The Glutamine-rich Region of the HIV-1 Tat Protein Is Involved in T-cell Apoptosis*

Human immunodeficiency virus (HIV) infection and the progression to AIDS are characterized by the depletion of CD4 (cid:1) T-cells. HIV-1 infection leads to apoptosis of uninfected bystander cells and the direct killing of HIV-infected cells. This is mediated, in part, by the HIV-1 Tat protein, which is secreted by virally infected cells and taken up by uninfected cells. We chemically synthesized two 86-residue subtype D Tat proteins, Ug05RP and Ug11LTS, from two Ugandan patients who were clini-cally categorized as either rapid progressor or long-term survivor, with non-conservative mutations located essentially in the glutamine-rich region. Structural het-erogeneities were revealed by CD, which translate into differing trans -activational and apoptotic effects. CD data analysis and molecular modeling indicated that the short (cid:2) -helix observed in subtype D Tat proteins from rapid progressor patients such as Tat Mal and Tat Ug05RP is not present in Ug11LTS. We show that Tat Ug05RP is more efficient than Tat Ug11LTS in its trans activational role and in inducing apoptosis in binding tubulin via the mitochondrial pathway. The glutamine-rich region of Tat appears to be involved in the Tat-mediated apoptosis of T-cells. a structure-function between two subtype D Tat proteins from rates of progression. Structural heterogene-ities are revealed by CD, which translate into differing trans activational and apoptotic effects. CD data analysis and molecular modeling indicate that the short (cid:1) -helix observed in subtype D Tat proteins from RP patients such as and Ug05RP is not present in Ug11LTS. We show that LTS Tat is less efficient than RP Tat in trans -activation and in inducing apoptosis in T-cells. The glutamine-rich region (region V) of Tat appears to be involved in the Tat-mediated apoptosis CCTGTTAAA-3 (cid:6) (antisense); and (cid:3) 2 -microglobulin, 5 (cid:6) -CCGACATTGA- AGTTGACTTAC-3 (cid:6) (sense) and 5 (cid:6) -ATCTTCAAACCTCCATGATG-3 (cid:6) (antisense). PCR was performed according to the manufacturer’s in- structions with 3 m M MgCl 2 , 0.5 (cid:2) M each primer, 5 (cid:2) l of sample, and 1-fold LightCycler FastStart DNA Master SYBR Green I mixture in a total volume of 20 (cid:2) l. The reaction mixture was initially incubated at 95 °C for 10 min to denature the cDNA. Amplification was performed for 45 cycles with the following cycle parameters: 10 s of denaturation at 95 °C, 10 s of primer annealing at 65 °C, and 15 s of fragment elongation at 72 °C. Amplified reverse transcription-PCR products were detected on-line via intercalation of the fluorescent dye. The Light- Cycler was then programmed to carry out a melting cycle to verify the specificity of the desired product. This consisted of a denaturing step at 95 °C, after which the LightCycler cooled to 65 °C and then increased the temperature at a rate of 0.2 °C/s until the temperature reached 95 °C. Fluorescence was acquired every 0.1 °C. The conversion of the melting curves into melting peaks (plot of the negative derivative of fluorescence versus temperature) allowed identification of the specificity of the reaction. Quantification and the melting curve were analyzed with the LightCycler analysis software. All results are expressed as the ratio of the copy number of the target gene to the copy number of (cid:3) 2 -microglobulin and were normalized so that FasL expression from Ug05RP-treated cells equals 1.00. in neutral aqueous buffer with an extinction coefficient of (cid:4) 275 nm (cid:7) 1.07 liter/g/cm. All experiments were done in PG buffer (20 m M sodium phosphate buffer and 0.1 m M GTP (pH 7.0)) at 25 °C. Fluorescence measurements and uncorrected spectra were obtained with a PerkinElmer Life Sciences Model 50 luminescence spectrometer with slit widths of 5 and 10 nm monitored by an IBM PS2 computer. Fluorescence spectra were obtained in PG buffer using 0.2 (excitation direction) (cid:4) 1-cm cells (Hellma) thermostatted at 25 °C by circulating water from an external water bath. Quenching of the intrin-sic tubulin fluorescence signal by Tat was employed to estimate the binding parameters. Tubulin (2 (cid:2) M ) was titrated by various concentrations of Tat. Fluorescence measurements were performed with an ex- citation wavelength of 295 nm to specifically excite the tryptophan residues. At the concentration used, spectra gave no appreciable inner filter effect ( A (cid:2) 0.05). The changes in fluorescence ( (cid:8) F ) at 338 nm between the Tat-tubulin complex and Tat ( (cid:8) F (cid:7) F complex (cid:5) F Tat ) were plotted versus total Tat concentrations. The curves ( (cid:8) F corr ) were in-verted and fitted to the saturation curve equation by nonlinear least- squares regression analysis using the following equation: (cid:8) F corr (cid:7) (cid:8) F max [Tat f ]/ K d (cid:3) [Tat f ], where [Tat f ] is the free Tat concentration and (cid:8) F max is the maximum difference in fluorescence. The concentrations of bound Tat ([Tat b ]) and free Tat and the dissociation constant ( K d ) were determined using the following equation: [Tat b ] (cid:7) 1/2(([Tat 0 (cid:3) [P 0 (cid:3) K d ) (cid:5)

