Secretion of the Human T Cell Leukemia Virus Type I Transactivator Protein Tax*

Human T cell leukemia virus type I (HTLV-I) is the etiologic agent of adult T cell leukemia and HTLV-I-associated myelopathy/tropical spastic paraparesis. The HTLV-I protein Tax is well known as a transcriptional transactivator and inducer of cellular transformation. However, it is also known that extracellular Tax induces the production and release of cytokines, such as tumor necrosis factor-α and interleukin-6, which have adverse effects on cells of the central nervous system. The cellular process by which Tax exits the cell into the extracellular environment is currently unknown. In most cell types, Tax has been shown to localize primarily to the nucleus. However, Tax has also been found to accumulate in the cytoplasm. The results contained herein begin to characterize the process of Tax secretion from the cell. Specifically, cytoplasmic Tax was demonstrated to localize to organelles associated with the cellular secretory process including the endoplasmic reticulum and Golgi complex. Additionally, it was demonstrated that full-length Tax was secreted from both baby hamster kidney cells and a human kidney tumor cell line, suggesting that Tax enters the secretory pathway in a leaderless manner. Tax secretion was partially inhibited by brefeldin A, suggesting that Tax migrated from the endoplasmic reticulum to the Golgi complex. In addition, combined treatment of Tax-transfected BHK-21 cells with phorbol myristate acetate and ionomycin resulted in a small increase in the amount of Tax secreted, suggesting that a fraction of cytoplasmic Tax was present in the regulated secretory pathway. These studies begin to provide a link between Tax localization to the cytoplasm, the detection of Tax in the extracellular environment, its possible role as an extracellular effector molecule, and a potential role in neurodegenerative disease associated with HTLV-I infection.

I-associated myelopathy/tropical spastic paraparesis (HAM/ TSP). The HTLV-I transactivating oncoprotein Tax plays an integral role in productive viral replication and disease progression. Tax has been studied with respect to its interaction with a number of cellular signaling pathways and transcription factor families, including activating transcription factor/cAMPresponse element-binding protein and NF-B (1)(2)(3)(4)(5). Specifically, Tax enhances cAMP-response element-binding protein binding within the HTLV-I long terminal repeat, which in turn enhances transcription of viral mRNA (1). With respect to the NF-B pathway, cytoplasmic Tax acts by binding directly to the IKK-␥ subunit of the IKK complex. This association induces the phosphorylation and degradation of IB␣, the inhibitor of NF-B, thereby allowing the NF-B complex to migrate to the nucleus and enhance gene expression (6 -8).
The pathogenesis of both ATL and HAM/TSP is coupled, at least in part, to the biological activity of Tax (9 -13). Tax has been demonstrated to be a key player in the malignant transformation of HTLV-I-infected T cells. In addition, several of the pathogenic processes associated with HAM/TSP have been postulated to occur as a result of the extracellular activity of Tax. First, in some individuals with HAM/TSP, Tax has been shown to promote hyperstimulation of the immune system. Specifically, these individuals harbor an extremely large number of Tax-reactive CD8 ϩ T cells that reside in the cerebrospinal fluid (14). Second, in some HTLV-I-infected individuals, antibodies directed against Tax also cross-react with the neuronal protein heterogeneous nuclear ribonuclear protein-A1, implicating the process of molecular mimicry in the progression of HAM/TSP (15,16). Finally, several studies have examined the effects of extracellular Tax on cells of the central nervous system (CNS). For example, microglial cells were previously shown to release tumor necrosis factor-␣ (TNF-␣), interleukin-6 (IL-6), and interleukin-1␤ (IL-1␤) in response to extracellular Tax (17). Additionally, neuronal cells were shown to release TNF-␣ in response to extracellular Tax (18). Each of these cytokines has been demonstrated to induce severe abnormalities in the CNS including, but not limited to, dysfunction of oligodendrocytes leading to demyelination of neurons, a pathology that is also observed in HAM/TSP patients.
There have been two cell types demonstrated to harbor HTLV-I proviral DNA in vitro and in vivo: the T lymphocyte and the astrocyte. Clearly, T lymphocytes represent the most commonly infected cell type and primary site of productive viral replication. Consequently, this cell type would represent an obvious source of extracellular Tax, since it infiltrates the CNS during the course of neurologic disease and has been associated with CNS lesions characteristic of HAM/TSP. Additionally, astrocytes have been demonstrated to harbor HTLV-I proviral DNA and exhibit an altered physiology after co-culture with HTLV-I-infected T cells (19 -21). Thus, it is possible that both T lymphocytes and astrocytes may serve as a source of extracellular Tax in the CNS. In 1990, Lindholm et al. (22) addressed the possibility that Tax may be secreted from HTLV-I-infected T lymphocytes. Tax was detected in the extracellular environment after purification of medium derived from MT-2 cell cultures, an HTLV-I-infected CD4 ϩ T lymphocyte cell line (22). The apparent release of Tax from MT-2 cells was most likely not the result of cell lysis or apoptosis, because the HTLV-I glycoprotein p24 could not be detected in the medium. Despite these interesting results, many questions still remain concerning the source of extracellular Tax and how the protein exits the cell.
In this report, we have investigated the secretion of Tax utilizing a cell biology system that has been previously used to examine the general properties of protein secretion. To this end, the baby hamster kidney (BHK-21) cell line was utilized for several reasons, including its previous use in numerous studies examining the biological process of protein secretion, its high transfection efficiency, and its ease of use in microscopic analysis. In addition to this, 293T cells were selected to study Tax secretion because of their human origin and because they were of kidney origin and therefore physiologically consistent with the BHK-21 cell line. Furthermore, the 293T cell line has previously shown utility in protein secretion studies and other microscopic analyses and is also transfected with high efficiency in many experimental systems.
