Association of Tat with purified HIV-1 and HIV-2 transcription preinitiation complexes.

The HIV-1 (human immunodeficiency virus type 1) and HIV-2 Tat proteins increase the level of transcription from their corresponding long terminal repeats. Tat activates transcription likely by interaction with components of the transcriptional initiation and elongation complexes during different stages of the transcription reaction. In the current study, two approaches were used to address the sites at which Tat becomes stably associated with the HIV transcription complex. First, we isolated column purified HIV-1 and HIV-2 transcription complexes that were competent for in vitro transcription and found that wild-type but not mutant Tat protein was specifically associated with this complex. An intact HIV TATA element and the presence of functional TATA-binding protein were necessary for Tat association. In contrast, the HIV-1 and HIV-2 TAR bulge sequences which serve as binding sites for Tat were not required for its association with the HIV preinitiation complex. A second complementary approach using immobilized HIV-1 and HIV-2 templates also demonstrated a functional association of Tat with HIV-1 and HIV-2 preinitiation complexes. Wild-type but not mutant Tat proteins associated with transcription complexes assembled on immobilized HIV-1 and HIV-2 templates and the association of Tat correlated with increases in the level of in vitro transcription. These results indicate that Tat can associate with HIV-1 and HIV-2 transcription complexes prior to the initiation of transcription by RNA polymerase II.

The HIV-1 (human immunodeficiency virus type 1) and HIV-2 Tat proteins increase the level of transcription from their corresponding long terminal repeats. Tat activates transcription likely by interaction with components of the transcriptional initiation and elongation complexes during different stages of the transcription reaction. In the current study, two approaches were used to address the sites at which Tat becomes stably associated with the HIV transcription complex. First, we isolated column purified HIV-1 and HIV-2 transcription complexes that were competent for in vitro transcription and found that wild-type but not mutant Tat protein was specifically associated with this complex. An intact HIV TATA element and the presence of functional TATA-binding protein were necessary for Tat association. In contrast, the HIV-1 and HIV-2 TAR bulge sequences which serve as binding sites for Tat were not required for its association with the HIV preinitiation complex. A second complementary approach using immobilized HIV-1 and HIV-2 templates also demonstrated a functional association of Tat with HIV-1 and HIV-2 preinitiation complexes. Wild-type but not mutant Tat proteins associated with transcription complexes assembled on immobilized HIV-1 and HIV-2 templates and the association of Tat correlated with increases in the level of in vitro transcription. These results indicate that Tat can associate with HIV-1 and HIV-2 transcription complexes prior to the initiation of transcription by RNA polymerase II.
HIV-1 1 and HIV-2 gene expression is regulated by multiple cis-acting sequences in the LTR and the transactivator protein Tat (reviewed in Refs. 1 and 2). Both HIV-1 and HIV-2 have similar regulatory elements in their LTRs including NF-B, SP1, TATA, and TAR (reviewed in Refs. 1 and 2). A variety of studies indicate that both the sequence and spacing of the NF-B, SP1, and TATA elements play an important role in regulating the levels of basal and Tat-induced transcription (3)(4)(5)(6)(7)(8)(9)(10)(11)(12). For example, changing the sequence of the HIV-1 and HIV-2 TATA elements to those found in several other viral and cellular promoters can result in marked decreases in the level of Tat activation (3,8,9). These results demonstrate that unique structural features in the HIV-1 and HIV-2 LTR facilitate activation by Tat.
Both the HIV-1 and HIV-2 LTRs contain an element known as TAR which extends downstream from the transcription initiation site and is essential for Tat activation (4,11,(13)(14)(15)(16)(17). TAR forms a stable stem-loop RNA structure which contains both bulge and loop sequences which are critical for its role in Tat activation. In contrast to the single HIV-1 TAR RNA structure which extends from ϩ1 to ϩ60 (15,17), the HIV-2 TAR RNA contains a duplicated RNA structure extending from ϩ1 to ϩ123 with distinct loop and bulge sequences (14,18). The TAR RNA loop sequences are critical for the binding of a cellular factor known as TRP-185 (19 -21), while the bulge serves as the binding site of Tat (22,23). The TAR RNA loop and bulge structures are also critical for the binding of RNA polymerase II (24). Thus both the TAR element and upstream regulatory elements are important in regulating the activation of gene expression by the HIV-1 and HIV-2 Tat proteins.
The full-length HIV-1 and HIV-2 Tat proteins are comprised of 86 and 130 amino acids, respectively, and have a similar organization including cysteine-rich, core, and basic domains which are critical for Tat function (reviewed in Refs. 1 and 2). The HIV-1 Tat protein can transactivate both the HIV-1 and HIV-2 LTRs to similar levels while the HIV-2 Tat protein transactivates the HIV-1 LTR to a lesser degree than the HIV-2 LTR (14,25). In the absence of HIV-1 and HIV-2 Tat proteins, so called short or nonprocessive transcripts are synthesized from the HIV-1 and HIV-2 LTR (26 -29), while in the presence of these proteins this polarity effect is abrogated (26 -32). One potential mechanism to explain Tat activation is that the structure of TAR RNA serves to pause the elongating RNA polymerase II and that Tat in conjunction with cellular transcription factors and kinases act on this paused polymerase to increase its processivity. Consistent with this model, a recent study demonstrated that Tat was stably associated with components of the HIV-1 transcriptional elongation complex (33).