Following infection with HIV-1, 1 the rate of clinical disease progression varies enormously between individuals. Some studies have reported that disease progression is more rapid in Africa (1-3), but others contest this view (4,5). In Uganda, a study has demonstrated that the time from seroconversion to death is very similar to that reported for patients in industri-alized countries prior to the introduction of anti-retroviral therapy (6). Those at the extremes of clinical progression have been termed rapid progressors (RP) and long-term survivors (LTS). Definitions of what constitutes RP and LTS vary between studies and cohorts. In Uganda, we have defined an LTS as a person who has a CD4 count Ͼ500 cells/mm 3 6 or more years postinfection, whereas an RP is a person who died or progressed to a CD4 count Ͻ200 cells/mm 3 within 5 years of seroconversion.
Many factors such as host susceptibility and immune function (7) and health care and co-infections (6) as well as factors relating to the viral strain (8) may affect the rate of progression to AIDS. The HIV-1 Tat (trans-acting transcriptional activator) protein is an important factor in the manifestation of immune dysfunction in many HIV-1-infected individuals before the substantial loss of CD4 ϩ T-cells (9). During HIV infection, transcripts of Tat are found before integration (10). Tat is a short HIV-encoded trans-activating regulatory protein (86 -106 amino acids) whose primary role is in regulating productive and processive transcription from the HIV-1 long terminal repeat (11)(12)(13)(14)(15). It is encoded by two exons and has six different regions, each having specific biochemical and functional characteristics (16). Region II (residues 22-37) has seven conserved cysteines, which are required for the trans-activational activity of Tat. Region III has a highly conserved Phe 38 and, combined with the end of Region II, is thought to be important in tubulin binding and apoptosis (17). Region IV (residues 49 -59) is rich in basic residues and has a highly conserved sequence that is important for trans-activation and binding to the nucleotide trans-activator-responsive region (TAR) sequence (11) and in the uptake of Tat by cells (18). Region V (residues 60 -72) is a glutamine-rich region that forms an ␣-helix in the subtype D Tat NMR structure (19). Tat-TAR binding induces the formation of an ␣-helix in region V (11). Region VI is the C terminus and is coded by the second exon. It has been proposed that the second coding exon of Tat is a functionally constrained epitope in which mutations produce a virus that replicates poorly in vivo (20). A similar folding pattern but with local structural variations is observed between subtype B and D variants (19 -21).
Despite the lack of a signal sequence, Tat is released by infected cells and is found in detectable levels (0.01-0.1 nM) in the culture supernatants of cells infected with HIV-1 (22)(23)(24)(25) and in the sera of HIV-1-infected patients (26). Tat is efficiently taken up by a variety of cells (23,25,(27)(28)(29) and may enter T-cells by clathrin-mediated endocytosis (30). Extracellular Tat has many functions, which are thought to play a major role in enabling HIV to escape immune surveillance and to act as a viral toxin in contributing to AIDS pa-thology (17,(31)(32)(33). Extracellular Tat is able to regulate cytokine gene expression and immune cell hyperactivation (34) and to stimulate the growth of Kaposi's sarcoma cells (22,35). Extracellular Tat, secreted from infected cells, can induce apoptosis in neighboring uninfected cells (24, 36 -39). Recently, the rate of apoptosis has been correlated with non-progression to AIDS (40). Extracellular Tat can up-regulate the expression of Fas ligand (FasL) mRNA in macrophages and increase the susceptibility of Tat-expressing cells to CD4 cross-linking-induced death (24,41,42). Apoptosis contributes to the massive depletion of CD4 ϩ T-cells and consequently to the loss of immune competence during HIV-1 infection (43,44). Other effects of extracellular Tat include inhibition of the proliferation of uninfected T-cells (24,37,45), possibly by repression of major histocompatibility complex class I transcription (46); regulation of the expression of the HIV-1 coreceptor CXC chemokine receptor-4 on T-lymphocytes (47); and repression of the expression of manganese-superoxide dismutase (48).
