Intracellular Inhibition of HIV-1 Replication Using a Dual Protein- and RNA-Based Strategy

Home , Integrase, Nef (protein), Tat (HIV)

L Duan, M Zhu, I Ozaki, H Zhang, DL Wei and RJ Pomerantz The Dorrance H Hamilton Laboratories, Center for Human Virology, Division of Infectious Diseases, Department of Medicine, Jefferson Medical College, Thomas Jefferson University, 1020 Locust Street, Suite 329, Philadelphia, Pennsylvania 19107, USA

Exporting unspliced human immunodeficiency virus type 1 function, thereby inhibiting viral replication. In the present (HIV-1) RNA from the nucleus to the cytoplasm, through studies, different HIV-1 RRE region-specific hammerhead an interaction between the viral regulatory Rev protein and ribozymes were constructed and their anti-HIV-1 repli- Rev response element (RRE) RNA, is a critical step in the cation effects were assayed in diverse RNA polymerase HIV-1 life-cycle. Disruption of either Rev or the RRE will (pol) II and III promoters and vector systems in cell culture. completely inhibit HIV-1 replication. As such, a strategy for Utilizing this combination of an SFv and a ribozyme as a somatic gene therapy to treat HIV-1 infection by intracellu- dual strategy to block HIV-1 replication, both at the protein lar expression of an anti-HIV-1 Rev single chain variable and RNA level, data from these studies demonstrated that fragment (SFv) and a ribozyme which specifically targets potent inhibition of HIV-1 replication can be achieved via the RRE was developed. The anti-Rev D8SFv, which this approach. Combination gene therapies hold promise, specifically targets the Rev activation domain, may be a analogous to combination chemotherapeutic regimens, for key component of combination intracellular immunization, the in vivo treatment of HIV-1 infections. as it has been previously shown to potently inhibit Rev

Keywords: HIV-1; gene therapy; ribozyme; antibody

Introduction human anti-HIV-1 immunoglobulins, such as HIV-1 Rev, reverse transcriptase (RT), integrase, Tat, Nef and Integrated into the host cell’s genome, the HIV-1 provirus gp120.6–9 Via inhibition of different viral targets, the HIV- appears to function as a series of genes and genetic 1 life-cycle can now be specifically inhibited at pre- or elements within HIV-1-infected patients. As such, the post-integration stages in cell culture. This protein-based acquired immune deficiency syndrome (AIDS) has simi- intracellular immunization strategy not only shows a larities to an acquired genetic disease. HIV-1-infected high potency in inhibiting HIV-1 replication, but also cells can be divided into three general categories: those demonstrates the specificity of the single chain variable with active viral production and high cellular turnover, fragment (SFv)-based approach.6–9 productive viral replication from long-lasting cells (eg Hammerhead ribozymes are RNA motifs capable of certain macrophage populations) and latently infected the magnesium-dependent, site-specific cleavage of RNA cells. With at least nine open read frames (orf), HIV-1 and are utilized by a number of small pathogenic plant uses several strategies to control its replication and virions and satellite RNAs for processing during rolling- escape host immune responses, with a high rate of cycle replication.10 The hammerhead ribozyme motif 1–3 mutagenesis. In the past, several strategies have been cleaves the phosphodiester bond downstream of a GUX developed for intracellularly expressing anti-HIV-1 moi- triplet, where X can be C, U or A. GUC is the triplet that eties, such as the antisense approaches, ribozymes, viral is most efficiently cleaved. Target specificity for this dominant-negative mutant proteins or multimers of the cleavage can be achieved by flanking the hammerhead Tat activation region (TAR) and the Rev response ribozyme motif with antisense sequences, complemen- 4,5 element (RRE) for inhibition of viral replication. It is tary to the target RNA.11 Recently, ribozymes have been apparent that to target the viruses’ highly conserved targeted to a wide variety of substrates and tested in bio- functional protein domains or viral RNA structures for logical systems, to achieve inhibition of cellular gene inhibition of viral replication may require prevention of expression or viral replication.12 escape mutations, as a key element for long-term thera- With viral RNA as the packaged genetic material, 4 peutic benefit of HIV-1-infected patients. using RT as a virion-encapsidated functional enzyme, Recently, in order to specifically inhibit HIV-1, we have HIV-1 shows a high mutation rate and easily generates developed several panels of protein-based anti-HIV-1 escape mutants from several chemotherapeutic treat- moieties, which are derived from modified murine and ments.3 Although we have developed strategies for SFvs to target highly conserved HIV-1 protein functional domains, to prevent escape mutations a dual-blocking Correspondence: L Duan strategy of targeting other HIV-1 proteins or targeting Received 15 October 1996; accepted 27 January 1997 HIV-1 proteins and RNAs at the same time, should be Protein-RNA-based-HIV-1 gene therapy L Duan et al 534 extremely useful. For testing this dual-blocking strategy, two HIV-1 RRE-specific hammerhead ribozymes have been generated. Due to the complexity of intracellular ribozyme function, our laboratories have set up several systems for functional analysis of the intracellular efficiency of RRE-specific ribozymes. Then, a combination construct of a functional ribozyme, with an anti- HIV-1 Rev SFv, was created. The data presented in this report indicate that it is possible to functionally combine the protein- and RNA-based HIV-1 inhibition strategies, in single constructs.

