Transcriptional activation of JC virus by human T-lymphotropic virus type I Tax protein in human neuronal cell lines.

Polyomavirus JC (JCV) causes the human demyelinating disease, progressive multifocal leukoencephalopathy (PML). The recent demonstration of cases of PML in association with human T-lymphotropic virus type I (HTLV-I) infection prompted us to examine whether the HTLV-I-encoded regulatory protein Tax activates JCV transcription. By employing a dual luciferase assay, we initially found that the expression of Tax activated the transcriptional potential of both early and late promoters of JCV in human neuronal but not in non-neuronal cells. We subsequently analyzed the mechanism of Tax-induced activation of the JCV promoter in neuronal cells with the following results: 1) the JCV promoter that lacks the NF-kappaB-binding motif could not be activated by Tax; 2) the overexpression of IkappaBalpha abolished Tax-induced transcriptional activation of the JCV promoter; 3) a Tax mutant (M22) lacking the potential for activation via the NF-kappaB pathway did not activate the JCV promoter. Furthermore, Tax enhances the gene expression of JCV T antigen and VP1. We examined mechanisms of the cell-specific activation of the JCV promoter by Tax. Electrophoretic mobility shift assay demonstrated the presence of Tax-bound protein(s) that were specifically present in non-neuronal cells. This study is the first demonstration of the activation of JCV promoter by HTLV-I Tax in an NF-kappaB-dependent manner.

JC virus (JCV) 1 is known to be a causative agent of the human demyelinating disease, progressive multifocal leukoencephalopathy (PML), that is observed mainly in immunosuppressive states such as in AIDS, advanced stage of malignant tumors, or following organ transplantation (1). JCV belongs to the polyomaviridae of double-stranded DNA viruses that also include simian virus 40 (SV40) and BK virus. Although the complete genome of JCV shares extensive homology with SV40 and BK virus, the nucleotide sequences of the transcriptional regulatory region of JCV are quite different. In addition, the JCV promoter sequence can be classified into two types, nonvirulent type (archetype) and pathogenic type (PML type). The regulatory region of JCV isolated from PML brain (PML type JCV) is generally composed of 98-base pair (bp) tandem repeats including TATA sequences, whereas archetype JCV that is excreted in the urine in asymptomatic individuals has only one copy of the 98-bp sequence flanked with additional 23-and 66-bp residues. Therefore, it has been postulated that the rearrangement of the regulatory region of JCV from archetypal sequence to that of 98-bp tandem repeats permits cytolytic proliferation of JCV in glial cells that results in PML (2). A number of transcriptional factors, such as NF-1, NF-B, YB-1, Pur␣, Tst-1, GF-1, and T antigen have been shown to bind JCV regulatory regions in in vitro analysis (3) and are thought to be involved in the species and tissue specificity of JCV propagation. Human immunodeficiency virus (HIV)-encoded Tat protein has also been reported to transactivate the JCV promoter (4). Despite the extensive efforts to clarify the transcriptional factors involved in the JCV activation, the precise mechanisms of its regulation in central nervous tissue and of pathogenesis of PML are still under the investigation.
Recently, we and others (5, 6) have described cases of PML in the setting of human T-lymphotropic virus type I (HTLV-I) infection. These have raised the question of whether HTLV-I proteins could have been involved in the transcriptional activation of the JCV promoter. HTLV-I regulatory protein Tax is a 40-kDa protein that is known to activate several cellular and viral genes including HIV, cytomegalovirus (CMV), and SV40. In this report, we demonstrate that expression of Tax activates the transcriptional potential of JCV promoter in human neuronal/glial cells, whereas transactivation of the virus promoter by Tax was not detected in non-neuronal cells.