Here, we compare a structure-function relationship between two subtype D Tat proteins from patients in Uganda with different rates of disease progression. Structural heterogeneities are revealed by CD, which translate into differing transactivational and apoptotic effects. CD data analysis and molecular modeling indicate that the short ␣-helix observed in subtype D Tat proteins from RP patients such as Tat Mal and Tat Ug05RP is not present in Ug11LTS. We show that LTS Tat is less efficient than RP Tat in trans-activation and in inducing apoptosis in T-cells. The glutamine-rich region (region V) of Tat appears to be involved in the Tat-mediated apoptosis of T-cells via mitochondria.

Tat Sequences-The Rural Clinical Cohort was established in South
West Uganda by the Medical Research Council (United Kingdom) Programme on AIDS in Uganda in 1990 (49). It was initially composed of prevalent cases of HIV-1 identified by a serosurvey of a larger General Population Cohort. The General Population Cohort continues to be annually serosurveyed, and any persons who have seroconverted to HIV-1-positive are invited to join as incident cases. Patients are invited to attend the study clinic every 3 months, at which time they provide a detailed medical history and undergo a full physical examination and laboratory investigation if required. In this way, the mean time from seroconversion to death has been estimated in this population to be 9.8 years (6). Tat sequences from directly PCR-amplified proviral DNA were obtained from two individuals from this cohort. The estimated date of seroconversion for Ug05RP was July 1993, and death occurred 4.5 years later in January 1998. This compares with Ug11LTS, for which the estimated date of seroconversion was November 1990 and on March 2004 was still at World Heath Organization Stage 2 with a CD4 count in excess of 400 cells/mm 3 .
Protein Synthesis, Purification, and Characterization-The Tat proteins were assembled according to the method of Barany and Merrifield (50) on (4-(hydroxymethyl)phenoxymethyl)copolystyrene (1% divinylbenzene)-preloaded resin (0.5-0.65 mmol; PerkinElmer Life Sciences) on an ABI 433A automated synthesizer (PerkinElmer Life Sciences) as described previously (51). Purification was carried out using a Beckman high-pressure liquid chromatography (HPLC) apparatus with a Beckman C 8 reverse-phase column (10 ϫ 150 mm). Buffer A was water with 0.1% trifluoroacetic acid, and Buffer B was acetonitrile (Merck) with 0.1% trifluoroacetic acid. The gradient was buffer B from 15 to 35% over 40 min with a flow rate of 2 ml/min. HPLC analysis was carried out using a Merck Chromolith TM Performance RP-8e (4.6 ϫ 100 mm) with similar buffers but using a gradient from 10 to 50% over 15 min with a flow rate of 1.8 ml/min. Amino acid analyses were performed on a Beckman Model 6300 analyzer, and mass spectrometry was carried out using an Ettan TM MALDI-ToF Pro (Amersham Biosciences).
trans-Activation with HIV Long Terminal Repeat-transfected Cells-The trans-activation function of the HIV-1 Tat variants was tested using the human HeLa P4 cells as described previously (51). Cells (2 ϫ 10 5 /well) were incubated in 24-well plates (Falcon) at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum (Invitrogen) and 100 g/ml neomycin. After 24 h, cells were washed with phosphate-buffered saline (PBS). Tat protein was dissolved in phosphate buffer at pH 6 (to avoid precipitation that occurs at neutral pH) and was directly mixed with Dulbecco's modified Eagle's medium supplemented with 0.01% protamine (Sigma) and 0.1% bovine serum albumin (Sigma) and added to the cells. Two concentrations of Tat (100 and 50 nM) were tested to verify that the level of ␤-galactosidase was dose-dependent regarding extracellular Tat. After 16 h at 37°C and 5% CO 2 , cells were washed with PBS and lysed; the ␤-galactosidase content was measured with a commercial antigen capture enzyme-linked immunosorbent assay (Roche Applied Science); and absorbance values (B) were measured at 405 nm. Results were normalized using Bradford reagent (Sigma). B 0 corresponds to the background ␤-galactosidase expressed by HeLa cells in Dulbecco's modified Eagle's medium supplemented with 0.01% protamine and 0.1% bovine serum albumin without Tat.
