Virion-Targeted Viral Inactivation of Human Immunodeficiency Virus Type 1 by Using Vpr Fusion Proteins

Abstract

Inactivation of progeny virions with chimeric virion-associated proteins represents a novel therapeutic approach against human immunodeficiency virus (HIV) replication. The HIV type 1 (HIV-1) Vpr gene product, which is packaged into virions, is an attractive candidate for such a strategy. In this study, we developed Vpr-based fusion proteins that could be specifically targeted into mature HIV-1 virions to affect their structural organization and/or functional integrity. Two Vpr fusion proteins were constructed by fusing to the first 88 amino acids of HIV-1 Vpr the chloramphenicol acetyltransferase enzyme (VprCAT) or the last 18 C-terminal amino acids of the HIV-1 Vpu protein (VprIE). These Vpr fusion proteins were initially designed to quantify their efficiency of incorporation into HIV-1 virions when produced in cis from the provirus. Subsequently, CD4⁺ Jurkat T-cell lines constitutively expressing the VprCAT or the VprIE fusion protein were generated with retroviral vectors. Expression of the VprCAT or the VprIE fusion protein in CD4⁺ Jurkat T cells did not interfere with cellular viability or growth but conferred substantial resistance to HIV replication. The resistance to HIV replication was more pronounced in Jurkat-VprIE cells than in Jurkat-VprCAT cells. Moreover, the antiviral effect mediated by VprIE was dependent on an intact p6^gag domain, indicating that the impairment of HIV-1 replication required the specific incorporation of Vpr fusion protein into virions. Gene expression, assembly, or release was not affected upon expression of these Vpr fusion proteins. Indeed, the VprIE and VprCAT fusion proteins were shown to affect the infectivity of progeny virus, since HIV virions containing the VprCAT or the VprIE fusion proteins were, respectively, 2 to 3 times and 10 to 30 times less infectious than the wild-type virus. Overall, this study demonstrated the successful transfer of resistance to HIV replication in tissue cultures by use of Vpr-based antiviral genes.

The assembly and maturation of human immunodeficiency virus (HIV) particles constitute a complex process in which the structural gag, pol, and env gene products are expressed in the form of polyprotein precursors. The production of infectious viral particles requires incorporation of the Env glycoproteins gp120 and gp41 during viral budding and processing of the Gag and Gag-Pol polyproteins by the viral protease (12, 34). In addition to gag, pol, and env gene products, HIV type 1 (HIV-1) virions have been shown to contain several accessory proteins, including Vpr (7), Vif (15), and Nef (35). Vpr, a 14-kDa, 96-amino-acid nuclear protein, is packaged into HIV virions in amounts similar to those of Gag, while Vif and Nef appear to be incorporated in quantities comparable to that of Pol (5, 6, 35). The HIV-1 Vpr protein is incorporated in trans into viral particles through an interaction with the p6 domain of the Gag polyprotein precursors (20, 23).

The resistance of AIDS to traditional drug therapy has prompted a search for alternative treatments for this disease. One potential approach, termed intracellular inhibition or immunization, is designed to render cells resistant to viral replication and to limit the spread of virus in a cell culture or an individual (1, 14). Inhibition of HIV replication by such a strategy has now been established with different antiviral genes, including those directed at a nucleic acid or protein target (14). Virus-targeted inactivation represents a novel genetic approach to interfering with HIV replication. In this strategy, a deleterious amino acid sequence or sequence with enzymatic activity is fused to a virion-associated component to prevent the production of infectious viral particles and the subsequent spread of de novo infection (3). Several studies have demonstrated the efficient packaging of Vpr fused with a polypeptide sequence. The incorporation of Vpr fused to integrase (VprIN), reverse transcriptase (RT) (VprRT), enzymatically active staphylococcal nuclease, chloramphenicol acetyltransferase (CAT), and a trans-dominant negative mutant of the HIV-1 protease into HIV particles has been demonstrated (3, 13, 31, 37–39). Moreover, the expression of Vpr proteins fused to HIV-1 protease cleavage sites was recently shown to affect HIV-1 replication (33). However, the possibility that the stable transfer of genes encoding Vpr fusion proteins in HIV target cells could provide genetic resistance to viral replication remains to be demonstrated. In this paper, we report that a reporter enzyme (CAT) or a short peptide sequence derived from the HIV-1 Vpu protein fused to the carboxyl terminus of Vpr can be targeted specifically and efficiently into HIV particles. Using retroviral vectors, we demonstrate that the stable expression of these Vpr fusion proteins in CD4⁺ T cells can confer substantial resistance to HIV-1 replication by affecting the infectivity of progeny virions.

