cDNA Encoding a Single-Chain Antibody to HIV p17 with Cytoplasmic or Nuclear Retention Signals Inhibits HIV-1 Replication

Avariety of gene therapy approaches are being explored to inhibit HIV-1 replication. One such approach involves the intracellular expression of Ab genes to target either the pre- or postintegration steps in HIV-1 replication (1, 2, 3, 4, 5). The HIV matrix protein, p17, is of particular interest because of its multiple roles in the HIV-1 replication cycle. It is synthesized from a gag precursor polyprotein after cleavage by a viral-coded protease (6, 7), and its membrane associations are necessary for viral assembly and release (8). p17 forms the inner capsid core of the virion and plays an important role in the incorporation of envelope proteins into virions via interaction with the cytoplasmic tail of glycoprotein 41 (9, 10, 11, 12); p17 is also involved in the transport of the viral preintegration complex into the nucleus (13, 14). Several p17 mutagenesis studies have already indicated that both N- and C-terminal mutations can inhibit HIV-1 replication (15, 16, 17, 18). Because of its multiple roles in the HIV-1 replication cycle, p17 is an attractive target for molecular intervention.

Recently, it has been shown that cells transfected with a cDNA encoding a single-chain (sc)² Ab to HIV-1 are resistant to HIV-1 infection (1, 2, 3, 4, 5). In this study, we constructed a cDNA encoding a sc Ab directed against p17. sc Abs consist only of the V regions of the heavy (V_H) and light (V_L) chains of the parental Ab; these regions are linked together by a flexible peptide linker. We also added specific localization signals to the cDNA construct to target the sc Ab fragment (scFv) to either the nucleus or the cytoplasm of the cell. T cells that had been transfected with these constructs were then infected with HIV-1, and the protective efficacy of the expressed scFv was evaluated.

The anti-gag p17 murine hybridoma cell line, MH-SVM33C9/ATCC HB 8975, was obtained from the American Type Culture Collection (ATCC, Manassas, VA); this cell line was shown to react only with gag p17 on HIV-1 Western blot strips (Dupont New England Nuclear, Boston, MA) and is known to bind a C-terminal epitope (DTGHSSQVSQNY) on p17 (19). mRNA was isolated from hybridoma cells by an mRNA isolation kit (Stratagene, La Jolla, CA), and the first strand of cDNA was synthesized as described previously (20) using Moloney murine leukemia virus reverse transcriptase (RT) (Life Technologies, Gaithersburg, MD). cDNA of the V_H and V_L chains were amplified by PCR. The primers used to amplify the V regions were based on degenerate sequences of the conserved regions of the hypervariable complementarity-defining regions (Ig Prime Methodologies, Novagen, Madison, WI). The PCR conditions and buffers were as described in the manufacturer’s protocol. The amplified fragments were cloned in the TA cloning system (Invitrogen, San Diego, CA) and sequenced using dideoxy chain termination reactions (Amersham, Arlington Heights, IL). The sequences obtained from the V_H and V_L regions were compared with published sequences. The V_H gene belonged to the miscellaneous group of mouse V_H genes that had a homology of 86% (21). The V_L gene belonged to group II of mouse V_L genes with a homology of 95% (22).

scFvs for bacterial expression were constructed using overlapping PCR. The primers for overlapping PCR were designed based on the obtained sequence data and included HindIII sites (indicated in bold) at the 5′ and 3′ ends of the V_H and V_L genes. Leader sequences were deleted, and a synthetic linker was used (Gly₄-Ser)₃ to link the V_H and V_L regions. The V_H and V_L regions were reamplified using primer 1 with primer 2 and primer 3 with primer 4, respectively. Primers 2 and 3 had a 17-bp overlap (indicated with italics). Subsequently, V_H and V_L were joined with primers 1 and 4 by an overlapping PCR. The products were sequenced, and the scFv construct was ligated into the pET25b+ vector (Novagen) using the HindIII restriction site (Fig. 1).