We could also demonstrate the following: 1) JCV that lacks NF-B motif in its regulatory region cannot be activated even in the presence of Tax; 2) overexpression of the inhibitory protein, IB␣, abolishes Tax-induced transcriptional activation of JCV promoter; 3) a Tax mutant (M22) lacking the potential for NF-B activation fails to transactivate the JCV promoter. Thus three independent experimental approaches demonstrate that Tax transactivation of the virus promoter occurs via the NF-B pathway. Furthermore, the detection of a protein com-plex bound by JCV NF-B oligonucleotides using EMSA has suggested there is an inhibitory protein function in non-neuronal cells that suppresses JCV promoter activity. These results provide new insights into the understanding of the activation of JCV and a possible role for HTLV-I infection in the pathogenesis of certain cases of PML.

EXPERIMENTAL PROCEDURES
Cells and Tissue Culture-Human neuroblastoma cell line IMR-32 (JCRB 9050), human embryonic kidney cell line HEK293 (JCRB 9068), human cervical carcinoma cell line HeLa (JCRB 9004), human malignant melanoma cell line MeWo (JCRB 0066), and human mixed glioma cell line KG-1-C (JCRB 0236) were provided from the Health Science Research Resources Bank (Japan). Human glioblastoma cell lines U-373 MG (ATCC HTB 17), U-87 MG (ATCC HTB14), and U-138 MG (ATCC HTB16) were purchased from the American Type Culture Collection. All cell lines were maintained in Dulbecco's minimal essential medium supplemented with 10% fetal bovine serum and antibiotics (penicillin and streptomycin) from Sigma. None of the cell lines used these experiments expressed T antigen which is known to activate JCV promoter (data not shown).
Plasmids-Tst-1/Oct6 cDNA was kindly provided from Dr. Wegner (7) and subcloned into the mammalian expression vector, pFLAG-CMV-2 (Sigma) named pFLAG-Tst-1. The pJC1-4-ϾpJCV plasmid (HSRRB VG015), the source for the PML type JCV regulatory region was provided from Health Science Research Resources Bank. The luciferase reporter vector, pGL3-Basic, and internal control vector, pRL-TK, were purchased from Promega. The IB␣ expression vector (pIB␣) was kindly provided from Dr. Peyron through Dr. Ouchi (8). As the JCV regulatory regions are known to be quite polymorphic, we chose the regulatory regions of both CY (9) and Mad1 (10) as archetype and PML types, respectively. Nucleotide sequence analysis of the pJC1-4-ϾpJCV plasmid revealed a 4-bp insertion between the 98-bp repeat sequence in the regulatory region compared with that of Mad1. The sequence flanking the inserted region is as follows: 325 CAAGG-ggtc-330 CTGTA (the lowercase letters represent inserted nucleotides and the uppercase letters represent original Mad1 sequences, with numbers corresponding positions on the viral genome (10)). To eliminate the insertion, two DNA segments flanking the insertion were separately amplified by PCR, subcloned into a cloning vector to ligate each other, and named pBR-Mad1. The fragment containing the complete regulatory region of Mad1 was used for generating both the wild-type and the NF-B and NF-1 motif deletion mutants. The CY regulatory region from pJC-CY (11) was also used to create the wild-type, the NF-B, and NF-1 motif deletion mutants. A schematic illustration of JCV regulatory regions containing the NF-B and NF-1 motifs and TATA sequences is shown in Fig. 1. The JCV is bidirectionally transcribed from the regulatory region and both regulatory regions in the anti-clockwise direction for early transcripts and clockwise direction for late transcripts, respectively. These NF-Bbinding sites indicated as (Fig. 1, solid boxes) -AAG GGA ATT TCC CTG G-(16 nucleotides) and the NF-1-binding sites (Fig. 1, shaded boxes) -TGG CTG CCA GCC AA-(14 nucleotides) were separately deleted from both Mad1 and CY regulatory regions by oligonucleotide site-directed mutagenesis using the Unique Site Elimination mutagenesis kit (Amersham Pharmacia Biotech). Subsequently, the regulatory regions of wild type and NF-B-deleted were separately subcloned upstream of the firefly luciferase gene in the reporter vector pGL3-Basic in both directions, and NF-1-deleted regulatory regions of Mad1 and CY were subcloned in the late direction. Finally, these 10 kinds of luciferase reporter vectors (named pJCV-Mad E or L -luc, pJCV-CY E or L -luc, pJCV-Mad⌬B E or L -luc, pJCV-CY⌬B E or L -luc, pJCV-Mad⌬1 L -luc, pJCV-CY⌬1 L -luc) were used for the transfection experiments described below. For construction of Tax mutant (M22), two amino acids, Thr 130 and Leu 131 of the wild-type Tax were substituted by Ala 130 and Ser 131 using oligonucleotide site-directed mutagenesis. Both the wild-type Tax and M22 fragments were separately subcloned into the mammalian expression vector, pCXN 2 -FLAG, generating N terminus FLAG epitope (DYK-DDDDK)-tagged fusion protein (12) and were named pFLAG-Tax WT and pFLAG-Tax M22 , respectively.