MATERIALS AND METHODS

Site-directed mutagenesis and plasmid DNA constructions.

HxBRU (long terminal repeat [LTR]-gag⁺ pol⁺ vif⁺ vpr⁺ tat⁺ rev⁺ vpu⁻ env⁺ nef⁻-LTR) and HxBRU-R- (LTR-gag⁺ pol⁺ vif⁺ vpr⁻ tat⁺ rev⁺ vpu⁻ env⁺ nef⁻-LTR) are two isogenic infectious molecular clones of HIV-1 that differ only in their ability to express Vpr (22). A unique XbaI site was inserted at position 5410 (position 1 is the site of transcription initiation of the HxBRU molecular clone) in HxBRU by use of two-step PCR-based mutagenesis as previously described (22). The nucleotide sequence of the sense mutagenic oligonucleotide was as follows: 5′-CAG AGG AGA TCT AGA AAT GGA GCC-3′. The resulting HxBRUXbaI mutant construct was confirmed by sequencing with a Sequenase kit (United States Biochemical Co.).

To generate a molecular clone of HIV-1 containing Vpr fusion proteins, an XbaI-XbaI fragment carrying the last 18 amino acids of the HIV-1 Vpu protein (IE) or the CAT enzyme from amino acids 2 to 219 was generated by PCR with, as a template, the previously described HxBH10 molecular clone of HIV-1 (40) or the SL3CAT expression plasmid (7). The XbaI primers used to perform the amplification step were as follows: IE, sense A, 5′-TCA TCT AGA GTG GAG ATG GGG-3′; IE, antisense B, 5′-CAT TCT AGA ATT ATA TCC TCA-3′; CAT, sense A, 5′-AAG GAA GCT TCT AGA GAG AAA AAA-3′; CAT, antisense B, 5′-TTA AGG GCA TCT AGA ACT GCC TTA-3′. Insertion of the PCR-generated XbaI fragment carrying IE or CAT into the newly created XbaI site in HxBRUXbaI yielded, respectively, HxBRU-VprIE and HxBRU-VprCAT.

To obtain the HxB89-R- provirus, the env gene sequence of HxBH10 (41) from nucleotides 6350 to 8470 was replaced by the corresponding region from proviral clone 89.6 (9) (nucleotides 6346 to 8468). The HxB89-R- vpr gene contains a single-base insertion at position 5773 which generates a frameshift in the vpr coding sequence and yields a truncated, 72-amino-acid protein. Double-amino-acid-substitution mutations in the p6 coding region of HxB89-R- were introduced by site-directed mutagenesis. The following oligonucleotide was used for the construction of mutant clone LF44,45PS: 5′-CTC AGA TCA CCC TCT GGC AAC GAC-3′. The presence of the mutations in the final proviral construct was confirmed by DNA sequence analysis.

To construct retroviral vectors and pCEP-4 (Invitrogen) expression vectors containing the fusion proteins, vprIE, vprCAT, or control vpr genes were PCR amplified with specific BamHI oligonucleotide primers surrounding vpr or vpr fusion protein sequences (sense primer A, 5′-GAA ACT GAC AGG ATC CAG ATG GAA C-3′; antisense primer B, 5′-GGA TGC TTC CAG GAT CCT AGT CTA GGA-3′) and HxBRU, HxBRU-R-, HxBRU-VprIE, or HxBRU-VprCAT as templates. The resulting fragments were cloned into the BamHI restriction site of pBabepuro (30) or pCEP4 to generate pBabepuro-VprIE, pBabepuro-VprCAT, pCEPVpr, pCEPVpr-R-, pCEPVprIE, and pCEPVprCAT. For pBabepuro-RevM10, RevM10 was PCR amplified from pM10 (26) with specific SalI oligonucleotide primers (sense primer A, 5′-CGA AGC TAG TCG ACT AGG CAT CTC C-3′; antisense primer B, 5′-CCT ATC TGT CGA CTC AGC TAC TGC-3′) and cloned into the SalI restriction site of pBabepuro.