In the experimental analysis contained herein, Tax was demonstrated to co-localize with several cytoplasmic organelles associated with exocytosis including the endoplasmic reticulum (ER) and the Golgi complex (GC), suggesting that Tax may migrate through these two organelles during the course of secretion. Second, time lapse video microscopy demonstrated that a large fraction of Tax localized to the cytoplasm and moved in a manner consistent with that of microtubule-associated proteins or secretory vesicles. Third, Tax was purified from the medium of Tax-transfected BHK-21 and 293T cells, strongly suggesting that Tax was secreted from these cell types. Fourth, Tax was shown to be secreted as a full-length protein, suggesting that it entered the cellular secretory pathway in a leaderless and uncleaved manner, a result similar to several other secreted proteins, including HIV-1 Tat, IL-1␤, and basic fibroblast growth factor (23,24). Finally, a small increase in Tax secretion was detected after combined treatment of cells with the secretagogues phorbol myristate acetate (PMA) and ionomycin, indicating that at least a fraction of Tax was present in the regulated secretory pathway.
Construction of Plasmids Encoding Fusion Proteins and Purification of Recombinant DNA-Full-length Tax cDNA coding sequence was cloned into pEYFP-N1 (Clontech, Palo Alto, CA) using PCR and Tax-specific primers, yielding a fusion protein in which yellow fluorescent protein (YFP) was fused to the carboxyl terminus of Tax. Tax and Tax mutant protein coding sequences were cloned into pIRES-EGFP (Clontech) using PCR and the Tax-specific primers that also encoded FLAG (DDDDK) or His 6 (HHHHHH) tags. Mutation of FLAG-Tax-His 6 to FLAG-Tax-COPII-His 6 ( 330 DHE 332 to 330 AHA 332 ) was performed using site-directed mutagenesis (Stratagene, La Jolla, CA). Plasmid DNA used for screening and automated sequencing was isolated using the Concert miniprep system as described by the manufacturer (Invitrogen). The nucleotide sequence of all plasmid constructs was confirmed by automated sequencing and subsequent bioinformatics analysis using Lasergene software (DNASTAR, Madison, WI). Plasmid DNA used for transient transfections was isolated using the HiSpeed Plasmid Midi DNA purification system as described by the manufacturer (Qiagen, Valencia, CA).
Construction of FLAG-Tax-His 6 (FTH 6 ) Plasmid and Purification of FTH 6 Protein-To facilitate the immunoprecipitation and quantitative estimation of secreted Tax, FLAG and His tags were added to the Tax. The FTH 6 plasmid was constructed by inserting the FLAG DNA coding sequence at the 5Ј end of the Tax-His 6 gene in the pTax-His 6 expression plasmid provided by Dr. Chao-Zen Giam (Uniformed Services University of Health Sciences, Bethesda, MD) (25). Insertion of the FLAG coding sequence was performed using QuikChange site-directed mutagenesis (Stratagene). FTH 6 was expressed in Escherichia coli (HB101) for 14 h, after which cells were pelleted in a JA-17 rotor (5000 rpm, 15 min, 4°C). FTH 6 protein was purified using the His-bind purification system as described by the manufacturer (Novagen, Madison, WI). Purified protein was subjected to SDS-PAGE (180 V, 45 min) and was stained using the Silver Stain Plus reagent (Bio-Rad). The concentration of FTH 6 protein was determined using densitometry with subsequent comparison against a set of silver-stained low range molecular weight marker proteins (Bio-Rad). The presence of both the FLAG and His 6 tags was verified by Western immunoblot analysis. Microscopic Analyses-Transfected cells were prepared for microscopy ϳ24 h post-transfection. For imaging of live cells transfected with Tax-YFP along with either CFP-nucleus (CFP-Nuc), CFP-ER, or CFP-Golgi, no preparation was performed prior to microscopic analysis. For imaging of Tax secretion constructs, medium was aspirated, and cells were washed once with 0.2 ml of phosphate-buffered saline (PBS). Cells were fixed for 30 min with paraformaldehyde (4%) and washed twice with PBS for 5 min. Cells were permeabilized with Triton X-100 (0.5%) in PBS for 15 min and again washed twice for 5 min with PBS. FLAG epitope constructs were immunostained using an anti-FLAG-Cy3 antibody (Sigma). Nucleic acids were stained with DAPI (100 pM, 2 min; Molecular Probes, Inc., Eugene, OR) diluted in PBS. Coverslips were added along with Fluormount-G mounting medium (Southern Biotechnology Associates, Inc., Birmingham, AL). Cells were visualized using an Olympus IX-81 automated microscope in conjunction with the appropriate filter cubes for visualizing enhanced YFP and ECFP or visualizing EGFP, DAPI, and Cy3. Images were obtained with a Cook CCD Sensicam digital camera controlled by Slidebook software (Intelligent Imaging Innovations, Denver, CO). All components of the microscopy system were controlled using an Apple Macintosh G4 dual 1-GHz processor computer (Apple, Cupertino, CA). Raw fluorescent images were deconvolved (no neighbors method) using Slidebook.
For time lapse microscopy, cells were plated onto round glass coverslips (60 mm) and transfected as described under "Transient Transfections." After transfection (8 h), the coverslip containing the transfected cells was inserted into the FCS2 live cell chamber system (Bioptechs, Butler, PA) and incubated at 37°C. Some cells were treated with either the microtubule polymerization inhibitor nocodozole (50 M, 90 min; Sigma) or the kinesin inhibitor monastrol (100 mM, 4 h; Sigma) just prior to insertion into the live cell chamber system. The live cell chamber was then attached to the Olympus IX-81 microscopy system as outlined above. Time lapse images (ϫ60) were obtained at intervals of 3 s as guided by the Slidebook software system.