Apoptosis-The different effects of Tat-induced apoptosis by Ug05RP and Ug11LTS were carried out using Jurkat T-cells. Cells were cultivated in RPMI 1640 medium with Ultraglutamine (Cambrex, East Rutherford, NJ) supplemented with 10% (v/v) heat-inactivated fetal calf serum, 100 g/ml streptomycin, and 100 units/ml penicillin. Cells were incubated in 24-well flat-bottomed plates (Falcon) at 37°C with the indicated amounts of Tat as described in the figure legends. To perform cell cycle analysis, cells were harvested at 19 h, fixed in cold methanol, and incubated with 120 g/ml propidium iodide (Sigma) immediately before analysis. DNA content was measured by flow cytometry using a BD Biosciences FACScan. To detect apoptosis by fluorescein isothiocyanate (FITC)-conjugated annexin V staining, cells were harvested at 19 h, and 5 ϫ 10 5 live cells were incubated with 4 g/ml FITC-conjugated annexin V and 5 g/ml propidium iodide for 15 min (annexin V-FITC kit, Bender Medsystems, Vienna, Austria) followed immediately by flow cytometry using the FACScan. Cytogram analysis was performed with Cell Quest Pro® software (BD Biosciences) as described previously (52). Cells were distributed into three populations: viable cells (low FITC-conjugated annexin V labeling and low propidium iodide, R4), early apoptotic cells (high FITC-conjugated annexin V labeling and low propidium iodide labeling, R2), and cells that had lost membrane integrity as a result of very late stage apoptosis (high FITCconjugated annexin V and high propidium iodide labeling, R1). Thus, cells in quadrants R1 and R2 were apoptotic cells.
Western Blot Analysis-After treatment with Tat, the culture medium was removed, and cells were washed with cold PBS. HeLa P4 cells were lysed for 30 min at 4°C in radioimmunoprecipitation assay buffer (20 mM Tris-HCl (pH 8.0), 200 mM NaCl, 1% Triton X-100, 1 mM EDTA, and protease inhibitors) without bovine serum albumin. Lysates were then centrifuged at 4000 ϫ g for 15 min at 4°C. We obtained a cytoplasmic supernatant and a pellet corresponding to the nuclear extract.
Jurkat T-cells were lysed for 1 min in digitonin, and cell lysates were centrifuged at 15,000 ϫ g for 15 min at 4°C. All supernatants were stored at Ϫ80°C. Samples (40 g) were heated at 100°C for 5 min in Laemmli sample buffer containing reducing agent, separated by 15% SDS-PAGE, and transferred onto a nitrocellulose membrane (53). Nonspecific sites were blocked by a 1-h incubation at room temperature with PBS, 0.1% Tween 20, and 10% dried nonfat milk (10% MPBS). Membranes were then incubated with rabbit anti-Tat sera at 1:500 dilution in 10% MPBS, mouse anti-cytochrome c antibody (1:500 dilution in 5% MPBS; Pharmingen), or mouse anti-porin antibody (1:1000 dilution in 5% MPBS; Calbiochem; the levels of the mitochondrial outer membrane porin were systematically measured to ensure that lysis of the mitochondria did not occur during Jurkat T-cell lysis); washed four times; and incubated with a horseradish peroxidase-labeled anti-rabbit (Sigma) or anti-mouse (Amersham Biosciences) secondary antibody (1: 2000 dilution) for 1 h at room temperature in 5% MPBS. The bound horseradish peroxidase was revealed with H 2 O 2 and 0.1% diaminobenzidine tetrahydrochloride (Sigma) in PBS as substrate. Bands were analyzed using NIH Image software.
Real-time PCR-The different effects of Tat-induced up-regulation of FasL by Ug05RP and Ug11LTS were carried out using Jurkat T-cells. Cells were cultivated as described for the apoptosis assay. Cells (2 ϫ 10 5 /ml) were incubated in 24-well flat-bottomed plates at 37°C with 2 M Tat for 19 h. Cells were harvested, washed with PBS, and stored at Ϫ80°C. Total cellular RNA was prepared with a total RNA isolation kit (Macherey-Nagel, Hoerdt, France) in accordance with the manufacturer's directions. 1 g of total RNA was used for reverse transcription with random primers. FasL mRNA expression in relation to ␤ 2 -microglobulin expression (internal standard) was determined using the LightCycler system and the FastStart DNA Master SYBR Green I kit (Roche Diagnostics). Human-specific primers of the two genes were then designed based on the sequence of each gene as follows: FasL, 5Ј-GTAGGATTGGGCCTGGGGAT-3Ј (sense) and 5Ј-AGTTGGACTTG-CCTGTTAAA-3Ј (antisense); and ␤ 2 -microglobulin, 5Ј-CCGACATTGA-AGTTGACTTAC-3Ј (sense) and 5Ј-ATCTTCAAACCTCCATGATG-3Ј (antisense). PCR was performed according to the manufacturer's instructions with 3 mM MgCl 2 , 0.5 M each primer, 5 l of sample, and 1-fold LightCycler FastStart DNA Master SYBR Green I mixture in a total volume of 20 l. The reaction mixture was initially incubated at 95°C for 10 min to denature the cDNA. Amplification was performed for 45 cycles with the following cycle parameters: 10 s of denaturation at 95°C, 10 s of primer annealing at 65°C, and 15 s of fragment elongation at 72°C. Amplified reverse transcription-PCR products were detected on-line via intercalation of the fluorescent dye. The Light-Cycler was then programmed to carry out a melting cycle to verify the specificity of the desired product. This consisted of a denaturing step at 95°C, after which the LightCycler cooled to 65°C and then increased the temperature at a rate of 0.2°C/s until the temperature reached 95°C. Fluorescence was acquired every 0.1°C. The conversion of the melting curves into melting peaks (plot of the negative derivative of fluorescence versus temperature) allowed identification of the specificity of the reaction. Quantification and the melting curve were analyzed with the LightCycler analysis software. All results are expressed as the ratio of the copy number of the target gene to the copy number of ␤ 2 -microglobulin and were normalized so that FasL expression from Ug05RP-treated cells equals 1.00.