Cells and DNA transfection.

MT4, Jurkat, COS-7, and 293T cells (all from the American Type Culture Collection) and HeLa-CD4-LTR/β-Gal indicator cells were maintained as described previously (18, 22). MT4 (5 × 10⁶) and Jurkat (1 × 10⁷) cells were transfected with 10 μg of plasmid DNA by the DEAE-dextran transfection method as described previously (22). COS-7 cells (10⁶) were transfected with 20 μg of proviral DNA and 293T cells (10⁶) were transfected with 20 μg of pCEP expression vectors by the calcium phosphate method as described previously (22).

Cell transduction and infection.

In order to generate Jurkat cell lines expressing Vpr fusion proteins or RevM10, retroviral vector technology was used. The ΨCRIP packaging cell line was cultured as described previously (10). Aliquots (10 μg) of pBabepuro, pBabepuro-VprIE, pBabepuro-VprCAT, or pBabepuro-RevM10 vectors were transfected into ΨCRIP packaging cells by the calcium phosphate method (22). ΨCRIP cells that stably produced amphotropically packaged pBabepuro, pBabepuro-VprIE, pBabepuro-VprCAT, or pBabepuro-RevM10 vectors were obtained by selection with 1.2 μg of puromycin (Calbiochem) per ml, generating ΨCRIP-puro, ΨCRIP-VprIE, ΨCRIP-VprCAT, or ΨCRIP-RevM10 cell lines, respectively. Transduction of Jurkat cells was performed as described previously (36). Puromycin-resistant Jurkat cell populations were isolated, characterized, and subsequently challenged with HIV-1. Infections were performed as follows: 3 × 10⁶ parental, control, or transduced Jurkat cells were absorbed for 2 h with different amounts of HIV in 1 ml of RPMI medium plus 10% fetal calf serum. Following infection, the culture medium was changed every 3 days and cells were resuspended in 10 ml of fresh RPMI medium plus 10% fetal calf serum at densities of 5 × 10⁵ viable cells per ml.

Metabolic labeling, immunoprecipitation, and CAT activity measurement.

At 72 h posttransfection, COS-7 cells were starved in methionine-free Dulbecco modified Eagle medium for 30 min. The cells were then metabolically labeled with 50 μCi of [³⁵S]methionine per ml for 16 h. For each sample, viral particles were pelleted by ultracentrifugation of an equal volume of supernatant fluid through a 20% sucrose cushion at 35,000 rpm for 2 h at 4°C in a Beckman 55.2 Ti rotor. The labeled pelleted virus and cells were lysed in radioimmunoprecipitation assay buffer and immunoprecipitated as previously described (22) with an HIV-1-positive human serum combined with either a rabbit anti-Vpr polyclonal serum (22), a rabbit anti-Vpu peptide serum (8), or a rabbit anti-CAT polyclonal serum (5′ Prime 3′ Prime) in a 1:1 ratio. HIV-associated proteins were separated on a sodium dodecyl sulfate–12.5% polyacrylamide gel and visualized by autoradiography. Quantitative analysis of immunoprecipitated proteins was performed by densitometric scanning of the autoradiographic signals on a Personal Densitometer (Molecular Dynamics) with ImageQuant software, version 3.22. In order to evaluate the level of VprCAT expressed in cells and incorporated into virions, CAT assays were performed (16). Cells and virus pelleted through a 20% sucrose cushion (see above) (virus was produced from HxBRU-VprCAT-transfected COS-7 cells) were harvested with 250 mM Tris-HCl (pH 7.4) and processed as previously described (16).

RNA extraction and Northern blot analysis.