RNA polymerase II can exist in a large multisubunit complex which includes a number of general transcription factors in addition to other factors involved in mediating transcriptional activation (34 -40). This so called RNA polymerase II holoenzyme is sufficient to mediate transcriptional activation in response to a variety of upstream activator proteins following their binding to the promoter element (35). Since Tat is capable of binding to the core RNA polymerase II (16), the holoenzyme is a potential target for Tat activation. Such an interaction would be consistent with a model that Tat could associate with both components of the preinitiation complex and the transcriptional elongation complex.
One of the major questions in understanding HIV-1 and HIV-2 Tat function is what point during HIV transcriptional * This work was supported in part by the National Institutes of Health and the Veterans Administration. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: HIV-1 and -2, human immunodeficiency virus types 1 and 2; LTR, long terminal repeat; TBP, TATAbinding protein; TAF, TATA-associated factor; CAT, chloramphenicol acetyltransferase. activation these proteins associate with the transcription complex. Two possibilities need to be considered. First, HIV-1 and HIV-2 Tat may bind to TAR RNA and then associate with the transcription complex to stimulate transcriptional elongation. Alternatively, HIV-1 and HIV-2 Tat may associate with the preinitiation complex and then later bind to TAR RNA to result in increases in transcriptional elongation (41). To test these possibilities, we assembled purified HIV-1 and HIV-2 preinitiation complexes using either column chromatography or immobilized templates and assayed for the association of the HIV-1 and HIV-2 Tat proteins. We demonstrated that wildtype but not mutant HIV-1 and HIV-2 Tat proteins specifically associated with transcriptionally active HIV-1 and HIV-2 preinitiation complexes. Furthermore, HIV-1 and HIV-2 Tat association with the HIV-1 and HIV-2 preinitiation complexes resulted in increased levels of in vitro transcription. These results indicate that the HIV-1 and HIV-2 Tat proteins initially interact with components of the preinitiation complex to facilitate the assembly or modify transcription complexes that are capable of processive transcriptional elongation.

Purification of HIV-1 and HIV-2 Tat 1 and 2 Proteins-Recombinant
HIV-1 and HIV-2 Tat proteins fused to a glutathione S-transferase moiety were produced in bacteria, purified using glutathione-Sepharose chromatography and cleaved with thrombin as described previously (5,20,42). The HIV-1 Tat mutant 50/57 contains mutations which change basic amino acids to the neutral amino acid glycine (20). The HIV-2 Tat mutant construct ⌬84 (5) is truncated in the basic domain and was a gracious gift from Dr. A. Rice. The wild-type and mutant HIV-1 Tat proteins both contained flag epitope sequences inserted at their carboxyl terminus that were recognized by the M2 monoclonal antibody (Kodak).
DNA Templates and Nuclear Extracts-HeLa nuclear extract was prepared as described (43) and pre-cleared by centrifugation at 30,000 rpm in a T-1270 Sorvall rotor. To inactivate TBP, heat treatment of the HeLa nuclear extract was performed at 47°C for 15 min as described (44). To restore transcription activity, 3 l of TBP which is equivalent to 3 footprint units (Promega) was added to 400 g of the HeLa nuclear extract. The HIV-2 plasmid constructs used as templates (wild-type and mutant) have been previously described (5). Briefly, a fragment from the HIV-2 Rod isolate (41) containing sequences extending from Ϫ254 (Asp-718) to ϩ156 (HindII) relative to the transcription initiation site (generous gift from Dr. G. Nabel) were cloned into a CAT expression vector and linearized for run-off transcriptions using the unique NcoI site in the CAT gene. The double bulge mutant construct contained site-directed mutants in which the bulge 1 (nucleotides ϩ27 to ϩ28) was changed from TT to AA and in bulge 2 (nucleotide ϩ62) which was changed from T to A. An HIV-1 DNA fragment (wild-type and mutants) from Ϫ154 (AvaI) to ϩ 1064 (SpeI) was end-filled and cloned into pGEM 3Z (SmaI) (Promega) and screened for promoter orientation close to the EcoRI site. The HIV-1 bulge mutant contained a mutation at nucleotide ϩ22 while the HIV-1 NTATA mutation in the HIV-1 promoter contained a mutation in the TATA box from TATAA to GCGAA (9). The plasmids were linearized using a unique XbaI site on the pGEM 3Z plasmid and have previously been used in in vitro transcription analysis (5).
