IκB-α Represses the Transcriptional Activity of the HIV-1 Tat Transactivator by Promoting Its Nuclear Export*

The long terminal repeat of human immunodeficiency virus, type 1 (HIV-1) contains an NF-κB enhancer and is potently inhibited by IκB-αS32/36A, a proteolysis-resistant inhibitor of NF-κB transacting factors. The evidence that NF-κB is dispensable for HIV-1 expression raises the question of whether IκB-α represses the HIV-1 transcription by mechanisms distinct from NF-κB inhibition. Here, we report that IκB-α negatively regulates the HIV-1 expression and replication in an NF-κB-independent manner by directly binding to Tat, which results in the nuclear export and cytoplasmic sequestration of the viral transactivator. The sequence of IκB-α required for Tat inhibition spans from amino acids 72 to 287 and includes the nuclear localization signal, the carboxyl-terminal nuclear export signal, and the binding site for the arginine-rich domain of Tat. This novel mechanism of cross-talk between Tat and IκB-α provides further insights into the mechanisms of HIV-1 regulation and could assist in the development of novel strategies for AIDS therapy.

The expression of HIV-1 3 is dependent on the RNA polymerase II complex and is regulated by cellular transacting factors that interact with the viral promoter (1). The HIV-1 LTR contains a basal promoter (nucleotides Ϫ78 to Ϫ1) with a TATAA box and three Sp1-binding sites, an enhancer (nucleotides Ϫ105 to Ϫ79) with two tandem B sites, and a negative regulatory region (nucleotides Ϫ454 to Ϫ106) (2). The HIV-1 encoded Tat protein dramatically increases the viral gene expression by binding to the transactivation-responsive region at the 5Ј leader HIV-1 RNA (nucleotides ϩ1 to ϩ59) (3) and promoting transcription initiation (4). In particular, Tat promotes the assembly of the preinitiation complex and nucleosomal remodeling through the interaction with several transcription factors and cofactors including TBP (5,6), NF-B (7), Sp1 (8), and the histone acetyltransferases p300/CBP and P/CAF (9 -11). In addition, Tat recruits the cyclin T1/CDK9 complex, which hyperphosphorylates the carboxyl-terminal domain of the large subunit of RNA polymerase II, resulting in processive transcription (12)(13)(14)(15).
The NF-B transacting factors enhance the HIV-1 transcription by binding to the B sites of the HIV-1 LTR (16,17). The NF-B proteins, namely p105/p50, p100/52, p65/RelA, c-Rel, and RelB, share a Rel homology domain that is required for subunit dimerization, nuclear localization, and DNA binding (18,19). The NF-B transcriptional activity is negatively regulated through the association with the IB proteins (20). IB-␣ (21), the best characterized and ubiquitous member of the IB family, contains six ankyrins, a nuclear localization signal (NLS) (22)(23)(24), and two nuclear export signals located at the amino terminus (N-NES) (25)(26)(27)(28) and carboxyl terminus (C-NES) (29,30). In unstimulated cells, IB-␣ associates with the p50/p65 NF-B complex and inhibits the DNA binding and transcriptional activity of NF-B (18,20). The IB-␣/NF-B complex shuttles in and out of the nucleus and is prevalently retained in the cytoplasm (25)(26)(27)31). Upon NF-B-activating stimuli, IB-␣ is phosphorylated at serines 32 and 36 by the IB kinase complex, ubiquitinated at lysines 21 and 22, and proteolyzed by the 26 S proteasome (18,20). Following the proteolysis of IB-␣, the free NF-B complex translocates to the nucleus, where it binds to the B sites and activates the transcription of the NF-B-dependent genes, including the IB-␣ gene (18,20). The newly synthesized IB-␣ migrates to the nucleus, where it displaces the NF-B complex from DNA and promotes its nuclear export, thus terminating the NF-B transcriptional activity (29,31,32).
IB-␣ represses the Tat-mediated transactivation of the HIV-1 LTR upon cell transfection (33). Moreover, the mutant IB-␣S32/36A, which is resistant to the signaling-induced proteolysis (34,35), inhibits the expression of HIV-1 (36). Consistent with these findings, we showed that NL-IB-M, a recombinant HIV-1 strain expressing the IB-␣S32/36A gene integrated in the nef region, is strongly repressed at the transcriptional level and highly attenuated (37). In that study, the inhibition of HIV-1 by IB-␣S32/36A was ascribed to the repression of the NF-B-dependent expression of HIV-1. However, this mechanism accounts partially for the HIV-1 transcriptional inhibition because viral strains lacking the NF-B enhancer are competent for transcription and replication (38 -40). The evidence that NF-B is dispensable for the transcriptional activation of HIV-1 raises the question of whether IB-␣ represses other transcription factors, which differ from NF-B and are required for HIV-1 expression. To test this possibility, we have analyzed the functional and physical interactions of IB-␣ with the HIV-1 Tat transactivator, which is indispensable for viral replication (41,42). Here, we report that IB-␣ binds to Tat and promotes its nuclear export to the cytoplasm. According to this novel evidence, IB-␣ acts as a potent repressor of HIV-1 transcription by inhibiting both the NF-B and Tat transacting factors, which are major players in the transcriptional activation and elongation of HIV-1 transcripts.

