Gene Therapy (1998) 5, 16651676 1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.

uk/gt

High level inhibition of HIV replication with combination RNA decoys expressed from an HIV-Tat inducible vector
C Fraisier1, A Irvine2, C Wrighton2, R Craig3 and E Dzierzak1
1

Erasmus University Rotterdam, Department of Cell Biology and Genetics, Rotterdam, The Netherlands; 2Cobra Therapeutics Ltd, and 3Therexsys Ltd, The Science Park, University of Keele, UK

Intracellular immunization, an antiviral gene therapy approach based on the introduction of DNA into cells to stably express molecules for the inhibition of viral gene expression and replication, has been suggested for inhibition of HIV infection. Since the Tat and Rev proteins play a critical role in HIV regulation, RNA decoys and ribozymes of these sequences have potential as therapeutic molecular inhibitors. In the present study, we have generated several anti-HIV molecules; a tat-ribozyme, RRE, RWZ6 and TAR decoys and combinations of decoys, and tested them for inhibition of HIV-1 replication in vitro. We used T cell specic CD2 gene elements and regulatory the HIV inducible promoter to direct high level expression and a 3 UTR sequence for mRNA stabilization. We show that HIV replication was most strongly inhibited with the

combination TAR+RRE decoy when compared with the single decoys or the tat-ribozyme. We also show that the Tat-inducible HIV promoter directs a higher level of steadystate transcription of decoys and inhibitors and that higher levels of expression directly relate to increased levels of inhibition of HIV infection. Furthermore, a stabilization of the 3 end of TAR+RRE inhibitor transcripts using a -globin 3 UTR sequence leads to an additional 15-fold increase in steady-state RNA levels. This cassette when used to express the best combination decoy inhibitor TAR+RRE, yields high level HIV inhibition for greater than 3 weeks. Taken together, both optimization for high level expression of molecular inhibitors and use of combinations of inhibitors suggest better therapeutic application in limiting the spread of HIV.

Keywords: HIV; RNA decoys; ribozyme; HIV-inducible vector; gene therapy

Introduction
Over the past 10 years, molecular approaches to combat HIV infection have been suggested as potential clinical therapies and are generally referred to as intracellular immunization.1 This method relies on the introduction of DNA into the cells which can be, or are already, HIVinfected. DNA constructs thought to be useful in intracellular immunization strategies are those which express inhibitor or antiviral molecules such as: dominant negative regulatory proteins or receptors; RNA inhibitors such as antisense molecules or ribozymes; RNA decoy; and anti-HIV-specic antibodies. With some exceptions, most of the antiviral strategies are based on molecules encoded by HIV itself, mainly the regulatory or structural proteins. Tat and Rev are two of the HIV regulatory proteins that are critical to HIV replication and have been most extensively characterized, thus offering great potential for molecular inhibition strategies.2 The Tat regulatory protein acts as a potent transactivator of HIV long terminal repeat (LTR) directed transcription. Tat binds directly to the Tat activation region (TAR) which is located immediately 3 of the site of initiation of
Correspondence: EA Dzierzak, Erasmus University Rotterdam, Department of Cell Biology and Genetics, PO Box 1738, 3000 DR Rotterdam, The Netherlands Received 20 May 1998; accepted 20 July 1998

HIV-1 transcription3 and increases both the initiation and elongation of HIV RNAs. Tat-mediated transactivation of HIV promoter-driven transcription has been shown to increase reporter gene expression 10- to 1000-fold.49 While the HIV-1 Tat protein binds with high afnity to both HIV-1 and HIV-2 TAR RNAs and consequently can transactivate both HIV-1 and HIV-2 LTRs, the HIV-2 Tat protein binds with high afnity to and transactivates only its own HIV-2 LTR.10,11 Rev is a small nuclear regulatory protein expressed from multiply spliced HIV RNA. In contrast to Tat, Rev acts post-transcriptionally by binding to a cis-acting RNA sequence, the Revresponse element (RRE). The Rev-RRE interaction strongly enhances HIV1 production by facilitating the extranuclear transport of unspliced (9 kb) and single spliced (4 kb) mRNAs that encode HIV-1 structural proteins.12 Since the Tat and Rev regulatory proteins play a critical role in the HIV infectious cycle, protein-based gene therapy strategies with transdominant mutants have been developed by several laboratories. These studies demonstrate that Tat and Rev mutant proteins inhibit HIV replication in lymphoid cell lines1316 as well as in CD4+ primary hematopoietic cells15,17,18 and CD34+ hematopoietic precursors.19,20 However, such mutant proteins if expressed in in vivo gene therapy applications may elicit an immune response against themselves. Thus, RNAbased inhibition strategies have been proposed2 so as to eliminate the risk of rejection.

Regulated expression of HIV molecular inhibitors C Fraisier et al

1666

RNA decoy strategies use short RNA oligonucleotides which mimic critical regulatory sequences in HIV. Decoys such as TAR21 and RRE22 could inhibit the action of Tat and Rev regulatory proteins in HIV replication by sequestering these RNA binding proteins. TAR RNA from HIV-1 forms a stable stem-loop structure (from +1 to +60), the maintenance of which is critical for Tatmediated transactivation.23,24 The HIV-2 TAR element is more complex than HIV-1 TAR and the secondary structure consists of an additional loop within the predicted Tat recognition sequence. The requirement for recognition of two such loops by HIV-2 Tat may explain the incomplete reciprocal transactivation by HIV-1 and HIV2 Tat.25,26 Located within the env gene is the HIV-1 RRE. RRE RNA is predicted to form a central stem-loop and ve stem-loop structures.27,28 Biochemical analyses have identied a high afnity Rev-binding site in stem-loop II of HIV-1 RRE.29,30 To date, several studies have shown the inhibition of HIV replication by the expression of either TAR or RRE RNA decoy sequences.21,22,31,32 Another category of RNA inhibitor consists of HIV RNA specic ribozymes. Ribozymes are described as catalytic RNA molecules capable of recognizing and cleaving a specic target RNA.33 Classically dened, hammerhead ribozymes are RNA molecules that hybridize to complementary RNA sequences in which the central part of the sequence forms a specic secondary structure where reactive groups are located that mediate specic cleavage of the target RNA at a consensus GUC target.34 Ribozymes are potent genetic therapies against HIV, as they cleave both incoming HIV genomic RNA and newly transcribed viral mRNA.35 HIV-specic ribozymes targeting the 5 U5 region,36,37 Tat,38,39 Rev, Env40 or Gag41 have been made and have been shown to inhibit HIV replication. To optimize the inhibiting effects of the protein-based and RNA-based strategies, many studies have used combinations of these diverse molecules: Tat and Rev transdominant mutants,14,42 transdominant mutant and RNA decoy43 and RNA decoy and ribozyme.44,45 In exploring the potential of RNA-based strategies for inhibiting the spread of HIV for possible clinical applications, we tested several molecular inhibitors; RNA decoys corresponding to RRE, RWZ6 (a part of RRE stemloop II which binds three molecules of Rev)46 and TAR and a combination of these. We also tested the inhibiting potential of a hammerhead ribozyme targeted to the entire Tat RNA. The aim of our study was to determine the potency of these inhibitors using CD2-based expression vectors designed to give high level virusinducible and T lymphocyte-specic expression. We present data showing that Tat-inducibility together with mRNA stabilization provide the highest level of expression of inhibitor molecules. Moreover, we demonstrate for the most effective inhibitor, the combination TAR+RRE decoy, a clear correlation between increased expression level and increased inhibition of HIV replication.

ecular constructs alone or in combination would efciently inhibit HIV production; (3) to attain the highest expression of the best single or combination HIV molecular inhibitor and determine quantitatively whether higher expression levels correlate with greater inhibition of HIV.