Results

Development and design of RRE-specific ribozymes One of the parameters which determine intracellular ribozyme function is the molecular ratio of target RNA (substrate) and ribozyme RNA. Intracellular ribozyme RNA expression levels can be manipulated by either increasing RNA transcription, such as by using RNA polymerase (pol) III promoters,13 fusing ribozyme RNA with an rRNA transcription unit,14 or modifying the structure of the expressed ribozymes to increase stab- ility.15 Of note, for inhibition of HIV-1, full-length unspliced viral RNA should be an ideal target for a combination anti-Rev SFv/ribozyme dual-blocking strategy. The HIV-1 Rev regulatory protein functions not only in unspliced viral RNA translocation from the cell nucleus to the cytoplasm, but also binds to the RRE region to stabilize viral RNA. Without functional Rev, the full-length viral RNA is reduced to nearly undetectable levels.16 The low-level of unspliced viral RNA may be important for ribozyme-based RNA target strategies. In the case of intracellular expression of anti-Rev SFv, the Figure 1 Sequences of targeted sites in the HIV-1 RRE RNA and the HIV-1 RNA expression pattern was the same as a Rev- predicted structures of ribozymes II, V and Vm. The target GUC deleted mutant viral RNA pattern.17 Thus, in order to sequences of HIV-1 RRE RNA are located as indicated by arrows combine a ribozyme with anti-HIV-1 Rev SFv, the HIV- (shadowed letters) (a). Ribozymes are designed to associate with RRE 1 RRE, which only exists in incompletely spliced viral RNA through their antisense sequences. Mutated ribozyme-Vm was created by specifically altering two nucleotides (AA to TT) in the ribozyme- RNAs, was selected for ribozyme targeting. V catalytic domain (b). The parameters which are involved in intracellular cleavage of target RNA by specific ribozymes are not fully understood.18 In order to search for the best cleav- in the presence of Mg+2. When these ribozymes were age site for an anti-RRE ribozyme, all of the GUC triplets incubated with the substrate RNA at a 1:1 molar ratio for in the HIV-1 IIIB and NL4–3 strains’ RRE region 20 min, the substrate RRE RNA was cleaved, at the sequences were analyzed by using a computer-assisted expected specific sites, into P1 and P2 fragments. Data in program, which predicts the secondary structure of RNA Figure 2 demonstrate that ribozymes (II and V) cleave the molecules.19 Figure 1 illustrates the two sequences of the target RRE RNA at specific sites in a cell-free cleavage trans-acting hammerhead ribozymes, designed to cleave system, equally well. By combining the two transcribed the selected GUC triplets in RRE RNA sequences near ribozymes (II and V) in a single construct, more efficient the loop structures, which were finally selected for testing in vitro cleavage of the targeted RRE RNA was obtained in the present studies (ribo-II and V). A mutant ribozyme (see Figure 2). Utilizing phosphorimager analysis with a two base alteration in its functional domain is also (Molecular Dynamics, Sunnyvale, CA, USA), this combi- illustrated in Figure 1 (ribo-Vm). Of importance, this tar- nation of ribozymes was demonstrated to be additive get region in the RRE is well-conserved among different (cleavage efficiency: ribozymes II, 46%; V, 57%; and II clades of HIV-1.36 and V, 82%).

In vitro ribozyme-mediated cleavage of HIV-1 RRE RNA Inhibition of HIV-1 replication in T lymphocytic cells Construction and cloning of ribozymes-II, V, II+V, the transduced with retroviral vectors expressing ribozymes HIV-1 RRE sequence for the substrate in ribozyme cleav- driven by a pol III promoter age assays, and in vitro transcription and puriﬁcation of In our recent ribozyme-mediated ␣1-antitrypsin gene the substrate and ribozymes, are described in the replacement studies, we have found that a hammerhead- Materials and methods. In vitro transcribed hammerhead structure ribozyme, similar to the anti-HIV-1 RRE ribozymes, ribo-II (118 bases) and ribo-V (124 bases), ribozymes, functions much more efﬁciently when were incubated with a 274 base substrate of RRE RNA expressed from a pol III promoter, as compared to a pol Protein-RNA-based-HIV-1 gene therapy L Duan et al 535

Figure 3 Structure of retroviral vectors. As examples of ribozyme constructs expressed from an internal CMV promoter, pSLXCMVribo-V and ribo-II+V are illustrated. The tRNA promoter cassette with RNA polymerase (pol) III termination signal (T) in pDCT-ribo-V is illustrated, as Figure 2 In vitro cleavage of HIV-1 RRE RNA substrate by ribozymes an example. The retroviral vector which expresses the anti-HIV-1 Rev SFv II, V or combination of II+V. Ribozyme-II (118 bases) and ribozyme-V cDNA from the CMV promoter is illustrated as vector pSLXCMV- (124 bases) were labeled with ␥-32P UTP (Rz) and cleaved RRE RNA D8SFv. A bifunctional retroviral vector, which expresses D8SFv from a substrate (S) into two fragments; P1 (5′-cleavage product) and P2 (3′- CMV promoter and tRNA-driving the anti-RRE ribozyme-V, is illus- cleavage product). Lane II: ribozyme-II; lane T: substrate RRE RNA; lane trated as pCMV-D8SFv/tRNA-V. II+T: incubation products of the substrate and ribozyme-II; lane V: + ribozyme-V; lane V T: incubation products of the substrate and ribozyme- HIV-1 strain, demonstrates that only the tRNA- V; lane II+V: ribozymes II+V; lane II+V+T: incubation products of the NL4–3 substrate and ribozymes II+V. ribozyme-V construct showed significant (40–90%) inhibition of HIV-1 replication in retrovirus-transduced, mixed SupT1 cell populations, in multiple independent II promoter.20 To study the effect of anti-RRE ribozymes experiments (Figure 4a). This inhibition of viral repli- on the inhibition of HIV-1 replication in human T lym- cation was solely due to ribozyme function, not antisense phocytic cells, SupT1 cells, which can be infected with effects, as the ribozyme-Vm construct, with two mutated HIV-1 and support high-level viral replication,6 were first nucleotides in the ribozyme catalytic domain, showed no transduced with the pSLXCMV retroviral vector contain- viral inhibition (Figure 4c). At the challenge dosage of ing either ribozyme-II, -V, or combined ribozyme-II+V, MOI (0.22) there was no difference in viral growth expressed under the control of an internal cytomegalo- between the control cells and all ribozyme-expressing virus (CMV) promoter (Figure 3). The same individual cells (Figure 4b). Thus, the anti-HIV-1 effects of these ribozymes were also constructed to be under the control ribozymes can be overwhelmed with a very high viral of a pol III promoter (tRNA) in the DCt retroviral vector input. Of note, with CMV promoter-driven anti-RRE (Figure 3). Due to the inserted DNA size limitation in the ribozymes, none of the tested constructs showed any pro- pol III transcription unit (eg for efficient transcription not tection in HIV-1 challenges, with even low MOIs of 0.004 more than 80 bases (unpublished data)), the ribozyme- or above (Figure 4a). II+V fragment was not inserted into the DCt vector. As a control, an anti-mouse immunoglobulin gene kappa Transcription of ribozyme RNA from CMV or tRNA chain (abV␬) ribozyme was cloned into both retroviral promoters vectors.21 The failure of ribozymes driven from an internal CMV In order to distinguish the mechanism(s) of inhibition promoter, but not tRNA-promoter derived ribozymes, to of HIV-1 replication, via antisense and/or ribozyme func- inhibit HIV-1 replication prompted us to investigate the tions, the RRE ribozyme-V sequence was mutated AA to transcription levels of these ribozymes in retroviral vec- TT, in the ribozyme catalytic domain, to disrupt tors. As indicated in Figure 3, there are at least two pro- ribozyme function but maintain antisense arms (Figure moters on each of the analyzed vectors. Previously, we 1). This mutated ribozyme DNA fragment was also have shown that the murine leukemia virus-long ter- inserted into the DCt vector as a control. After transduc- minal repeat (MLV-LTR) can maintain reasonable tran- tion of each of the ribozyme constructs into SupT1 cells, scriptional activity in SupT1 or CEM cells for more than transduced cells were selected in 1 mg/ml of G418 for 3 6 months.7 Thus, all of tRNA promoters’ transcriptional weeks. Utilizing G418-resistant clones, limiting dilution orientations in the constructs were designed in the same was performed 2 days after transduction in G418 selec- direction as the 5′ MLV-LTR, so antisense effects of the tion for 3 weeks. Then, both pooled mixed populations LTR-driven transcripts would not alter tRNA-driven and individual resistant clones were analyzed in HIV-1 transcripts. pSLXCMV and pDCT constructs (Figure 5a) challenge experiments. The data summarized in Figure 4, were packaged and transduced into SupT1 cells for 2 at different multiplicities of infection (MOIs) with the months and the total cellular RNA was extracted for Protein-RNA-based-HIV-1 gene therapy L Duan et al 536