Isolation of Preinitiation Complexes by Gel Filtration Chromatography-Preinitiation complexes were assembled by incubating 125 l of pre-cleared nuclear extract in buffer D (43) with either 5 g (HIV-2 LTR) or 8 g (HIV-1 LTR) of template in a total volume of 250 l. The final conditions were 20 mM HEPES pH 7.9, 10% glycerol, 50 mM KCl, 0.1 mM EDTA, 0.25 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, and 5 mM MgCl 2 . Approximately 0.5 g of double-stranded DNA oligonucleotides extending from Ϫ35 to Ϫ10 in the HIV-1 LTR with a mutation in the TATA sequences or other unrelated oligonucleotides were added to the binding reactions to decrease nonspecific binding of nuclear proteins to the HIV-1 and HIV-2 LTR templates. After a 10-min preincubation at 30°C, 400 ng of Tat protein was added and the mixture was further incubated at 30°C for 30 min. The reaction was cooled on ice for 2 min, loaded onto a Sepharose CL-2B gel filtration column (0.7 ϫ 43 cm) equilibrated with 20 mM HEPES pH 7.9, 10% glycerol, 50 mM KCl, 0.1 mM EDTA, 5 mM MgCl 2 , 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride at a 8-ml/h flow rate, and a total of 70 fractions were collected (300 l each). The HIV-1 and HIV-2 DNA templates were detected predominantly between fractions 22 and 26 with a peak in the amount of DNA seen in fractions 23 and 24. Three of the retained column fractions containing the peak levels of HIV-1 or HIV-2 DNA were pooled as were each of the three fractions surrounding this peak and analyzed by Western blot analysis using antibodies directed against TBP, HIV-1 and HIV-2 Tat, and general transcription factors.
To analyze for DNA content, a 40-l aliquot of each fraction was extracted with phenol-chloroform and loaded onto an 0.8% agarose gel. For Western analysis, column fractions were precipitated with acetone using bovine serum albumin as carrier. Proteins were resolved on an SDS-polyacrylamide electrophoresis gel, blotted onto a nitrocellulose matrix (Hybond C, Amersham), and analyzed with the appropriate antibodies using standard techniques (anti-TBP, Promega; rabbit polyclonal anti-HIV-2 Tat; anti-flag (Kodak) monoclonal for HIV-1 Tat; anti-TFIIH and anti-TFIIE subunits, Santa Cruz Biotechnology). For in vitro transcription reactions, a nucleotide mixture (final concentration: 0.06 mM ATP, CTP, UTP, and 0.006 mM GTP) and 1 l of 10 mCi of [␣-32 P]GTP was added to a 96-l aliquot of each column fraction and the reaction was incubated at 30°C for 1 h. The products were separated using a 6% denaturing polyacrylamide gel and subjected to autoradiography.

Biotinylation and Coupling of HIV-1 and HIV-2 Templates and in Vitro Transcription
Reactions-A DNA fragment containing HIV-2 LTR sequences from Ϫ254 (Asp-718) to ϩ156 (HindIII) was fused to the CAT gene and cloned into pUC19 with Asp-718 and NcoI. An SstI-NcoI fragment was subcloned into pGEM 5Z and an SstI-ApaI fragment was then purified. This fragment was subsequently biotinylated using terminal deoxynucleotidyl transferase (Life Technologies, Inc.) and biotin-dCTP. To select for the specific orientation of the fragment, it was digested with NcoI and the double stranded oligomer from the 5Ј portion of the fragment was removed by centrifugation on a Microcon-30 (Amicon). The HIV-1 LTR construct containing an AvaI (Ϫ160)/HindII (ϩ80) fragment fused to the CAT gene was restricted with AvaI/NcoI and cloned into pGEM 5Z. An SstI/ApaI fragment was isolated and digested with NcoI to select for the specific orientation of the fragment. Alternatively, bio-dCTP tailed HIV-1 and HIV-2 DNA templates (1-3 mg) were incubated with nuclear extract and Tat in the presence of poly(dA-dT) or other competitors to decrease nonspecific binding of TBP to the beads. In this case, the templates were immunoprecipitated with goat anti-biotin antibodies (ϳ1 g, Sigma) for 30 min at 30°C, and bound to protein G-Sepharose 4 fast flow beads (Pharmacia) for 1 h at room temperature.