EXPERIMENTAL PROCEDURES
Plasmids-pLTRluc contains the U3 and R regions of the pNL4-3 molecular clone of HIV-1 upstream of the luciferase gene (37). pSV-␤-gal was purchased from Promega. To generate p3XFLAG-CMV-Tat, p3XFLAG-CMV-Tat C (22,25,27)A, and p3XFLAG-CMV-Tat R(49 -57)A, the sequence of Tat was amplified from the pGEX-2T-Tat expressing vectors (43) and ligated to EcoRI/XbaI-digested p3XFLAG-CMV-7.1 (Sigma). pRc/CMV-HA-IB-␣S32/36A was previously described (44). The plasmids expressing the IB-␣ mutants 120 -317, 1-280, 1-269, 1-242, 72-287, and 72-269 were generated by PCRmediated amplification of the IB-␣ sequence from pCMV4-HA-IB-␣ with appropriate forward and reverse primers followed by ligation to the HindIII/XbaI-digested pCMV4-HA. The mutant IB-␣ 72-287 L(272,274,277)A was generated by site-directed mutagenesis of the IB-␣ 72-287 sequence using the forward primer ATACAGCAGCAGCTGGGCCAGG-CGACAGCAGAAAACGCGCAGATGCTGCCAGAGA and reverse primer CTGGCCCAGCTGCTGCTGTATCCGGGT-GCTTGGGCGGCC with the mutated triplets indicated in bold type. The mutant IB-␣ N/C NES was generated by site-directed mutagenesis of IB-␣ 1-317 at the level of the N-NES with the forward primer AAGGAGCTGCAG GAG GCG CGC GCG GAG CCG CAG GAG GTG and reverse primer CTCC-TGCAGCTCCTTGACCATGGAGTCCA, and at the level of C-NES with the same primers described for IB-␣ 72-287 L(272,274,277)A. The mutated nucleotides are shown in bold type. The pGEX-2T-IB-␣ plasmids were generated by PCRmediated amplification of the IB-␣ sequences from the plasmid pCMV4-HA-IB-␣ followed by ligation to BamHI/EcoRIdigested pGEX-2T (Amersham Biosciences). The pcDNA3 plasmids expressing the IB-␣ mutants under the T7 promoter were generated by PCR amplification of the IB-␣ genes from the pCMV4-HA-IB-␣ plasmids followed by ligation to KpnI/ XbaI-digested pcDNA3 (Invitrogen). All of the constructs were verified by automated DNA sequencing. To generate the NF-B-deleted pNL4-3.Luc.R Ϫ E Ϫ , the XhoI/HindIII fragment of pNL4-3.Luc.R Ϫ E, which contains the 3Ј LTR, was replaced with the corresponding fragment from the NF-B-deleted pLTRluc. To generate pNL-⌬B-IB-M and pNL-⌬B-IB-as, the viral plasmids pNL-IB-M and pNL-IB-as (37) were digested with NaeI to isolate the 2.35-and 1.61-kb fragments, which contain the viral sequence from the unique NaeI site within the IB-␣S32/36A-FLAG insert in the nef region in sense or antisense orientation, respectively, to the unique NaeI site of pNL4-3 (10,346 nucleotides) in the flanking region downstream to the 3Ј LTR. The DNA fragments were ligated to the NaeI site of pBlueScript K ϩ (Stratagene) to generate pBSK-IB-M and pBSK-IB-as, respectively. The two tandem B sites within the 3Ј LTR of pBSK-IB-M and pBSK-IB-as were deleted by site-directed mutagenesis using the forward primer FNLDKB CGAGCTTGCTACAAGGGATCTAGATCCAGG-GAGGCGTGGCCTGGGC and the reverse primer RNLDKB TCCTTGTAGCAAGCTCGATGTCAGCAGTTCTTGAA-GTAC to generate pBSK-⌬B-IB-M and pBSK-⌬B-IB-as, respectively. The mutated sequence of B sites is shown in bold type in the forward primer. The viral plasmids pNL-⌬B-IB-M and pNL-⌬B-IB-as were generated by replacement of the NaeI-digested 2.35-and 1.61-kb DNA fragments with the corresponding region from pBSK-⌬B-IB-M and pBSK-⌬B-IB-as, respectively.