Figure 4 Continued.

size of and lack of a poly-A tail on tRNA-derived tran- Figure 4 Inhibition of HIV-1 replication in anti-RRE ribozyme-trans- scripts may make the ribozyme RNA more efficient in duced T lymphoid cells. SupT1 cells were transduced with CAT-, anti- targeting the RRE RNA. RRE ribozyme II-, ribozyme V- and II+V ribozyme-expressing retroviral vectors, using an internal CMV promoter, or ribozymes II, V and Vm in Dual-functioning retroviral construct and inhibition of a tRNA cassette (mixed cellular populations). SupT1 cells were infected HIV-1 replication in cell culture with HIV-1NL4–3 (MOI: 0.01 to 0.22). (a) Transduced SupT1 cell-lines HIV-1 challenge experiments demonstrated that infected with NL4–3 (MOI: 0.01). (b) Transduced SupT1 cells infected with NL4–3 (MOI: 0.22). (c) Ribozyme-Vm transduced SupT1 cells ribozyme-V was significantly more potent in inhibiting infected with NL4–3 (MOI: 0.03). HIV-1 replication was quantified by viral replication, as compared with ribozyme-II with a determining HIV-1 p24 antigen levels in the culture supernatants, using tRNA promoter (Figure 4a). In our previous anti-HIV-1 an ELISA (Dupont). The data are representative of at least three sets Rev SFv studies, D8SFv has been shown to possess the of independent experiments for (a) and two sets of experiments for (b) ability to inhibit HIV-1 replication through specific alter- and (c). ation of Rev function.7 As such, we next combined the protein- and RNA-based intracellular strategies to Northern blot analysis, using a ribozyme DNA fragment attempt to improve viral inhibition and prevent possible as a probe (Figure 5b). Of note, all of the promoters on escape mutations. The MluI–BglII DNA fragment, which the retroviral vectors expressed transcripts. The tran- contains the tRNA promoter-ribozyme-V and pol III ter- scription level from the internal CMV promoter actually minal function unit, was inserted into the pSLXCMV- showed stronger expression than the tRNA promoter in D8SFv 3′ LTR NaeI site, with the same transcriptional these experiments. Several repeat experiments demon- orientation as the CMV promoter in this vector (see Fig- strated the CMV promoter yielding better transcriptional ure 3). This dual-functioning retroviral vector was then activity, as compared to the tRNA promoter (Figure 5b). packaged in PA317 cells and transduced into SupT1 cells This may be explained by higher transcription activity of and human peripheral blood mononuclear cells (PBMC) the CMV promoter, or CMV-derived pol II transcripts, for HIV-1 challenge experiments. The protein expression which carry a poly-A tail, may be more stable than tran- of D8SFv from this vector was confirmed using Western scripts from the tRNA promoter. Nevertheless, the small blotting, as indicated in Figure 6. In the acute HIV-1 Protein-RNA-based-HIV-1 gene therapy L Duan et al 537 of anti-Rev SFv, anti-RRE-ribozyme, and the dual- functioning expression vector on HIV-1 chronically infected cells, each of the packaged retroviral vectors was transduced into H9/IIIB cells. After 2 weeks of G418 selection, the mixed populations of cells were maintained in G418-free medium for 3 weeks, then HIV-1 p24 antigen production from the transduced H9/IIIB cells was quantified. As indicated in Figure 8, anti-Rev SFv and the trans-dominant negative mutant Rev protein, RevM10, demonstrated 40–50% inhibition of viral replication, but the ribozyme alone only reduced the HIV-1 p24 antigen production level by 18–27%. When the anti-Rev and ribozyme dual retroviral vector was transduced into the H9/IIIB cells, 75–84% inhibition of viral replication inhibition was achieved (Figure 8). As such, a dual protein- and RNA-based approach may be efficient for genetic therapeutics targeting chronically HIV-1-infected cells.