Transcription reactions were done by mixing 1 g of biotinylated DNA template, 0.5 g of poly(I-C), 25 l of pre-cleared nuclear extract, 50 mM KCl, 5 mM MgCl 2 , 20 mM HEPES pH 7.9, 0.5 mM dithiothreitol, 0.2 mM dATP, and either 200 ng of wild-type or mutant Tat. The reactions were incubated at 30°C for 30 min. Streptavidin-coated magnetic beads (Dynabeads, Dynal) (5 l) pre-equilibrated in buffer (20 mM HEPES pH 7.9, 1 mM dithiothreitol, 5 mM MgCl 2 ) were then added to the reaction and the mixture was further incubated at 30°C for 15 min. The bound templates were then harvested using a magnetic stand and the bound complexes were washed twice with 200 l of the same buffer. Transcription reactions were initiated by the addition of nucleotides (60 M ATP, UTP, CTP and 6 M GTP), 10 Ci of [␣-32 P]GTP, and 10 units of RNasin (Boehringer Mannheim). The reaction was incubated at 30°C for 1 h. To assay in vitro transcription, the labeled RNA products were separated on a 6% denaturing gel and the products were detected by autoradiography. Beads were washed twice with changing of the Eppendorf tubes with 100 l of the nuclear extract buffer. For Western blot analysis, the beads were boiled in Laemmli loading buffer for 5 min, and the proteins were subject to electrophoresis on 15% SDS-polyacrylamide gel electrophoresis. Blotting was done using hybrid-C extra membrane (Amersham). Western blots were analyzed with anti-TBP (Promega), anti-flag M2 (Kodak) for the flag-tagged HIV-1 Tat or rabbit anti-HIV-2 Tat primary antibodies. Horseradish peroxidase-conjugated goat anti-mouse and goat anti-rabbit secondary antibodies and ECL reagent system were also used for the Western analysis. The results of Western blot analysis and in vitro transcription with both the Dynabead and protein G-Sepharose analysis of the biotinylated HIV templates were similar.

RESULTS
Tat Is Associated with the HIV-2 Transcription Complex-Our previous in vitro transcription analysis indicated that the HIV-2 LTR could be strongly activated by both the HIV-1 and HIV-2 Tat proteins (5). The HIV-2 LTR template resulted in higher overall levels of basal and Tat-induced transcription than the HIV-1 LTR so that it was a useful template with which to perform in vitro studies to analyze the assembly of preinitiation complexes. A previous study demonstrated that when the adenovirus E4 promoter with upstream GAL4-binding sites was incubated with HeLa nuclear extract and fractionated on a Sepharose CL-2B column, specific column fractions could be isolated in which this template was stably associated with general transcription factors in a preinitiation complex that could be separated from unbound general transcription factors (45,46). These column purified preinitiation complexes were competent for in vitro transcription indicating that functional transcription complexes could be assembled (45,46).
To determine if a similar system could be used to study the cellular factors comprising HIV transcription complexes, a linearized HIV-2 LTR CAT template was incubated with HeLa nuclear extract and gel filtration chromatography was performed. The column fractions were then analyzed for DNA content, in vitro transcriptional activity, and the presence of different general transcription factors using Western blot analysis. Two protein peaks were detected in fractions eluted from the Sepharose CL-2B column with DNA being present within the first protein peak (Fig. 1A). In vitro transcription of these column fractions was assayed by the addition of unlabeled and labeled ribonucleotides and demonstrated that the transcriptional activity of the HIV-2 LTR correlated with the presence of the DNA (Fig. 1A). No transcriptional activity was present in column fractions obtained from the second protein peak (data not shown). Western blot analysis was then performed with antibodies directed against different general transcription factors to analyze for their presence in pooled column fractions (Fig. 1B). Proteins present in pooled column fractions which were active for HIV-2 transcription contained TBP in addition to other components of the preinitiation complex including both subunits of TFIIE (p57 and p34) and two of the components of TFIIH (p62 and p89) (47). Other components of the basal transcription complex including RNA polymerase II and TATAassociated factors or TAFs (TAF250) were also present in the transcriptionally active fractions isolated from the Sepharose CL-2B column (data not shown). Basal transcription factors were also retained in the column indicating that only a portion of these proteins associated with the HIV-2 template in a preinitiation complex (Fig. 1B).
HIV-2 transcription complexes were next isolated and analyzed to determine whether Tat was able to associate with general transcription factors comprising the HIV-2 preinitiation complex. An HIV-2 LTR template was incubated in the presence of HeLa nuclear extract and the HIV-2 Tat protein and following column chromatography was analyzed for in vitro transcription and Western blot analysis (Fig. 2). Two protein peaks were detected using the different column fractions ( Fig. 2A). Western blot analysis indicated that a portion of TBP and HIV-2 Tat were associated with pooled column fractions containing transcriptionally active HIV-2 templates, while the remainder of the TBP and Tat proteins were retained in later pooled column fractions that were not associated with the HIV-2 template (Fig. 2B). Quantitation of this and other gels by densitometry indicated that approximately 15-20% of the TBP and HIV-2 Tat were associated with the DNA containing fractions. These results indicate that the HIV-2 Tat protein can associate with transcriptionally active HIV-2 preinitiation complexes.