Transfection and Luciferase Assays-Transient transfections were carried out by the Effectene method (Qiagen) following the manufacturer's instruction. In some experiments 1.0 ϫ 10 5 cells were plated onto 12-well plates, and for other experiments 3-10 ϫ 10 5 cells were seeded onto 60-mm plates and grown overnight. Vectors, including the reporter vectors, the internal Renilla luciferase control vector (pRL-TK), and other protein expression vectors were co-transfected as indicated in the figure legends. The total amount of transfected DNAs for the 60-mm plate was up to 1 g. All assays for firefly and Renilla luciferase activity were performed using one reaction tube sequentially (Promega). Briefly, at 48 h post-transfection, the cells were washed with phosphate-buffered saline and lysed with Passive Lysis Buffer. After a freeze/thaw cycle, samples were mixed with Luciferase Assay Reagent II, and the firefly luminescence was measured with a Luminometer (Turner Designs, CA). Next, samples were mixed with the Stop & Glo reagent, and the Renilla luciferase activity was measured as an internal control. Finally, luciferase activity was calculated as follows: (firefly luciferase activity of the sample/Renilla luciferase activity of the sample) Ϫ (firefly luciferase activity of the background (control vector)/ Renilla luciferase activity of the background (control vector)). The results of each experiment were confirmed by three independent transfections. The data are presented as mean values Ϯ S.D. Comparisons between groups were made with the t test for paired observations.
Immunoblotting-Anti-FLAG monoclonal antibody (M2) was purchased from Sigma, and anti-IB␣ (C-21), anti-p50, and anti-p65 (sc-1190X and sc-109X, respectively) were purchased from Santa Cruz Biotechnology. Cell lysates from each experiment were resolved by 11% SDS-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto nitrocellulose filters, and immunoblotted with the antibody. Immunoreactive bands were detected with an anti-Ig conjugated with horseradish peroxidase followed by ECL (Amersham Pharmacia Biotech) and analyzed with the LAS-1000 plus (Fuji Film, Japan).