Discussion A number of strategies have recently been explored to render human cells resistant to HIV-1 replication.1,2 As a further development of intracellular immunization strategies against HIV-1 infection, suitable combinations of different anti-HIV-1 moieties, which specifically inhibit different key HIV-1 proteins or viral RNA elements, will not only be able to enhance antiviral efficiency, but also prevent possible escape mutants.3 The viral early-gene product, Rev, which controls the expression of the structural genes of HIV-1, exerts its function through binding to an RNA element, RRE. HIV- 1 RRE, located in the envelope’s RNA sequence, consists of 234 nucleotides and can assume a highly complex secondary structure consisting of a central stem and five stem-loops.2,22 By binding to RRE, specifically stem-loop II, HIV-1 Rev functions as a chaperone to transport unspliced viral RNA into the cytoplasm, for viral struc- Figure 4 Continued. tural protein translation and viral genome packaging.22 As RRE functions in the productive HIV-1 life-cycle, disrupting the RRE structure using minimal RRE oligomers infection experiments, this dual-functioning vector as a molecular decoy has been explored, as an intracellu- showed improved viral inhibition, as compared with lar immunization strategy against HIV-1 replication.23,24 D8SFv alone. In three HIV-1 challenge experiments, In vivo binding of multimers of HIV-1 Rev protein to the increased inhibition by this dual-functioning retroviral RRE during HIV-1 replication may form a protected vector was observed, dependent on the challenge HIV-1 RRE–Rev protein complex. As such, this complex may dosage (MOI). Figure 7a and b illustrates representative not be so easily targeted by ribozymes. As we have dem- experiments using SupT1 cells and human PBMC, which onstrated in our anti-HIV-1 Rev SFv studies, functional were challenged with HIV-1NL4–3 and a primary HIV-1 Rev can be trapped in a cytoplasmic location by com- isolate (SI 92006262), respectively. Although there was plexing with a cytoplasmic SFv protein, and this SFv significant variability in independent experiments of reduces intracellular Rev half-life more than four-fold, to absolute HIV-1 p24 antigen expression, there was always less than 4 h.17,25 Thus, the HIV-1 RRE RNA region will a 0.5 to 1.0 log decrease in HIV-1 replication, comparing not be efficiently bound by the Rev protein in the pres- the D8SFv plus the ribozyme-V with D8SFv alone. ence of the SFv and, thus, the open structure (ie loop II of the RRE) may allow further targeting of the RRE RNA HIV-1 replication in chronically infected cells can be region by specific ribozymes. Fine mapping has shown inhibited with combination intracellular immunization that sequences within stem-loop II encode the primary strategies determinants for Rev function and are the major binding The concept of anti-HIV-1 gene therapy, using intracellu- sites for Rev.23 Importantly, data have also demonstrated lar immunization strategies, should not only protect the that the interaction between Rev and the RRE is absol- cells which have not yet been infected by HIV-1, these utely required for HIV-1 replication.26 We now describe approaches should also be able to inhibit the replication the use of an HIV-1 RRE-specific hammerhead-like of HIV-1 in chronically infected cells to reduce viral load ribozyme as an anti-HIV-1 intracellular immunization in different stages of disease. To test the inhibitory effects methodology, and by combination with anti-Rev SFv, Protein-RNA-based-HIV-1 gene therapy L Duan et al 538

Figure 5 (a) Retroviral vector constructs and predicted promoter transcripts. RNA (a): transcripts from MLV-LTR promoter in pSLXCMV vector; RNA (b): transcripts from internal CMV promoter in pSLXCMV; RNA (c): transcripts from MLV-LTR promoter in pDCT-ribo-V vector; RNA (d): transcripts from tRNA promoter in pDCT-ribo-V. (b) Northern blot analysis of ribozyme RNA and GAPDH mRNA in G418-selected SupT1 mixed population cells transduced with retroviral vectors, carrying ribozymes under the control of the tRNA or internal CMV promoter. SupT1 cells were transduced with different ribozyme-carrying retroviral vectors and selected for 2 weeks. Then the cells were maintained in G418-free medium for 6 weeks. 1 × 107 cells from each of the transduced populations was used for RNA extraction. Transcription of ribozyme RNA: lane 1: cells transduced with pSLXCMV-ribo-II+V; lane 2: cells transduced with pDCT-ribo-V; lane 3: transduced SupT1 cells with pSLXCMV vector, as a negative control. The blot was hybridized with a ribo-II+V DNA probe. As an RNA loading control, the same blot was rehybridized with a GAPDH probe. Transcripts a, b, c and d are as shown in panel a.

forms an HIV-1-specific dual-blocking strategy to inhibit blotting for more than 2 months (Figure 5 and Ref. 20). acute and chronic HIV-1 infection. Of note, RNA transcripts from a CMV promoter are more In the present study, we have first targeted HIV-1 RNA than 2 kb in size. Recent data have suggested that the 3′ using a ribozyme-based strategy. Considering the speci- ribozyme structure is important for both ribozyme stab- ficity of ribozymes, the hammerhead-like structure was ility and function.15 A large ribozyme construct may not chosen to target HIV-1 RNA, with each arm sequence be suitable for intracellular functioning of the moiety. approximately 14–20 bases.18 For a combined strategy to Modification to a pol III-type promoter expression system interrupt the Rev–RRE interaction, our laboratories had may not result in a much higher expression level for already developed the anti-Rev SFv, and now extend ribozyme moieties, as compared to pol II promoters, but these on-going studies to target the RRE as an HIV-1 the smaller size of the ribozymes may more efficiently RNA motif. To target GUC sites, shown in Figure 1, both access and bind the target RNA.28 In the present studies, ribozyme-II and ribozyme-V GUC motifs were chosen, ribozyme-II and ribozyme-V were also cloned into a DCt- near the RRE loop regions, which have potential accessi- based retroviral vector for evaluation of antiviral bility for ribozyme cleavage. efficiency in transduced SupT1 cells. In this vector, Using the in vitro ribozyme cleavage reactions, both ribozyme-V alone demonstrated inhibition of HIV-1 rep- ribozyme-II and ribozyme-V showed similar efficiency lication. When the ribozymes-II and -V were combined and a combination of ribozyme-II and -V could enhance in vitro for cleavage of RRE RNA, they showed an addi- the in vitro efficiency. These moieties were cloned as tive effect on the targeted RNA, but CMV-driven single or combined ribozymes downstream of the CMV ribozymes-II+V demonstrated no improvement of anti- promoter in pSLXCMV, which has been shown to be suit- viral function. So far in our laboratory, we have found able for high-level protein expression.27 There were only that data from an in vitro assay generally do not corre- minimal anti-HIV-1 effects detected after the vectors spond to the in vivo efficiency of ribozymes (data not were transduced into SupT1 cells, even though the CMV- illustrated). driven ribozymes could easy be detected by Northern Ribozyme-V, when placed in the tRNA functional cas- Protein-RNA-based-HIV-1 gene therapy L Duan et al 539 One of the major benefits of the dual-blocking HIV-1 strategy is to prevent possible escape mutants during long-term HIV-1 infection, which every anti-HIV-1 therapy has to approach. However, we have not yet found an escape mutation which is generated from the anti-Rev D8SFv-based method, which specifically targets the highly conserved Rev activation domain.17 Of importance, these data demonstrate an additive modest effect of the combination of a specific RNA- and protein-based approach to anti-HIV-1 gene therapy. This effect, although modest in absolute numbers, may hold importance if in vivo HIV-1 RNA are already low, based on other therapeutics. Certainly, only in vivo phase II studies will be useful in the final analysis. Improvement of long-term expression vectors, the combination of blocking the HIV-1 life-cycle at pre- and post-integration sites, or dual strategies to inhibit HIV-1 replication targeting viral protein and RNA in host cells, clearly pro- vide reason for cautious optimism that such approaches might succeed as molecular therapeutics for clinical HIV- 1 infection.

