to ϩ 77) with either wt or mutated Sp1-binding sites (Fig. 2) were employed in the co-

FIG. 4. HBx synergizes with T-cell activation signals to induce HIV-1 B enhancer through both NF-B-and NF-AT-binding sites.
A, EMSA was performed using a 32 P-labeled probe containing the B enhancer sequence from HIV-1 LTR and nuclear extracts from Jurkat cells either unstimulated (left lane) or stimulated with PMA (10 g/ml) plus calcium ionophore (1 M). Where indicated nuclear extracts were preincubated with specific antibodies before the addition of the radiolabeled probe. The arrows indicate the four major specific complexes. P.I., preimmune antiserum. B, Jurkat cells were co-transfected with 1 g of pBTK-Luc or pmBTK-Luc along with 2 g of either pSV-HBx or the control plasmid pSV-hygro and 2 g of either pGFP-VIVIT or the control plasmid pGFP. 36 h after transfection cells were either left untreated or stimulated with PMA (10 ng/ml) plus calcium ionophore (0.5 M) for 16 h prior to luciferase assay. The data are represented as fold induction over the value of untreated Jurkat cells transfected with each reporter plasmid along with the control vectors, and only the results of activated cells are shown. The effect of GFP-VIVIT is indicated as percentage of inhibition. The values shown represent the mean fold inductions Ϯ S.E. of three independent experiments. transfection assays in Jurkat cells. Mutation of the most distal Sp1 sequence (SpIII) did not significantly affect either the induction by PMA ϩ Io or the synergistic activation exerted by HBx over the effect of PMA ϩ Io (Fig. 6A). In contrast, mutation of the other two Sp1 sequences, SpII and SpI, markedly decreased the induction by PMA ϩ Io and partially prevented the cooperative induction by HBx (Fig. 6A). Furthermore, mutation of two (SpII and SpI) or three (SpIII, SpII, and SpI) Sp1-binding sites almost completely abolished both the induction by PMA ϩ Io and the synergistic activation by HBx over the effect of PMA ϩ Io (Fig. 6A). In agreement with previous results (9), mutation of the Sp1-binding sites SpI and/or SpII reduced the Tat-induced transcriptional activity of the HIV-1 LTR, as well as the synergistic effect of HBx over the activation by PMA ϩ Io plus Tat (data not shown). These results indicate that the Sp1 sequences, especially SpI and SpII, are sufficient to mediate the cooperative activation of the HIV-1 LTR by HBx and PMA ϩ Io.
To investigate the effect of Sp1 mutations in the presence of the B enhancer, the plasmids pXP1BSpwt and pXP1BSpIϩIIϩIII (Fig. 2) were employed in the co-transfection experiments. As shown in Fig. 6B, mutation of the three Sp1-binding sites in the presence of B enhancer markedly diminished the induction by PMA ϩ Io in comparison with the construct pXP1BSpwt. However, the cooperative activation by HBx over the effect of PMA ϩ Io was restored in the triple Sp1 mutant in the presence of the B enhancer (Fig. 6B), indicating that the Sp1-binding sites are sufficient but not necessary to mediate the synergistic activation by HBx and PMA ϩ Io. Similar results were obtained in the presence of Tat protein (data not shown). Fig. 6C summarizes the mean induction by HBx over the effect of PMA ϩ Io of this set of constructs.