Tat-Tubulin Interaction-Lamb brain tubulin was purified by ammonium sulfate fractionation and ion-exchange chromatography, stored in liquid nitrogen, and prepared for use as described (54 -57). Its concentration was determined in SDS in neutral aqueous buffer with an extinction coefficient of ⑀ 275 nm ϭ 1.07 liter/g/cm. All experiments were done in PG buffer (20 mM sodium phosphate buffer and 0.1 mM GTP (pH 7.0)) at 25°C. Fluorescence measurements and uncorrected spectra were obtained with a PerkinElmer Life Sciences Model 50 luminescence spectrometer with slit widths of 5 and 10 nm monitored by an IBM PS2 computer. Fluorescence spectra were obtained in PG buffer using 0.2 (excitation direction) ϫ 1-cm cells (Hellma) thermostatted at 25°C by circulating water from an external water bath. Quenching of the intrinsic tubulin fluorescence signal by Tat was employed to estimate the binding parameters. Tubulin (2 M) was titrated by various concentrations of Tat. Fluorescence measurements were performed with an excitation wavelength of 295 nm to specifically excite the tryptophan residues. At the concentration used, spectra gave no appreciable inner filter effect (A Ͻ 0.05). The changes in fluorescence (⌬F) at 338 nm between the Tat-tubulin complex and Tat (⌬F ϭ F complex Ϫ F Tat ) were plotted versus total Tat concentrations. The curves (⌬F corr ) were inverted and fitted to the saturation curve equation by nonlinear leastsquares regression analysis using the following equation: With these values, bound and free Tat concentrations are calculated, and then the nonlinear least-squares regression analyses are executed. The initial set is corrected in the next step by a Newton-Gauss procedure. This iterative procedure is continued until the minimum sum of squared deviations between experimental and calculated values of ⌬F corr is obtained.
Circular Dichroism-CD spectra were measured with a 50-m path length from 260 to 178 nm on a Jasco J-810 spectropolarimeter. Data were collected at 0.5-nm intervals using a step autoresponse procedure (Jasco). CD spectra are presented as ⌬⑀ per amide. The samples were prepared in 20 mM phosphate buffer (pH 4.5). The protein concentration was 1 mg/ml. The CD data were analyzed to determine the secondary structure content according to the method of Manavalan and Johnson (58) using a set of 32 reference proteins.
Molecular Modeling-Models were built using the Insight II 2000 package including Discover, Biopolymer, and Homology software (Accelrys, Inc., San Diego, CA) running on a Silicon Graphics O2 workstation. The Tat Mal structure was obtained from NMR studies (Protein Data Bank code 1K5K) (19). Models were optimized with CVFF (consistent valence force field) in terms of the internal energies using the van der Waals energy to monitor each step of the model. Minimization was performed with steepest descent and conjugate gradient algorithms.
Statistical Analysis-Statistical analysis was performed using a twotailed Student's t test unless otherwise stated. p values Յ0.05 were considered to be statistically significant.