The primers used for the different amplification reactions were as follows: 1) 5′-gcctaagcttcagatccagttggtgcagtc-3′, 2) 5′-cggacccgccacctccagaccctccgccacctgcagagacagtgaccagagtc-3′, 3) 5′-tggaggtggcgggtccggcggtggagggtcggatgttgtgatgacccagactc-3′, 4) 5′-gagcaagctttttgatttccagcttggtac-3′, 5) 5′-cgctctagactaacactcattccgttga-3′, 6) 5′-cgctctagactataccttcctcttattcggtggagtacactcattcctgttgaagctc-3′, and 7) 5′-gcctaagcttatcatggattggcagatccagttggtgcagtc-3′.

For mammalian expression, the V_H and V_L regions and the C region of κ light chain (C_κ) were reamplified using primer 7 with primer 2 and primer 3 with primer 5, respectively. The amplified fragments were joined by overlapping PCR strategy using primer pairs 7 and 5. The primers were designed to include the HindIII site at the 5′ end for the V_H chain and the XbaI site at the 3′ end for the C_R chain (indicated in bold). The 5′ end also includes a Kozak sequence, and the first six nucleotides from the leader sequence. The 3′ end has a stop signal. The SV40 nuclear localization signal (NLS) sequence (i.e., TPPKKRKV) was included at the C_R terminus of the scFv in an overlapping PCR reaction; this reaction was similar to the one described for the cytoplasmic construct, except primer 5 was replaced with primer 6. Subsequently, the scFv constructs were ligated at the HindIII and XbaI sites into a modified version of pRcCMV (Invitrogen), in which the TCR enhancer (TCE) (23) was inserted into the BglII site (20).

The Escherichia coli strain BL21(DE)₃ was transformed with both sense and antisense orientation pET25b+ scFv constructs. The transformed bacteria were induced with 0.4 mM of isopropyl β-d-thiogalactoside (IPTG); the lysates were analyzed by 12% SDS-PAGE and visualized after Coomassie blue staining for protein expression. For analysis of the induced proteins, both soluble and insoluble fractions from the bacterial lysates were resolved on 10% SDS-PAGE and blotted onto polyvinylidene difluoride membranes (Novex, San Diego, CA). The blotted proteins were blocked with milk diluent solution (Kirkegaard and Perry, Gaithersburg, MD) and then reacted with herpes simplex virus (HSV)-Tag Ab (1:1000; Novagen) for 2 h followed by anti-mouse peroxidase conjugate (1:2000, Sigma, St. Louis, MO). Blots were washed three times and then reacted with 3,3′,5,5′-tetra methyl benzidine (TMB) substrate (Kirkegaard and Perry). To show that the induced bacterial scFv bound to HIV-1 p17, the bacterial proteins were purified on Ni²⁺ columns according to the manufacturer’s recommendations (Novagen). A total of 5 μg of E. coli expressed recombinant fusion protein (i.e., HIV-1 p17 and maltose-binding protein) (Intracel, Cambridge, MA); an equal amount of lysozyme protein (Sigma) was separated on a 10% Tris glycine SDS gel (Novex), blotted, and reacted with the bacterial purified scFv followed by anti-HSV Tag Ab and anti-mouse Ab peroxidase conjugate. The specific bands were visualized by 3,3′, 5,5′-tetra methyl benzidine substrate staining (Kirkegaard and Perry). The p17 bacterial fusion protein migrated as a 52-kDa band.

Expression and targeting to different cell compartments were studied by the transient expression of the scFv proteins. Briefly, 5 × 10⁵ COS-7 cells were grown in four chamber slides and then transfected with 2.5 μg of supercoiled DNA (vector alone, scFvCyt, or scFvN) plus N-(1-[2,3-dileoyloxy]propyl)-N,N,N-trimethylammonium methyl sulfate reagent according to the manufacturer’s recommendations (Boehringer Mannheim, Indianapolis, IN). The transfected cells were analyzed for their expression of scFv at 48 h posttransfection by staining with FITC-labeled anti-mouse κ-chain mAb (Sigma).

To generate stable transfectants, 20 μg of ScaI-linearized plasmid DNA was transfected into the Jurkat cell clone E-6 (ATCC) by electroporation. Briefly, 2 × 10⁶ cells in 0.8 ml of RPMI 1640 were mixed with DNA from vector alone, scFvCyt, or scFvN; electroporated (960 μF and 250 V); selected with G418 (1.5 mg/ml; Life Technologies); and expanded into cell lines.