Cell Culture, Transfection, and Luciferase Assay-HeLa and MEFs were cultured in Dulbecco's modified Eagle's medium (Invitrogen), Jurkat cells in RPMI (Invitrogen). The culture media were supplemented with 10% heat-inactivated fetal calf serum and 2 mM L-glutamine at 5% CO 2 and 37°C. The cells were transfected with DNA by using FuGENE 6 (Roche Applied Science), and the total amounts of DNA were equalized by transfection of pRc/CMV empty vector (Invitrogen). For luciferase assays, pSV-␤-gal plasmid (0.2 g) was co-transfected with the pLTRluc plasmids to monitor the transfection efficiency. Forty-eight hours post-transfection, the cells were lysed in lysis buffer of Dual Light Luciferase System (Tropix, Bedford, MA). The luciferase and ␤-galactosidase activities were evaluated by using the Dual Light luciferase system (Tropix, Bedford, MA) in a bioluminometer (Turner Biosystem, Sunnyvale, CA). The ratio of firefly luciferase activity to ␤-galactosidase activity was expressed as relative light units.
Viral Integration-Genomic DNA was extracted from aliquots of infected cells (2 ϫ 10 6 ) using TRIzol (Invitrogen) and amplified with primers that annealed in the U5 region of the LTR (MH 531) and in the 5Ј end of the gag gene (MH 532). The reaction mixture (25 l) contained genomic DNA (200 ng), primers (600 nM), and 1ϫ iQ SYBR Green Supermix (Bio-Rad). Real time PCR was performed by using iCycler Apparatus (Bio-Rad). After an initial denaturation step (95°C for 8 min), the cycling profile for total HIV-1 DNA was 50 cycles consisting of 95°C for 10 s, 60°C for 10 s, and 72°C for 6 s. Viral DNA was normalized to cellular genomic glyceraldehyde-3-phosphate dehydrogenase. Primers were as follows: MH531, TGTGTGC-CCGTCTGTTGTGT; MH532, GAGTCCTGCGTCGAGAG-AGC; glyceraldehyde-3-phosphate dehydrogenase forward, GAAGGTGAAGGTCGGAGTC; and glyceraldehyde-3-phosphate dehydrogenase reverse, GAAGATGGTGATGGGAT-TTC. The HIV-1 DNA copy number was measured as reported (46). Viral Stocks and Cell Culture Infection-293-T cells were transfected with viral plasmids, and the viral production was measured by p24 enzyme-linked immunosorbent assay. Jurkat cells (5 ϫ 10 4 cells) were infected with p24 (0.3 ng) of viral stocks, and the cell supernatants were collected every 3 days for p24 assay; equal volumes of fresh medium were replaced into the cultures at the same time.
Statistical Analysis-The data were reported as the means Ϯ S.E. and the statistical significance of differences between means was assessed by using the two-tail unpaired Student's t test. The differences between the means were accepted as statistically significant at the 95% level (p ϭ 0.05).
GST Pulldown-GST fusion proteins were produced in Escherichia coli strain BL21 as previously described (43). Bacterial cultures (500 ml) were grown to exponential phase and induced with 0.25 mM isopropyl-␤-D-thiogalactopyranoside (Sigma-Aldrich) for 3 h to express GST fusion proteins. The bacteria were lysed by sonication in buffer A (1ϫ PBS, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, and 1ϫ protease inhibitor mixture EDTA-free), and the lysate was clarified by centrifugation at 27,000 ϫ g for 30 min at 4°C. The supernatant was incubated with 1 ml of a 50% (v/v) slurry of glutathione-Sepharose beads (Amersham Biosciences) previously equilibrated in buffer A. After incubation on a rotating wheel at 4°C for 2 h, the beads were washed five times with buffer A and subjected to a high salt wash (0.8 M NaCl) to free the fusion proteins from contaminating bacterial nucleic acids (47). GST fusion proteins were eluted with 500 l of 50 mM Tris-HCl containing 10 mM glutathione and 1 mM DTT. The eluted GST fusion proteins were dialyzed against dialysis buffer (1ϫ PBS, 1 mM DTT, 10% glycerol), and aliquots (5-10 g) were conjugated with glutathione-Sepharose (20 l) in 500 l of binding buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% sodium deoxycholate, 0.2% SDS, 2 mM EDTA, 3% Triton X-100, 5 mM DTT, 1ϫ protease inhibitor mixture EDTA-free). The GST fusion proteins conjugated with glutathione-Sepharose were collected by centrifugation at 700 ϫ g for 5 min at 4°C, and aliquots (5-10 g) were incubated with cell extracts (200 g) in 500 l of binding buffer supplemented