HIV-1 Tat Is Also Associated with the Transcription Complex-Similar experiments were performed to determine whether HIV-1 Tat was also capable of associating with the HIV-1 transcription preinitiation complex. A linearized HIV-1 LTR template was incubated with HeLa nuclear extract and the HIV-1 Tat protein and then fractionated on Sepharose CL-2B as described above. Two protein peaks were present in eluted column fractions and ethidium staining of agarose gels demonstrated HIV-1 DNA in the fractions that were active for in vitro transcription of the HIV-1 LTR (Fig. 3A). No in vitro transcriptional activity from the HIV-1 LTR was detected in the second protein peak (data not shown). Western blot analysis indicated that 19% of the TBP was present in pooled column fractions containing transcriptionally active HIV-1 templates while the remainder of the TBP was detected in fractions which were retained in the column (Fig. 3B). Likewise 15% of the HIV-1 Tat protein was also present in the same column fractions which contained TBP and transcriptionally active HIV-1 templates (Fig. 3B). The remainder of the HIV-1 Tat protein was retained in column fractions that did not contain the HIV-1 template (Fig. 3B). Thus, both the HIV-1 and HIV-2 Tat proteins can associate with active HIV-1 and HIV-2 transcription complexes.
Wild-type but Not Mutant Tat Proteins Are Associated with HIV Transcription Complexes-It was critical to address whether there was specificity for HIV-1 and HIV-2 Tat association with preinitiation complexes. The ability of wild-type HIV-1 and HIV-2 Tat to associate with the HIV-1 and HIV-2 preinitiation complexes was compared to that of mutants in the HIV-1 Tat basic domain (Tat 50/57) and the HIV-2 Tat basic domain (Tat ⌬84) (5,20). The HIV-1 Tat mutant 50/57 substitutes glycine residues for basic amino acids while the HIV-2 Tat mutant ⌬84 is a truncation mutant which deletes several basic amino acids. Previous results indicated that HIV-1 and HIV-2 Tat basic domain mutants failed to activate transcription (5,16,20). The reasons for this may be due to their inability to bind to TAR RNA or to their failure to associate with RNA polymerase II which is a key component of the preinitiation complex (5,16). First, wild-type or mutant HIV-2 Tat proteins were incubated with a linearized HIV-2 LTR CAT template and HeLa nuclear extract followed by column chromatography on a Sepharose CL-2B column using the conditions described in Figs. 1 and 2. Column fractions within both the first protein peak which were active for in vitro transcription and fractions in the second protein peak were analyzed by Western blot analysis with TBP and HIV-2 Tat antibodies. This analysis demonstrated that 24% of the wild-type HIV-2 Tat protein and 19% of the TBP were present in fractions contain- ing HIV-2 DNA (Fig. 4A). However, under identical chromatographic conditions, the HIV-2 Tat basic domain mutant ⌬84 was not associated with fractions containing HIV-2 DNA and was exclusively retained in the column while 20% of TBP was present in the DNA containing column fractions (Fig. 4B).
Similar experiments were then performed with wild-type and mutant HIV-1 Tat proteins and the HIV-1 LTR template. Following column chromatography, 19% of the wild-type HIV-1 Tat protein and 17% of the TBP were present in column fractions in the first protein peak that was active for in vitro transcription of the HIV-1 LTR while the remainder of these proteins were present in retained column fractions (Fig. 4C). In contrast, an HIV-1 Tat basic domain mutant between residues 50 and 57 was not present in column fractions that were active for in vitro transcription of the HIV-1 LTR, although 23% of the input TBP was present in these fractions (Fig. 4D). Thus, the wild-type but not mutant HIV-1 and HIV-2 Tat proteins were associated with transcriptionally active HIV-1 and HIV-2 transcription complexes and this association was dependent on the basic domain of Tat.

Tat Association with HIV-1 and HIV-2 Transcription Complexes Does Not Require TAR RNA Bulge Elements-HIV-1
Tat binds to the HIV-1 TAR RNA bulge structure which extends between ϩ22 and ϩ24 (22, 23) while HIV-2 Tat binds to each of the duplicated HIV-2 TAR RNA bulge structures extending between ϩ27 and ϩ28 and ϩ62 and ϩ63 (5). One possibility to explain the association of Tat with the HIV-1 and HIV-2 preinitiation complexes is that during the incubation of the HIV template with HeLa nuclear extract prior to in vitro transcription analysis that TAR RNA is synthesized due to small amounts of contaminating nucleotides in the extract. The ability of wild-type but not mutant HIV-1 and HIV-2 Tat to bind to this TAR RNA could explain the association of these Tat proteins with the HIV-1 and HIV-2 preinitiation complex. To rule out this possibility, we compared the ability of HIV-1 and HIV-2 Tat to associate with HIV-1 and HIV-2 LTR templates that were mutant in their TAR RNA bulge elements.
An HIV-1 TAR RNA mutant in the bulge at ϩ22 (20) and an HIV-2 mutant in both TAR RNA bulge structures at ϩ27 and ϩ28 and ϩ62 (5) that did not bind to Tat in gel retardation analysis were used in these studies. HIV-1 and HIV-2 templates which transcribe these mutant TAR RNAs were incubated with HeLa nuclear extract and wild-type HIV-1 or HIV-2 Tat proteins followed by column chromatography (Fig. 5). Column fractions from both the first and second protein peaks were analyzed by Western blot using antibodies against TBP, HIV-2 Tat, or flag-tagged HIV-1 Tat. A portion of both HIV-1 Tat (20%) and HIV-2 Tat (12%) and 24 and 33% of the input TBP, respectively, were present in column fractions which contained the transcriptionally active HIV-2 and HIV-1 mutant DNA templates (Fig. 5, A and B). Both TBP and HIV-2 and HIV-1 Tat were also present in the retained column fractions (Fig. 5, A and B). These results indicate that the ability of HIV-1 and HIV-2 Tat to associate with the HIV-1 and HIV-2 transcription complexes was not due to the binding of these Tat proteins to TAR RNA synthesized prior to in vitro transcription analysis.