Materials and methods

Ribozyme designs and constructions The sequences of the targeted HIV-1 (IIIB strain) RRE region and designs of the ribozymes are illustrated in Fig- ure 1. As the target sites for ribozymes, two GUC sequences in the highly conserved HIV-1 RRE stem-loop II and V regions have been chosen, due to their locations near the RNA loop structures, which may be quite approachable for ribozyme cleavage and also are highly conserved regions in most HIV-1 isolates. The two ribozymes were named for RRE domains II and V, which are based on a computer-generated RRE RNA structure, utilizing the RNA-Fold program.19 For the design of trans-acting HIV-1 ribozymes, desig- nated ribozyme-HIV-II and ribozyme-HIV-V, respect- Figure 6 Western blot analysis of anti-Rev D8 SFv expression in mixed ively, 14 to 20 bases of each antisense arm sequence populations of SupT1 cells transduced with the dual-functioning retrovi- against the RRE were flanked on both sides of the ham- ral vector, pCMV-D8SFv/tRNA-V. SupT1 cells were transduced with the merhead motif, to allow the ribozymes to associate CMV-D8SFv/tRNA-V vector. The cells underwent 2 weeks of G418 selection and then were maintained in G418-free medium for 3 weeks. Total specifically with RRE RNA through their complementary cellular protein was extracted and separated on 10% SDS-PAGE. D8SFv sequences, but still maintain reasonable flexibility of protein expression was evaluated with a rabbit anti-mouse kappa chain RNA structure.18 For the construction of the ribozymes, antibody. ␤-Actin was probed with a rabbit anti-human ␤-actin antibody. two complementary oligonucleotides were synthesized. Lane 1: SupT1 control cells; lane 2: SupT1 cells transduced with dual- MluI and XhoI sites and BamHI and MluI sites were expression vector. added to the 5′ ends of the oligonucleotides used to create ribozymes-V and -II, respectively. The sequences of sette, can reproducibly inhibit HIV-1 replication. This pol oligonucleotides were as follows: sense primer for III-ribozyme unit was further inserted into ribozyme-II-1: 5′-AGGATCCTGTACCGTCAGCGTCA pSLXCMVD8SFv within the 3’ MLV-LTR U3 region, to TTCTGATGAGTCCGTGAGGACG-3′, antisense for generate the dual-functioning anti-HIV-1 construct. Of ribozyme-II-2: 5′-CACGCGTGCACTATGGGCGCAGCG note, the ‘double copy’ vector is thought to enhance TTTCGTCCTCACGGACTCATC-3′, sense for ribozyme- expression from the promoter by duplicating the tRNA V-1: 5′-CACGCGTGGAGCTGCTTGATGCCCCACTGAT cassette in target cells.23 By targeting HIV-1 Rev with GAGTCCGTGAGGACG-3′, antisense for ribozyme-V-2: anti-Rev SFv and disrupting the RRE with ribozyme-V, 5′-CACTCGAGCATCTGTTGCAACTCACAGTTTCGTCC an increased anti-HIV-1 effect can now be illustrated in TCACGGACTCATC-3′. For the construction of a mutant both acute HIV-1 infection and most importantly chron- ribozyme-V as control, a two nucleotide AA to TT switch ically infected human T lymphocytes in cell culture. was created in the catalytic domain to disrupt ribozyme Although utilizing both a combined RNA- and protein- function but maintain the antisense arm. This oligonucle- based approach to inhibit HIV-1 is attractive for several otide, ribo-Vm2 was: 5′-CACTCGAGCATCTGTTGCAA reasons noted above, the scheme also has several com- CTCACAGTAACGTCCTCACGGACTCATC-3′. The ribo- plexities to overcome, including promoter types and zymes in DNA form were synthesized by incubating 1 locations of therapeutic moieties within vector constructs. ␮g of two oligonucleotides to form a hemiduplex. Then Protein-RNA-based-HIV-1 gene therapy L Duan et al 540

Figure 7 Inhibition of HIV-1 replication in anti-HIV-1 dual-functioning vector-transduced SupT1 cells and human PBMC. SupT1 cells (mixed populations) (a) and PBMC-stimulated using PHA/IL-2 (b), were transduced with CAT-, anti-Rev D8SFv-, tRNA-derived ribozyme V-, and the dual- functioning CMV-D8SFv/tRNA-V vectors, three times with daily changes of fresh vector-containing medium.7 Cells were then infected with either × 6 HIV-1NL4–3 (MOI 0.03; for SupT1 cells) or a primary HIV-1 isolate (SI 92006262; HIV-1 p24 antigen 150 pg/1 10 cells per 1 ml medium per well in 6-well plates for PBMC) for 2 h. HIV-1 replication was quantiﬁed by assaying HIV-1 p24 antigen levels in the culture supernatants, using an ELISA (Dupont). The data are representative of at least two sets of independent experiments, performed in duplicate.