Tat from an RP Patient Has a Superior trans-Activational
Potential Compared with Tat from an LTS Patient-In this study, we have used two subtype D Tat sequences that were derived directly from identified patients in Uganda. Both sequences come from patients with abnormal disease progression. Previous data have shown a correlation between progression rates and Tat activity (21). Ug05RP came from a patient that showed rapid progression, and Ug11LTS came from a patient described as an LTS. These two Tat variants show 91% homology (Fig. 1). It has already been demonstrated that Tat can be taken up by uninfected bystander cells (23,29). We tested the ability of these Tat variants to cross the cell membrane and to trans-activate the HIV-1 long terminal repeat in transfected HeLa P4 cells (Fig. 2). Initially, HeLa P4 cells were incubated with 1 M extracellular synthetic Tat that was added to the culture medium. This is the minimum amount of Tat that can be used because of the sensitivity of Western blotting. After immunoblotting, we observed that both Ug11LTS and Ug05RP were taken up by the transfected cells, that they were found predominantly in the cytoplasmic region (data not shown), and that only a small portion reached the nucleus to activate transcription ( Fig. 2A), which agrees with previous data (23). Importantly, both Ug11LTS and Ug05RP demonstrated a similar efficiency in their cytoplasmic and nuclear uptake. Thus, the mutations in the two proteins do not play a role in the cellular uptake of Tat. However, we observed a significant difference between the trans-activational potential of these two proteins at both 100 nM (p ϭ 0.0033) and 50 nM (p ϭ 0.0189) (Fig. 2B). Ug05RP had twice the trans-activational potential compared with Ug11LTS at 100 nM.
Ug11LTS Induces Apoptosis at an Inferior Rate Compared with Ug05RP-It is known that the addition of extracellular recombinant Tat Bru and Tat HXB2, both subtype B, induces apoptosis and increases sensitivity to apoptotic signals in both primary CD4 ϩ T-cells and T-cell lines, thus contributing in part to the progressive loss of T-cells associated with HIV-1 infection (24,37,61). However, little is known about the link between the effect of different Tat proteins and disease progression. We investigated the ability of our synthetic subtype D Tat proteins to induce apoptosis. We used an assay in which Jurkat T-cells were incubated with different Tat proteins at different  (59), and Tat HXB2 came from a strain isolated in France (60) and is the isolate that is used most frequently in Tat and HIV research, although it is a subtype B strain. The boxes indicate the location of the six Tat regions.
concentrations. The apoptotic effects were evaluated by flow cytometry after labeling the cells with propidium iodide and FITC-conjugated annexin V. Treatment of cells with 10 M Ug05RP (but not Ug11LTS) caused a 5-fold increase in the amount of hypodiploid nuclei (cells in ϽG 1 phase) (Fig. 3A). We verified this result for apoptosis by annexin staining (Fig. 3B). At 2 M Tat, there was a significant difference between the levels of apoptosis, with Ug11LTS inducing almost 41% less apoptosis than Ug05RP (p ϭ 0.0307). At 0.2 M Tat, the difference was not significant (p ϭ 0.1965). Treatment of Jurkat T-cells with Ug05RP (but not Ug11LTS) also caused the release of cytochrome c from the mitochondria (Fig. 3C), suggesting that the death signals initiated by Ug05RP transmit through the mitochondrial pathway.
Ug05RP Induces Greater Up-regulation of FasL mRNA Compared with Ug11LTS-Previous reports have suggested that Tat may induce apoptosis by up-regulating the expression of FasL (24). To investigate the contribution of this pathway in Tat-induced apoptosis, quantitative analysis of FasL mRNA in Jurkat T-cells treated with 2 M Ug05RP and Ug11LTS was carried out. Treatment of cells with Ug05RP induced twice the amount of FasL mRNA expression compared with treatment with Ug11LTS (p ϭ 0.004 with the Mann-Whitney test) (Fig. 3D).
Tat-Tubulin Interactions-It has been shown that Tat interacts with tubulin and microtubules in the cytoplasm of nonmitotic cells and that this interaction is important in Tatinduced apoptosis of Jurkat T-cells via the mitochondrial pathway that is facilitated by the pro-apoptotic Bcl-2 family member Bim (17). Both Tat and tubulin have tryptophan and tyrosine residues. During complex formation, the fluorescence emission spectrum can either shift in the wavelength of maxi-mum fluorescence emission or shift in fluorescence intensity. These shifts can therefore be used to evaluate the association constant. Fig. 4A shows a decrease in fluorescence intensity without any special maximum shift in the presence of 1 and 8 M Ug11LTS. This confirmed the existence of an Ug11LTStubulin complex and provides a convenient means to determine the affinity of the interaction between tubulin and Ug11LTS. The intensity changes at 338 nm were plotted versus Ug11LTS concentrations and analyzed as described under "Materials and Methods" (Fig. 4A, inset). We found, in three independent experiments, that the K d was (4 Ϯ 2) ϫ 10 Ϫ7 M. With high concentrations of Ug11LTS (Ͼ12 M) (Fig. 4A, O ⅐ O), we observed a shift in the maximum fluorescence intensity and an increase in fluorescence emission at wavelengths Ͼ400 nm due to turbidity of the samples. Turbidity was not seen with Ug11LTS alone in PG buffer (data not shown), so we deduce that aggregation of the Tat-tubulin complex was responsible for this effect. In the case of Ug05RP, the turbidity of the Tattubulin complex was observed at 4 M (Fig. 4B). However, as with Ug11LTS, this was not seen with Ug05RP in PG buffer alone. We therefore conclude that aggregation of the Tat-tubulin complex was responsible for this effect. However, as this aggregation occurred at lower concentrations, we were unable to calculate the K d , but we could deduce that the K d for Ug05RP was smaller than that for Ug11LTS.