HBx Induces HIV-1 LTR Transcriptional Activity by Acting as a Nuclear Co-activator-A dual mechanism for HBx function has been proposed. HBx has been shown to activate cytoplasmic signal transduction pathways (35)(36)(37)(38)(39)(40)(41) and to exert a nuclear co-activation function (26 -34). To assess these possibilities, the reporter plasmid pXP1LTR wt was co-transfected with the expression vectors pNLSHBx-FLAG and pSLNHBx-FLAG or the empty vector. The plasmid pNLSHBx-FLAG encoded a HBx protein fused to a nine-amino acid NLS, and pSLNHBx-FLAG encoded a HBx fused to a related sequence in which three amino acids of NLS were mutated to render it nonfunctional for nuclear import (SLN). Previous studies have shown that SLN-HBx is mostly localized in the cytoplasm but also to some extent in the nucleus, whereas NLS-HBx is relocalized exclusively to the nucleus and is no longer able to activate cytoplasmic signal transduction pathways, but it retains certain nuclear transcriptional activities (37). In Jurkat cells these variants of HBx displayed the expected pattern of distribution (data not shown). Nuclear-targeted HBx was able to cooperate with PMA ϩ Io in the induction of HIV-1 LTR even to a greater extent as compared with SLN-HBx (Fig. 7A), suggesting that HBx may exert its function by acting as nuclear coactivator of HIV-1 LTR. Mutation analyses of HBx have defined two separate regions necessary for its trans-activation function (51, 52) that overlapped the sequences involved in the interaction of HBx with RPB5 (an RNA polymerase subunit) and with TFIIB, respectively (27, 29 -31). To analyze the relevance of these regions of HBx in the activation of the HIV-1 LTR, substitution mutants of HBx affected either in its functional interaction with RPB5 (NLS-HBx m93) or with TFIIB (NLS-HBx m138), were included in the co-transfection assays. As shown in Fig. 7A, both mutations abrogated the synergistic activation of the HIV-1 LTR by HBx.
To further confirm that the interaction of HBx with components of the basal transcription machinery was necessary for the activation of the HIV-1 LTR, a series of nested deletion mutants of HBx (Fig. 7B) were co-transfected into Jurkat cells along with pXP1LTR wt. Deletion of the N-terminal portion of HBx up to amino acid 51 (5D3, 5D1) did not affect the induction of the HIV-1 LTR (Fig. 7B). Further removal of N-terminal sequences up to amino acid 72 or 102 (5D4 and 5D2), affecting the binding to RPB5 (30), abolished the co-activation function of HBx (Fig. 7B). In a similar manner, removal of C-terminal sequences of HBx, which contained, at least in part, the binding sequences for TFIIB (3D39 and 3D99), resulted in loss of coactivation function of HBx (Fig. 7B). Therefore, these results reinforce the hypothesis of a nuclear function of HBx in the transcriptional activation of HIV-1 LTR. DISCUSSION HIV-1 infection gives rise to a progressive disease with three clinical phases (1). The acute phase occurs within the first weeks after HIV-1 exposure and is characterized by a high level of plasma viremia. This initial phase is followed by an extended asymptomatic phase, which varies among different HIV-1 patients. During this latent phase, there is a slow but progressive FIG. 5. Synergistic activation of the HIV-1 LTR by HBx is maintained in the absence of the B regulatory element. A, Jurkat cells were co-transfected with 1 g of the reporter plasmid indicated (pXP1LTR wt, pXP1LTR ⌬B, pXP1LTR ⌬TAR, or pXP1LTR ⌬B ⌬TAR) along with 2 g of pSV-HBx or the control plasmid pSV-hygro and 0.5 g of pcDNA3. 36 h after transfection cells were either left untreated or stimulated with PMA (10 ng/ml) plus calcium ionophore (0.5 M) for 16 h prior to luciferase assay. B, luciferase assay similar to that described in A except for including pCMV-Tat instead of pcDNA3. The luciferase activities are represented as fold inductions over the expression of each reporter plasmid in the absence of any stimuli and only the results of activated cells are shown. The values shown represent the mean fold inductions Ϯ S.E. of at least four independent experiments. The numbers under each pair of columns represent the mean fold induction by HBx over the effect of PMA ϩ Io or PMA ϩ Io plus Tat. F.I., fold induction. deterioration of the immune system that leads to the final disease phase, characterized by the appearance of AIDS symptoms. The large interindividual variability in the AIDS incubation period is poorly understood, probably because many factors such as genetic background of the infected host and the virus, age, nutritional state, access to medical care, and coinfection with opportunistic pathogens may influence the HIV-1 outcome. In this context, co-infection of HIV-1 patients with HBV is frequently observed (17,18), which may cause complex bidirectional interactions among both viruses. It has been shown that HIV-1-induced impairment of cell-mediated immunity may cause higher HBV replication and a higher risk of cirrhosis without an evident increase of the necroinflammatory processes (53). On the other hand, it has been reported that HBV and HIV-1 co-infection may lead to a more rapid progression to AIDS (21) and to a reduced survival rate in patients already suffering from AIDS (22,23). In contrast, other studies have not revealed an association of HIV-1 and HBV co-infection with disease progression or with reduced survival rate of AIDS patients (24,25). The reasons for these discrepancies are not well understood, but they may reflect the complex and multifactorial nature of this disease.