Detection of Native Tax and FTH 6 in Cell Culture Medium-BHK-21 cells (1 ϫ 10 7 ) were seeded in a 100-mm culture dish and transfected (in triplicate) with pCMV4, pCMV-Tax, or pCMV-FTH 6 using Lipofectamine 2000 as described under "Transient Transfections." After transfection (24 h), fresh medium containing protease inhibitors aprotinin and leupeptin (1.0 g/ml each; Sigma) was added. After an additional 24 h, the medium was collected and subjected to centrifugation (600 ϫ g, 5 min) to pellet cell debris. Cells remaining in culture were washed once with ice-cold PBS (1.0 ml) and lysed at 4°C with shaking for 10 min using M-PER extraction reagent supplemented with HALT protease inhibitor (Pierce). Both cell lysate and supernatant were concentrated using a nanosep 10K liquid concentrator (Pall Life Sciences, Ann Arbor, MI). Total protein in the supernatant was precipitated by the addition of ice cold 10% trichloroacetic acid for 20 min on ice. The precipitate was collected by centrifugation at 14,000 rpm for 15 min at 4°C. The protein pellet was washed twice with 100% acetone (10-min incubation on ice), dried, and resuspended in Laemmli Sample Buffer (Bio-Rad). Cell lysates were diluted 1:2 with sample buffer. All of the samples were denatured at 95°C for 5 min and loaded onto a 12% Tris-HCl SDS-polyacrylamide gel (Bio-Rad) and subjected to electrophoresis (125 V, 90 min). Samples were then blotted onto polyvinylidene difluoride membrane (Immobilon-P; Millipore Corp., Bedford, MA) for 2 h (100 mA). Western blot analysis was performed at room temperature by first blocking with 5% bovine serum albumin dissolved in a solution of PBS with 0.05% Tween 20 (PBST; 1 h) and then rinsed once with PBST for 10 min and twice with PBS for 5 min. Blots were then incubated for 1 h with a 1:50 dilution (PBST plus 1% bovine serum albumin) of anti-Tax monoclonal antibody (TAB 170), a generous gift from Dr. Fatah Kashanchi (George Washington University, Washington, D. C.). After primary antibody incubation, blots were washed as before and incubated with a 1:10,000 dilution (PBST plus 1% bovine serum albumin) of Protein G-peroxidase (Sigma) for 1 h. After the final wash, blots were developed using Western Lightening (PerkinElmer Life Sciences) and exposed to x-ray film.
Purification of Secreted FTH 6 -Six plates (35 mm) of BHK-21 cells were transfected with each plasmid construct used (pIRES-EGFP and FTH 6 ) as described under "Transient Transfections." After transfection, medium in each well was replaced with new medium containing the protease inhibitors aprotinin and leupeptin (1.0 g/ml; Sigma). After 24 h, the medium (12 ml/sample) was collected and subjected to centrifugation (600 ϫ g, 1 min) to pellet all cell debris. Medium was then transferred to a 15-ml conical tube, and 250 l of anti-FLAG-agarose beads (Sigma) was added. Samples were then placed on an end-over-end rotor overnight at 4°C. Tubes were subjected to centrifugation (400 ϫ g, 5 min), and the medium was removed. Agarose beads were washed three times (500 l) with 1ϫ wash buffer (Sigma), resuspended in 50 l of 2ϫ SDS loading buffer (Sigma), and heated at 95°C for 5 min. The presence of FTH 6 was assayed using Western immunoblot analysis as described above using either anti-GFP antibody ab-290 (Abcam Ltd.), the M2 anti-FLAG antibody (Sigma), or the anti-His 6 -horseradish peroxidase antibody (Sigma) as primary antibody and anti-rabbit horseradish peroxidase-conjugated secondary antibody (Amersham Biosciences).
Quantitation of Tax Secretion by ELISA-BHK-21 cells were transfected with FTH 6 as described above, and 24 h post-transfection, new medium containing the protease inhibitors aprotinin and leupeptin (1 g/ml; Sigma) was added. After an additional 24 h, the medium was collected and subjected to centrifugation (16,000 ϫ g, 1 min) to pellet all cell debris. The supernatant was separated into two fractions. One fraction (100 l) was utilized for the detection of lactate dehydrogenase (LDH) using the CytoTox 96 nonradioactive cytotoxicity assay as described by the manufacturer (Pierce). The other fraction (1 ml) was concentrated to a volume of 200 l using a nanosep 10K liquid concentrator (Pall Life Sciences, Ann Arbor, MI). Cells remaining in the culture well were washed once with ice-cold PBS (1 ml) and lysed on ice for 10 min using 200 l of M-PER extraction reagent supplemented with HALT protease inhibitor (Pierce). Cell debris was pelleted by centrifugation (16,000 ϫ g, 10 min). Cell extract and corresponding concentrated medium were loaded onto a 96-well plate precoated with anti-FLAG antibody (Sigma). Purified FTH 6 protein was also loaded onto a 96-well plate as a series of quantification reference standards (1000, 100, 10, and 1 ng/well). The plate was incubated at 37°C for 2 h, at which time each well was washed three times with 300 l of PBS with 1% Tween (PBST). A solution of PBST and anti-His 6 -horseradish peroxidase antibody (ab1187, 200 l, 1:1000 dilution; Abcam), was then added to each well and incubated at 4°C for 2 h. Each well was then washed three times with 300 l of PBST (1%). Detection of bound anti-Tax-horseradish peroxidase was detected using tetramethyl benzidine substrate as described by the manufacturer (Pierce) and measured using a microplate reader (450 nm). This procedure was repeated for a total of three times with each cell type.