Results
The aims of our study were: (1) to construct efcient expression cassettes, including a Tat-inducible expression cassette, for the high level transcription of HIV RNAbased molecular inhibitors; (2) to determine which mol-

Generation of expression cassettes, HIV inhibitors and stably transfected cell lines Generally, promoters directing ubiquitous and constitutive transcription have been utilized for expression of HIV inhibitors.14,21,22,36,3840,42,43,45 However, control elements known as locus control regions (LCRs) have been shown to confer high level tissue specic, chromosomal position independent expression.47 In the case of the human -globin locus LCR, high level, erythroid-specic expression is obtained.48 Likewise the human CD2 (hCD2) gene LCR yields high level, sustained copy number dependent T cell-specic gene expression which is independent of the chromosomal integration site of the transgene.49 Thus, the use of strong LCRs for high level tissue specic transcription holds great promise for sustained expression of transgenes. Since CD4 T cells are the predominant target of HIV infection we used the hCD2 gene promoter, 3 untranslated region (UTR) and LCR as our rst generation expression vector: pVA-CD2 described by Zhumabekov et al.50 In a strategy to obtain inducible higher level expression of HIV inhibitory molecules, an HIV-Tat inducible vector was constructed by substitution of the hCD2 promoter by the HIV-2 LTR to yield pVA-HIV. To increase steady state levels of RNA further, a derivative of pVA-HIV was used which contains the human -globin second intron (betaIVS-II) and polyadenylation sequences in place of the endogenous CD2 3 UTR to provide RNA stabilization.5153 Hence a third vector was constructed, pVA-HIV-3, which contains the HIV2LTR,8 the human -globin 3 UTR and the hCD2 LCR. The construction and characterization of this construct is described elswhere (Wrighton et al, manuscript in preparation). These vectors pVA-CD2, pVA-HIV and pVA-HIV-3 (Figure 1a) were used to express HIV molecular inhibitors in stably transfected CEM human T cells. Several RNA-based molecular inhibitor sequences of HIV were constructed as described in the Materials and methods section and in Figure 1b. The Tat regulatory protein serves as a direct transcriptional transactivator of HIV through its physical association with its cis-acting target, TAR. Overexpression of TAR RNA has been shown previously by others to inhibit HIV replication.21 Similarly, regulatory protein Rev acts through a physical association with its cis-target, RRE, to induce the expression of unspliced HIV-1 mRNAs. Hence TAR and RRE decoys (and RWZ6, a shortened version of RRE) were constructed. Also, since a hammerhead ribozyme of the HIV tat gene was shown previously to cleave specically targeted HIV-RNA sequences, we constructed a Tatspecic ribozyme from the HIV-1 tat coding region. These molecular inhibitor constructs were inserted singly or in combination into the above-described expression cassettes containing the puromycin selectable marker gene. Human CD4-positive CEM T cells were stably transfected. Transfectant populations were selected with puromycin and used for HIV inhibition studies.

Regulated expression of HIV molecular inhibitors C Fraisier et al

1667

Figure 1 (a) Schematic representation of the pVA vector used for expression of the HIV inhibitory genes. HIV inhibitory genes encoding RNA decoys or Tat-ribozyme (1) were inserted in the multiple cloning site of the pVA vector containing the human CD2 (hCD2) promoter (pVA-CD2) or the HIV2LTR (pVA-HIV) promoter (2), a 3 untranslated region (UTR) from the hCD2 gene or from the human -globin gene so as to stabilize RNAs (3) and the hCD2 locus control region (LCR) (4) to direct copy number-dependent and position-independent specic expression in T cells. (b) Schematic representation of the RNA decoys (RRE, RWZ6, TAR) and the tat-specic ribozyme stably expressed in CEM cells for infection experiments. Critical nucleotides forming the high afnity binding site in RRE and RWZ6, and the consensus GUC sequence at the cleavage site for the ribozyme are highlighted.

Regulated expression of HIV molecular inhibitors C Fraisier et al

1668

Combination RNA decoys inhibit HIV infection of CEM cell transfectants most efciently To determine if the inhibitor sequences generated in our laboratory were effective and to determine which inhibitors or combination inhibitors were best in preventing the spread of HIV, puromycin-selected stable populations of transfected cells were infected with HIV at several MOIs and tested for the production of either p24 or reverse transcriptase (RT) at various times after infection. Control and transfectant populations were analyzed for the presence of CD4 surface antigen. All cell lines were 91% to 99% CD4-positive with similar intensity of staining (data not shown). Transfected CEM populations carrying pVA-CD2RWZ6, pVA-CD2-TAR or pVA-CD2-TAR+RWZ6 were infected with HIV-1IIIB/LAI at an MOI of 0.0004. Figure 2a shows that CEM transfectants carrying the RWZ6 decoy are almost as infectable as control CEM cells as determined by RT activity at day 9 (2.0 0.29 c.p.m./105 ml compared with 3.8 0.67 for the control cells). The TAR decoy on its own was able to reduce RT levels only up to day 9 (0.68 0.07 c.p.m./105 ml) when compared with CEM control cells. In contrast, transfectants carrying a TAR decoy in addition to RWZ6 (TAR + RWZ6) keep RT levels to less than 50% of maximum for at least 3 days as compared with the control (8.3 0.48 c.p.m./105 ml at day 13 compared with 17.1 3.0 for the control cells). To verify this nding, these transfectant populations were infected with ve-fold more HIV-1IIIB/LAI (MOI, 0.002) (Figure 2b). At this MOI the RWZ6 and TAR transfected cells only slightly resist HIV infection. At day 9 of culture CEM control cells are maximally infected (RT activity 15.1 0.7 c.p.m./105 ml) while RWZ6 and TAR transfectant populations are 75% infected (11.6 1.9 c.p.m./105 ml and 11.0 1.1 c.p.m./105 ml, respectively). Similar to the result at the lower MOI, the TAR+RWZ6 transfectants inhibit best. At day 9 only 40% RT level (5.8 2.6 c.p.m./105 ml) was found in supernatants of TAR+RWZ6 cells as compared with that of control cells. All three transfectant populations reach peak level infection at day 13 of culture. Hence, the TAR+RWZ6 combination decoy appears to be the best inhibitor of HIV-1 replication. Molecular inhibitors were next tested in the context of the pVA-HIV expression cassette. Since the single RNA decoys were unable to inhibit HIV replication efciently at the high MOIs, we compared these and other molecular inhibitor constructs in cell transfectants infected with HIV-1IIIB/LAI at a very low MOI of 0.0001 (Figure 3). HIV infection was determined by sensitive p24 immunoassay on samples taken every 34 days up to 23 days after infection. CEM transfected populations carrying RRE, RWZ6 or tat ribozyme sequences all inhibit HIV infection to a similar degree, giving 9095% inhibition at day 20 as compared with control CEM infected cells. Infection in these populations increases at day 23 giving similar levels of p24 (2.7 to 4.4 g/ml). In contrast, the CEM population carrying the RRE sequence in combination with a TAR decoy (TAR+RRE) shows complete inhibition, with no or barely measurable p24 up to day 23 (6 ng/ml). This strong inhibition was conrmed when the TAR+RRE transfectant population was tested at higher MOIs (see next section). Thus, the TAR+RRE double RNA decoy combination sequence is the best molecular inhibitor of HIV infection. It is interesting to note that HIV