polymerase chain reaction (PCR) amplifications were per- USA) to generate a plasmid, pGRRE, which contains two formed in a 100 ␮l volume containing: 10 mm Tris-HCl GUC sites as the targets for the HIV-1 RRE ribozyme-II m m ␮m (pH 8.3), 50 m KCl, 2.5 m MgCl2, 200 of each and V. The proper construction of the ribozymes and dNTP and 5 Units of Taq polymerase (Perkin-Elmer, Fos- RRE RNA were confirmed by DNA sequencing, using the ter City, CA, USA). The cycle conditions were as follows: PRISM Ready Reaction DyeDeoxy Terminator Cycle 94°C for 1.5 min, 50°C for 1.5 min and 72°C for 2 min for Sequencing Methodology (ABI, Foster City, CA, USA). 10 cycles. Then, PCR products were directly cloned into the pT7Blue-T vector (Novagen, Madison, WI, USA) to In vitro transcription and ribozyme cleavage reactions generate plasmids pT7ribo-II and pT7ribo-V. For cloning pT7-ribo-II and pT7-ribo-V, containing the ribozyme the RRE RNA, to generate the substrate RNA for sequences, were digested with SmaI to linearize the plas- ribozyme-II and ribozyme-V in in vitro cleavage reac- mids, then transcribed with bacteriophage T7 RNA poly- tions, two oligonucleotides were synthesized (sense merase in the presence of ␥-32P UTP to generate primer, RREp-1: 5′-TTCTTGGAGCAGCAGGGAG-3′; ribozymes-II and V of 117 and 127 bases, respectively. antisense primer, RREp-2: 5′-CTAGGATTCTTGCTT pGRRE, containing the RRE RNA sequence, was lin- GGA-3′), and subjected to PCR, using the HIV-1 HXB2 earized with SalI and transcribed with T7 RNA poly- plasmid (R7) as a template. PCR amplification for 35 merase in the presence of ␥-32P UTP to generate a sub- cycles consisted of 94°C for 1.5 min, 50°C for 1.5 min, and strate RRE RNA of 292 bases. The ribozyme and substrate 72°C for 2 min. The 226 base-pair PCR product was then RRE RNA were incubated in a 10 ␮l reaction volume con- cloned into the pGEM-T vector (Promega, Madison, WI, taining: 50 mm Tris-HCl (pH 7.5), 1 mm EDTA and 10 Protein-RNA-based-HIV-1 gene therapy L Duan et al 541 For subcloning of the ribozymes, ribozyme-II was excised from pT7ribo-II with BamHI, and cloned into the BglII site downstream of the internal CMV promoter in the pSLXCMV vector, to generate pSLXCMVribo-II (Figure 3). Ribozyme-V was excised from pT7ribo-V with MluI and XhoI, and cloned into the MluI–XhoI sites in the pSLXCMV vector to create pSLXCMVribo-V. For cloning both ribozyme fragments downstream of the CMV promoter, firstly, the EcoRI–XbaI fragment from pT7ribo-II was inserted into pSP72 vector’s (Boehringer Mannheim) EcoRI–XbaI sites to generate pSP-RRE-II, then the XbaI– SmaI fragment which contains ribozyme-V from pT7 ribo-V was inserted into pSP-RRE-II’s XbaI–PvuII sites to produce pSP72-RRE-II+V. To place the double ribozyme downstream of the CMV promoter, the BglII–XhoI fragment was inserted into pSLXCMV’s BglII–XhoI sites (Figure 3). Retroviral vector pDCt2T (Figure 3) was used to insert ribozyme-V into the tRNA promoter cassette; tRNA/Met (D3–2) gene with an RNA polymerase (pol) III termination signal.13,29 This construct is called a ‘double copy’ vector because the tRNA promoter is inserted in the U3 region of the 3′-long terminal repeat (LTR). When the recombinant virus is transduced into a target cell-line, the U3 region in the 3′-LTR is used as a primer for synthesis of the 5′-LTR. Then, this tRNA promoter cassette is dupli- cated in the 5′-LTR. Briefly, for ribozyme subcloning, the 74 bp fragment of ribozyme-V was removed using digestions with SacII and BamHI from the pSP72-ribozyme-V vector and was then inserted into SacII–BamHI sites of pDCt2T, to generate pDCT-RRE-ribozyme-V. As a control for the RRE ribozyme constructs, the abV␬ ribozyme developed in our laboratory,21 which specifically targets the mouse Figure 8 Inhibition of HIV-1 replication in chronically infected H9/IIIB immunoglobulin gene kappa chain variable regions, was cells with RevM10 and the dual-functioning vector (ribozyme-V and anti- cloned into pSLXCMV or pDCt2T, to generate Rev D8SFv). H9/IIIB cells were transduced with pLSXN-M10, D8SFv, pSLXCMVabV␬ or pDCt2abV␬T, respectively. The HIV- ribozyme-V, the dual-functioning retroviral vector, pCMV-D8SFv/tRNA- V, selected in G418 (800 ␮g/ml) for 3 weeks and then further cultured 1 Rev trans-dominant inhibitor, M10, was also cloned into 24 in G418-free medium for 3 weeks. The cells were standardized to 1 × 106 the pLSXN vector for control experiments. Using cells per ml and split 1:3 every third day. HIV-1 production was quantified pCMVM10 as a template,24 a M10 cDNA fragment was by assaying HIV-1 p24 antigen levels in the culture supernatants, using PCR amplified with two oligonucleotides: Rev-1: 5′- an ELISA (Dupont). The data are representative of at least two sets of CGAATTCGCCATGGCAGGAAGAAGCGGA-3′; Rev-2: independent experiments. 5′-GAGATCTATTCTTTAGCTCCTGA-3′ and the 372 bp Rev M10 DNA fragment was cloned into pT7Blue(R) for m ° m MgCl2. These RNAs were heated to 95 C for 2 min DNA sequencing. Then, the EcoRI–BglII M10 fragment and cooled on ice. The reactions were performed at 37°C was inserted into pLSXN via EcoRI–BamHI sites to gener- for 20 min, then the reactions were stopped by the ate the M10 expression plasmid, pLSXN-M10. addition of an equal volume of 95% formamide, 20 mm For construction of the dual-functioning retroviral con- EDTA, 0.05% bromophenol blue, and 0.05% xylene struct, pSLXCMV-D8SFv-tRNA-V, which carries both a cyanol, and heated to 65°C for 5 min. The reaction pro- ribozyme and the anti-Rev D8SFv, the tRNA promoter- ducts were separated on a 6% polyacrylamide-7 m urea ribozyme-V-terminal function fragment was recovered gel in TBE buffer. The labeled RNAs were then visualized from pDCT-RRE-ribozyme-V with BglII and MluI diges- by autoradiography. tions, and blunted via the Klenow reaction. The fragment was then inserted into the NheI site of the MLV 3′ LTR Retroviral vectors by partial digestion of pSLXCMV-D8SFv (see Figure 3).7 