Treatment of Cells with Brefeldin A or a Combination of PMA and Ionomycin-Cells were plated in 6-well plates and transfected with secretion constructs as described under "Transient Transfections." After transfection (24 h), the medium was removed and replaced with 2 ml of medium containing either brefeldin A (10 g/ml) or a combination of PMA (100 nM; Sigma) and ionomycin (1 M; Sigma). Cells were incubated for 4 h, after which time 1 ml of medium was removed and replaced with 1 ml of new medium containing either brefeldin A or a combination of PMA and ionomycin. After initial medium replacement (24 h), the medium was removed and subjected to centrifugation at (400 ϫ g). Medium and cells were then prepared as outlined under "ELISA Detection of Tax." Detection of Apoptosis and Necrosis of Transfected BHK-21 Cells-Cells were plated in 8-well glass slides and either not transfected or transfected (pUC18, pTax, FTH 6 ) as described under "Transient Transfections." Untransfected cells incubated in medium containing 10 M camptothecin (Sigma) for 4 h were used as the positive control for the experiments designed to detect apoptosis. Untransfected cells incubated in medium containing 50 M ebselen (Sigma) for 6 h were used as the positive control for experiments to detect necrosis. Reactions with only ethanol and Me 2 SO were utilized to control for each solvent used with ebselen and camptothecin, respectively. After transfection (24 h), as well as after the stated length of time for each positive control, apoptosis and necrosis were detected using the Vybrant apoptosis detection kit 7 as described by the manufacturer (Molecular Probes, Inc., Eugene, OR). Cells were viewed as described under "Microscopic Analyses."

Tax-YFP Co-localizes with Organelles Associated with Pro-
tein Secretion-Several studies have examined the intracellular localization of the HTLV-I Tax protein (19, 26 -29). Most of the investigations, which examined Tax intracellular localization in HeLa and COS-7 cells and the HTLV-I-infected cell lines C8166-45 and MT2, have concluded that Tax resides mainly in the nucleus in the form of interchromatin granules and spliceosomal speckles (30), sites of high rates of gene transcription. As a result, most studies of Tax function have focused on the role of nuclear Tax and its involvement in gene regulation. However, recent studies have demonstrated that there is a significant amount of Tax localized to the cytoplasm, an amount that is cell type-dependent. For example, recent studies utilizing HeLa cells and primary astrocytes or astrocytic cell lines infected with HTLV-I demonstrated that Tax accumulated to significant levels in the cytoplasm as well as the nucleus (19,(31)(32)(33)(34). Furthermore, the localization of Tax in the cytoplasm of these cells was observed in both small and large punctate particles. These punctate structures suggested that Tax interacted with cytoplasmic proteins or localized to specific cytoplasmic organelles. In either case, many of the proteins involved in interfacing with cytoplasmic Tax have not been characterized in detail. Many studies have examined the adverse effects of extracellular Tax, including the stimulation of the production and release of cytokines. However, the pathway whereby Tax gains access to the extracellular environment has not been identified. Theoretically, Tax may exit the cell through at least three avenues, cell necrosis, cell apoptosis, or via the cellular secretory pathway. Based on our observations, we hypothesized that a portion of the particulate forms of Tax observed in the cytoplasm represent Tax protein localized to the formal cellular secretory pathway. To begin to examine this hypothesis, we proceeded to determine whether cytoplasmic Tax was localized to cellular organelles associated with the secretory pathway, specifically the ER and GC. To allow for rapid determination of the intracellular localization of Tax in live cells, Tax was fused to the amino terminus of yellow fluorescent protein (YFP). Tax-YFP and Tax-GFP constructs have been previously used in intracellular localization assays and are effective reagents to detect Tax in live cells (27). To study the interaction of Tax with the cellular secretory pathway, cells were selected because of their utilization in many other studies concerning protein secretion (35)(36)(37)(38)(39). After transient transfection, Tax-YFP accumulated in both the nucleus and cytoplasm of BHK-21 cells (Fig. 1), similar to previous reports utilizing other cell types. In the nucleus, Tax-YFP was found in the characteristic interchromatin granules and spliceosomal speckles (Fig. 1). To characterize specific localization of cytoplasmic Tax, Tax-YFP was co-transfected into BHK-21 cells with either of two recombinant plasmids encoding cyan fluorescent protein (CFP), CFP-ER or CFP-Golgi, that localize to the endoplasmic reticulum and Golgi apparatus, respectively. Utilizing deconvolution microscopy, it was demonstrated that Tax-YFP colocalized with both the ER and Golgi markers, strongly suggesting that Tax was localized within both of these organelles (Fig. 1). YFP alone was localized diffusely throughout the cell and was not concentrated to any specific organelle (data not shown). Given that most secretory proteins travel through the ER and Golgi, this evidence implicated Tax localization to the secretory pathway.