with 1 g/l of bovine serum albumin on a rocking platform for 3 h at 4°C. To remove nucleic acids, the cell   DECEMBER 21, 2007 • VOLUME 282 • NUMBER 51 extracts were treated with micrococcal nuclease (0,2 unit/l) for 30 min at 28°C. Protein complexes were collected by centrifugation at 700 ϫ g for 5 min at 4°C, washed in binding buffer, and resuspended in loading buffer (125 mM Tris-HCl, pH 6.8, 5% SDS, 1% bromphenol blue, 10% ␤-mercaptoethanol, 25% glycerol). The proteins were resolved on 10% SDS-PAGE, transferred to polyvinylidene difluoride membranes, and analyzed by immunoblotting with the indicated antibodies. The pcDNA3 plasmids expressing the IB genes under the T7 promoter were used as templates to translate in vitro the [ 35 S]methionine-labeled IB proteins by using the TNT Quick Coupled transcription/translation systems (Promega). Aliquots (10 l) of translation mixture were incubated with GST-Tat or GST proteins (10 g) in 500 l of binding buffer supplemented with 1 g/l of bovine serum albumin on a rocking platform for 3 h at room temperature. Following GST pulldown, the proteins were separated by 12% SDS-PAGE and analyzed by autoradiography and immunoblotting with antibodies.

Physical and Functional Interaction of IB-␣ with HIV-1 Tat
Co-immunoprecipitation-Cell extracts were performed in PBS containing 1% Triton X-100 and 1ϫ Protease Inhibitor Mixture EDTAfree. Antibodies (2.5 g) were preincubated with protein G-Sepharose (Amersham Biosciences) (20 l) in 50 l of immunoprecipitation buffer (PBS containing 2% Triton X-100, 300 mM NaCl, 5 mM DTT, 1ϫ Protease Inhibitor Mixture EDTA-free) overnight at 4°C on a rocking platform. The protein G-Sepharose-coupled antibodies were incubated with cell extract (500 g) in 500 l of immunoprecipitation buffer overnight at 4°C on a rocking platform. The immunocomplexes were collected by centrifugation at 700 ϫ g for 5 min at 4°C, washed in immunoprecipitation buffer, and resuspended in SDS gel loading buffer. The proteins were separated on 10% SDS-polyacrylamide gel, transferred to polyvinylidene difluoride membrane, and analyzed by immunoblotting with antibodies.
Confocal Microscopy-Confocal microscopy was performed as previously described (48). HeLa cells were seeded on poly-L-lysinetreated glass coverslips, fixed, and permeabilized using Cytofix/Cyto-Perm kit (BD Biosciences Pharmingen, San Diego, CA). To visualize FLAG-Tat and HA-IB-␣, the immunostaining was performed with anti-FLAG-M2-FITC mAb (F-4049; Sigma-Aldrich) and anti-HA rabbit antiserum (SC-805; Santa Cruz Biotechnologies) followed by goat anti-rabbit Alexa Fluor568 (A11011; Molecular Probes, Eugene, OR). The nuclei were stained with TO-PRO-3 iodide (T3605; Molecular Probes). The coverslips were mounted on glass slides by using ProLong Antifade Kit (P7481; Molecular Probes). The images were collected on a Leica TCS-SP2 confocal microscope (Leica Mycrosystems, Wetzlar, Germany) with a 63ϫ Apo PLA oil immersion objective (NA 1.4) and 60-m aperture. Z stacks of images were collected using a step increment of 0.2 m between planes. FLAG-Tat was visualized by excitation with an argon laser at 488 nm and photomultiplier tube voltage of 420 mV. HA-IB-␣ was detected using a krypton laser at 568 nm and photomultiplier tube voltage of 650 mV. The nuclei were detected using a krypton laser at 613 nm and photomultiplier tube voltage of 450 mV. Single optical sections using 4X averaging were acquired by sequential scanning to collect the images in three channels. For quantitative analysis of the nuclear and cytoplasmic protein levels, horizontal sections scanned through the nucleus and cytoplasm of thirty representative cells were evaluated. Fluorescence-based assessment of protein levels was performed by image analysis using the LEICA Scan TCS-SP2 software (Leica Mycrosystems). Quantization was performed on 8-bit gray scale images with no saturated pixels. The mean nuclear or cytoplasmic fluorescence was measured as the ratio between total fluorescence/total pixels at the nuclear or cytoplasmic level in individual cells. The relative nuclear or cytoplasmic fluorescence was calculated as the ratio between the mean nuclear or cytoplasmic fluorescence and the mean fluorescence of the whole cell (49).