A total of 2 × 10⁶ cells were incubated for 30 min with 1.2 ml of methionine-free RPMI 1640 medium supplemented with 10% FBS, 2 mM glutamine, and 1 mM sodium pyruvate. Next, cells were metabolically labeled with [³⁵S]methionine (125 μCi/ml; Express protein labeling mix, Dupont) for 4 h. After labeling, supernatants were collected; cells were washed three times with HBSS and then placed in lysis buffer (10 mM Tris-HCl (pH 7.4), 1% Nonidet P-40 (w/v), 150 mM NaCl, 1 mM EDTA, 1 mg/ml BSA, and 1 mM PMSF). Radioactivity was determined by TCA precipitation. The lysed fractions were precleared with protein A plus G (Calbiochem, La Jolla, CA) for 30 min and then reacted with anti-κ-chain mAb overnight at 4°C. The complex was subsequently precipitated with protein A plus G Sepharose for 1 h. The pellet was dissolved in SDS-PAGE buffer, electrophoresed on 12% SDS-PAGE, and subjected to autoradiography following fluorography for 30 min.

Jurkat CD4⁺ cell lines that had been transfected with scFvCyt, scFvN, or vector alone were monitored for both cell growth and surface markers as described previously (20). A total of 1 × 10⁵ cells/ml in 24-well plates were assessed for cell growth, and live cells were counted by trypan blue staining. Cells in PBS containing 1% BSA and 0.05% sodium azide were incubated at 4°C with FITC-labeled Abs to CD4 or CD3 (Becton Dickinson, Mountain View, CA) or with a mouse IgG1-FITC Ab control (Sigma) and then analyzed with a FACScan (Becton Dickinson).

Human CD4⁺ cell lines that had been stably transfected with scFvCyt, scFvN, or vector alone were infected with either HIV-1 IIIB (100 or 500 tissue culture ID₅₀ (TCID₅₀)/10⁶ cells) or the 3′-azido-3′-deoxythymidine-resistant primary isolate 5AO12 (24) in a final volume of 0.5 ml and incubated for 2 h at 37°C. Cells were then washed with HBSS; resuspended in 5 ml of DMEM supplemented with 10% FCS, 2 mM glutamine, and 1 mM sodium pyruvate; and incubated at 37°C in T25 flasks. At 3-day intervals, both media and cells were collected and centrifuged at 400 × g for 10 min; supernatants were stored. Fresh medium was supplemented to maintain a constant cell number. Viral titers in supernatants were measured by ELISA with an HIV-p24 kit (Dupont) according to the manufacturer’s recommendations or by infectivity titration on CEM174 cells. The viability of cells and the number of syncytia were counted by trypan blue staining. The RT activity in the supernatants was measured as described previously using poly(rA)-oligo(dT) as a primer template (25).

For p17 detection, cells that had been infected for 9 days with a high viral dose (500 TCID₅₀ or 300,000 RT activity/10⁶ cells) were radiolabeled and lysed. Samples with equal TCA-precipitable counts were immunoprecipitated with either 5 μg of sheep anti-p17 Ab (AIDS Reagent and Reference Program, National Institutes of Health) or 5 μg of anti-κ-chain mAb overnight after preclearing with sheep and mouse serum, respectively. Following incubation, lysates were reacted with recombinant protein G (Life Technologies) for 1 h, washed, pelleted, and solubilized in SDS-PAGE buffer. Samples were subsequently electrophoresed on 15% SDS-PAGE and subjected to autoradiography after fluorography.

cDNA encoding V regions from a murine hybridoma expressing Ab to p17 was cloned and sequenced. Both the deduced amino acid sequences of the V_H and V_L chains and a flexible linker (GGGGS)₃ are shown in Figure 1 A. The secretory leader sequence has been removed based on available consensus sequences.

scFvs were then constructed and cloned into a pET25b⁺ vector (Fig. 1,B) and expressed in the E. coli strain BL21(DE)₃ with both sense (+) and antisense (−) scFv gene constructs. The transformed bacteria were induced with IPTG, and bacterial lysates were analyzed on 12% SDS-PAGE. Figure 2,A shows a major band of 35 kDa which corresponds to the predicted molecular mass of the scFv protein in bacterial lysates having a sense (+) but not an antisense (−) construct. The 35-kDa protein was found in both the soluble and insoluble fractions (Fig. 2,B) and bound specifically to HIV p17 (Fig. 2,C). The binding pattern of the scFv to both the 52-kDa rHIV-1 p17 fusion protein (Fig. 2 C) and the native viral p17 (data not shown) was similar to that observed with the original mAb.