Movement of Cytoplasmic Tax-YFP in BHK-21 Cells Is Similar to Proteins Contained within Secretory Vesicles-Proteins that are secreted from the cell and associated with the secretory pathway are usually packaged into secretory vesicles within the Golgi complex and then migrate to and reside within the cytoplasm until their contents are released into the extracellular environment. Since it was demonstrated that Tax was associated with the ER and Golgi, it was necessary to determine whether some of the Tax-YFP cytoplasmic particles were associated with secretory-like vesicles to complete the link between ER localization and cytoplasmic Tax. Studies performed previously have examined the movement of secretory proteins and secretory vesicles utilizing time lapse video microscopy and inhibitors of microtubule polymerization or motor proteins to halt secretory vesicle movement. Consequently, these methods were also utilized to track Tax-YFP in BHK-21 cells co-transfected with CFP-Nuc to demarcate the nucleus. A representative cell from a population of Tax-YFP-transfected cells is shown in Fig. 2A. Each cell transfected with Tax-YFP contained vesicle-like particles present within the cytoplasm. Time lapse photographic analysis of these particles indicated that many of the Tax-YFP-containing structures moved in a manner similar to microtubule-associated structures and other secretory proteins. The stop-and-go and change-of-direction motion of several Tax-YFP-containing cytoplasmic structures is shown in Fig. 2B. Additionally, many of the Tax-YFP particles migrated significant distances (4 -10 M) over the time period examined (Fig. 2B), another characteristic feature of mictotubule-associated vesicles. To provide additional evidence that cytoplasmic Tax-YFP may be associated with the secretory pathway, Tax-YFP-transfected cells were incubated with either nocodazole or monastrol, two compounds known to inhibit movement of secretory vesicles. Specifically, nocodozole inhibits microtubule polymerization, whereas monastrol inhibits the microtubule-associated motor protein kinesin. In both treated cell cultures (Fig. 2, C and E), movement of the cytoplasmically localized Tax-YFP (Fig. 2, D and F) was abrogated or severely retarded. These results provided further evidence that at least some fraction of the cytoplasmic Tax-YFP particles were associated with the microtubule-mediated secretory pathway.

Full-length Tax Is Secreted from BHK-21 Cells-The co-localization of Tax-YFP to the ER and GC and the association of
Tax-YFP-containing particles in the cytoplasm with secretory vesicles suggested Tax movement through the cellular secretory pathway. To continue the investigation of Tax secretion, it was important to determine whether Tax is released from the cell into the extracellular environment. In order to effectively detect, immunoprecipitate, and quantify Tax in an efficient manner, two amino acid tags were fused to the protein. Specifically, a FLAG tag (DDDDK) was fused to the amino terminus, and a His 6 tag was fused to the carboxyl terminus of Tax (Fig.  3). To ensure, by visualization, that each Tax construct was effectively transfected into the target cell population, the FTH 6 construct was cloned into the pIRES-EGFP vector (pIRES-FTH 6 ) where the expression of GFP was promoted by an internal ribosome entry site (IRES). Thus, any green cell within a culture could also be expected to express FTH 6 .
Numerous studies have demonstrated that even small changes in amino acid structure or the addition of long amino acid tags to Tax may significantly alter many functional properties of Tax, including its intracellular localization and/or its ability to activate cellular or viral gene expression (20). As a result, several functional properties of FTH 6 were examined. First, FTH 6 expression was examined using transfected BHK-21 cell lysates for SDS-PAGE analysis and Western immunoblotting (Fig. 3C). Tax was detected using antibodies against both the amino terminus FLAG tag and the carboxyl terminus His 6 tag to ensure that full-length Tax was being expressed. In addition, Western immunoblotting with an anti-GFP antibody demonstrated that GFP was expressed evenly throughout all transfected cultures. Furthermore, immunofluorescence microscopy was utilized to determine whether the FLAG and His 6 tags affected intracellular localization of Tax following transfection of BHK-21 cells (Fig. 3B). As shown, FTH 6 localized to both the nucleus and cytoplasm, consistent with the nucleocytoplasmic localization of Tax-YFP in BHK-21 cells (Fig. 1).
Confident that the additional FLAG and His 6 peptide tags did not affect Tax expression and intracellular localization, experimentation was performed to analyze the secretion of In an effort to determine which intracellular organelle(s) Tax localizes to, Tax-YFP was transiently co-transfected with either CFP-Nuc, CFP-ER, or CFP-Golgi in BHK-21 cells. After a 24-h incubation, cells were viewed on an inverted fluorescent microscope using a ϫ60 objective. After image capture, each image was deconvolved using the nearest neighbors method within Slidebook (Intelligent Imaging Innovations). Green, Tax-YFP; blue, CFP constructs; yellow, co-localization. FTH 6 . Secreted FTH 6 was immunoprecipitated from either the medium or the cellular lysate utilizing an anti-FLAG antibody and subsequently examined by Western immunoblot analysis using both anti-His 6 and anti-FLAG antibodies (Fig. 4, top). This method resulted in the purification and detection of a 43-kDa protein, the predicted size of FTH 6 protein, from both the lysate and medium, and the protein was detected using both anti-His 6 and anti-FLAG tag antibodies. This result is significant for two reasons. First, it is the most conclusive evidence to date that Tax can be detected in the medium of cells transfected with a recombinant plasmid encoding Tax. Second, since extracellular Tax was purified and detected utilizing amino acid tags found at both the amino and carboxyl terminus of the protein, Tax was probably released from the cell in a leaderless manner, a mechanism common to only a few proteins, such as IL-1␤ and HIV-1 Tat. Finally, it was important to ensure that the FLAG and His 6 tags were not the causal reason for Tax secretion. To assess this possibility, proteins contained in the medium of Tax (pCMV-Tax) and FTH 6 transfected cells were precipitated and subjected to SDS-PAGE. Western blot analysis was then performed using an anti-Tax monoclonal antibody (TAB 170) (Fig. 4, bottom). This method resulted in the detection of both FTH 6 and native Tax, strong evidence that native Tax was secreted and that the addition of the FLAG and His 6 was not the basis for this observation.