Figure 2 Inhibition of HIV replication in CEM transfectants populations containing the combination TAR+RWZ6 RNA decoy. CEM cells stably transfected with pVA-CD2-TAR (.), RWZ6 (), TAR+RWZ6 () were in vitro infected with HIV-1IIIB/LAI and at MOI of 0.0004 (a) and MOI of 0.002 (b). Cell-free supernatants were tested every 34 days for the presence of reverse transcriptase activity (RT). HIV replication was compared with that in control infected CEM cells (). Uninfected CEM cells () were used as the negative control. Triplicate samples were analyzed for all time-points and the average of the mean is shown. At the MOI of 0.0004 (a), values of RT activity (c.p.m./105 ml) at day 13 after infection are 17.1 3.0 for CEM cells, 17.7 3.1 for RWZ6, 14.6 1.8 for TAR and 8.3 0.5 for TAR+RWZ6 cells. At the MOI of 0.002 (b), values of RT activity (c.p.m./105 ml) at day 9 after infection are 15.1 0.7 for CEM cells, 11.6 1.9 for RWZ6, 11.0 1.1 for TAR and 5.8 2.6 for TAR+RWZ6 cells.

infection is inhibited in pVA-HIV vector transfectants as compared with the pVA-CD2 vector transfectants and nontransfected CEM cell controls. This inhibition is most likely due to the presence of the TAR sequence in pVAHIV, acting as a molecular decoy as suggested by others.54 Our results showing slight inhibition of HIV infection with a single TAR decoy in the CD2 cassette further support this suggestion (Figure 2a and b).

Regulated expression of HIV molecular inhibitors C Fraisier et al

Figure 3 Inhibition of HIV replication in CEM transfectant populations containing the combination TAR+RRE RNA decoy. CEM cells stably transfected with pVA-HIV-Rbz tat (), RWZ6 (), RRE () and TAR+RRE () were challenged with HIV-1IIIB/LAI at an MOI of 0.0001. Cell-free supernatants were tested every 34 days for the presence of HIV1 p24 core protein. HIV replication was compared with that in control CEM cells () and with that in pVA-CD2 () and pVA-HIV ()-transfected cells. Uninfected CEM cells () were used as the negative control.

Quantitative expression of RNA decoys and the tatribozyme in stably transfected cells While initial experiments demonstrated the effectiveness of the molecular inhibitors in delaying HIV infection, we sought to determine the quantitative levels of expression of each of the inhibitors in the various cassettes. We focused on the inhibitors tat-ribozyme, RRE and TAR+RRE. To conrm the presence of the inhibitor constructs in the puromycin-resistant transfectant cell populations and to determine the copy number, Southern blot hybridization was performed. Using a puromycin genespecic probe to detect a 1.4 kb XhoI fragment contained in the vector sequence of pVA-CD2 and pVA-HIV constructs, we found copy numbers in the transfectant populations ranging from 1.1 to 2.0 (Figure 4a). The CEM cell populations were found to be similarly heterogeneous as determined by a smear of hybridizing bands in Southern blot analysis using a restriction enzyme which cut only once within the transfected constructs (not shown). RNA was isolated from the stably transfected cells, treated with RNase-free DNase to eradicate potentially contaminating DNA and reverse-transcribed using oligo(dT) oligonucleotides. We performed semi-quantitative RT-PCR analysis by using serially diluted cDNAs. The cDNAs were amplied in the presence of RRE, Tat and human -actin specic primers. As shown in Figure 4b, the presence of a 184 bp PCR product detected after hybridization with a Tat specic probe indicates the expression of the tat-ribozyme in pVA-CD2-Rbztat and pVA-HIV-Rbztat CEM transfectants. The presence of a 245 bp PCR product hybridizing with an RRE probe indicates the expression of the RRE RNA decoy in CEM transfectants containing pVA-HIV-RRE, pVA-HIV-TAR+RRE or pVA-HIV-3-TAR+RRE. No corresponding fragments

were amplied from untransfected CEM cells, from CEM cells transfected with vector sequences or from samples that were not subjected to reverse-transcriptase treatment. As an internal control for RNA quantication, the steady state RNA levels of -actin were determined similarly by RT-PCR. All samples showed a 590 bp PCR product characteristic of -actin expression. After phosphorimager analysis, normalization with actin and copy number consideration, RT-PCR analysis demonstrates that the pVA-HIV-Rbztat transfectant population expresses the Rbz tat RNA at a 5.4-fold higher level than the pVA-CD2-Rbztat transfectants (Figure 4 and Table 1) suggesting that the HIV promoter is more active than the CD2 promoter. Since the stability of mRNA may also inuence steady state levels of mRNA, we exchanged the hCD2 3UTR in pVA-HIV with the human -globin gene 3 UTR to make the pVA-HIV-3 cassette and directly compared the steady state levels of RRE mRNA in pVA-HIV-TAR+RRE transfectants with pVAHIV-3-TAR+RRE transfectants. As shown in gure 4 and Table 1, the level of steady state RRE mRNA in the transfectants with the 3 -globin UTR, as determined after normalization to the -actin control (Table 1) was 9.6-fold higher than that found in the transfectants with the 3 CD2 UTR. After further normalization to the transgene copy number, a 15.8-fold increase of TAR+RRE expression was obtained with the pVA-HIV-3 vector as compared with the pVA-HIV vector. Thus, in comparison to the other vectors, the pVA-HIV-3 expression cassette promotes the highest levels of decoy mRNA production. A further advantage to the pVA-HIV-3 expression cassette is that the HIV promoter is inducible. HIV promoter mediated transcription has been shown to be increased 10 to 1000 times in the presence of Tat.49 Thus, we tested for an increase in transcription from pVA-HIV-TAR+RRE and pVA-HIV-3-TAR+RRE in the presence of Tat. A Tat peptide corresponding to the rst exon of Tat protein (aa 172) was added to cultures of CEM transfectants and RNA was prepared after 3 days. Semi-quantitative RTPCR analysis was performed as described above, using the RRE specic oligo primers (Figure 5). In the presence of Tat we nd a four-fold increase in inhibitor gene expression in pVA-HIV-TAR+RRE transfectants and a two-fold increase in inhibitor gene expression in pVAHIV-3-TAR+RRE transfectants. Therefore, the HIV-2 promoter in these expression constructs is Tat-inducible and leads to higher levels of RNA decoy expression.

1669

Inhibition of HIV infection is related to levels of inhibitor expression After determining the relative steady state mRNA levels of the transfectant populations, we sought to determine whether quantitatively higher expression levels of the molecular inhibitors would correlate with greater inhibition of HIV. To ensure that each transfectant population was equally infectible with HIV, CD4 levels were determined by FACScan analysis. Before infection, control and transfectants CEM cells were 91% to 99% CD4-positive with equivalent intensity of staining (data not shown). Inhibition of HIV infection due to Rbztat expression from the pVA-CD2 and pVA-HIV cassettes was compared in the transfectant populations (Figure 6). After challenge with an MOI of 0.002, a substantial inhibition of HIV replication was observed in the pVA-HIV-Rbztat transfectant population. HIV replication was reduced by a factor of 8