IB-␣ Represses the Tat-mediated Transactivation and Replication of HIV-1 Independently of NF-B
Activity-To determine the effect of IB-␣ on the transcriptional activity of Tat, HeLa cells were transiently transfected with the luciferase gene under the control of the wild type or NF-Bor Sp1-deleted HIV-1 LTR in the presence or absence of Tat and IB-␣. In agreement with previous observations (39), the deletion of the NF-B or Sp1 sites significantly reduced the basal expression (Fig. 1A) and the Tat-mediated transactivation (Fig. 1, B-D) of the HIV-1 LTR. IB-␣ inhibited the Tat transcriptional activity in a dose-dependent manner up to 80% for the wild type LTR (Fig. 1B) and 60% in the case of NF-B-deleted LTR (Fig.  1C). The evidence that IB-␣ inhibited the Tat-mediated transactivation of the LTR in the absence of the NF-B enhancer underscored the existence of mechanisms of LTR inhibition distinct from NF-B repression. IB-␣ completely re-  DECEMBER 21, 2007 • VOLUME 282 • NUMBER 51 pressed the Tat-mediated transactivation of the Sp1-deleted LTR (Fig. 1D); this strong inhibition was likely caused by repression of both NF-B-dependent and independent transactivation of the LTR. The LTR inhibition was not a consequence of pro-apoptotic activity of IB-␣ because the cleavage of caspase-3 and poly(ADP-ribose)polymerase was undetected in IB-␣-transfected cells (supplemental Fig. S1).

Physical and Functional Interaction of IB-␣ with HIV-1 Tat
Next, we analyzed the effect of IB-␣ on Tat in the absence of NF-B activity. To this end, the expression of the NF-B-deleted LTR was analyzed in MEFs lacking the p50 and p65 subunits of NF-B (50 -52). Because the murine cyclin T1 does not allow the generation of the P-TEFb/Tat/transactivation-responsive region complex for efficient transcriptional elongation (53), p50 Ϫ/Ϫ p65 Ϫ/Ϫ MEFs were transfected with or without the hCycT1. IB-␣ significantly inhibited the Tat-mediated transactivation of the NF-B-deleted LTR in a dose-dependent manner in presence or absence of hCycT1 ( Fig. 2A), which rules out the possibility that IB-␣ repressed the Tat activity by interaction with hCycT1.
Further, we analyzed the effect of IB-␣ on the expression of the single-cycle replication virus NL4-3.Luc.R Ϫ E Ϫ carrying the wild type or NF-B-deleted LTR. Jurkat cells were transfected with the proteolysis-resistant mutant IB-␣S32/36A or with IB-␣ siRNA to up-regulate or down-regulate the intracellular levels of IB-␣, respectively. Transfected cells were infected with VSV-G-pseudotyped NL4-3.Luc.R Ϫ E Ϫ virions that carry the wild type or NF-B-deleted LTR. The virion production was significantly reduced by hyperexpression of IB-␣ and increased by knocking down the endogenous IB-␣ with IB-␣ siRNA in both infections with the wild type (Fig.  2B, left panel) or the NF-B-deleted virus (Fig. 2B, right panel). No difference in the number of integrated virus among the different samples was observed (supplemental Fig.  S2). These results suggest that the levels of endogenous IB-␣ inversely affected the expression of the integrated HIV-1 genome independently of the presence of the NF-B-binding sites in the HIV-1 LTR.
To analyze the effect of IB-␣ on the HIV-1 replication in the absence of the NF-B-binding sites of the HIV-1 LTR, we generated the viral plasmids pNL-⌬B-IB-M and pNL-⌬B-IB-as, which carry the IB-␣S32/36A-FLAG cDNA inserted into the nef region in sense or antisense orientation, respectively, and were deleted of the two tandem B sites in the LTR (Fig. 2C). These recombinant HIV-1 plasmids were the NF-Bdeleted derivatives of pNL-IB-M and pNL-IB-as (37), which express or do not express, respectively, IB-␣S32/36A-FLAG. Jurkat cells were infected with the wild type LTR viruses (NL-IB-M and NL-IB-as) or the NF-B-deleted LTR viruses (NL-⌬B-IB-M and NL-⌬B-IB-as), and the viral production was measured by p24 detection in culture supernatants. As previously reported (37), NL-IB-M was potently attenuated as compared with the control NL-IB-as because of the IB-␣S32/ 36A expression (Fig. 2D, left panel). In the case of NF-B-deleted viruses, a significant attenuation of NL-⌬B-IB-M was also observed as compared with the control NL-⌬B-IB-as (Fig. 2D, right panel). These results indicate that IB-␣ inhibited the HIV-1 replication independently of the NF-B enhancer in the HIV-1 LTR and supported the evidence of additional mechanisms of HIV-1 inhibition by IB-␣ other than NF-B repression.