HIV-1 and HIV-2 Preinitiation Complex Formation Is Necessary for Tat Association-Formation of preinitiation complexes may be regulated by the association of TFIID with the promoter followed by the stepwise association of other general transcription factors or alternatively by the association of TFIID with the RNA polymerase II holoenzyme (34, 36 -39, 47, 48). To determine whether components mediating the assembly of the HIV-1 and HIV-2 preinitiation complex were critical for Tat association, the TBP component of the basal transcription complex was inactivated and the ability of Tat to associate with this complex was assayed. Heat treatment of HeLa nuclear extract at 47°C for 15 min has been demonstrated to inactivate TBP and prevent TFIID activity in in vitro transcription assays (44). Heat treatment of the HeLa nuclear extract prevented in vitro transcriptional activity with the HIV-2 LTR (data not shown). This heat-treated HeLa nuclear extract was then tested for its ability to form HIV-2 transcription complexes that could associate with HIV-2 Tat. HIV-2 Tat did not associate with column fractions containing HIV-2 DNA incubated with heat-treated HeLa nuclear extract, although HIV-2 Tat was present in the retained column fractions (Fig. 6A). To test whether the addition of bacterial produced TBP restored the ability of the heat-treated HeLa nuclear extract to facilitate HIV-2 Tat association, TBP was added to the heat-treated extract and incubated with the HIV-2 LTR and HIV-2 Tat. Following column chromatography, fractions within the first protein peak which contained the HIV-2 DNA template had 12% of the TBP and 10% of the HIV-2 Tat protein in the retained column fractions (Fig. 6B), indicating that the formation of the HIV-2 preinitiation complex was critical for Tat association.
Since TBP binds to the TATA box and is critical for both transcriptional activity and for HIV-2 Tat association with the HIV-2 transcription complex, we next tested whether mutations in the HIV-1 TATA element prevented association of HIV-1 Tat with the HIV-1 transcription complex. Previous studies have demonstrated that the HIV-1 TATA element is critical for both in vitro and in vivo activation by HIV-1 Tat (3,8,9,49). The wild-type HIV-1 template and an HIV-1 TATA mutant were each incubated with HeLa nuclear extract and HIV-1 Tat prior to fractionation on a Sepharose CL-2B column (Fig. 7). Only 5% of the TBP but less than 1% of the HIV-1 Tat was associated with column fractions containing the HIV-1 TATA mutant with the majority of both TBP and Tat being found in the retained column fractions (Fig. 7A). In contrast, column fractions containing the transcriptionally active wildtype HIV-1 LTR template contained 9% of the input TBP and 11% of the input HIV-1 Tat (Fig. 7B). These results indicate that an intact HIV-1 TATA element and functional TBP are critical for the formation of HIV-1 and HIV-2 preinitiation complexes that can associate with the HIV-1 and HIV-2 Tat proteins. However, this analysis does not address whether TBP, TAFs, the RNA polymerase II holoenzyme or additional factors are the targets for Tat association.
Tat Association with the HIV-1 and HIV-2 Preinitiation Complex Results in Increased Levels of in Vitro Transcription-The studies described in the previous sections demonstrated that the HIV-1 and HIV-2 Tat proteins were able to associate with HIV transcription complexes. Since this association required the presence of TBP and the TATA element and occurred prior to transcriptional initiation, it was likely that HIV-1 and HIV-2 Tat associated with the HIV-1 and HIV-2 preinitiation complexes. However, these studies did not address whether association of HIV-1 and HIV-2 Tat with the HIV-1 and HIV-2 preinitiation complexes led to increased levels of in vitro transcription (13, 30, 49 -52). To determine whether HIV-1 and HIV-2 Tat association with the preinitiation complex increased in vitro transcription, HIV-1 and HIV-2 LTR CAT fragments were first biotinylated and incubated with HeLa nuclear extract in the presence of wild-type or mutant Tat proteins. The templates were then coupled to streptavidin magnetic beads, washed extensively, and in vitro transcription was performed (45,53,54). The run-off transcripts were analyzed following gel electrophoresis and autoradiography. The beads were extensively blocked with bovine serum albumin and nonspecific competitor oligonucleotides were included during the incubation of the HIV-1 and HIV-2 templates with HeLa nuclear extract so that neither TBP nor HIV-1 and HIV-2 Tat gave background binding to the beads in Western blot analysis (data not shown).
