Regulation of transcription of human immunodeficiency virus type-1 (HIV-1) requires specific interaction of Tat protein with the trans-activation response region (TAR). Inhibition of replication of HIV-1 has previously been achieved with a TAR decoy, namely a short RNA oligonucleotide that corresponded to the sequence of the authentic TAR RNA. Since TAR RNA has the potential to interact with cellular factors, we examined the effect of TAR RNA on efficiency of transcription in nuclear of HeLa cell extracts. We performed an in vitro transcription assay in the presence of authentic TAR RNA using a template that was driven by the CMV (cytomegalovirus) early promoter in a HeLa nuclear extract and found, for the first time, that TAR RNA inhibited transcription by ~60-70% independently of the Tat-TAR interaction. Furthermore, we evaluated inhibition of transcription by variants of TAR RNA and found that the TAR RNA loop, bases surrounding the loop, the triple base bulge and the `lower' stem region of TAR RNA were responsible for the inhibition of transcription. Taken together, earlier reports on proteins that bind to TAR RNA and the present results suggest that integrity of TAR RNA is important for efficient binding to cellular transcription factors. As judged from the significant inhibition observed in this study, the TAR decoy might sequester transcription factors and thus it might potentially be able to inhibit transcription of housekeeping genes that are unrelated to Tat function.
Expression of genes encoded by human immunodeficiency virus type-1 (HIV-1) is regulated by interaction of cellular factors and viral trans-activator protein, Tat, with specific regulatory elements in the HIV-1 long terminal repeat (LTR) (1 ). The HIV-1 regulatory protein Tat binds to one of the regulatory elements in the LTR region called the trans-activating response region (TAR) (2 -4 ). TAR is an RNA element consisting of 59 nt and is the minimal motif sufficient for formation of the stable hairpin structure that is responsible for binding of Tat in vivo (2 ,5 -7 ). Tat stimulates transcription substantially after binding to TAR RNA (8 -10 ). Deletion studies of TAR RNA revealed that `bulge' residues are required for both specific binding to Tat and trans-activation, whereas loop sequences are necessary for trans-activation but are not essential for binding of Tat in vivo (5 ,11 -15 ). Studies in vitro showed that the three base bulge (UCU or UUU) and two specific base pairs adjacent to the bulge of TAR RNA constitute the core elements for binding of Tat (15 -17 ). Tat has two major domains, a cysteine-rich region and a highly basic region. The latter region is specific for binding to TAR RNA. The product of the tat gene not only plays a key role in trans-activation of HIV-1 genes but also exerts a variety of effects on growth and metabolism of the host cell (18 ,19 ). In addition, it was recently shown that Tat is important for efficient reverse transcription of HIV-1 (20 ). Since Tat protein has diverse functions in the life cycle of HIV-1, as well as in viral proliferation, it is an important and attractive target in efforts to develop weapons against HIV.
Several genetic strategies have been examined in attempts to repress proliferation of HIV. Trans-dominant proteins, single chain antibodies, antisense molecules, ribozymes, decoys (for a review see 21 ) and use of the LTR of HIV to produce inducible and toxic gene products have all been tested in cells infected by HIV (22 ). Combinations of these strategies (for example a ribozyme and a decoy) have also been examined (23 ,24 ). Although expression and regulation of such therapeutic molecules might be possible in vivo, their constitutive expression could lead to cellular toxicity or to an immune response by the host against the engineered cells. This problem is especially significant in the case of toxins and suicide genes. Among various RNA-based strategies against HIV infection, the decoy strategy has a potential advantage over the use of other RNA inhibitors, such as short antisense RNAs and ribozymes, because generation of escape mutants might be less frequent: alterations in Tat or Rev (HIV-1 protein) that prevent binding to a decoy would also prevent binding to native elements [such as the Rev responsive element (RRE) and TAR sequences]. Both RRE and TAR RNAs have been exploited as decoys and found to inhibit replication of HIV by as much as 80-97% (25 -27 ).