The Sequence of IB-␣ from Amino Acids 72 to 287 Is Required for Tat Inhibition-The sequence of IB-␣ encompassing amino acids 1-317 contains six ankyrins (amino acids 72-287), the NLS (amino acids 110 -120), the N-NES (amino acids [45][46][47][48][49][50][51][52][53][54][55], and the C-NES (amino acids 265-277) (Fig. 3A). To map the IB-␣ domains required for Tat inhibition independently of NF-B repression, the activity of IB-␣ mutants was analyzed in p50 Ϫ/Ϫ p65 Ϫ/Ϫ MEFs by transient expression of the NF-B-deleted LTR and Tat. The IB-␣ mutants that were progressively deleted of the carboxyl-terminal from amino acids 317 to 280 significantly inhibited the Tat activity, whereas no inhibition was induced by IB-␣ 1-269 deleted of the C-NES (Fig. 3B). Further, deletions of the carboxyl-terminal of IB-␣ from amino acids 269 to 242 did not affect the Tat activity (Fig. 3B). IB-␣ 72-317 lacking the amino-terminal sequence from amino acids 1 to 72 significantly inhibited Tat, whereas IB-␣ 120 -317, which was deleted of the NLS, lost the inhibitory activity (Fig. 3B). These results indicated that the sequences of IB-␣ from amino acids 72 to 120 (overlapping the NLS) and from amino acids 269 to 280 (overlapping the C-NES) were both required for Tat inhibition. This was confirmed by experiments where the mutant IB-␣ 72-287, which contains both the NLS and C-NES, inhibited Tat, whereas the mutants IB-␣ 72-269 and IB-␣ 72-287 L(272,274,277)A, which carry deletion or base pair substitutions of critical leucine residues of the C-NES sequence (29), respectively, failed to inhibit Tat (Fig. 3B). Lack of inhibition was confirmed at higher doses of IB-␣ 120 -317, IB-␣ 72-269, and IB-␣ 72-287 L(272,274,277)A (supplemental Fig. S3). The IB-␣ mutants were all expressed in cell extracts, and no correlation was found between the level of expression and the inhibitory activity (Fig. 3C). These results demonstrated that the minimal sequence of IB-␣ required for Tat inhibition spanned from amino acids 72 to 287; this region encompasses the six ankyrins of IB-␣ including the NLS and C-NES.
IB-␣ Binds to the Arginine-rich Domain of Tat-To test whether IB-␣ physically interacts with Tat, the GST pulldown assay was performed with extracts from cells transfected with pCMV4-HA-IB-␣. GST-Tat retained IB-␣ expressed in HeLa and MEFs (Fig. 4A, lanes 1 and 2). The binding of Tat with IB-␣ was also observed in p50 Ϫ/Ϫ p65 Ϫ/Ϫ MEFs (Fig. 4A, lane  3), which ruled out that IB-␣ and Tat were recruited in the same complex by associating with the p50 and p65 subunits of NF-B. IB-␣ was not retained by GST protein (Fig. 4A, lanes  4 -6). The association of endogenous IB-␣ with GST-Tat was also observed in HeLa extracts (Fig. 4B, lane 1). The treatment of the cellular extracts with micrococcal nuclease did not affect the binding of IB-␣ with Tat (Fig. 4C, lane 2), thus ruling out the possibility that the association of the two proteins was bridged by nucleic acids.
To map the Tat domain that binds to IB-␣, GST-IB-␣ was incubated with extracts from HeLa cells transfected with the wild type Tat or the mutants Tat C (22,25,27)A and Tat R(49 -57)A fused to the FLAG epitope (Fig. 4D). In pulldown assay, GST-IB-␣ retained the wild type Tat and Tat C(22,25,27)A (Fig. 4E, lanes 2 and 4), whereas it did not bind to Tat R(49 -57)A (Fig. 4E, lane 3); Tat was not retained by GST alone (Fig.  4E, lanes 6 -8). The association of IB-␣ with Tat was further tested by in vivo immunoprecipitation with extracts from HeLa cells transfected with the plasmids expressing FLAG-Tat and HA-IB-␣. IB-␣ immunoprecipitated with the wild type Tat and Tat C (22,25,27)A (Fig. 4F, lanes 2 and 3), whereas it did not associate with Tat R(49 -57)A (Fig. 4F, lane 4). Altogether, these results indicate that the arginine-rich region of Tat encompassing amino acids 49 -57 is required for the association with IB-␣.