Human CD4⁺ Jurkat T cells were stably transfected with either vector alone or scFv constructs for cytoplasmic (scFvCyt) or nuclear (scFvN) expression (Fig. 1,B). Stable cell lines were obtained by G418 selection. The cell line JV and two clones each of scFvCyt (no. 3 and 4) and scFvN (no. 3 and 4) were subsequently labeled with [³⁵S]methionine and immunoprecipitated with anti-mouse κ-chain Ab. The predicted 38-kDa band (i.e., scFv) was found in CD4⁺ T cells that had been transfected with constructs containing either the cytoplasmic or nuclear retention signal but was not found in cells that had been transfected with vector alone (data not shown). To determine the localization of the expressed proteins, COS-7 cells were transiently transfected with scFvCyt, scFvN, or vector alone. Figure 3 shows that scFvs were found primarily in the cytoplasm of COS-7 cells that had been transfected with scFvCyt and primarily in the nucleus of COS-7 cells that had been transfected with scFvN. No changes in the growth rate or expression of CD3 or CD4 were observed in CD4⁺ T cell lines that had been stably transfected with either scFvCyt or scFvN as compared with vector controls (Fig. 4, A and B).

To determine whether scFvCyt or scFvN can bind specifically to p17 within HIV-1 infected CD4⁺ T cells, lysates from cells that had been infected with a high infectious dose (5 times higher than the HIV-1 challenge dose of 100 TCID₅₀) were examined for their capacity to immunoprecipitate radiolabeled p17 with either anti-mouse κ-chain mAb, which can immunoprecipitate scFv-bound p17, or sheep anti-HIV-1 p17 antiserum, which can precipitate both bound and unbound p17. Figure 5,A shows that p17 was immunoprecipitated with anti-mouse κ-chain mAb from both scFvCyt and scFvN cell lines but to a much greater extent from cells that had been transfected with scFvCyt than with scFvN. Immunoprecipitation bands were not detected in either HIV-infected or uninfected JV cells that served as controls. Although scFvN bound less p17 than scFvCyt, approximately the same amount of p17 was present in the three different infected T cell lines as measured by immunoprecipitating the cell lysates with sheep anti-p17 Ab (Fig. 5 B). The nearly equal levels of p17 in the three different HIV-1-infected cell lines also argue that the inhibition in virus replication occurred mainly after the infection was established.

To study the effect of anti-p17 scFv expression in different cell compartments on HIV-1 replication, stably transfected human Jurkat CD4⁺ T cells were challenged with HIV-1 IIIB. Figure 6,A shows that RT activity was not detected or was only slightly above background levels for the 21 days of the experiment in cells expressing scFvCyt. RT activity was inhibited for as long as 12 to 15 days in cells expressing scFvN, after which it approached the level seen in cells that had been transfected with vector alone. Similarly, p24 gag protein (Fig. 6,B) was not detected in any appreciable amount until 18 to 21 days postinfection in cells expressing scFvCyt; at this time, only at low concentrations were detected compared with the control JV cells. In cells expressing scFvN, p24 gag was inhibited for 12 to 15 days, after which it increased substantially. Overall, scFvCyt cells showed continuous HIV-1 inhibition in terms of syncytium formation, p24 and RT levels, and the maintenance of cell viability during the 21-day experiment. The data in the Figure 6 are representative of seven separate experiments. Figure 7 shows that the replication of the drug-resistant primary HIV-1 isolate 5AO12 was also inhibited in scFv cell lines.