Although decoys might act as much more efficient inhibitors, with possible Ki values in the subnanomolar range, than other molecules such as antisense RNAs and ribozymes, decoys might potentially be toxic to cells if they were to sequester cellular factors, in particular when the decoy RNA happens to include regions that can interact with cellular proteins. Several previous studies showed that cellular factors, such as TRP-185 (28 ), Tat-SF1 (29 ), polymerase II (28 ,30 ) and others (31 -33 ), bind efficiently to TAR RNA. However, to the best of our knowledge the effects of the TAR RNA sequence on the cellular machinery have not yet been examined. In the present study we performed in vitro cell-free transcription assays and demonstrate that authentic TAR RNA inhibits transcription that is independent of the Tat-TAR interaction. Furthermore, we identify important regions of TAR RNA responsible for inhibition of transcription, namely the loop, residues surrounding the loop, the triple base bulge and the lower stem region of TAR RNA.
Oligodeoxyribonucleotide templates containing the T7 promoter and sequences that correspond to the RNAs shown in Figure 1 were synthesized with a DNA synthesizer (model 392A; Applied Biosystems, USA). In the presence of the reverse primer (5'-GGGTTCCCTAGTTAGCCAGA-3') single-stranded DNA oligonucleotides were converted to double-stranded DNA by Taq DNA polymerase (Nippon Gene, Japan). The reaction was carried out in a 100 [mu]l mixture that contained 10 mM Tris-HCl, pH 8.8, 50 mM KCl, 1.5 mM MgCl2, 0.1% Triton X-100, 0.2 mM dNTPs, 100 pmol reverse primer, 78 pmol DNA oligonucleotide and 2.5 U Taq DNA polymerase. The reaction mixture was cycled at 94oC for 1 min, 45oC for 1 min and 68oC for 2 min, until a product of the desired size was obtained. The resulting double-stranded DNA template was precipitated in ethanol and transcribed by T7 RNA polymerase to generate TAR RNA or mutant TAR RNAs. Transcription in vitro was completed during incubation at 37oC for 2 h using a T7 Ampliscribe kit (Epicentre Technologies, USA). After synthesis of RNAs and treatment with DNase I, reaction mixtures were fractionated by electrophoresis on a 10% denaturing polyacrylamide gel. RNAs were extracted and recovered from the gel after ethanol precipitation.
In order to investigate the effects of TAR RNA on the cellular machinery at the transcription level we used a CMV (cytomegalovirus) immediate early promoter that either contained or lacked enhancer elements. We chose CMV DNA as the template, as an example, for evaluation of the effect of TAR RNA on LTR- independent transcription of a template. The CMV early promoter region (from nt -238 to 364) was amplified by Taq DNA polymerase (Takara, Japan) with specific primers (5'-TTAGTCATCGC TATTACCATGG-3' and 5'-AGGCCTGGATTCACAGGACGGGTG-3') by PCR (94oC for 3 min, 50oC for 1.15 min and 72oC for 3 min, 30 cycles). The resulting product of PCR (602 nt) was recovered by ethanol precipitation and used in the transcription assay. The transcription reaction was carried out with a HeLa nuclear extract (Promega, USA) in the presence of [[alpha]-32P]CTP. Initially, 13 U HeLa nuclear extract, 3 mM MgCl2, 0.4 mM each ATP, GTP and UTP and 16 [mu]M CTP plus 10 [mu]Ci [[alpha]-32P]CTP (3000 Ci/mmol; Amersham) were combined in buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 0.2 mM EDTA, 0.5 mM DTT and 20% glycerol), mixed with 100 pmol either TAR RNA or a variant or yeast total tRNA (Boehringer, Germany) and allowed to equilibrate for 15 min at 30oC. To this reaction mixture was added 100 ng template DNA for PCR to give a final reaction volume of 25 [mu]l, with further incubation at 30oC for 45 min. The reaction was terminated by addition of 175 [mu]l stop solution (0.3 M Tris-HCl, pH 7.4, 0.3 M sodium acetate, 0.5% SDS, 2 mM EDTA and 3 [mu]g/ml tRNA) and the products extracted once with phenol and chloroform before ethanol precipitation. The newly synthesized RNAs were denatured in loading buffer (25 mM EDTA and 4.5 M urea) at 90oC for 5 min, loaded on a 6% polyacrylamide gel containing 7 M urea and fractionated by electrophoresis. Bands on the gel were quantitated with an image analyzer (BAS 2000; Fuji Film, Japan).
In vitro transcription was also performed with a nuclear extract of Jurkat cells (Santa Cruz Biotechnology, USA) using a CMV early promoter that contained enhancer elements, either in the presence or absence of authentic TAR RNA. All other conditions were similar to those described above for the HeLa nuclear extract.