Three murine leukemia virus (MLV) vectors, containing the neomycin-resistance (neo) gene as a selectable marker, Cell cultures and transfections were utilized for the expression of anti-HIV-1 moieties The SupT1 and CEM cells are CD4+ human T lympho- in target cells. Retroviral vector pLXSN, and pSLXCMV cytic cell lines, susceptible to HIV-1 infection,7 which are which contains an internal human immediate–early cyto- grown in RPMI-1640 medium supplemented with 10% megalovirus (CMV) promoter, were used.27 For fetal calf serum (FCS). H9/IIIB is a stable T cell line which 30 expression of anti-HIV-1 Rev SFvD8 and chloram- has been chronically infected with the HIV-1IIIB strain. phenicol acetyl transferase (CAT) as controls, retroviral The retroviral packaging cell-line, PA317,27,31 was grown vectors encoding these moieties have been described in DMEM supplemented with 10% FCS. All the cells were 7 ° previously. grown at 37 C in a humidified incubator with 5% CO2. Protein-RNA-based-HIV-1 gene therapy L Duan et al 542 For production of helper-free, recombinant MLV-based duced SupT1 cells was extracted by a modification of the retroviral vectors, subconfluent PA317 cells were trans- method of Chomczynski and Sacchi.34 Expression of fected with 10 ␮g of retroviral expression plasmids, using CMV or tRNA promoter-driven ribozymes was detected a standard calcium-phosphate transfection method by Northern blot hybridization analyses, as described (Promega), according to the manufacturer’s instructions. previously,20 employing a ribozyme cDNA probe and After 48 h, medium containing recombinant retrovirus glyceraldehyde 3-phosphate dehydrogenase (GAPDH) particles was collected and titers of the retroviral shuttle cDNA as a control. Samples of total RNA (20 ␮g) were vectors were determined, as described previously.7 Per- denatured in buffer containing 0.5 mg/l glyoxal, 50% ipheral blood mononuclear cells (PBMC) were obtained dimethyl sulfoxide, 10 mm phosphate, and then elec- by Ficoll–Hypaque separation of phlebotomized blood trophoresed in 1% agarose gels, transferred to Gene- from HIV-1-seronegative individuals. Screen filters (New England Nuclear, Boston, MA, USA), and baked for 2 h at 80°C. The filters were prehybridized Transduction of T lymphocytic cells and viruses and were subsequently hybridized under stringent con- Each target cell type was infected with recombinant retro- ditions with the ribozyme or GAPDH DNA fragments viral vectors, overnight, in the presence of 8 ␮g/ml labeled with ␣-32P dCTP, using a primer extension meth- polybrene. On day 3 after transduction, the cells were odology (Amersham, Arlington Heights, IL, USA). After selected with 1 mg/ml of G418 (Gibco-BRL, Gaithers- hybridization, the filters were washed and the signals burg, MD, USA) for 3 weeks. G418-resistant clones were were visualized by autoradiography. generated by limiting dilution or pooled. Both individual clones and mixed pooled cell populations were subjected Protein extraction and Western blot analysis to further functional analyses. Cells transduced with retroviral vector pCMV- For human PBMC, after 3 days of phyto- D8SFv/tRNA-V, or nontransduced cells, were lysed as hemagglutinin/interleukin II stimulation (PHA, 5 ␮g/ml, described previously.35 Proteins were subjected to Sigma, St Louis, MO, USA; IL-2 100 U/ml, United Cross electrophoresis in 10% sodium dodecyl sulfate polyacryl- Int’l Biotech, Beijing, PRC), 1 × 106/ml cells were cultured amide gels and subsequently transferred to a Poly Screen with 10 ml of transfected PA317 supernatant, for 3 days, (PVDF) membrane (Dupont). After the membrane was with daily replacement of fresh packaging cell-line super- blocked with 5% non-fat dry milk, the specific protein natant. The transduced PBMC, cultured only in IL-2 (50 expression was detected with rabbit polyclonal anti- U/ml), were challenged with HIV-1. mouse kappa chain (Sigma) for D8SFv. As a control, rab- For chronically infected H9/IIIB cells, transfected cells bit anti-human ␤-actin antibodies were used. The Dupont were first selected in G418 for 2 weeks and maintained Western Blot Chemiluminescence methodology was util- in the G418-free culture medium for 2 more weeks. Then, ized, with the manufacturer’s suggested protocol HIV-1 replication was assayed by measuring HIV-1 p24 (Dupont). antigen level in the supernatants, with standard cell cultures at 0.5–1 × 106 cells per ml, and subcultured every 3 Acknowledgements days. The levels of HIV-1 p24 antigen were determined via an enzyme-linked immunosorbant assay (ELISA- The authors would like to thank Dr Eli Gilboa for kindly Dupont, Boston, MA, USA). providing the DCt2T vector, Dr James A Hoxie for pro- The HIV-1 viral strains used in this study include NL4– viding the H9/IIIB cells, Dr Stephen Spector for provid- 3, which contains all open reading frames,32 and R7- ing primary HIV-1 isolates, Drs Omar Bagasra and Mark HXB2.33 The preparation of viral stocks and their A Laughlin for the many helpful discussions, SL Liu for titration, using tissue culture infection doses (TCID), technical assistance and Ms Rita M Victor and Ms Brenda were described previously in detail.7 A primary clinical O Gordon for excellent secretarial assistance. This work syncytia-inducing HIV-1 isolate was provided by Dr was supported in part by US PHS grant AI36552 and a Stephen A Spector (University of California at San Diego) Pediatric AIDS Foundation grant (50605) to RJP. (SI 92006262). References HIV-1 challenge studies The G418-selected mixed cell populations were first 1 Levy JA. Pathogenesis of human immunodeficiency virus infec- maintained in G418-free medium for at least 2 weeks, tion. Microbiol Rev 1993; 57: 183–289. 