We have demonstrated herein that the hepatitis B virus HBx protein, either alone or in synergy with cellular activation signals, induces HIV-1 replication in an in vitro system, in which HIV-1 derived proteins, including Tat protein, are expressed. We have also shown that HBx synergizes with both Tat and mitogenic stimuli in the induction of HIV-1 LTR transcriptional activity. Interestingly, HBx is able to further induce the activity of HIV-1 LTR when co-stimulated with Tat plus T-cell activation signals, suggesting that in the cellular environment where HIV-1 replication is ongoing, the presence of HBx may accelerate this process.
The B enhancer of the HIV-1 LTR mediates the response to a wide variety of stimuli, including specific T-cell activation signals. This regulatory element interacts and responds to NF-B/Rel family of transcription factors. More recently, it has been reported that members of the NF-AT family of transcription factors can also bind to an overlapping but distinct sequence of the B enhancer. It has been shown that NF-AT2 (also called NF-ATc) positively regulates HIV-1 replication in T-cells (54) and cooperates with NF-B and Tat in HIV-1 LTR transcriptional activation (6). The studies regarding the functional role of NF-AT1 (or NF-ATp) on HIV-1 replication and gene expression showed discordant results. One report showed that in Jurkat cells NF-AT1 decreases Tat-mediated HIV-1 LTR activation and competes for binding to the B enhancer with NF-B (7). In contrast, another study has demonstrated that NF-AT1 enhances HIV-1 replication and HIV-1 LTR transcriptional activity in primary CD4-positive T-cells (8). We have confirmed that the B enhancer interacts with both NF-B and NF-AT family members upon stimulation of the cells with PMA ϩ Io. Our results indicate that the B enhancer functions mainly as a NF-B-responsive element when the cells are treated with stimuli that trigger nuclear translocation of both transcription factors. However, it is plausible that NF-ATmediated transactivation of the HIV-1 LTR may be important in instances when the T-cells receive stimuli that preferentially activate NF-AT (55) or when the relative abundance of NF-AT is higher than NF-B, as occurs in naive primary T-cells (56). It has been demonstrated previously that HBx is able to induce NF-B-and NF-AT-dependent transcription in different cell types (39,40,57). We have demonstrated that HBx synergizes with T-cell activation signals to induce the B enhancer both in a NF-B-and NF-AT-dependent manner in Jurkat cells. Surprisingly, in the context of the HIV-1 LTR, the B enhancer does not seem to be necessary for the cooperative activation by HBx in the presence of PMA ϩ Io and Tat. Our results are seemingly discordant with previous reports, which have implicated, at least in part, the B enhancer in mediating the response of the HIV-1 LTR to HBx (58 -60). In one of these studies (59), the HBx-responsive region of the HIV-1 LTR was localized by nested deletion analysis within the region spanning from Ϫ104 to Ϫ57. However, the deletion of these sequences not only eliminated the B enhancer, but also removed the Sp1-binding sites SpIII and SpII, although this was not specifically mentioned. The other studies identified the B enhancer, by deletion and point mutation analysis, as the cisacting element that mediated, at least in part, the response to HBx in Jurkat cells (58) and in HepG2 cells (60). The discrepancies between our findings and these studies could be due to different conditions in which the effect of HBx was analyzed. In these studies the effect of HBx on HIV-1 LTR activity was analyzed in the absence of any other stimuli. Moreover, one of these studies was performed in the hepatocyte-derived cell line HepG2 (60), which does not include representative host cells for HIV-1 and HBV co-infection. In contrast, our study was carried out in a T-cell-derived cell line and focused on the cooperative effect of HBx with Tat protein and/or mitogenic stimuli.