Regulated expression of HIV molecular inhibitors C Fraisier et al

1670

Figure 4 (a) Determination of copy number of the constructs in transfected cells. Cellular DNA was digested with BamHI and XhoI, separated on a 0.8% agarose gel, transferred onto a nylon membrane and hybridized with a puro-specic probe, giving a 1.4 kb signal. A 3.4 kb BamHI human globin LCR fragment from pLCR was used as probe for internal single copy loading control. Increasing amounts of pVA-HIV plasmid DNA digested with BamHI and XhoI were also added to negative control genomic DNA from nontransfected CEM cells and analyzed simultaneously as copy number controls. Quantication was performed by Phosphorimager analysis. *The DNA in this lane is from the pVA-HIV-3-TAR+RRE transfectant population. (b) Expression of Tat ribozyme and RRE in stably transfected cells. RNA was extracted from control CEM cells, CEM cells transfected with pVA-CD2 and pVA-HIV control plasmid DNAs, and from CEM cells stably transfected with pVA-CD2-Rbz tat, pVA-HIV-Rbz tat, pVA-HIV-RRE, pVA-HIVTAR+RRE and pVA-HIV-3-TAR+RRE. Specic RNA expression was determined by RT-PCR using tat- and RRE-specic primers, and primers specic to human -actin mRNA as internal control, giving PCR products of 184, 245 and 590 bp, respectively. No corresponding Rbz tat or RRE fragment was amplied from negative cell lines or from samples not subjected to the RT (not shown). For semi-quantitative RT-PCR, serial two-fold (Rbz) and four-fold (RRE and -actin) dilutions of cDNA samples were amplied. PCR products were electrophoresced on a 2% agarose gel and stained with ethidium bromide. For Southern blotting, PCR products were transferred on to a nylon membrane and hybridized with 32P-labeled Tat-, RRE- and actin-specic probes. Relative signal intensitites were quantitated with a Phosphorimager.

at day 6 and an 84% inhibition was observed at day 9 (RT activity 2.5 0.25 c.p.m./105 ml compared with 15.1 0.73 for control cells). Infection peaked in this population at day 13, compared with day 9 for control cells. In contrast, the pVA-CD2-tat ribozyme transfectant population showed only a very slight inhibition at day 6 (RT levels 1.52 0.31 c.p.m./105 ml compared with 2.15 0.09 for control cells) and infection peaked at day 9 (14.6 1.05 c.p.m./105 ml). Thus, inhibition of HIV replication appears to be directly related to increased expression of the Rbztat from the pVA-HIV cassette; a 5.4fold increase in basal level RNA expression from the HIV cassette as compared with the CD2 cassette and a further two- to four-fold increase in Tat-inducible RNA expression from the HIV promoter giving an estimated

total 10- to 20-fold increase in Rbztat steady-state RNA levels. To examine this correlation between expression levels and inhibition further, we compared the inhibition of HIV infection in transfectant populations carrying pVAHIV-RRE, pVA-HIV-TAR+RRE and pVA-HIV-3TAR+RRE (Figure 7a). When infected at an MOI of 0.002, all three transfectant populations show inhibition. The expression of RRE alone in the pVA-HIV cassette reduced RT activity by 60% at day 6 (0.8 0.26 c.p.m./105 ml compared with 2.15 0.09 for control cells) and peaked at day 13 after infection (as compared with day 9 for control CEM cells). The expression of RRE in combination with TAR in pVA-HIV-TAR+RRE transfectants yielded a further reduction in RT activity, resulting in 95% inhi-

Regulated expression of HIV molecular inhibitors C Fraisier et al

Table 1 Quantication values for Rbz tat and RRE expression in CEM transfectant populations Transfectant populations

-actina Rbz Normalization Copy Normalization to copy tatb to -actind number numbere
192 177 13.7 88.5 1.7 2.0 8.1 44.2

pVA-CD2-Rbz 1:14 pVA-HIV-Rbz 1:2

-actina RREc Normalization Copy Normalization to copy to -actind number numbere

pVA-HIV-RRE 1:13.5 220 pVA-HIV1:16.5 440 TAR+RRE pVA-HIV-31:3.5 900 TAR+RRE 16.3 26.7 257.1 1.7 1.8 1.1 9.6 14.8 233.7

After RT-PCR using serial dilutions of cDNA and Southern blotting with Rbz tat-, RRE- and -actin specic probes, quantication was performed by Phosphorimager analysis and normalization to the -actin control. (a) Dilution of cDNA giving the same value for -actin signal and falling within the linear range. Quantitation value for Rbz tat (b) and RRE (c) signal at a 1:2 dilution of cDNA for each sample. All values fall within the linear range. These values were then normalized to the actin control (d) and to the copy number of the construct in the transfected cell population (e).

bition of infection at day 9 (0.71 0.08 c.p.m./105 ml compared with 15.1 0.73 for control cells) and peaking only at day 16. Most impressively, the TAR+RRE combination in the pVA-HIV-3 cassette yielded the greatest inhibition. RT levels were reduced by 9095% up to day 13 (0.97 0.21 c.p.m./105 ml) and never approached even the 50% level of RT seen in the other transfectants or the control CEM cells. RT levels in pVA-HIV-3TAR+RRE transfectants reached a maximum at day 23 (4.9 0.47 c.p.m./105 ml) and the peak level of RT production was only 30% of the peak levels seen in the other cell populations. Thus, TAR+RRE in the pVA-HIV-3 cassette is the most effective inhibitor of HIV infection. As the level of mRNA expression is greatest in the pVAHIV-3-TAR+RRE population (9.6 times higher than that in the pVA-HIV-TAR+RRE population), the levels of inhibition with these decoys directly correlates with levels of decoy mRNA.

Discussion
The use of RNA decoys and ribozymes specic to HIV regulatory proteins Tat and Rev has been proposed as a possible therapeutic method for inhibition of HIV replication. In this study, we have examined the ability of several RNA decoys corresponding to TAR and RRE regions of HIV RNA and a ribozyme specic for tat mRNA to inhibit HIV-1 replication. These inhibitory molecules were stably produced in a human T lymphoid cell line (CEM) using three different plasmid expression vectors in order to examine which expression cassette yielded the highest levels of inhibitor RNA and whether expression level correlated with degree of inhibition of HIV infection. Rbztat, TAR, RRE, RWZ6, TAR+RWZ6 and TAR+RRE

inhibitors were generated and tested for inhibition of HIV-1 infection. The TAR region from HIV-2 was used since both HIV-1 and HIV-2 Tat are able to bind and transactivate the HIV-2 LTR.25 Hence the HIV-2 TAR decoy should be an effective inhibitor of both HIV-1 and HIV-2 infection. While others have shown that HIV-1 TAR and multiple TAR decoys are effective in inhibiting HIV infection,21,55 our studies with the HIV-2 TAR decoy show that it acts as an inhibitor when used on its own. However, we also show that it is most effective when used in combination with the RRE decoy. In biochemical assays, small Rev-binding sequences, RWZ2 and RWZ6, have been demonstrated to cooperatively bind several molecules of Rev protein.46 We sought to determine if these subunits are more effective in inhibiting HIV infection than the entire RRE. In Figure 1c, our results show that RWZ6 is slightly less or equal in effectiveness to RRE, and RWZ2 possesses no HIV inhibiting activity (data not shown). Hence the entire RRE sequence was used for quantitative expression and inhibition studies. Finally, the Rbztat inhibitor was found to be only as active as the RRE or RWZ6 decoys when examining CEM cell transfectants infected at a very low MOI of HIV-1 (0.0001). To enable intracellular immunization strategies to be successful therapies for in vivo anti-HIV treatment, molecular inhibitors must be expressed at high levels indenitely in virally susceptible cells, chiey T cells. The human CD2 expression cassette containing the CD2 LCR promotes such high level, long-term expression in T cells. In vivo studies have shown that the CD2 LCR confers chromosomal site position-independent, copy numberdependent expression of exogenous genes in T cells.49 Hence, the CD2 LCR offers great advantage over conventional expression cassettes for long-term therapeutic purposes. Therefore we used an expression cassette containing the CD2 gene regulatory elements, 3 UTR and LCR. To promote higher level expression in HIV-infected cells we also generated a Tat-inducible expression cassette containing the HIV-2 LTR promoter sequences, the CD2 3 UTR and the CD2 LCR. Since Tat-mediated transactivation has been shown to increase transcription signicantly from this promoter, we sought to determine whether higher expression would yield greater inhibition after HIV infection. Finally, a third expression cassette was used in which the 3 UTR of CD2 was replaced with the 3 UTR of the -globin gene so as to stabilize mRNAs. Our studies demonstrate that the pVA-HIV-3 expression cassette yields the highest levels of inhibitor transcription. In a comparison between pVA-CD2 and pVA-HIV directed expression, a ve- to six-fold increase in expression was observed for Rbztat with the pVA-HIV vector. The ability of the pVA-HIV expression cassette to direct high level T cell specic transcription in vivo has been recently tested in transgenic mice. In preliminary results we have found lymphoid-specic basal and high level Tat-inducible expression (D Abraham, personal communication). Moreover, a 15.8-fold increase in TAR+RRE expression was revealed when the pVA-HIV cassette was modied with the -globin 3 UTR, as determined after normalization to the transgene copy number (Table 1). Thus, higher levels of steady-state mRNA can be obtained with basal expression from the HIV promoter. Additionally the inducibility of the HIV promoter to highly transactivate transcription of HIV