Quantification of Tax Secretion by ELISA-Since we have demonstrated that secretion of FTH 6 behaves in the same manner as native Tax, FTH 6 was utilized for the quantitative analysis of secreted Tax. Using an ELISA-based system, FTH 6 secretion from both BHK-21 and 293T cells was demonstrated (Fig. 5). This system is advantageous over previous systems that have been described (22) in that it provides a simple, yet rapid and sensitive analysis of Tax secretion. In addition, this system has also allowed examination of Tax protein mutants that may be of specific relevance to Tax protein secretion. For example, examination of the Tax primary amino acid sequence determined that Tax contained a putative DXE signal (amino acids 300 -302; Fig. 3). The DXE amino acid signal was previously reported to be important in binding to the COPII complex, a complex important in the progression of secretory proteins from the ER to the GC, and concentration of these proteins into secretory vesicles (40). This report also demonstrated reduced secretion of the protein after the DXE signal was mutated to AXA. Thus, a Tax protein construct AXA, FLAG-Tax-COPII-His 6 (FTCOPIIH 6 ; Fig. 3), containing a mutation in the DXE signal was designed for use in these studies. To compare the amount of Tax released between experimental repetitions, the relative amount of Tax secreted was converted to percentage of total Tax secreted in each transfected culture. Surprisingly, a large percentage of Tax was released from BHK-21 cells into the medium, an average of 53.1% of all Tax detected in both the medium and cell lysate fractions (Fig. 5A). Additionally, 293T cells secreted 16.2% of Tax detected in both the medium and cell lysate. The difference in the percentage of Tax secreted by BHK-21 versus 293T cells may be the result of a factor(s) that would result in a general difference in the secretion output of either cell type and any of a number of cellular or species differences that could specifically alter the ability of Tax to enter the secretory pathway as a result of changes in cellular recognition of secretory signal sequences within Tax. Importantly, these results demonstrated that Tax secretion could be observed in cells of hamster and human origin. Purified FTH 6 protein was utilized as both a positive control for detection of the chimeric protein and a protein concentration standard that could be used to calculate the specific amounts of FTH 6 released in each culture. Utilizing this method, an average of 132.5 and 82.3 ng of Tax was detected in each 2 ml of BHK-21 and 293T cell culture medium, respectively. Surprisingly, the amount of FTCOPIIH 6 mutant protein secreted from BHK-21 cells was only slightly below that detected with the parental protein. This result was somewhat unexpected, given that our hypothesis predicted that Tax in-teraction with the COPII machinery would be necessary for the migration of Tax to the GC. The results from 293T cell cultures were more consistent with the hypothesis suggesting that the putative DXE signal within Tax was more functional in this cell type. More recent reports have suggested that other amino acid signals surrounding the DXE motif are just as important, or more so, than that of the DXE motif itself for migration of proteins from the ER to the GC (41). Thus, study of the putative DXE signal within Tax must be examined in context with other signals at the carboxyl-terminal end of the protein.

Tax Secretion Is Inhibited by Brefeldin A-Brefeldin
A is a chemical inhibitor of protein secretion that specifically blocks secretory vesicle migration from the ER to the GC. Thus, to provide additional evidence that Tax utilized the cellular secretory system, analysis of FTH 6 secretion in BHK-21 cells was performed in the absence or presence of brefeldin A (10 g/ml). After 24 h of treatment, the ELISA secretion assay was performed to determine the effects of brefeldin A. The results (Fig.   FIG. 3. Intracellular localization and expression analysis of plasmid constructs utilized in secretion analyses. A, pictoral representations of the plasmid constructs utilized in secretion analyses of Tax and Tax mutant proteins. Each construct contains an amino-terminal FLAG tag as well as a carboxyl-terminal His 6 tag. Additionally, each construct was inserted into a plasmid with a GFP coding sequence expressed under control of an IRES. Tax-COPII contains a DXE to AXA mutation at amino acids 330 -332. The DXE motif was previously described as a domain important in protein localization to the secretory pathway. B, the intracellular localization of secretion constructs was determined to ensure that the additional amino acid tags did not alter the normal intracellular distribution of Tax in BHK-21 cells. Secretion construct plasmids were transiently transfected into BHK-21 cells. Subsequent to the transfection procedure (24 h), cells were fixed and stained as outlined under "Experimental Procedures." Images of cells representative of the entire population are shown. Red, anti-FLAG; blue, nucleus (DAPI). Cells were viewed on an inverted fluorescence microscope using a ϫ40 objective as described under "Experimental Procedures." After image capture, each image was deconvolved using the no neighbors method within Slidebook (Intelligent Imaging Innovations). C, Western immunoblot analysis of constructs used in this study. Cell lysates of transfected BHK-21 cells were subjected to electrophoresis on a 10% Tris-HCl SDS-polyacrylamide gel and then blotted onto nitrocellulose. Constructs were detected using an anti-FLAG and anti-His 6 polyclonal antibody. Expression of GFP protein was detected using a GFP polyclonal antibody. The dots represent the bands corresponding to construct names listed at the top of each lane. 5B) demonstrate that the amount of detectable extracellular Tax was reduced by a small amount, suggesting that brefeldin A at least partially blocked Tax secretion. This result was not unexpected based on the co-localization of Tax with the GC.