Structural Studies-Ug05RP and Ug11LTS share 87% sequence homology with Tat Mal, the two-dimensional NMR structure of which has already been defined (19), and 91% sequence homology between themselves. We compared the CD spectra of Tat Mal with those of the Ugandan Tat variants (Fig.  5). CD measurements were made between 178 and 260 nm, corresponding to the -* and n-* transitions of the amide chromophore located in polypeptide chains (62). The three spectra are characterized by negative bands at 200 nm, which could be due to non-organized structures and/or ␤-turns (62). The low intensity of the negative bands at 200 nm suggests that the three proteins are not random-coil but rather structured with ␤-turns as the main secondary structures. This is in agreement with the Tat Mal structure determined by two-dimensional NMR (19). Tat Ug05RP and Tat Mal have similar CD spectra, whereas differences are observed in the CD spectrum of Tat Ug11LTS (Fig. 5). This observation suggests that there is a difference in the secondary structure between the LTS Tat variant and the two RP Tat variants. This interpretation is confirmed by the results of the CD data analysis shown in Table I ( 58). Interestingly, we found that there is just 5% ␣-helix structure (three to four residues) in both Tat Mal and Tal Ug05RP, but not in Tat Ug11LTS. Molecular modeling of Ug05RP using the NMR structure of Tat Mal (19) showed that it is possible to obtain a structure similar to that of Tat Mal despite the mutations in these two proteins, in agreement with the CD spectra (Fig. 5). This indicates that the mutations in Tat Mal and Tat Ug05RP do not induce structural changes (Fig. 6). However, the molecular modeling of Tat Ug11LTS showed that the other mutations could induce a structural heterogeneity in region V (Fig. 6, light blue), the location of the ␣-helix in Tat Mal and Tat Ug05RP (Fig. 6). The two mutations Q63H and T64A in Ug11LTS appear to distort the ␣-helix structure. Overall, the two models are not markedly different (Fig. 6), and both adopt a folding comparable with that of Tat Mal (19). DISCUSSION In this study, we found a significant difference in the apoptotic and trans-activational activities of HIV-1 Tat proteins from RP and LTS patients. The CD spectra (Fig. 5) and molecular modeling (Fig. 6) indicated that much of the sequence is conserved to maintain structure, and the few mutations in Tat Ug05RP and Tat Ug11LTS (Fig. 1) therefore indicate the Tat regions directly involved in Tat-induced apoptosis. Interestingly, these mutations are not located in the usual functional Tat regions, i.e. regions II and IV (9). Moreover, only four mutations can explain the different activities of Tat Ug05RP and Tat Ug11LTS: R19S, A58S, Q63H, and A64T (Fig. 1). Mutation A58S in Ug11LTS is interesting because it is in agreement with a report indicating that mutation A58T is linked to LTS (63). Another study showed that when Tat Bru peptide-(47-72) binds to TAR, it adopts an ␣-helix structure located at the C terminus (11). Molecular modeling of Tat Ug05RP and Tat Ug11LTS showed that it is possible to obtain an ␣-helix similar to that of Tat Mal with Tat Ug05RP, both of which are RP. However, the restriction of the rotation freedom of and angles due to His 63 and Thr 64 in Ug11LTS reduces the ability to form the ␣-helix required for Tat-TAR binding, which is necessary for trans-activational ability (11).
Previous studies on Tat-induced apoptosis of Jurkat T-cells have shown that mutation of Val 36 , Cys 37 , Phe 38 , and Ile 39 to Ala residues inhibits the ability of Tat to bind tubulin (17). It is probable that these mutations induce conformational changes in Tat and disrupt the orientation of the Tat binding site for tubulin. In our studies, we have shown that both Tat proteins can bind to tubulin, but with different affinities. The CD spectra showed a conformational change with just a few mutations existing in Ug11LTS and Ug05RP (Fig. 5). CD data analysis showed that the content of the ␣-helix, which is located in region V, is modified by the mutations (Table I). Molecular modeling of Ug11LTS and Ug05RP showed that region III is similar, whereas region V is modified, as the ␣-helix is not present in Ug11LTS (Fig. 6). We deduce that the conformation of this region is important in tubulin binding.