1671

Regulated expression of HIV molecular inhibitors C Fraisier et al

1672

Figure 5 Increase of TAR+RRE inhibitor expression in pVA-HIV vector in presence of Tat protein. CEM cells stably transfected with pVA-HIVTAR+RRE and pVA-HIV-3-TAR+RRE were resuspended at 106 cell/ml in culture medium in the absence or presence of 5 g/ml Tat peptide (aa 1 72). After a 3-day culture, RNA was extracted and expression of RRE RNA decoy was determined by semi-quantitative RT-PCR, Southern blotting and Phosphorimager analysis (see Figure 4b).

Figure 6 Inhibition of HIV replication in CEM transfectant populations containing the tat ribozyme. CEM cells stably transfected with pVA-CD2Rbz tat ( ) and pVA-HIV-Rbz tat () were infected with HIV-1IIIB/LAI at an MOI of 0.002. Cell-free supernatants were tested every 34 days for the presence of reverse transcriptase activity (RT). HIV replication was compared with that in control infected CEM cells (). To ensure that each transfectant population was equally infectable by HIV, CD4 levels and percentages were determined by FACScan analysis. Before infection, pVA-CD2-Rbz tat, pVA-HIV-Rbz tat and control CEM cells were 99%, 97% and 91% CD4-positive, respectively, and the intensity of CD4 staining between each of these populations was equivalent. Uninfected CEM cells () were used as the negative control. Triplicate samples were analyzed for all time-points and average of the mean is shown. Values of RT activity (c.p.m./105 ml) at day 9 after infection are 15.1 0.7 for control CEM cells, 14.6 1.1 for pVA-CD2-Rbz tat and 2.5 0.2 for pVA-HIVRbz tat cells.

inhibitory genes in the presence of Tat has been shown5557 and is conrmed for the constructs used in this study. Quantication of steady-state mRNA from pVA-HIV vectors after Tat-mediated induction in the presence of extracellular Tat peptide shows a two- to four-fold increase of expression of RRE in cells carrying pVA-HIV-3TAR+RRE and pVA-HIV-TAR+RRE, respectively. Hence taken together (1) the 5.4-fold increase of Rbz tat expression in pVA-HIV as compared with pVA-CD2; (2) the 15.8-fold increase of TAR+RRE expression in pVAHIV-3 compared with pVA-HIV; and (3) the two- to four-fold increase of RNA decoy expression in the presence of extracellular Tat, the inducible pVA-HIV-3 vector compared with the pVA-CD2 vector yields levels of steady-state inhibitor RNA increased by at least 170- to 340-fold. This pVA-HIV-3 expression cassette should facilitate in vivo application of molecular-based antiviral treatments. While it appears intuitive that higher expression of molecular inhibitors should result in greater inhibition of viral infection, we sought to determine this specically for the expression vectors and inhibitors described in this study. In initial experiments using MOIs of 0.002 and 0.0004, either no or minimal inhibition was observed for pCD2-TAR, RWZ6 and Rbztat transfectant cells. Only at a very low MOI of 0.0001 and in the pVA-HIV cassette was signicant inhibition seen for these inhibitors. In all cases the degree of inhibition of HIV infection is greater when the molecular inhibitor is expressed in the pVAHIV cassette. This is directly demonstrated in Figure 6 for Rbztat. As Rbztat RNA is found in the pVA-HIV transfectants at a ve- to six-fold higher level than in the pVACD2-Rbztat transfectants, signicant inhibition resulted only with pVA-HIV-Rbztat at an MOI of 0.002. No inhibition resulted with pVA-CD2-Rbztat. Thus, the expression of the ribozyme, and similarly the RNA decoys in the pVA-CD2 cassette, may be too low to cleave tat mRNAs or sequester Tat or Rev proteins efciently.

Regulated expression of HIV molecular inhibitors C Fraisier et al

Only with higher basal expression from the HIV promoter or Tat-induced expression from this promoter are functionally effective levels of these inhibitors present. This is more dramatically demonstrated by the comparison of steady-state mRNA levels and inhibition of infection (MOI 0.002) in transfectant populations carrying pVA-HIV-TAR+RRE and pVA-HIV-3-TAR+RRE (Figure 7a). The 9.6-fold increase in steady-state mRNA in the pVA-HIV-3 transfectants leads to a much prolonged period of low level RT production, as well as protecting the cells from producing peak levels of RT activity as seen for the pVA-HIV transfectants. The replacement of the CD2 3 UTR with the -globin 3 UTR results in clear evidence that the inhibition of HIV replication is dependent upon the level of the RNA inhibitor. It is interesting that these same cells infected at a high MOI of 0.01 (5000 TCID50/ml) show even greater inhibition of HIV infection (Figure 7b). While the peak of infection in the control cells is slightly earlier than in the control cells infected at an MOI of 0.002, the highest level of RT activity in the pVA-HIV-3-TAR+RRE transfectants is found at day 2023 and is only 20% of maximal levels of control infected cells. Hence, infection with a higher viral dose may provide an initial and added amount of Tat to allow transactivation of transcription from the pVA-HIV3-TAR+RRE inhibitor construct. At this time it is unclear in these experiments whether all the pVA-HIV-3-

TAR+RRE cells are infected and producing low levels of HIV, or whether a few cells are highly infected. In future experiments we will sample the cultures at various times after infection and analyze CD4 and gp120 levels on the surface of infected cells. Also in situ hybridization with probes detecting other RNAs encoded by HIV will further clarify this point. Successful inhibition of HIV infection has been obtained in CD4+ cells from noninfected individuals transduced with vectors expressing anti-tat ribozymes, tat antisense molecules and TAR RNA decoys.39 Suppression of HIV replication and prolonged survival were also observed in CD4+ cells from HIV-positive subjects, ex vivo transduced with anti-HIV genes expressing an anti-U5 ribozyme.36 Hence, to prove efcacy of the RNA decoys and the tat-ribozyme described further, experiments should be carried out to determine whether these anti-HIV molecules are protective against different laboratory strains and against primary eld isolates of HIV, and whether they can inhibit HIV infection in primary cells. Finally, the antiviral approach taken in these studies will require an efcient method to deliver these sequences to the appropriate cells (the rst choice delivery is to hematopoietic stem cells) for therapeutic application. While viral,2 as well as nonviral2,55,58 delivery systems have been proposed and tested, there is at this time