Reverse Transcription (RT)-PCR Analysis of Viral Proteins-The closed circular CY DNA was generated from the plasmids, pJC-CY, after EcoRI digestion and self-ligation. This closed circular viral DNA was co-transfected with Tax expression vector, pFLAG-Tax WT , into the KG-1-C cells plated onto 6-well plates the day before transfection. At two days post-transfection, the serum concentration of conditioned media was decreased from 10 to 2%. Five days after transfection, cells were harvested and used for RT-PCR of mRNAs of the viral early and late proteins, T antigen and VP1, respectively, and the Tax protein. Total RNA was isolated with TRIzol reagent (Life Technologies, Inc.), and 1 g of total RNA was reverse-transcribed using Superscript II (Life Technologies, Inc.) according to the instructions of manufacturer. PCR was performed using AmpliTaq Gold DNA polymerase (Perkin-Elmer). Primers were as follows: T(t)Ag-F (5Ј-ATG GAC AAA GTG CTG AAT AGG G-3Ј) and T(t)Ag-R (5Ј-TTA AAG CTT TAG ATC CCT GTA GG-3Ј) were used to amplify the JCV early protein, T antigen (nucleotides 5013 to 4495); VP1-F (5Ј-ATG GCC CCA ACA AAA AGA AAA GG-3Ј) and VP1-R (5Ј-CAC TGT GGC ATT CTT TGG A-3Ј) were used to amplify the JCV late protein, VP1 (nucleotides 1469 to 1999); Tax-566F (5Ј-AGC GAA TAG AAG AAC TCC TC-3Ј) and Tax-R (5Ј-TCA GAC TTC TGT TTC TCG GA-3Ј) were used to amplify HTLV-I Tax (nucleotides 566 to 1062). The PCR conditions were as follows: for JCV proteins, 95°C for 10 min: 3 cycles of 95°C for 1 min, 50°C for 1 min, 72°C for 1 min: 27, 37, or 47 cycles of 94°C for 1 min, 58°C for 1 min, 72°C for 1 min, and final extension of 72°C for 10 min, for Tax protein, 95°C for 10 min: 3 cycles of 95°C for 1 min, 55°C for 1 min, 72°C for 1 min: 42 cycles of 94°C for 1 min, 62°C for 1 min, 72°C for 1 min, and final extension of 72°C for 10 min. Amplified DNAs were run on 1.2% agarose gel and quantified by Imaging Densitometer (Bio-Rad).
Immunoprecipitation and Western Blot Analysis-Tax-and/or Tst-1-

FIG. 2. Activity of the JCV promoters in the presence of Tax.
A, influence of Tax on the JCV promoter activities in various cell lines. Transient transfections were carried out in IMR-32, KG-1-C, HEK293, U-87 MG, and U-138 MG cell. The following plasmids were used for the experiments: pJCV-Mad E or L -luc, pJCV-CY E or L -luc, pFLAG-Tax WT , pCXN 2 -FLAG, and pRL-TK. Each reporter vector pJCV-Mad E or L -luc or pJCV-CY E or L -luc was co-transfected with either mock vector, pCXN 2 -FLAG, or Tax expression vector, pFLAG-Tax WT , into the five different human cell lines. Luciferase activities in extracts from transfected cells were determined in at least three independent experiments. The values of each JCV promoter reporter vectors were calculated by dividing the amount of luciferase activity of the pJCV-CY E -luc in the absence of Tax in each cell line. The activity of the pJCV-CY E -luc in the absence of Tax is therefore 1 in all cell lines tested. Data are presented as relative luciferase activity Ϯ S.D. The fold activation (indicated below) was calculated for each reporter plasmid by comparing values from cells transfected with pFLAG-Tax WT with values from cells transfected with mock vector pCXN 2 -FLAG. Comparisons between two groups were made with the t test for paired observations. *, p Ͻ 0.05, and **, p Ͻ 0.01 statistically difference from cells in the absence of Tax. B, upper panel, various amounts (indicated below) of the pFLAG-Tax WT plasmid were introduced into both HEK293 cells and IMR-32 cells co-transfected with 60 ng of pJCV-CY E -luc, 3 ng of pRL-TK vector and mock plasmid to adjust the total amounts of transfected DNA to 300 ng. At 48 h post-transfection, extracts were prepared, and both firefly and Renilla luciferase intensities were determined. Luciferase activities in extracts from transfected cells were determined in at least three independent experiments. Data are presented as mean fold activation Ϯ S.D., which were calculated for each reporter plasmid by comparing values from cells transfected with pFLAG-Tax WT with values from cells transfected with mock vector pCXN 2 -FLAG. Lower panel, the expression levels of FLAG-Tax fusion protein were estimated by the Western blotting with the anti-FLAG antibody. Solid arrows indicate the signals of FLAG-Tax protein.
transfected HEK293 cells and IMR-32 cells at 48 h post-transfection were lysed with hypertonic buffer containing 50 mM Hepes-KOH (pH 7.8), 420 mM KCl, 0.1 mM EDTA, 5 mM MgCl 2 , 20% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, and 2 g/ml each of aprotinin, pepstatin, and leupeptin. 200 ng of lysates were incubated with anti-Tax monoclonal antisera (13) on ice for 1 h, bound to 15 l of protein G-Sepharose beads (Amersham Pharmacia Biotech). The bound proteins were eluted with SDS-sample buffer from the beads after washing with ice-cold lysis buffer. Western blot analysis were carried out with a rabbit anti-FLAG polyclonal antibody (Zymed Laboratories Inc.).