Tax Secretion Is Promoted after Incubation with PMA and Ionomycin-Release of proteins localized to the regulated secretory pathway is promoted by a variety of extracellular signals, also known as secretagogues. These signals vary in both their ability to induce secretion and the distinct regulated secretory pathway they induce. Two chemical compounds, PMA and ionomycin, have been shown to be effective secretagogues in a variety of cell types (42,43). Therefore, studies were designed to determine whether combined treatment with PMA and ionomycin induced the secretion of FTH 6 . After transfection with secretory plasmid constructs (24 h), cell culture medium was removed and replaced with medium containing PMA and ionomycin as well as protease inhibitors. After 24 h, the presence of FTH 6 was again assessed by ELISA. Treatment with PMA and ionomycin resulted in a small increase in the percentage of Tax released into the medium (Fig. 5B). Interestingly, the overall amount of FTH 6 detected in PMA and ionomycin cultures was more than those without treatment (ϳ200 ng more per culture). Since PMA and ionomycin induced overall protein production, more Tax was released in the supernatant, whereas the overall ratio of Tax in the supernatant as compared with the cells was the same.
Tax Secretion Is Not the Result of Increased Plasma Membrane Permeability-Tax has been associated with cellular dysfunction and apoptosis, and it has been generally thought that the primary source of extracellular Tax stemmed from a loss of cell membrane integrity instead of through the cellular secretory pathway. To determine whether the presence of Tax in the medium of Tax-transfected BHK-21 cells was the result of a loss in cell membrane integrity rather than release through the secretory pathway, two assays were performed. First, the LDH activity, a measurement of plasma membrane integrity loss, of each transfected cell culture was determined. Consequently, detection of LDH in Tax-transfected cells above that of mock-transfected cells would suggest that at least some extracellular Tax would be attributed to cell membrane integrity loss. Utilizing this method, it was determined that the level of LDH activity in FTH 6 transfected cultures was 1.42-fold over that of mock-transfected cultures (Fig. 6A). However, this level was similar to mock and vector-only controls. Furthermore, the level of LDH release ob-

FIG. 4. Full-length native Tax and FTH 6 is secreted from BHK-21 cells.
Top, BHK-21 cells were transiently transfected with either pIRES-EGFP or FTH 6 . After transfection (24 h), medium was removed and replaced with medium containing protease inhibitors. Subsequently (24 h), medium was removed and cleared of cells and cell debris by centrifugation. Then both secreted and intracellular FTH 6 was immunoprecipitated from the medium and cell lysate, respectively, using anti-FLAG-agarose beads. Immunoprecipitated proteins were subjected to electrophoresis on a 10% Tris-HCl SDSpolyacrylamide gel and then blotted onto nitrocellulose. His 6 tag-containing proteins were detected using a His 6 polyclonal antibody. Blots were stripped of antibody and reprobed with an anti-FLAG antibody. Purified FLAG-Tax-His 6 protein was also used as a positive control. Bottom, BHK-21 cells (1 ϫ 10 7 ) were seeded in a 100-mm culture dish and transfected with pCMV4, pCMV-Tax, and pCMV-FTH 6 as described under "Experimental Procedures." After 24 h, fresh medium containing protease inhibitors aprotinin and leupeptin (1.0 g/ml each; Sigma) was added. After an additional 24 h, Tax was detected in both medium and cell lysate by Western blotting using anti-Tax monoclonal antibody (TAB 170). Both native Tax and FTH6 were detected at the expected size by the Tax-specific monoclonal antibody. Each experiment was performed in triplicate and repeated at least twice.
tained with the experimental sample was only a fraction of the LDH value determined by total lysis of each positive control cell culture (data not shown). Cumulatively, these observations have suggested that general cellular lysis was not the primary route of Tax release into the cell medium.
A more specific assay was also utilized to visually determine the effect of Tax on the survival of BHK-21 cells and the release of Tax into the extracellular environment via loss of cell membrane integrity. The Vybrant apoptosis assay kit 7 (Molecular Probes) was utilized to stain for either apoptotic or necrotic cells (Fig. 6B). The compounds camptothecin and ebselen were utilized on untransfected BHK-21 cells as positive controls to induce apoptosis and necrosis, respectively. Several negative control reactions, drug vehicle (ETOH and Me 2 SO), cells alone, or cells transfected with a blank vector construct (pUC18) were utilized to determine whether any portion of the assay procedure induced either apoptosis or necrosis. Indeed, none of the negative control conditions promoted levels of apoptosis or necrosis above background. Two Tax constructs, pCMV-Tax and FTH 6 , were utilized to determine whether apoptosis or necrosis could be a contributing factor to the presence of extracellular Tax. The pCMV-Tax expression plasmid was utilized to determine whether the presence of a FLAG or His 6 tag in FTH 6 altered the capability of Tax to induce apoptosis or necrosis. Whereas Tax has previously been shown to induce apoptosis in some cell populations, utilization of the assay described above resulted in very few cells positive for either apoptosis or necrosis. Again, these results demonstrate that the presence of extracellular Tax is not the result of the loss of cell membrane integrity. DISCUSSION The role of extracellular Tax in the progression of HTLV-Iassociated pathogenesis has not been a major focus of research in recent years. The primary reason for this may be the belief that Tax does not enter the extracellular environment to any biologically relevant level. However, several studies have reported that extracellular Tax may impact a number of cellular functions. For example, Tax has been shown to induce the production and release of TNF-␣, a potent cellular cytokine that has been demonstrated to have effects on cells similar to those present in pathogenic lesions observed in HAM/TSP (17). Additionally, both neurons and microglia have been demon- FIG. 5. Quantitative determination of FLAG-Tax-His 6 secretion from BHK-21 and 293T cells by ELISA. A, BHK-21 or 293T cells were transiently transfected with secretion constructs. After transfection (24 h), medium was removed and