1673

Figure 7 Inhibition of HIV replication is related to the level of expression of the inhibitor TAR+RRE. (a) Cells stably transfected with pVA-HIV-RRE (), pVA-HIV-TAR+RRE () and pVA-HIV-3-TAR+RRE () were infected with HIV-1IIIB/LAI at an MOI of 0.002. Cell-free supernatants were tested every 34 days for the presence of reverse transcriptase activity (RT). HIV replication was compared with that in control infected CEM cells (). To ensure that each transfectant population was equally infectable by HIV, CD4 levels and percentages were determined by FACScan analysis. Before infection, pVA-HIV-RRE, pVA-HIV-TAR+RRE, pVA-HIV-3-TAR+RRE and control CEM cells were 93%, 92%, 93% and 91% CD4-positive, respectively, and the intensity of CD4 staining between each of these populations was equivalent. Triplicate samples were analyzed for all time-points and average of the mean is shown. Values of RT activity (c.p.m./105 ml) at day 9 after infection are 15.1 0.7 for control CEM cells, 11.2 4.8 for pVAHIV-RRE, 0.71 0.08 for pVA-HIV-TAR+RRE and 0.06 0.007 for pVA-HIV-3-TAR+RRE cells. (b) pVA-HIV-TAR+RRE () and pVA-HIV3TAR+RRE () cells were also infected with HIV-1IIIH/LAI at MOI of 0.01. Cell-free supernatants were tested every 34 days for the presence of reverse transcriptase activity (RT). At each time of sample collection, cells were counted using trypan blue and replated in fresh medium at 5 105 cell/ml; no difference in cell proliferation and viability was observed between the two transfectant populations. HIV replication was compared with that in control CEM cells () and with that in CEM cells transfected with control pVA-CD2 () and pVA-HIV () plasmid vectors. Before infection, pVA-HIVTAR+RRE and pVA-HIV-3-TAR+RRE transfectant cells were 94% and 92% CD4-positive. Control pVA-CD2 and pVA-HIV transfected cells and non-transfected CEM cells were 91%, 96% and 95% CD4-positive, respectively. Uninfected CEM cells () were used as the negative control. Triplicate samples were analyzed for all time-points and the average of the mean is shown. Values of RT activity (c.p.m./105 ml) at day 9 after infection are 15.0 1.8 for control CEM cells, 16.2 0.5 for pVA-CD2 and 12.2 1.6 for pVA-HIV control transfected cells, and 0.97 0.13 for pVA-HIV-TAR+RRE and 0.16 0.02 for pVA-HIV-3-TAR+RRE cells.

Regulated expression of HIV molecular inhibitors C Fraisier et al

1674

no method that is highly reliable. As these methods become improved, the specicity, prolonged nature and efciency of expression of therapeutic molecules will become important. We have shown in the studies presented here that high levels of molecular inhibitors can be expressed in human T cells using the HIV promoter. The use of the CD2 LCR will ensure long-term position-independent expression, as found in vivo in mouse transgenic models.59 Together with the increase in steady-state mRNA due to the use of the -globin 3 UTR in pVA-HIV-3 and the Tat-induced increase in HIVmediated transcription during HIV infection, these studies suggest the efcacy of this HIV-inducible vector to direct the in vivo expression of the combination decoy TAR+RRE in possible therapeutic applications.

and then transferred in 15 ml of RPMI 1640 15% FCS. Two days later, debris and dead cells were removed by density gradient (Lymphoprep; Nycomed, Oslo, Norway) and viable cells were resuspended in RPMI 1640 15% FCS containing 0.5 g/ml puromycin (Sigma, St Louis, MO, USA).

Tat-inducible RNA decoy expression CEM cells stably transfected with pVA-HIV-TAR+RRE and pVA-HIV-3-TAR+RRE Tat-inducible vectors were washed and resuspended at 106 cell/ml in culture medium containing 8 g/ml polybrene, in absence or presence of 5 g/ml Tat peptide (aa 172) (kindly provided by Dr E Blair, Wellcome, Beckenham, UK). After 3 days, RNA was extracted and analyzed by RT-PCR for expression of RRE (see below) HIV-1 challenge Control and stably transfected CEM cells were grown without puromycin at least 3 days before infection. The cells were challenged at 5 105 cell/ml with HIV-1IIIB/LAI viral doses from 50 to 5000 TCID50/ml (MOI from 0.0001 to 0.01) for 24 h at 37C. Cells were washed, resuspended in fresh medium and incubated in triplicate in 12-well plates (1 ml per well). Cell-free supernatants were collected every 34 days and tested for the presence of reverse transcriptase (RT) or the presence of the HIV-1 core p24 protein using a commercial ELISA (Dupont, Boston, MA, USA). At the time of sample collection, cells were counted and replated in fresh medium at 5 105 cell/ml. Southern blot analysis For determination of copy number, cellular DNA (10 g) was digested with BamHI and XhoI, separated on a 0.8% agarose gel, transferred on to a nylon membrane (Hybond-N+, Amersham, Amersham, UK), and hybridized with a 32P-labeled 1.4 kb XhoI puro-specic probe. A 3.4 kb BamHI human -globin fragment from pLCR plasmid DNA containing a 20 kb -globin LCR fragment was also used as an internal loading control. Increasing amounts of pVA-HIV plasmid DNA digested with BamHI and XhoI were also added to negative control genomic DNA from untransfected CEM cells and analyzed simultaneously as copy number controls. Quantication was performed by Phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA) analysis. RNA analysis RNA samples were extracted with LiCl/Urea or RNA isolation reagent (Ultraspec; Biotecx, Houston, TX, USA). To remove potentially contaminating DNA, all the RNA samples were digested with DNase. Fifty micrograms of RNA was mixed with 4 units of RNase-free DNase I (RQ1; Promega, Madison, WI, USA), 50 mm Tris pH 7.5, 10 mm MgCl2 in 200 l reaction, incubated at 37C for 30 min. DNase-treated RNA was reverse transcribed. Two micrograms of RNA were incubated with 1 l of oligo(dT) in 20 l reaction, incubated at 100C for 2 min and stored on ice. Ten microliters of the reaction was reverse transcribed in the presence of reverse transcriptase (RT) buffer (Enzyme Biotechnologies, Cambridge, UK), 1 mm dNTPs, 0.5 l (28 900 U/ml) RNA guard (Pharmacia, Milwaukee, WI, USA), 1 l (21 U/l) RT (Enzyme Biotechnologies) in 20 l reaction, incubated

Materials and methods

Cells and virus CEM cells were grown in RPMI 1640 (Gibco BRL, Life Technologies, Paisley, UK) containing 5% heat-inactivated fetal calf serum (FCS), and supplemented in antibiotics and glutamine, at 37C 5% CO2 (culture medium). HIV-1IIIB/LAI strain was kindly provided by the MRC AIDS Reagent Project. Virus stocks were prepared by infecting CEM cells. The virus titer of cell-free supernatant from infected CEM cells was determined to be 106 tissue culture 50% infective dose (TCID50) per ml. Generation of ribozyme and RNA decoys constructs The ribozyme specic to the tat RNA (Rbztat) was designed so that it cleaves the tat RNA at the rst GUC at position 49 (Figure 1b). The Rbztat was cloned into EcoRI and SmaI sites of the pVA-CD2 expression cassette and the equivalent pVA-HIV vector where the human CD2 promoter BglII-EcoR721 fragment has been substituted by the HIV-2 LTR BglII-blunted SpeI fragment (Figure 1a). The HIV-2 TAR sequence (Figure 1b) was cloned into the EcoRI site of the pVA-CD2 cassette. The RNA decoy corresponding to the HIV-1 RRE sequence (Figure 1b) was cloned into EcoRI and SmaI sites of the pVA-HIV cassette. The RWZ6 sequence in a EcoRV-SmaI fragment (kindly provided by R Zemmel and J Karns) (Figure 1b) was cloned into the SmaI site of pVA-CD2 and pVA-HIV. The combination TAR+RRE RNA decoy was cloned into the EcoRI and SmaI sites of pVA-HIV and the equivalent pVA-HIV-3 where the 3 untranslated region (UTR) of the human CD2 gene has been substituted by the 3 UTR of the human -globin gene (Wrighton et al, manuscript in preparation). For generation of TAR+RWZ6 expression vector, the EcoRV-SmaI RWZ6 fragment was cloned into the SmaI-digested pVACD2-TAR plasmid DNA. Electroporation of CEM cells CEM cells were washed twice with cold RPMI 1640 and resuspended at 107 cell/ml in cold RPMI 1640. For each sample, cells were mixed with 2030 g of XbaI-linearized plasmid DNA and placed into disposable electroporation cuvettes (0.4 cm gap width, Bio-Rad, Richmond, CA, USA) for 10 min on ice. Electroporations were then performed with the BioRad Gene Pulser at 500 F and 300 V for 20 ms. After electroporation the cells were maintained in the cuvettes for 10 min on ice