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared from HEK293 cells, Tax-transfected HEK293 cells, IMR-32 cells, and Tax-transfected IMR-32 cells employing Dignam's methods (14) with slight modifications. Binding reactions were carried out as follows: 10 5 cpm of 32 P-labeled oligonucleotide probe, 5049 GGG AAA AAC AAG GGA ATT TCC CTG GCC TCC 5078 (the underline represents the NF-B-binding site, and numbers correspond to positions on the viral genome), was incubated with 20 g of nuclear extracts in a binding buffer containing 10 mM Tris-HCl (pH 7.4), 50 mM NaCl, 1 mM EDTA, 5% glycerol, 1 mM dithiothreitol, 4 g of poly(dI-dC)⅐poly(dI-dC), and 0.5 g of bovine serum albumin in a total volume of 25 l. Following a 30-min incubation on ice, the resulting complexes were resolved on a 5% polyacrylamide gel containing 0.25ϫ Tris borate EDTA (TBE) buffer at 4°C. For supershift band assays, samples were incubated with anti-p50, anti-p65 antisera (sc-1190X and sc-109X, respectively), and anti-VSVG-Tag antisera (MBL, Japan) as negative control on ice for 1 h before addition of the probe.

Activation of the JCV Promoter by Tax in Human Neuronal
Cells-To evaluate the transcriptional activity of the JCV promoter in the presence of Tax, each of the luciferase reporter vectors with the JCV promoter derived from archetypal type CY or PML-type Mad1 strain were co-transfected with the Tax expression vector, pFLAG-Tax WT , into five different human cell lines. As the regulatory domains of JCV are comprised of two independent promoters, which drive the viral early and late proteins of both archetype and PML type, we examined the transcriptional activity of all four promoters: CY early, CY late, Mad1 early, and Mad1 late. As shown in Fig. 2A, we found that Tax could activate the JCV promoter in the neuronal cells, including neuroblastoma cells, IMR-32 cells, and neuroglial cells, KG-1-C, U-87 MG, and U-138 MG. In particular, in IMR-32 and KG-1-C cells, the activity of the JCV promoters was significantly increased with 2.46 -95.60-fold activation by Tax. In contrast, luciferase activity was not enhanced by Tax in the human embryonic kidney HEK293 cells. In addition, we have examined other non-glial cell lines, HeLa and MeWo, and have found that the basal transcriptional activities of JCV in these cell lines was from 1/1,000 to 1/100 less than those of HEK293 cells, and the activities were not significantly changed even in the presence of Tax (data not shown).
To confirm that Tax did not activate the JCV promoters in HEK293 cells, we carefully compared the correlation of Tax protein expression level and transcriptional activity in both HEK293 and IMR-32 cells, and we found that significant expression of Tax failed to activate the CY promoter in HEK293 cells (Fig. 2B). These results suggest that in the neuronal cell lines, Tax significantly activates all four JCV promoters used in this experiment CY early, CY late, Mad1 early, and Mad1 late.