The Sp1-binding sites of HIV-1 LTR are necessary for basal and Tat-induced transcriptional activity (9). However, it has been shown that other transcription factors such as NF-B, AP-1, or ATF-1 can compensate for Sp1 function in the replication of HIV-1 virions that contain deletion in the Sp1 sites (61,62). In agreement with these data, we have demonstrated that, in the absence of the B enhancer, the two most proximal Sp1 sites are sufficient to mediate the synergistic activation by HBx, but they are not necessary in the presence of the B enhancer. Transcriptional up-regulation by HBx through Sp1binding sites has also been reported for the promoter 4 of the IGF-II-encoding gene (50). In addition, it has been shown that HBx can stimulate the transcription mediated by a Gal 4-Sp1 chimeric protein (50). Mechanistically, the induction of Sp1dependent transcription by HBx appears to be mediated by enhanced phosphorylation and DNA binding activity of Sp1, but the exact molecular processes to explain how HBx could influence Sp1 phosphorylation state and function remain to be clarified.
It is well known that treatment of T-cells with mitogenic stimuli such as PMA ϩ Io activates the cytoplasmic signal transduction pathways that lead to nuclear translocation and activation of NF-B and NF-AT and to activation of Sp1 (63). Because HBx is able to further augment the HIV-1 LTR activity induced by PMA ϩ Io through both the B enhancer and Sp1 sites, it can be speculated that HBx exerts its trans-activation function, at least in part, by acting as a nuclear coactivator as has been proposed by other groups (26 -33). In this context, we have demonstrated that a nuclear-targeted HBx protein is also able to cooperate with T-cell activation signals to induce HIV-1 LTR activity and that mutations of HBx affecting FIG. 7. HBx induces HIV-1 LTR transcriptional activity by acting as a nuclear co-activator. A, Jurkat cells were co-transfected with 1 g of the reporter plasmid pXP1LTR wt along with 2 g of pSLNHBx-FLAG, pNLSHBx-FLAG, pNLSHBx-FLAG m93, pNLSHBx-FLAG m138, or the empty vector. 36 h after transfection cells were either left untreated or stimulated with PMA (10 ng/ml) plus calcium ionophore (0.5 M) for 16 h prior to luciferase assay. The luciferase activities are represented as fold inductions over the expression of each reporter plasmid in the absence of any stimuli and only the results of activated cells are shown. The values shown represent the mean fold inductions Ϯ S.E. of three independent experiments. B, Jurkat cells were co-transfected with 1 g of the reporter plasmid pXP1LTR wt along with the indicated deletion mutants of HBx and then stimulated or not with PMA ϩ Io. The ability of each HBx mutant to synergistically activate the HIV-1 LTR is indicated by ϩ or Ϫ.
its interaction with general transcription machinery abrogate the synergistic activation by HBx. This co-activation function might also account for the synergistic effect of HBx observed when only one of the elements from HIV-1 LTR susceptible to mediate HBx transactivation, the B enhancer or the Sp1binding sites, is present.
The mechanism by which HBx further increases Tat-mediated induction of HIV-1 LTR activity is not clear. Although we have been able to demonstrate protein-protein interaction between HBx and Tat, 2 the multiplicative activation of HIV-1 LTR by HBx and Tat is shown to be dependent on the TAR element. Thus, the HBx-Tat association does not seem to be sufficient to recruit Tat protein to the transcriptional machinery in the absence of TAR as has been suggested to occur with other viral transactivatiors (64,65). We suggest that once Tat has targeted the TAR element, HBx may collaborate with Tat in the stabilization of the transcription complexes.
Transcriptional up-regulation of the HIV-1 LTR, in synergy with Tat and/or cellular activation signals, has also been described for transactivators of other viruses, such as herpes simplex virus type-1, cytomegalovirus, human T-cell leukemia virus, and human herpesvirus 8, which are also commonly found in HIV-1 patients (64 -67). Therefore, in each HIV-1 carrier, the disease progression could be influenced by the co-infection with multiple viruses, making it difficult to establish the relative contribution of a particular co-infecting virus in the progression to AIDS. The results presented herein add new insight into the complex interaction established between HBV and HIV-1 replication cycles. Although clinical and epidemiological studies should confirm this, our findings provide molecular evidence that HBV may be a co-factor for HIV-1 disease progression.