HIV-2 Tat increased the level of transcription from the immobilized HIV-2 LTR when it was associated with components of the preinitiation complex (Fig. 8A). Similar levels of HIV-2 transcription were detected when HIV-2 Tat was present during both the formation of the preinitiation complex and the in vitro transcription reaction (Fig. 8A, lane 1) or when present only during the formation of the preinitiation complex (Fig. 8A,  lane 2). In contrast, the level of transactivation of the HIV-2 LTR was reduced when HIV-2 Tat was present only during the in vitro transcription reaction (Fig. 8A, lane 3). Only basal levels of in vitro transcription from the HIV-2 LTR were detected in the presence of the HIV-2 Tat protein ⌬84 (Fig. 8A,  lane 4). These results suggested that HIV-2 Tat associated with the HIV-2 preinitiation complex to increase in vitro transcription. Next we addressed whether wild-type or mutant HIV-2 FIG. 6. TBP is critical for the association of HIV-2 Tat with the HIV-2 preinitiation complex. A, HeLa nuclear extract (400 g) was heat-treated at 47°C for 15 min and then used to form preinitiation complexes with the HIV-2 LTR (5 g) and HIV-2 Tat (400 ng). B, these components were also assayed following the addition of 3 footprinting units of recombinant human TBP to the heat-treated HeLa extract. Following chromatography on a Sepharose CL-2B column, each fraction was analyzed for the presence of the DNA template. The pooled DNA containing and retained column fractions are indicated. Western blot analysis using antibodies directed against TBP and HIV-2 Tat were performed on the different column fractions and the recombinant control proteins. Tat proteins were directly associated with the immobilized HIV-2 template. Western blot analysis was performed using antibodies to detect TBP (Fig. 8B, lane 1) and wild-type and mutant HIV-2 Tat control proteins (Fig. 8B, lanes 2 and 3). TBP was associated with the HIV-2 template in the presence of HeLa nuclear extract (Fig. 8B, lanes 4 and 5). Wild-type HIV-2 Tat but not HIV-2 Tat ⌬84 associated with the HIV-2 preinitiation complex (Fig. 8B, lanes 4 and 5). Thus, the association of HIV-2 Tat with the HIV-2 preinitiation complex correlated with increases in in vitro transcription.
Similar experiments were also performed with a biotinylated HIV-1 LTR template which was incubated with HeLa nuclear extract and HIV-1 Tat prior to coupling to streptavidin magnetic beads and extensive washing. The presence of wild-type HIV-1 Tat but not the HIV-1 Tat mutant 50/57 increased the level of in vitro transcription from the immobilized HIV-1 template (Fig. 9A, lanes 1 and 2). Western blot analysis with antibodies against TBP and HIV-1 Tat demonstrated that TBP was able to associate with the HIV-1 preinitiation complex in the presence of HeLa nuclear extract (Fig. 9B, lanes 4 and 5). Wild-type HIV-1 Tat also associated with the HIV-1 preinitiation complex in contrast to the results with HIV-1 Tat 50/57 (Fig. 9B, lanes 4 and 5). These results indicate the direct association of HIV-1 Tat with the preinitiation complex was involved in increasing the level of in vitro transcription from the HIV-1 LTR. DISCUSSION HIV-1 and HIV-2 Tat proteins activate gene expression primarily by increasing the level of transcriptional elongation though effects on transcriptional initiation have also been noted (5, 8, 10, 13, 26 -28, 30 -33, 49, 50 -52, 55). Both HIV-1 and HIV-2 Tat require an RNA element downstream of the transcription initiation site known as TAR which is capable of binding HIV-1 and HIV-2 Tat (22,23) and cellular factors including RNA polymerase II (16). In the absence of HIV-1 and HIV-2 Tat, only short or nonprocessive transcripts are synthesized from the HIV-1 and HIV-2 LTR (26 -29, 49, 51) while in the presence of Tat processive transcripts are synthesized (27, 28, 30 -32, 49, 51, 52, 55). However, deletion of TAR does not result in the synthesis of processive transcripts indicating that TAR itself does not function as a classic transcription terminator (26). Both HIV-1 and HIV-2 Tat may regulate a process known as promoter clearance in which RNA polymerase II stalls at short distances from the transcription initiation site and efficient transcription elongation results from the action of viral or cellular regulatory proteins such as Tat (5,41,56,57). This mechanism could explain the dual effects of HIV-1 and HIV-2 Tat on increasing both transcription initiation and elongation.