Deletion of the NF-B Motif in the JCV Promoter Reduces Its Transcriptional Activity That Is Not Rescued by Tax-It is well established that Tax activates NF-B enhancer signals and can promote viral gene expression of HIV, CMV, and SV40 through the NF-B pathway. Since both the Mad1 and CY type of JCV also have the NF-B motif in the 5Ј region upstream of the 98-bp box in their regulatory regions (Fig. 1), we hypothesized that Tax activated the JCV promoter via the NF-B-binding motif. We constructed JCV promoter-driven reporter plasmids that lack NF-B-binding motif, pJCV-Mad⌬B E or L -luc, or pJCV-CY⌬B E or L -luc, and these were co-transfected into U-87 MG cells with the Tax expression vector, pFLAG-Tax WT . As shown in Fig. 3A, in the absence of Tax the activities of JCV promoters, CY early, CY late, Mad1 early, and Mad1 late, were greatly reduced by deletion of the NF-B-binding motif, and in particular, the activity of the CY late promoter was completely abrogated. Moreover, overexpression of Tax did not rescue the reduced transcriptional activities of these mutated JCV promoters lacking the NF-B motif. In these experiments, the expression levels of Tax were confirmed by Western blotting. Representative data are shown in Fig. 3B. These results indicate that the NF-B motif is necessary for Tax-induced transactivation of the JCV promoter and also for the basal level of transcription of the JCV promoters with the exception of the CY early promoter.

Deletion of the NF-1 Motif of the JCV Promoter Does Not Diminish Tax-induced Activation of Transcription-
To confirm the specificity of the NF-B motif in Tax activation, we also prepared a deletion mutant of the NF-1 sequence that has been shown to be important in the JCV regulatory region. When the NF-1 deletion mutant reporter vector pJCV-Mad⌬1 L -luc was co-transfected with Tax expression vector pFLAG-Tax WT into U-87 MG cells, the transactivation of the NF-1 deletion mutant promoter was similarly enhanced to that of the wild type (Fig.  4A). These results suggest that the NF-1 motif is not necessary for Tax-induced transactivation and also not necessary for basal level transcription of the JCV promoter in neuroglial cells.
Overexpression of IB␣ Inhibits Tax-induced Activation of the JCV Promoter-Recent reports have demonstrated that the overexpression of IB␣ provides a dominant negative effect on the NF-B pathway diminishing the signals from IKK that results in the retention of NF-B in the cytoplasm (15). We investigated whether overexpression of IB␣ could inhibit the transactivation of the JCV promoters by Tax. The IB␣ expression vector (pIB␣) was co-transfected with Tax expression vector, pFLAG-Tax WT , and each of the luciferase reporter vectors derived from either CY or Mad1-strain into U-87 MG cells. As shown in Fig. 5A, overexpression of IB␣ completely inhibited the enhancement of transcriptional activity of the JCV promoters by Tax. In these experiments, the protein expression levels of both the Tax with FLAG tag and IB␣ were confirmed by Western blotting using the anti-FLAG antibody and anti-IB␣ antibody, respectively. Representative data are shown in Fig. 5B.
The M22 Tax Mutant Does Not Activate the JCV Promoter-The M22 Tax mutant (T130A and L131S) has been shown to lack the ability to activate the NF-B-dependent promoters (16). For further confirmation that Tax activates the JCV promoter through the NF-B pathway, the M22 Tax mutant was also examined for its ability to transactivate the JCV promoter. The wild-type reporter vector, pJCV-CY E -luc was co-transfected into U-87 MG cells with either pFLAG-Tax WT or pFLAG-Tax M22 which express the wild-type Tax and the M22 Tax mutant, respectively. As shown in Fig. 6, the M22 Tax mutant failed to activate, whereas the wild-type could activate the JCV promoter. These results clearly demonstrate that HTLV-I Tax activates the JCV promoter via the NF-B-dependent pathway.
Enhancement of Gene Expression of JCV Early and Late Proteins by Tax-We investigated if Tax enhances the transactivation of the JCV promoters in a native form of JCV genome, and we measured the levels of mRNA encoding the JCV early and late proteins. Closed circular CY DNA was co-transfected with the Tax expression vector pFLAG-Tax WT into KG-1-C cells, and 5 days after the transfection, the expression levels of mRNA for the JCV early and late proteins were analyzed by RT-PCR. As shown in Fig. 7, Tax enhanced the expression levels in both early and late transcripts with 4.3-and 9.3-fold increases, respectively. Expression of Tax was demonstrated by RT-PCR assay because its protein expression could not be detected by Western blotting over 3 days post-transfection (data not shown).