The intracellular expression of an Ab that binds to a specific protein has many potential uses ranging from elucidating the function of the target protein to gene therapy. Intracellular Abs act by reducing the effective concentration and/or function of the target protein. The ease in making Abs to most proteins and to many different epitopes on the protein compares favorably with the difficulty in preparing chemicals or drugs that react with specific proteins. Moreover, in the case of viral infections such as HIV, Abs directed to the various viral proteins involved in different stages of the viral replication cycle also can be made with ease. Consequently, intracellular Abs enormously broaden the potential range of viral targets over that of neutralizing Abs that react only with extracellular virus. Intracellular Abs may also prove useful against drug-resistant viral mutants.

In the present study, we prepared an scFv from the cDNA of a mAb to p17 and expressed it in bacterial and mammalian cells. Our anti-HIV p17 scFv retained the specificity of the parental Ab molecule (26). In addition we used retention signals to direct scFvs to the nucleus or cytoplasm (2, 4, 27, 28). Murine C_κ was added to the scFv to increase its stability in the cytoplasm and the SV40 NLS was used to direct the scFv to the nucleus. Our studies showed that the anti-HIV-1 p17 scFvs were expressed in the cytoplasm or nucleus of the cell and were functionally active.

Our experiments have shown that HIV-1 replication is inhibited in human CD4⁺ T cells that express either anti-p17 scFvCyt or scFvN. The viral titer, as measured by p24 and RT, was considerably lower in these cells than in those expressing the vector alone. The stronger inhibitory effect of scFvCyt as compared with scFvN is consistent with our finding that scFvCyt binds more p17 than does scFvN. This effect is also consistent with the report stating that the majority of the total p17 within a cell is in the cytoplasm as compared with the nucleus (14). Some of the interactions of scFvN with p17, however, may also occur in the cytoplasm, since scFvN is first synthesized in the cytoplasm and then translocated to the nucleus. In this context, Leavy-Mintz et al. (2), recently used an scFv-based approach to target HIV integrase, which is a component of the preintegration viral complex that interferes only with integration events in the nucleus. They, too, found greater protection with cytoplasmic rather than nuclear scFv expression. In our experiments, the inhibition of HIV-1 replication was found with both high and low multiplicities of infection in both scFvCyt and scFvN cell lines (our unpublished observations). However, the degree of inhibition varied, and was most pronounced at a low infectious dose. Similar observations have been made in other systems (5, 29). Although all of the scFv-transfected cell lines used in our experiments were selected with G418 and cloned several times, the possibility that scFv cDNA may have been extruded from some of the cells has not been excluded. Extrusion of transfected cDNA might be one of the factors contributing to the breakthrough in infectivity. Another factor might be the generation of escape mutants in which the Ab-specific epitope is mutated and therefore results in a loss of susceptibility to HIV-1 Ab.

The present study extends the very recent study by Levin et al. (29), in which p17 was similarly targeted with an intracellular Ab. In contrast to Levin et al., who used a cDNA encoding a Fab fragment, we used an scFv that was targeted to a different and highly conserved epitope on p17. scFvs have some advantages over Fabs. First, the size of the foreign genetic information that is introduced into the cell is smaller and is less likely to be toxic. Second, the cytoplasm and nucleus are thought to be nonconducive environments for the assembly of the heavy and light chains to form a Fab (27). This problem does not exist with an scFv, since the heavy and light chains are already linked. In addition, we used retention signals to specifically target the Ab to the cytoplasm or nucleus. Cell lines expressing scFvCyt were more effective in inhibiting HIV than cell lines expressing scFvN. However, both scFvCyt and scFvN formed complexes with and retained p17 within the cell. Our study also adds to the increasing body of evidence that Abs can be expressed stably and without toxicity in a variety of cell types (27, 28, 30, 31, 32), targeted to different viral proteins involved in the replication cycle (1, 2, 3, 4, 5), and directed to specific compartments within the cell (2, 3, 4, 27, 28, 30); we also show that Abs act on drug-resistant isolates. The possibility that the simultaneous expression of several scFvs directed against different viral proteins and targeted to specific cellular compartments may be more efficacious than one scFv directed against a single viral protein alone is currently under investigation.

We thank Drs. Krishna Devadas, Andrew Yeudall, George Chen, Swapan De, and Ms. Aleksandra Jegorow for suggestions and for their critical reading of the manuscript.