A Tat-derived peptide (CQ, amino acids 37-72) was titrated against 5'-labeled TAR and its variants in 8 [mu]l binding reactions [10 mM Tris-HCl, pH 8.0, 70 mM NaCl, 2 mM EDTA, 40 nM total tRNA from yeast (Boehringer) and 0.01% Nonidet P-40 (Shell Chemicals)]. The peptide was chemically synthesized and purified and the composition of the preparation was analyzed by HPLC on a reverse phase column. Initially, each labeled TAR RNA was denatured at 94oC for 2 min and allowed to equilibrate at room temperature for 10 min before mixing with various concentrations of the CQ peptide. The mixtures were incubated at 30oC for 1 h and the complex and free RNAs were separated by PAGE on a 15% non-denaturing gel. The amount of complex was quantified directly on the gel with the image analyzer.
Several studies have shown that cellular factors bind specifically to TAR RNA and play a crucial role in trans-activation (28 -33 ). Studies both in vitro and in vivo with LTR-based templates and Tat have suggested that addition of exogenous TAR RNA and/or overexpression of TAR RNA can significantly inhibit trans- activation. Such significant inhibition might originate from a combination of two effects; (i) the expressed or added TAR RNA might act as a decoy by sequestering Tat and interfering directly with binding of Tat to the TAR in the LTR region, transcribed from LTR templates; (ii) the TAR decoy might sequester transcription factors together with other important proteins, such as RNA polymerase II, that are unrelated to the Tat-TAR interaction. Since earlier decoy studies relied on LTR-based vectors, it was not possible to distinguish between the two effects.
In order to differentiate between the two possible effects of TAR RNA, we used an in vitro transcription assay with HeLa nuclear extract, which has been routinely used to study Tat-mediated trans-activation of HIV-1 genes (34 ). This method can provide insight into Tat-mediated trans-activation and also allows us to screen various inhibitors that might interfere with Tat-TAR interactions. To examine the interaction between TAR RNA and cellular factors, we performed transcription assays using a CMV early promoter-based template in the presence and absence of authentic TAR RNA (Fig. 1 ).
As seen in Figure 2 A, which shows a representative autoradiogram, the basal level of transcription (lane 1) from the CMV early promoter was greatly affected by addition of 100 pmol authentic TAR RNA (lanes 3 and 4). Quantification of the results of four independent transcription experiments in the presence of TAR RNA revealed that transcription was inhibited by 60-70% (Fig. 2 B). In contrast, after addition of a similar amount of tRNA (total tRNA from yeast) the basal level of transcription remained either unaffected or was marginally affected (~10-20%; see Fig. 2 A and B). These results demonstrate that factors that are important in transcription bind to TAR RNA and, therefore, that the TAR decoy can inhibit transcription in vitro.
Although the concentration of TAR RNA (100 pmol) tested in the above studies in vitro might be achievable in vivo by use of various expression vectors, we were interested in determining the minimum concentration of TAR RNA for inhibition. Therefore, we performed the in vitro transcription reaction in the presence of concentrations of TAR RNA of from 0.01 to 100 pmol. As seen in Figure 3 , inhibition of transcription depended directly on the concentration of TAR RNA. As expected, lower concentrations (0.01-1.0 pmol) of TAR RNA resulted in moderate inhibition (~30%), while a 10 pmol concentration of TAR RNA caused a similar inhibition of transcription (60-70%) to that observed in the presence of 100 pmol TAR RNA (data not shown). Although the TAR RNA concentrations used here appear to be on the high side, nevertheless, similar concentrations of TAR RNA have been used in the past to observe TAR decoy effects in cell-free transcription assays (35 ). As observed in Figure 2 A and B, exogenous addition of 100 pmol tRNA produced either no affect or only a marginal affect (data not shown). These results clearly demonstrated that inhibition of transcription by TAR RNA depended on the concentration of the RNA.