replaced with medium containing protease inhibitors. Subsequently (after 24 h), medium and cell lysates were utilized in the ELISA-based secretion assay as outlined under "Experimental Procedures." Results are presented as the percentage of total Tax secreted into the medium. B, BHK-21 cells utilized in ELISA secretion assays were also exposed to either brefeldin A (BFA), a known inhibitor of the secretory process, or PMA and ionomycin, known inducers of some regulated secretory vesicles. BHK-21 cells were transiently transfected with FLAG-Tax-His 6 or FLAG-Tax-COPII-His 6 . Tax-COPII contains a DXE to AXA mutation at amino acids 330 -332. The DXE motif was previously described as a domain important in protein localization to the secretory pathway. After 24 h, cell medium was replaced with medium containing protease inhibitors, protease inhibitors plus brefeldin A, or protease inhibitors plus PMA and ionomycin. Cells were incubated for an additional 24 h, at which time the media and cell lysates were utilized in the ELISA-based secretion assay as outlined under "Experimental Procedures." Results are represented as either a percentage of total FLAG-Tax-His 6 released or as the average total amount (ng) of FLAG-Tax-His 6 released per 2 ml of culture. Assays were repeated for a total of three times for each cell type. strated to release TNF-␣ in response to extracellular Tax (17,18). Each of these cell types coexists with oligodendrocytes, cells that form the protective myelin sheath surrounding CNS neurons. Upon exposure to TNF-␣, oligodendrocytes enter a process of demyelination, an effect subsequently followed by neuronal defects and abnormalities similar to those observed in individuals with HAM/TSP. Thus, the study of the effects of extracellular Tax is likely to be of great importance with respect to understanding the pathogenesis of HTLV-I. As a result, studying the process by which Tax exits the cell and migrates into the extracellular environment will improve our understanding of the role of extracellular Tax in neurologic disease.
The results presented herein begin to solidify a link between HTLV-I-infected cells, the release of Tax into the extracellular environment, the reported effects extracellular Tax has on cells, and the progression of HAM/TSP. First, these results substantiate previous observations suggesting that Tax is released into the extracellular environment by a process other than apoptosis or lysis of the cell. This process may occur through a formal cellular secretory pathway. This was evident based on several results including the co-localization of Tax with cellular organelles associated with the cellular secretory pathway including, but not limited to, the ER and the GC. Second, the real time movement of Tax-containing punctate structures within the cytoplasm was congruent with the move-FIG. 6. Tax-transfected cells do not show signs of loss of cell membrane permeability by either necrosis or apoptosis. A, medium from transfected cells, BHK-21 or 293T, was assayed for LDH activity levels. The presence of LDH in the culture medium is a known marker for necrosis and cell membrane integrity loss. Results are represented as -fold over the LDH activity of mock-transfected cells. B, the detection of apoptosis or necrosis in BHK-21 Tax-transfected cultures was also determined by utilizing the Vybrant apoptosis detection kit 7 (Molecular Probes) as described under "Experimental Procedures." Necrosis was detected utilizing propidium iodide, and apoptosis was detected utilizing YO-PRO-1 and viewed with a fluorescence microscope fitted with TRITC and fluorescein isothiocyanate detection cubes. Ebselen (50 M) and camptothecin (10 M) were utilized as positive controls. ment of microtubule-associated secretory vesicles. Third, the Tax molecules released from these cells and detected using our system were full-length proteins, demonstrating that Tax enters the secretory pathway utilizing a leaderless system. Fourth, the inhibition of Tax secretion using brefeldin A supports observations of Tax localization to both the ER and GC, since brefeldin A blocks migration of proteins from the ER to the GC. Fifth, the increase in Tax secretion observed after treatment with PMA and ionomycin indicated that at least some Tax resided in the regulated secretory pathway.
Two final results noted here will assist the study of Tax secretion as well as provide a source of Tax made from mammalian cells instead of the bacterial source now used by many investigators. First, an ELISA-based system for detection of full-length Tax released from the cell has been established. Second, a new method for the purification of full-length Tax made within the context of a eukaryotic system has been identified. Traditionally, biochemical analysis of Tax has utilized Tax purified from bacterial cell cultures. The system presented herein utilized mammalian cells for production of Tax protein and has several advantages over purification from bacterial cultures. For example, mammalian cells are able to add secondary modifications to proteins that are not achieved in a bacterial background.
Whereas the CD4 ϩ T lymphocyte is the primary cell targeted by HTLV-I, the role of other HTLV-I-infected cell types has become an area of great interest. For example, HTLV-I-infected astrocytes have been demonstrated to produce and release many cytokines that have dramatic effects on the immediate cellular environment (19 -21, 34). These results are congruent with observations of spinal cord lesions in patients with late stage HAM/TSP. These lesions are the result of several consequences including, but not limited to, immune system interactions, molecular mimicry, and apoptosis. Additionally, it is possible that extracellular Tax is taken up by cells through a receptor-mediated process, similar to that suggested for HIV-1 Tat. If this is the case, extracellular Tax may be able to induce intracellular signaling pathways and gene transcription in preparation for infection by whole virus. A combination of cutting edge techniques including proteomics and mass spectrometry would now make such a study feasible. Thus, the release of cytokines and Tax, especially from HTLV-I-infected astrocytes, has tremendous effects on neighboring cells and the progression of HAM/TSP. With these pathogenic mechanisms in mind, the study of Tax secretion from infected astrocytes, CNS-infiltrating HTLV-I ϩ T cells, or a yet to be identified cell type becomes important in the elucidation of the exact mechanisms of HAM/TSP progression.