Physical and Functional Interaction of IB-␣ with HIV-1 Tat
By CLUSTALW-based multiple sequence alignment (align. genome.jp), the amino acid sequence of the IB-␣ sixth ankyrin is very divergent from the other five ankyrins of IB-␣ and the ankyrins of the human IB family (p100, p105, IB-␥, IB-⑀, and Bcl-3) (supplemental Fig. S4, A and B). In particular, the sequence TRIQQQL (amino acids 263-269 of IB-␣), which is present in the sixth ankyrin and is required for the binding to Tat (Fig. 5B), is absent in ankyrins 1-5 of IB-␣ as well as in the ankyrins of the IB family members (supplemental Fig. S4, A  and B). A more extended analysis by using FUZZPRO (bioweb. pasteur.fr/seqanal/interfaces/fuzzpro.html) failed to identify the TRIQQQL motif in the ankyrins of the human proteome except in the sixth ankyrin of IB-␣. Accordingly, p100 and p105, two members of the IB family showing the highest identity with the sixth ankyrin of IB-␣, were unable to bind to Tat (supplemental Fig. S4C, lanes 1 and 4). These results suggest that the sixth ankyrin of IB-␣ contains a unique diverged sequence as compared with other ankyrins, which might represent a privileged target site for Tat binding. Alternatively, this sequence might contribute to stabilize a peculiar structural domain required for the binding to Tat.
IB-␣ Exports Tat from the Nucleus to the Cytoplasm-The cellular distribution of IB-␣ and Tat was visualized by confocal fluorescence microscopy. HeLa cells were transfected with plasmids expressing FLAG-Tat and HA-IB-␣. When singularly transfected, Tat was nuclear, whereas IB-␣ 1-317 was mostly cytoplasmic (Fig. 6A); this was confirmed by the fluorescence-based analysis of 30 cells for each transfection (Fig. 7). When co-transfected, Tat and IB-␣ 1-317 co-localized within the cytoplasmic and perinuclear regions (Figs. 6A and 7). IB-␣ 120 -317, lacking both the N-NES and the NLS, and IB-␣ 1-269, lacking the C-NES, were prevalently cytoplasmic and did not affect the nuclear location of Tat (Figs. 6A and 7). IB-␣ 72-287, lacking the N-NES, was mostly cytoplasmic and promoted the translocation of Tat from the nucleus to the cytoplasm in 50% of the analyzed cells (Figs. 6A and 7). IB-␣ 72-269 and IB-␣ 72-287 L(272,274,277)A, which lacked the N-NES and C-NES, were distributed both in the nucleus and cytoplasm and did not affect the nuclear location of Tat (Figs. 6A and 7). No significant differences in the intracellular expression levels of the IB-␣ mutants were observed in transfected cells (Fig. 6B). These results suggested that IB-␣ promoted the displacement of Tat from nucleus to cytoplasm and that this activity required the integrity of the NLS and C-NES of IB-␣.
To analyze the role of the nuclear export activity of IB-␣ in Tat inhibition, we generated the mutant IB-␣ 1-317 N/C NES, which carries crucial base pair substitutions of both the N-NES (I52A,L54A) and C-NES (L272,274,277A), which inactivate the nuclear export activity (25,29). IB-␣ N/C NES was prevalently distributed in the nucleus and did not affect the nuclear location of Tat (Fig. 8, A and B). Moreover, IB-␣ N/C NES did not repress the Tat-mediated transactivation of the NF-B-deleted LTR (Fig. 8C), although it was able to bind to Tat in GST-pull down (Fig. 8D, lane 2). These results confirmed that IB-␣ inhibited Tat through the nuclear export to the cytoplasm.
We also verified the effect of leptomycin B, a nuclear export inhibitor (54), on the inhibition of Tat by IB-␣. In p50 Ϫ/Ϫ p65 Ϫ/Ϫ MEFs, leptomycin B did not affect significantly the level of Tat-mediated transactivation of the NF-B-deleted LTR (supplemental Fig. S5, lanes 2 and 5), whereas it caused the loss of Tat inhibition by the transfected IB-␣ (supplemental Fig. S5, compare lanes 2 and 3 with lanes 5 and 6). These results indicate that the leptomycin B-mediated arrest of nuclear export released Tat from the IB-␣ inhibition.