Regulated expression of HIV molecular inhibitors C Fraisier et al

at 37C for 90 min, or at 50C for 20 min, and stored at 20C before PCR. Control reactions lacking RT were performed to conrm that contaminating DNA was eliminated by DNase treatment. For semi-quantitative RT-PCR, three four-fold (RRE) and two-fold (Rbztat), and ve four-fold (-actin) diluted cDNA samples were amplied. A 50 l reaction mixture containing 1 l cDNA, 10 l PCR buffer (Tris-HCl 300 mm, NH4SO4 75 mm, MgCl2 17.5 mm, pH 9.5) for Rbztat and RRE primers (TatS1: 5-ATGGAGCCAGTAGATCCTAG-3, TatS2: 5-CTCCGCTTCTTCCTGCCA-3; RRE-S: 5-GAGCAGTGGGAATAGGAGC-3, RRE-AS: 5GGAGCTGTTGATCCTTTAGG-3), or 5 l SuperTaq buffer (Enzyme Biotechnologies) for human -actin primers (-actin1: 5-ATGGATGATGATATCGCCGC-3, -actin2: 5-GCGCTCGGTGAGGATCTT-3), 200 mm dNTPs, 0.1 unit SuperTaq (Enzyme Biotechnologies) polymerase and 200 ng each appropriate primer was used for amplication as follows; 94C 3 min, one cycle; 94C 45 s, 58C 45 s, 72C 45 s, 30 cycles; 72C 10 min, one cycle. RT-PCR products were electrophoresced on 1.5% (-actin) and 2.5% (Rbztat, RRE) agarose gels and analyzed by Southern blot hybridization using a nylon membrane (Hybond-N+; Amersham). Rbztat, RRE and actin DNA probes were 32P-labeled by random priming method using Klenow enzyme (Boehringer, Mannheim, Germany) and -32P-dATP (Amersham). Relative levels of specic RNAs were quantied by Phosphorimager (Molecular Dynamics) analysis and normalized to the actin internal control. The dilutions of cDNA were determined so that quantication values after hybridization fell within a linear range.

Acknowledgements
We thank Professor ADME Osterhaus (EUR, Rotterdam, NL) for access to the category-3 facilities in the Department of Virology (EUR, Rotterdam, NL). The HIV-1IIIB/LAI strain was obtained from the MRC AIDS Reagent Project and was kindly provided by Dr R Daniels and Dr C Vella (NIMR, London. UK). We thank Dr E Blair (Wellcome Laboratories, Beckenham, UK) for the Tat peptide, Dr D Kioussis (NIMR, London, UK) for the pVA-CD2 vector and Dr J Karn and Dr R Zemmel (MRC Laboratory of Molecular Biology, Cambridge, UK) for the RWZ6 sequence. We thank Dr D Abraham, Dr D Drabek and other members of the laboratory for advice and critical comments. CF was supported by a grant from Therexsys.

References
1 Baltimore D. Intracellular immunization. Nature 1988; 335: 395396. 2 Yu M, Poeschla E, Wong-Staal F. Progress towards gene therapy for HIV infection. Gene Therapy 1994; 1: 1326. 3 Rosen CA, Sodroski JG, Haseltine WA. Location of cis-acting regulatory sequences in the human T cell lymphotropic virus type III (HTLVIII/LAV) long terminal repeat. Cell 1985; 41: 813823. 4 Cullen BR. Trans-activation of human immnunodeciency virus occurs via a bimodal mechanism. Cell 1986; 46: 973982. 5 Frankel AD, Biancalana S, Hudson D. Activity of synthetic peptides from the Tat protein of human immunodeciency virus type 1. Proc Natl Acad Sci USA 1989; 86: 73977401. 6 Rice AP, Carlotti F. Mutational analysis of the conserved cysteine-rich region of the human immunodeciency virus type 1 Tat protein. J Virol 1990; 64: 18641868.

7 Tiley LS, Brown PH, Cullen BR. Does the human immunodeciency virus Tat trans-activator contain a discrete activation domain? Virology 1990; 178: 560567. 8 Brady HJM, Miles CG, Pennington DJ, Dzierzak EA. Specic ablation of HIV-Tat expressing cells by conditionally toxic retroviruses. Proc Natl Acad Sci USA 1994; 91: 365369. 9 Westendorp MO et al. HIV-1 tat potentiates TNF-induced NFB activation and cytotoxicity by altering the cellular redox state. EMBO J 1995; 14: 546554. 10 Emerman M et al. The specicity of the human immunodeciency virus type 2 transactivator is different from that of human immunodeciency virus type 1. EMBO J 1987; 6: 37553760. 11 Rhim H, Rice AP. TAR RNA binding properties and relative transactivation activities of human immunodeciency virus type 1 and 2 Tat proteins. J Virol 1993; 67: 11101121. 12 Cullen BR, Greene WC. Regulatory pathways governing HIV-1 replication. Cell 1989; 58: 423426. 13 Caputo A et al. Inhibition of HIV-1 replication and reactivation from latency by tat transdominant negative mutants in the cysteine rich region. Gene Therapy 1996; 3: 235245. 14 Ulich C, Harrich D, Estes P, Gaynor RB. Inhibition of human immunodeciency virus type 1 replication is enhanced by combination of transdominant Tat and Rev proteins. J Virol 1996; 70: 48714876. 15 Plavec I et al. High transdominant RevM10 protein levels are required to inhibit HIV-1 replication in cell lines and primary T cells: implications for gene therapy of AIDS. Gene Therapy 1997; 4: 128139. 16 Fraisier C et al. Inhibition of Tat-mediated transactivation and HIV replication with Tat mutant and repressor domain fusion protein. Gene Therapy 1998; 5: 946954. 17 Woffendin C et al. Expression of a protective gene prolongs survival of T cells in human immunodeciency virus-infected patients. Proc Natl Acad Sci USA 1996; 93: 28892894. 18 Rossi C et al. Inhibition of HIV-1 replication by a Tat transdominant megative mutant in human peripheral blood lymphocytes from healthy donors and HIV-1-infected patients. Gene Therapy 1997; 4: 12611269. 19 Su L et al. Hematopoietic stem cell-based therapy for acquired immunodeciency syndrome: efcient transduction and expansion of RevM10 in myeloid cells in vivo and in vitro. Blood 1997; 89: 22832290. 20 Bauer G et al. Inhibition of human immunodeciency virus-1 (HIV-1) replication after transduction of granulocyte colony-stimulating factor-mobilized CD34+ cells from HIV-1-infected donors using retroviral vectors containing anti-HIV-1 genes. Blood 1997; 89: 22592267. 21 Sullenger BA, Gallardo HF, Ungers GE, Gilboa E. Overexpression of TAR sequences renders cells resistant to human immunodeciency virus replication. Cell 1990; 63: 601608. 22 Lee SW, Gallardo HF, Gilboa E, Smith C. Inhibition of human immunodeciency virus type 1 in human T cells by a potent Rev response element decoy consisting of a 13-nucleotide minimal Rev-binding domain. J Virol 1994; 68: 82548264. 23 Feng S, Holland EC. HIV-1 tat trans-activation requires the loop sequence within tat. Nature 1988; 334: 165167. 24 Berkhout B, Gatignol A, Rabson AB, Jeang KT. TAR-independent activation of the HIV-1 LTR: evidence that tat requires specic regions of the promoter. Cell 1990; 62: 757767. 25 Arya SK, Gallo RC. Human immunodeciency virus type 2 long terminal repeat: analysis of regulatory elements. Proc Natl Acad Sci USA 1988; 85: 97539757. 26 Berkhout B, Gatignol A, Silver J, Jeang KT. Efcient trans-activation by the HIV-2 Tat protein requires a duplicated TAR RNA structure. Nucleic Acids Res 1990; 18: 18391846. 27 Malim MH et al. The HIV-1 Rev trans-activator acts through a structured target sequence to activate nuclear export of unspliced viral mRNA. Nature 1989; 338: 254257. 28 Mann DA et al. Co-operative rev binding to stem I of the Revresponse element modulates human immunodeciency virus type-1 late gene expression. J Mol Biol 1994; 241: 193207.