Tst-1/Oct6 Binds to Tax but Does Not Enhance the Taxinduced JCV Transactivation in HEK293 Cells-To analyze the mechanisms of cell type specificity of Tax-induced transactivation of JCV, we hypothesized that one transcriptional factor, Tst-1/Oct6 which is specifically expressed in neurons and glial cells and which is well known to stimulate the JCV promoter (7,17), may play a critical role. Specifically we examined whether the overexpression of Tst-1/Oct6 could convert HEK293 cells to permit Tax-dependent transactivation of the JCV promoter. As shown in Fig. 8A, overexpressed Tst-1/Oct6 was co-immunoprecipitated with Tax in HEK293 cells. However, the overexpression of Tst-1/Oct6 did not allow Tax to activate the CY early promoter in HEK293 cells (Fig. 8B).
HEK293 Cell-specific Proteins Bound to the JCV NF-B Motif-HTLV-I Tax transactivates variety of genes by the Rel/ NF-B family of transcriptional factors such as p50 and p65 which are differentially regulated in a cell type-dependent manner on the JCV promoter. Specifically, p50 has been reported to suppress the JCV transcriptional activity, whereas p65 has the potential for enhancement (18). To investigate the mechanism of cell type specificity of JCV transactivation, we initially examined the expression levels of the NF-B subunits, p50 and p65, in HEK293 and IMR-32 cells. As shown in Fig. 9A, Tax significantly enhanced the p50 expression in HEK293 cells compared with that in IMR-32 cells, whereas there was a similar expression levels of p65 by Tax in both cell lines. Thus, the significant increase in p50 expression levels induced by Tax in HEK293 cells might be involved in the suppression of Tax transactivation of the JCV promoter.
Next, we employed an EMSA to analyze JCV NF-B motifbinding proteins in HEK293 and IMR-32 cells. The EMSA demonstrated that in nuclear extracts from both IMR-32 and HEK293 cells Tax enhanced a major band (A, Fig. 9B). Furthermore, it could be demonstrated that an additional Tax enhanced band B exhibiting a lower mobility than band A could be detected only in HEK293 cells (Fig. 9B). The elimination of these bands by a cold probe containing the JCV NF-B motif in a dose-dependent manner demonstrates the specific interaction of the proteins with the probe. To examine the components of these protein complexes, the supershift assay was performed using anti-Tax, -p50, and -p65 antibodies. For both cell lines, the anti-Tax antibody supershifted the band A, and for HEK293 cells, the anti-p50, -p65, and -Tax eliminated band B. No effect was observed on band A in the supershift assay with anti-p50 or p65 antibodies, and this is presumably due to the fact that their potential binding sites are blocked by Tax (19). DISCUSSION It has been clearly demonstrated that JCV infection is increasingly important in HIV-1-infected individuals, and the incidence of PML has risen dramatically with the epidemic of AIDS. Evidence for a direct role for HIV in JCV activation has been provided by studies showing transactivation of the JCV late promoter by the HIV-1 Tat protein (4). Recently, we documented two autopsy-proven cases of PML without immunosuppression in the setting of co-infection of another retrovirus HTLV-I, and this raised the question of whether HTLV-I could activate JCV in a similar manner as HIV-1. We investigated if HTLV-I Tax could play a pivotal role in JCV activation, as Tax has been shown to alter the expression levels of a large number of cellular genes (20), and it has been reported that Tax can transactivate the promoters of several viral genes, including those of HIV, SV40, and CMV (21)(22)(23).