A variety of transcriptional activator proteins such as VP16 can stimulate both transcriptional initiation and elongation when fused to the DNA binding domain of GAL4 and bound to upstream promoters elements (58). Both HIV-1 and HIV-2 Tat primarily activate transcriptional elongation when delivered to the promoter via upstream binding sites (58,59). It is hypothesized that a variety of transcriptional activator proteins may be involved in recruiting cellular elongation factors or stimulating the activity of general transcription factors such as TFIIH to facilitate transcriptional elongation (47,52,58). Thus HIV-1 and HIV-2 Tat association with the preinitiation complex may facilitate the interaction with the same cellular factors that Tat interacts with when introduced to the promoter via GAL4-binding sites. The conclusion that HIV-1 and HIV-2 Tat interacts with components of the preinitiation complex is also supported by previous observations that distinct upstream promoter elements in the HIV-1 and HIV-2 LTR elements are critical for efficient Tat activation (3)(4)(5)(6)(7)(8)(9)(10)(11)(12). For example, SP1binding sites in the HIV-1 LTR increase the level of HIV-1 Tat activation to a much greater degree than other upstream binding sites which were inserted in place of the SP1-binding sites (58,59). Furthermore, substitution of TATA elements from other viral and cellular promoters in place of the HIV-1 TATA element results in little change in basal transcription but markedly reduces HIV-1 Tat activation (8,9,49,58). This suggests that factors binding to upstream regulatory elements in the HIV-1 and HIV-2 LTR are critical for efficient Tat activation.
To address the steps involved in HIV-1 and HIV-2 Tat activation of transcription, we determined at what point Tat could first become associated with the HIV-1 and HIV-2 transcription complex. Two approaches were used including the analysis of Tat association with column purified HIV-1 and HIV-2 preinitiation complexes and immobilized HIV-1 and HIV-2 templates. The analysis of column purified transcription complexes demonstrates that both the HIV-1 and HIV-2 Tat proteins are associated with active HIV transcription complexes. Column chromatography has previously been used to demonstrate that transcriptionally active preinitiation complexes can be isolated which contain cellular factors necessary for both transcriptional initiation and elongation (45,46). Cellular factors associated with RNA polymerase II, basal transcription factors, or TBP and TAFs are potential candidates for mediating HIV-1 and HIV-2 Tat association with the preinitiation complex. Although we do not address the cellular targets by which HIV-1 and HIV-2 Tat becomes associated with the preinitiation complex, our recent demonstration that HIV-1 and HIV-2 Tat can associate with the core RNA polymerase II would be consistent with the RNA polymerase II holoenzyme as a potential target for Tat association (16).
The RNA polymerase II holoenzyme contains RNA polymerase II and associated cellular factors that are required for activator-mediated transcription although it contains little or no TFIID and TFIIB and reduced levels of TFIIE and TFIIH  (34). Association of these factors with the preinitiation complex increases the levels of activator-mediated transcription. The requirement for TFIID and an intact TATA element for HIV-1 and HIV-2 Tat association with the transcription complex is consistent with a requirement for the assembly of a functional preinitiation complex for Tat association (48). HIV-1 Tat has previously been shown to associate with several components comprising the HIV-1 preinitiation complex including SP1 (60), TATA-binding protein (TBP) and associated factors (TAFs) (61), the p62 subunit of TFIIH (58) in addition to RNA polymerase II (16). Further studies will be needed to ascertain which step in the assembly of the preinitiation complex that Tat association occurs and whether Tat interacts with one or more previously characterized cellular targets or in fact may interact with novel factors.
The ability of HIV-1 and HIV-2 Tat to associate with preinitiation complexes assembled on immobilized HIV-1 and HIV-2 templates leads to increased levels of in vitro transcription. This result suggests that Tat association with the preinitiation complex is of functional significance. Since HIV-1 and HIV-2 Tat remain associated with HIV-1 and HIV-2 preinitiation complexes following extensive washing and is able to transactivate these promoters, it is likely that a number of cellular factors necessary for increasing HIV-1 and HIV-2 transcription directly associate with the preinitiation complex. These results would be consistent with a model in which the structure of the HIV-1 and HIV-2 LTR may preferentially assemble preinitiation complexes that are able to efficiently associate with HIV-1 and HIV-2 Tat (8). HIV-1 and HIV-2 Tat may exert their effects on increasing transcription at two separate steps including their association with the preinitiation complex and following their binding to TAR RNA.
Recent studies indicate that the HIV-1 Tat protein directly associates with the HIV-1 transcriptional elongation complex (33). During transcriptional elongation there is a dynamic recycling of factors bound to the promoter and the elongation complex (62). For example, TFIID remains associated with the DNA template while TFIIB is released from the promoter after transcriptional initiation and later reassociates to reinitiate transcription. Other basal transcription factors dissociate from the elongation complex, although TFIIF can reassociate with the transcription elongation complex (62). Given these results, we would suggest that HIV-1 and HIV-2 Tat associate with the preinitiation complex and subsequently that Tat binds to TAR RNA while RNA polymerase II is paused during its passage through TAR RNA. This process may facilitate the association of the HIV-1 and HIV-2 Tat proteins with additional components of the transcriptional elongation complex (52,63) to trigger a process such as phosphorylation of the RNA polymerase II CTD (64) or the recycling of transcription factors (62) to increase the processivity of RNA polymerase II. Identification of factors comprising both the transcriptional initiation and elongation complexes that interact with the HIV-1 and HIV-2 Tat proteins will be critical for a better understanding of their ability to activate HIV-1 and HIV-2 transcription.