DISCUSSION
This study provides further insight into the mechanisms of HIV-1 inhibition by the IB-␣ repressor. We have shown that IB-␣ represses the Tat activity independently of the NF-B inhibitory activity by physical association and displacement of Tat from the nucleus to the cytoplasm. The association of IB-␣ with the argininerich domain of Tat is not sufficient to interfere with the nuclear distribution and the transcriptional activity of Tat. In fact, the mutants IB-␣ 120 -317 and IB-␣ 1-269 bind to Tat without affecting the nuclear location and transcriptional activity of the viral transactivator (Fig. 9A). Instead, the inhibition of Tat correlates with the nuclear export activity of IB-␣, which requires both the NLS (amino acids 110 -120) and the C-NES (amino acids 265-277) together with the binding site for Tat (amino acids 263-269) (Fig.  9A). Consistent with this evidence, the mutant IB-␣ N/C NES, which contains the full-length sequence of IB-␣ but lacks the nuclear export signals, does not affect the nuclear location and the transcriptional activity of Tat (Fig. 9A). Altogether, these results suggest that IB-␣ binds to Tat in the nucleus and exports the viral transactivator to the cytoplasm, where the complex IB-␣/Tat is mostly retained (Fig. 9B).
Relevance of IB-␣/Tat Interaction in the Viral Cycle-The evidence that IB-␣ inhibits the transcriptional activity of Tat raises the question of why the endogenous IB-␣ does not counteract the viral expression in HIV-1-infected cells. Indeed, IB-␣ is subjected to persistent proteolysis in the course of HIV-1 infection (55)(56)(57). The HIV-1 entry through the gp120 envelope protein binding to CD4 receptor activates the IB kinase complex, which promotes the proteolysis of IB-␣ (58). This event leads to the transcriptional activation of NF-B-dependent genes, FIGURE 8. The nuclear export activity of IB-␣ is required for nuclear export and inhibition of Tat. A, HeLa cells (5 ϫ 10 5 ) were transfected with p3XFLAG-CMV-Tat (3 g) and pCMV4-HA-IB-␣ 1-317 N/C NES (3 g). The cells were analyzed by confocal microscopy as described under "Experimental Procedures." Scale bar, 10 m. B, the fluorescence-based analysis of FLAG-Tat and HA-IB-␣ was performed as detailed in Fig. 7. C, p50 Ϫ/Ϫ p65 Ϫ/Ϫ MEFs (3 ϫ 10 5 ) were transfected with the NF-B-deleted LTRluc (0.5 g) in presence or absence of p3XFLAG-CMV-Tat (0.5 g), pRc/CMV-HA-hCycT1 (0.5 g), and pCMV4-HA-IB-␣ 1-317 or pCMV4-HA-IB-␣ N/C NES (2 g). The luciferase activity was measured in cell extracts 48 h post-transfection. Fold activation was calculated relative to transfection in the absence of Tat, hCycT1, and IB-␣ expression plasmids. The mean values Ϯ S.E. (n ϭ 4) are shown. The asterisk indicates a statistically significant inhibition according to Student's t test (p ϭ 0.0017). D, [ 35 S]methionine-labeled IB-␣ wild type and IB-␣ N/C NES were incubated with GST-Tat or GST. Following GST pulldown, the protein complexes were separated by 12% SDS-PAGE and analyzed by autoradiography (top panel) and by Western blotting with anti-GST antibody (bottom panel).
including the HIV-1 genome and pro-inflammatory genes, which in turn sustain the proteolysis of IB-␣ and the activation of NF-B (59). In particular, Tat activates NF-B by inducing the degradation of IB-␣ (47), the up-regulation of NIK (60), and the transactivation of inflammatory cytokines (61)(62)(63).
The physical and functional interaction of IB-␣ with Tat discloses a novel mechanism of HIV-1 transcriptional regulation. In fact, the ratio between IB-␣ and Tat could determine the level of expression of the target genes, including HIV-1. In this scenario, whereas the endogenous IB-␣ does not block the viral expression because it undergoes proteolysis in the course of HIV-1 infection (55-57), a proteolysis-resistant IB-␣ mutant, such as IB-␣S32/36A, would subvert this equilibrium and repress HIV-1 by constitutive inhibition of both NF-B and Tat transcriptional activities (36,37). In this regard, the novel mechanism of Tat inhibition by IB-␣ described here could assist in the development of a novel class of HIV-1 inhibitors. In particular, the inhibitory sequence of IB-␣ (amino acids 72-287) represents a model structure to design peptide-based inhibitors acting at the transcriptional step of the HIV-1 life cycle.
The mechanism of Tat inhibition here described mimics the one of NF-B inhibition, because in both cases IB-␣ associates with these transcriptional factors in the nucleus and exports them to the cytoplasm. These events result in the down-regulation of the NF-B and Tat transcriptional activities and explain the strong attenuation of HIV-1 by IB-␣. In this regard, compounds that increase the stability of endogenous IB-␣, such as inhibitors of IB kinase and proteasomes, could be additional tools of conventional anti-retroviral therapies by inhibiting the NF-B-Tat-dependent HIV-1 transcription.