1675

Regulated expression of HIV molecular inhibitors C Fraisier et al

1676

29 Heaphy S et al. HIV-1 regulator of virion expression (Rev) protein binds to an RNA stem-loop structure located within the Rev-responsive element region. Cell 1990; 60: 685693. 30 Iwai S et al. Recognition of the high afnity binding site in revresponse element RNA by the human immunodeciency virus type-1 rev protein. Nucleic Acids Res 1992; 20: 64656472. 31 Lee SW, Gallardo HF, Gaspar O, Gilboa E. Inhibition of HIV1 in CEM cells by a potent TAR decoy. Gene Therapy 1995; 2: 377384. 32 Bahner I et al. Transduction of human CD34+ hematopoietic progenitor cells by a retroviral vector expressing an RRE decoy inhibits human immunodeciency virus type 1 replication in myelomonocytic cells produced in long-term culture. J Virol 1996; 70: 43524360. 33 Rossi JJ, Sarver N. RNA enzymes (ribozymes) as antiviral therapeutic agents. Trends Biotechnol 1990; 8: 179183. 34 Ruffner DE, Stormo GD, Uhlenbeck OC. Sequence requirements of the hammerhead RNA self-cleavage reaction. J Biochem 1990; 29: 1069510702. 35 Pomerantz RJ, Trono D. Genetic therapies for HIV infections: promise for the future. AIDS 1995; 9: 985993. 36 Leavitt MC, Yu M, Wong-Staal F, Looney DJ. Ex vivo transduction and expansion of CD4+ lymphocytes from HIV+ donors: prelude to a ribozyme gene therapy trial. Gene Therapy 1996; 3: 599606. 37 Li X et al. Gene therapy targetting cord blood-derived CD34+ cells from HIV-1-exposed infants: preclinical studies. Gene Therapy 1998; 5: 233239. 38 Lo KMS et al. Inhibition of replication of HIV-1 by retroviral vectors expressing tat-antisense and anti-tat ribozyme RNA. Virology 1992; 190: 176183. 39 Sun LQ et al. Resistance to human immunodeciency virus type 1 infection conferred by transduction of human peripheral blood lymphocytes with ribozyme, antisense, or polymeric trans-activation response element constructs. Proc Natl Acad Sci USA 1995; 92: 72727276. 40 Ramezani A, Ding SF, Joshi S. Inhibition of HIV-1 replication by retroviral vectors expressing monomeric and multimeric hammerhead ribozymes. Gene Therapy 1997; 4: 861867. 41 Lisziewicz J et al. Inhibition of human immunodeciency virus type 1 replication regulated expression of a polymeric Tat activation response RNA decoy as a strategy for gene therapy in AIDS. Proc Natl Acad Sci USA 1993; 90: 80008004. 42 Aguilar-Cordova E et al. Inhibition of HIV-1 by a double transdominant fusion gene. Gene Therapy 1995; 2: 181196. 43 Caputo A et al. Studies on the effect of the combined expression of anti-tat and anti-rev genes on HIV-1 replication. Gene Therapy 1997; 4: 288295. 44 Yamada O et al. A chimeric human immunodeciency virus type 1 (HIV-1) minimal Rev response element-ribozyme molecule exhibits dual antiviral function and inhibits cell-cell transmission of HIV-1. J Virol 1996; 70: 15961601.

45 Gervaix A, Li X, Kraus G, Wong-Staal F. Multigene antiviral vectors inhibit diverse human immunodeciency virus type clades. J Virol 1997; 71: 30483053. 46 Zemmel RW, Kelley AC, Karn J, Butler PJ. Flexible regions of RNA structure facilitate co-operative Rev assembly on the Revresponsive element. J Mol Biol 1996; 258: 763777. 47 Kioussis D, Festenstein R. Locus control region: overcoming heterochromatin-induced gene inactivation in mammals. Curr Opin Genet Dev 1997; 7: 614619. 48 Grosveld F, van Assendelft GB, Greaves D, Kollias G. Positionindependent, high level expression of the human -globin gene in transgenic mice. Cell 1987; 51: 975985. 49 Greaves DR, Wilson FD, Lang G, Kioussis D. Human CD2 3anking sequences confer high level, T cell-specic, positionindependent gene expression in transgenic mice. Cell 1989; 56: 979986. 50 Zhumabekov T, Corbella P, Tolaini M, Kioussis D. Improved version of a human CD2 minigene based vector for T cell-specic expression in transgenic mice. J Immunol Meth 1995; 185: 133140. 51 Antoniou M, Geraghty F, Hurst J, Grosveld F. Efcient 3-end formation of human beta-globin mRNA in vivo requires sequences within the last intron but occurs independently of the splicing reaction. Nucleic Acids Res 1998; 26: 721729. 52 Kabnick KS, Housman DE. Determinants that contribute to cytoplasmic stability of human c-fos and beta-globin mRNAs are located at several sites in each mRNA. Mol Cell Biol 1988; 8: 32443250. 53 Russel JE, Liebhaber SA. The stability of human-globin mRNA is dependent on structural determinants positioned within its 3 untranslated region. Blood 1996; 87: 53145323. 54 Liu J, Woffendin C, Yang Z, Nabel GJ. Regulated expression of a dominant negative form of Rev improves resistance to HIV replication in T cells. Gene Therapy 1994; 1: 3237. 55 Lori F et al. Rapid protection against human immunodeciency virus type 1 (HIV-1) replication mediated by high efciency non-retroviral delivery of genes interfering with HIV-1 tat and gag. Gene Therapy 1994; 1: 2731. 56 Cara A et al. Inhibition of HIV-1 replication by combined expression of gag dominant mutant and a human ribonuclease in a tightly controlled HIV-1 inducible vector. Gene Therapy 1998; 5: 6575. 57 Sanhadji K et al. Experimental gene therapy: transfer of Tatinducible interferon genes protects human cells against HIV-1 challenge in vitro and in vivo in severe combined immunodeciency mice. AIDS 1997; 11: 977986. 58 Woffendin C et al. Nonviral and viral delivery of a human immunodeciency virus protective gene into primary human T cells. Proc Natl Acad Sci USA 1994; 91: 1158111585. 59 Brady HGM et al. Altered cytokine expression in T lymphocytes from human immunodeciency virus Tat transgenic mice. J Virol 1995; 69: 76227629.