In order to identify the regions of TAR RNA that are responsible for interactions with cellular transcription factors in vitro, we synthesized and tested four variants. Mutant TAR-1 RNA with altered bases in the loop, mutant TAR-2 RNA with a substituted base pair (mutated bases are boxed), mutant TAR-3 RNA with a deletion of two bulge bases and mutant TAR-4 with deletion of the lower stem region are shown in Figure 1 . These variants were initially tested for their ability to bind to a Tat-derived peptide (CQ peptide; see Materials and Methods). Only TAR RNA variants with either deletion or substitution of conserved bases (such as TAR-2 and TAR-3 RNA) had significantly reduced affinity for the CQ peptide (Fig. 4 ). Both TAR-1 and TAR-4 bound efficiently to the CQ peptide (Fig. 4 ). However, the affinity for the CQ peptide of TAR-4 RNA was lower than the affinity of authentic TAR RNA. From these results it appeared, in accord with previous results (16 ,17 ), that binding ability to the CQ peptide was abolished only when the conserved residues in TAR RNA were replaced or missing.
All the studies described above were performed with a CMV promoter that lacked an enhancer element. Deletion of enhancer elements reduced the efficiency of transcription from the CMV promoter. One might argue, then, that inhibition by TAR RNA of transcription was observed only because of reduced efficiency of transcription from the CMV promoter (in particular in the absence of enhancer elements). In order to examine this possibility, we used a CMV promoter that contained enhancer elements. We again observed ~50-60% inhibition in the presence of authentic TAR RNA (Fig. 6 , lane 2). On the other hand, the presence of tRNA (100 pmol) during transcription remained without affect, as observed above (data not shown). Similarly, we also examined the effects of TAR RNA on transcription from other promoter elements, such as those of thymidine kinase (tk). Unfortunately, the efficiency of transcription from the tk promoter in HeLa nuclear extracts was very low in vitro, at least with our preparation of nuclear extract. Nevertheless, our studies with the CMV promoter suggest that TAR RNA interfered to the same extent with transcription, irrespective of its efficiency.
It was important to determine next whether inhibition of transcription by TAR RNA was limited exclusively to HeLa nuclear extracts. To address this question, we performed in vitro transcription reactions using a CMV promoter containing enhancer elements and a nuclear extract of Jurkat cells. The amount of transcript synthesized was significantly lower in the presence of authentic TAR RNA (Fig. 7 , lane 2) than in the control (Fig. 7 , lane 1). The synthesis of transcript in the nuclear extract of Jurkat cells was less efficient than that in HeLa extracts with either a CMV promoter containing enhancer elements or the tk promoter (data not shown). Despite the fact that our nuclear extract of Jurkat cells only weakly supported transcription in vitro, the extent of inhibition of transcription by authentic TAR RNA was similar to that observed in HeLa nuclear extracts.
In conclusion, the results obtained in this study clearly demonstrate, for the first time, that authentic TAR RNA inhibits transcription from a template based on the CMV early promoter that is not related to the Tat-TAR interaction. The integrity of authentic TAR RNA is important for interaction with several cellular factors. TAR RNA inhibited transcription in nuclear extracts of human cell lines, namely Jurkat and HeLa cells. The results of the present study can be summarized as follows: (i) host cellular proteins bind to authentic TAR RNA even in the absence of Tat protein; (ii) earlier published results on TAR decoy-mediated inhibition of Tat-dependent trans-activation, either in vitro or in vivo, might have been due, at least in part, to inhibition of transcription rather than exclusively inhibition of Tat function; (iii) it is possible that the TAR decoy functions as a decoy for some transcription factors or important proteins of the cell; (iv) since cellular factors were able to bind to TAR RNA in the absence of Tat, TAR RNA decoys might potentially be toxic to cells; (v) for specific inhibition of Tat function of HIV-1 in vivo, the present authentic TAR decoy might not be the most suitable antagonist. Clearly, in order to develop more specific inhibitors of Tat we must design an alternative to TAR or engineer changes in authentic TAR RNA: a second generation of Tat decoy needs to be developed in the wake of the present results.
It is now important to document in vivo inhibition of transcription by TAR RNA, since the concentrations of TAR used in the present study should also be achievable in cells. Sullenger et al. (26 ) found that expression of TAR RNA had no adverse effects on the cell and, in particular, on growth rate. We are now directly examining the levels of expression of housekeeping genes in HeLa cells in the presence and absence of TAR RNA. In addition, it is also important to systematically analyze whether previously identified transcription factors can relieve inhibition by TAR RNA in nuclear extracts upon addition of each individual factor exogenously.
*To whom correspondence should be addressed. Tel: +81 298 54 6085; Fax: +81 298 54 6095; Email: pkrkumar@nibh.go.jp
REFERENCES