Co-culture experiments of HIV-1-infected and uninfected cells have shown that whereas HIV-infected cells are resistant to HIV-induced death, uninfected CD4 ϩ cells die by apoptosis (65). In our assay for apoptosis, we observed a significant difference in quadrants R1 (cells that have lost membrane integrity as a result of very late stage apoptosis) and R2 (early apoptotic cells) between the two Tat variants as well as an overall difference (data not shown). When the cells were treated with Ug05RP, the percentage of cells in quadrant R1 was always significantly higher than when the cells were treated with Ug11LTS, suggesting that Ug05RP induces apoptosis faster than Ug11LTS.
A recent study has suggested that a pathway by which Tat induces apoptosis is by directly targeting the microtubule cytoskeleton, preventing depolymerization, and thus liberating Bim, leading to apoptosis through the mitochondrial pathway  (17). In support of this hypothesis, it has been shown that disruption of microtubule network functions liberates Bim, which translocates to the mitochondria, where it neutralizes Bcl-2 or activates Bax and therefore constitutes an initiating event in apoptotic signaling (66,67) and is independent of the Fas pathway (17). The central characteristic of this pathway is the release of cytochrome c from mitochondria (68,69). Treatment of Jurkat T-cells with Ug05RP (but not Ug11LTS) induced liberation of cytochrome c (Fig. 3C), suggesting that the death signals initiated by Ug05RP transmit via the mitochondrial pathway. We have observed the interaction of other synthetic Tat proteins (HXB2, Eli, and Oyi) with microtubules and   their correlation with apoptosis and cytochrome c liberation. 2 Thus, the difference in the speed of apoptosis could be due to the efficiency with which Tat interacts with microtubules, and the in vitro data presented here show that Ug05RP bound to tubulin with a higher affinity compared with Ug11LTS. However, in the study that implicated Bim in Tat-induced apoptosis, Tat could still induce apoptosis in Bim-negative cells, showing that Tat also initializes an apoptotic pathway in a microtubule-independent manner (17). Another pathway by which Tat has been shown to induce apoptosis is by up-regulating FasL expression (24). The ability of Ug05RP to upregulate this expression and its correlation with the level of apoptosis observed in our assays indicate that this receptormediated pathway is also important in Tat-induced apoptosis. Moreover, cytochrome c was not found to be present in the cytosol of Ug11LTS-treated cells (Fig. 3C), which also showed an increase in FasL mRNA production (Fig. 3D), suggesting that the mitochondrial pathway is not implicated in Ug11LTSinduced apoptosis.
Tat targets the microtubule cytoskeleton to induce apoptosis (17), but apoptosis does not occur in the very cells in which Tat is produced (38). It is not plausible that HIV has evolved a highly specific mechanism by which Tat kills off only bystander cells; thus, it is possible that two HIV-encoded proteins control the activities of the main regulators of the mitochondrial apoptotic pathway in infected cells. HIV-1 Nef activates the phosphatidylinositol 3-kinase⅐p21-activated kinase complex, which inactivates the pro-apoptotic Bad protein by phosphorylation (70), and HIV-1 Vpr activates the down-modulation of the pro-apoptotic Bax protein and up-regulates the levels of the anti-apoptotic Bcl-2 protein (71). Thus, Nef and Vpr possibly counteract the effects of Tat in infected cells and protect the infected cells from apoptosis.
In summary, the data presented here show the functional importance of the secondary structure of Tat and that few mutations can disrupt its functions. These Tat proteins came from two patients with distinct variations in disease progression. It is possible that such mutations lead to functional dif-ferences that may be related to disease progression. These results emphasize the tremendous importance of Tat in the progression to AIDS in HIV-infected patients and outline the importance to develop therapies that target Tat. FIG. 6. Molecular modeling of Tat Ug05RP and Tat Ug11LTS. Molecular modeling was carried out using Tat Mal two-dimensional NMR structure (19). Region I is red; region II (cysteine-rich region) is orange; region III is yellow; region IV (basic region) is green; region V is light blue; and region VI is blue. A and B, ribbon representations of Tat Ug05RP and Tat Ug11LTS, respectively; C, superimposition of Ug05RP and Ug11LTS calculated using backbone C-␣ atomic coordinates of each Tat. The folding is similar, but local variations are observed mainly in region V. The main difference between the two structures is the presence in Tat Ug05RP of a short ␣-helix observed in region V, which is not present in Tat Ug11LTS.