In our studies we could demonstrate that in neuronal/glial cells, the transcriptional activities of the JCV promoters were significantly enhanced by Tax. Furthermore, it could be shown  8. A, co-immunoprecipitation (IP) of nuclear extracts from Tax and Tst-1/Oct6-transfected cells by anti-Tax antibody. Nuclear extracts from Tax-and/or Tst-1-tranfected HEK293 cells at 48 h post-transfection were immunoprecipitated with anti-Tax monoclonal antibody and detected with anti-FLAG polyclonal antibody. B, influence of Tst-1 on the Tax-induced JCV transactivation. We examined whether Tst-1/Oct6 could transactivate the JCV CY early promoter in non-neuronal HEK293 cells using the luciferase assay. Transient transfections were carried out in HEK293 cells. The following plasmids were used: pJCV-CY E , pFLAG-Tax WT , pFLAG-Tst-1, pCXN 2 -FLAG, and pRL-TK. Luciferase activities in extracts from transfected cells were determined in at least three independent experiments. The activity of the pJCV-CY E -luc in the absence of neither Tax nor Tst-1 is therefore 1 in both cell lines tested. Data are presented as relative luciferase activity Ϯ S.D.
that Tax could not transactivate mutants of the JCV promoter lacking the NF-B motif. In contrast, Tax could enhance the promoter activity of another mutant of JCV promoter with a deletion of the NF-1 motif sequence. In addition, we have shown that overexpression of IB␣ dramatically inhibited the enhancement of JCV promoter induced by Tax and that the M22 Tax mutant that abrogates the ability of the protein to activate via NF-B could not transactivate the JCV promoters. Taken together, we have demonstrated that HTLV-I Tax activates the JCV promoters, CY early, CY late, Mad1 early, and Mad1 late, via the NF-B-dependent pathway. We also investigated whether Tax could enhance the expression of JCV proteins. In this experiment, both CY and Mad1 DNAs were cotransfected with the Tax expression vector into KG-1-C cells. In contrast to the CY-transfected cells, in which the gene expression was markedly enhanced by Tax, over 70% of the Mad1transfected cells showed cytoplasmic vacuolating change leading to cell death from the 2nd day after transfection in the presence of Tax (data not shown). As such it was impossible to detect the mRNAs of the JCV proteins in Mad1-transfected cells.
It is well known that JCV has a highly restricted cell type specificity and that transactivation of JCV by HTLV-I Tax occurs only in neuronal/glial cells. To analyze the mechanism of the neuronal specificity, we examined whether Tax could transactivate the JCV promoter in the setting of overexpression of the brain-specific transcriptional factor, Tst-1/Oct6. We could demonstrate using immunoprecipitation assays that Tax bound Tst-1/Oct6 in vivo. Because Tst-1/Oct6 is known to combine with a DNA motif in the vicinity of the TATA box (24), it was hypothesized that Tax might form a complex with NF-B, Tst-1/Oct6, and the TATA-binding protein and activate the JCV promoter. However, despite overexpression of Tst-1/Oct6 in HEK293 cells Tax failed to transactivate the JCV promoter.
We have also examined the expression levels of the NF-B family, p50 and p65 in the presence of Tax, because p50 inhibits whereas p65 activates the JCV promoter via the NF-B motif (18). In the presence of Tax, there was a significant increase of p50 expression levels in nuclear extracts from HEK293 cells compared with IMR-32 cells, whereas p65 was equally expressed in both cell lines. This result suggests the possibility that a surplus of p50 induced by Tax might be involved in suppression of the transactivation of JCV in HEK293 cells.
(B) was detected with the JCV NF-B probe. A second band (A) was recognized in both cell types. The latter band (A) probably represents a complex of Tax and NF-B, because this band was shifted by anti-Tax antibody and Tax could interact with both p65 and p50 in immunoprecipitation assays (data not shown). These findings raise the possibility that HEK293 cell-specific protein(s) function suppressively in Tax-induced activation of the JCV promoter. The narrow tissue-tropism of JCV is thought to be dependent on its regulatory region which a number of transcriptional factors bind, and most of these proteins function as up-regulators of the JCV promoter. In this study, we have shown the neuronal/glial cell-specific transactivation of the JCV promoter by HTLV-I Tax, and we demonstrate the existence of a non-neuronal cell-specific protein complex, which interacts with NF-B and which may be associated with the inhibition of the transactivation activity of Tax. Ongoing studies should identify the components of the latter and of the neuronal-specific factors involved in transactivation.