2 Cullen BR. Human immunodeficiency virus as a prototypic before HIV-1 infection. The nontransduced SupT1 or complex retrovirus. J Virol 1991; 65: 1053–1056. SupT1 cells which were transduced with CAT and differ- 3 Feinberg MB. The molecular biology and pathogenesis of HIV- ent SFvs and ribozymes were incubated with infectious 1 infection. Curr Opin Infect Dis 1992; 5: 214–220. NL4–3 or R7-HXB2 virus at various input MOI for 4 h. 4 Pomerantz RJ, Trono D. Genetic therapies for HIV infections: Cells were then washed four times with prewarmed, promise for the future. AIDS 1995; 9: 985–993. serum-free media. Cells were maintained in growth 5 Duan L-X, Pomerantz RJ. Intracellular antibodies for HIV-1 gene medium. Every 3 to 5 days, cells were split 1:2 to main- therapy. Sci Med 1996; 3: 24–33. tain a cell density of approximately 0.5–1 × 106 cells per 6 Mintz PL et al. Intracellular expression of single-chain variable ml and the culture supernatants were determined by fragments (SFv) to inhibit early stages of the viral life-cycle by HIV-1 p24 antigen ELISA. Cell viability was monitored targeting HIV-1 integrase. J Virol 1996; 70: 8821–8832. 7 Duan L-X et al. Intracellular immunization against HIV-1 infec- by trypan blue exclusion staining. tion of human T-lymphocytes: utility of anti-Rev single-chain variable fragments. Hum Gene Ther 1995; 6: 1561–1571. Ribozyme expression in cell culture 8 Pilkington GR et al. Recombinant human Fab antibody frag- Total RNA from different promoter-driven, ribozyme- ments to HIV-1 Rev and Tat regulatory proteins: direct selection expressing retroviral vector-transduced or nontrans- from a phage library. Mol Immunol 1996; 33: 439–449. Protein-RNA-based-HIV-1 gene therapy L Duan et al 543 9 Shaheen F et al. Targeting human immunodeficiency virus type 24 Malim MH et al. Stable expression of transdominant rev protein 1 reverse transcriptase by intracellular expression of single- in human T cells inhibits human immunodeficiency virus repli- chain variable fragments to inhibit early stages of the viral life- cation. J Exp Med 1992; 176: 1197–1201. cycle. J Virol 1996; 70: 3392–3400. 25 Kubota S et al. Nuclear preservation and cytoplasmic degra- 10 Symons RH. Small catalytic RNAs. Ann Rev Biochem 1992; 61: dation of human immunodeficiency virus type I Rev protein. 641–671. J Virol 1996; 70: 1282–1287. 11 Haseloff J, Gerlach WL. Simple RNA enzymes with new and 26 Pomerantz RJ, Seshamma T, Trono D. Efficient replication of highly specific endoribonuclease activities. Nature 1988; 334: human immunodeficiency virus type 1 requires a threshold 585–591. level of Rev: potential implications for latency. J Virol 1991; 66: 12 Sarvar N et al. Ribozymes as potential anti-HIV-1 therapeutic 1809–1813. agents. Science 1990; 247: 1222–1225. 27 Miller AD, Roseman GJ. Improved retroviral vectors for gene 13 Lee SW et al. Inhibition of human immunodeficiency virus type transfer and expression. BioTechniques 1989; 7: 980–990. I in human T cells by a potent Rev responsive element decoy 28 Yu M et al. A hairpin ribozyme inhibits expression of diverse consisting of 13-nucleotide minimal Rev-binding domain. J Virol strains of human immunodeficiency virus type I. Proc Natl Acad 1994; 68: 8254–8264. Sci USA 1993; 90: 6340–6344. 14 Rossi JJ. Controlled, targeted, intracellular expression of 29 Adeniyi-Jones S et al. Generation of long read-through tran- ribozymes: progress and problems. Trends Biotechnol 1995; 13: scripts in vivo and in vitro by deletion of 3′-termination and pro- 301–306. cessing sequences in the human tRNAiMet gene. Nucleic Acids 15 Thompson JD. Improved accumulation and activity of Res 1984; 12: 1101–1115. ribozymes expressed from a tRNA-based RNA polymerase III promoter. Nucleic Acids Res 1995; 23: 2259–2269. 30 Popovic M et al. Detection, isolation and continuous production 16 Pomerantz RJ, Seshamma T, Trono D. Efficient replication of of cytopathic retrovirus (HTLV-III) from patients with AIDS and human immunodeficiency virus type I requires a threshold level pre-AIDS. Science 1984; 224: 497–500. of Rev: potential implications for latency. J Virol 1992; 66: 31 Miller AD, Buttimore C. Redesign of retrovirus packaging cell- 1809–1813. lines to avoid recombination leading to helper virus production. 17 Duan L-X et al. Potential inhibition of human immunodeficiency Mol Cell Biol 1983; 6: 2895–2902. virus type I replication by an intracellular anti-Rev single-chain 32 Adachi A et al. Production of acquired immunodeficiency syn- antibody. Proc Natl Acad Sci USA 1994; 91: 5075–5079. drome-associated retrovirus in human and nonhuman cells 18 Snorri Th et al. Structure–function relationships of hammerhead transfected with an infectious molecular clone. J Virol 1986; 59: ribozyme from understanding to applications. TIBTECH 1995; 284–291. 13: 280–285. 33 Feinberg MB et al. HTLV-III expression and production involves 19 Denman RB. Using RNAFOLD to predict the activity of small complex regulation at the levels of splicing and translation of catalytic RNAs. Biotechniques 1993; 15: 1090–1094. viral RNA. Cell 1986; 46: 807–817. 20 Ozaki I et al. Ribozyme-mediated specific gene replacement of 34 Chomczynski P, Sacchi N. Single-step method of RNA extrac- the ␣1-antitrypsin gene in human hepatoma cells. Hum Gene tion by acid guanidinium thiocyanate-phenol-chloroform Ther (submitted). extraction. Anal Biochem 1987; 162: 156–159. 21 Duan L-X, Pomerantz RJ. Elimination of endogenous aberrant 35 Kevin K et al. Total protein extraction from cultured cells for kappa chain transcripts from sp2/0-derived hybridoma cells by use in electrophoresis and Western blotting. BioTechniques 1996; specific ribozyme cleavage: utility in genetic therapy of HIV-1. 20: 662–668. Nucleic Acids Res 1994; 22: 5433–5438. 36 Meyers G et al. Human Retroviruses and AIDS: A Compilation and 22 Cullen BR, Greene WC. Regulatory pathways governing HIV-1 Analysis of Nucleic Acid and Amino Acid Sequences. HIV Sequence replication. Cell 1989; 58: 423–426. Database, T-10, MSK710, Los Alamos National Laboratory: Los 23 Lee TC et al. Over-expression of RRE-derived sequence inhibits Alamos, NM, 1996. HIV-1 replication in CEM cells. New Biol 1992; 4: 66–74.