Cytoplasmic and nuclear RNAs were isolated from 10⁷ parental or transduced Jurkat cells with TRIzol reagent as described by the manufacturer (GIBCO/BRL). Poly(A) RNA was then purified with a QuickPrep Micro mRNA purification kit as instructed by the manufacturer (Pharmacia). RNA was resolved on a 1% agarose gel containing formaldehyde and transferred to a nitrocellulose membrane (Bio-Rad). The Vpr probe was generated by PCR amplification with BamHI primers (see above) and pNL4.3 as a template. The purified DNA fragment was labeled with [α-³²P]dCTP by use of a random-primer DNA-labeling system (Pharmacia). Prehybridization and hybridization were performed as previously described (21).

RESULTS

Construction of Vpr fusion proteins.

Recent studies reported that the expression of the HIV-1 Vpr protein induced the accumulation of cells in the G₂ phase of the cell cycle (17, 32). Mutational analysis revealed that the highly charged carboxy-terminal domain appears to be critical for Vpr stability and cell cycle arrest, while the integrity of the N-terminal amphipatic alpha helix is required for virion incorporation and nuclear localization (11, 24, 25). Since the expression of therapeutic molecules should not affect cell homeostasis in vitro and more importantly in vivo, we constructed Vpr fusion proteins by fusing antigens to the carboxyl-terminal end of a truncated Vpr mutant (88 amino acids). Two Vpr fusion proteins were initially generated to explore the possibility of targeting a sequence with enzymatic activity or an epitope tag into HIV-1 virions. For the first fusion partner, we used the C-terminal 18 amino acids of the Vpu protein, thus generating the VprIE protein. We selected this amino acid sequence because it contains an immunodominant epitope (IE) well recognized by a rabbit anti-Vpu peptide serum (8). Moreover, in order to study the efficiency of incorporation of larger proteins and to investigate whether an enzyme could be active in virions, we used the CAT protein as the second fusion partner (VprCAT). DNA fragments encoding the two Vpr fusion proteins were generated and subsequently cloned into the HxBRU infectious molecular clones as well as into the pBabepuro retroviral vector and the pCEP4 expression vector as described in Materials and Methods.

Protein expression and replication potential of HIV-1 virions carrying VprCAT and VprIE fusion proteins.

To test the expression and incorporation of the Vpr fusion proteins into HIV-1 virions, HxBRU-VprIE, HxBRU-VprCAT and, as a control, HxBRU were transfected into COS-7 cells. At 72 h posttransfection, labeled cell lysates and pelleted virus were immunoprecipitated with an HIV-1-positive human serum combined with specific anti-Vpr, anti-Vpu, or anti-CAT antibodies. Results indicated that antibodies directed against Vpr or Vpu immunoprecipitated the predicted 18-kDa VprIE protein from virions and cell lysates (Fig. 1, lanes 3, 4, 9, and 10). Similarly, antibodies directed against Vpr or CAT detected the predicted 35-kDa VprCAT protein in virions and cell lysates (Fig. 1, lanes 5, 6, 11, and 12). This result indicates that VprIE and VprCAT fusion proteins are expressed from recombinant proviruses in transfected COS-7 cells. We observed with HxBRU-VprIE a second band corresponding to the molecular mass of the 16-kDa Vpr protein. This Vpr-related product appears to result from a deletion of the Vpu-derived sequences, possibly by recombination rearrangement in COS-7 cells, since the product comigrated with Vpr and was immunoprecipitated with the anti-Vpr antibody but not with the anti-Vpu antibody (Fig. 1, lanes 9 and 10). Under our conditions, low-molecular-mass products derived from VprCAT were not detected in significant amounts (Fig. 1, lanes 11 and 12).

FIG. 1.

Expression and incorporation of Vpr fusion proteins into HIV-1 virions. HxBRU, HxBRU-VprIE, and HxBRU-VprCAT were transfected into COS-7 cells. At 72 h later, cell lysates and pelleted virions were analyzed by radioimmunoprecipitation. Samples were probed with an HIV-1-positive human serum combined with either anti-Vpr, anti-Vpu, or anti-CAT antibodies as indicated.

Figure 1 also demonstrates that VprIE and VprCAT can be incorporated into HIV-1 virions (lanes 9 to 12). The efficiency of incorporation of the Vpr fusion proteins was evaluated by densitometric scanning of the virion-associated Vpr and p6 RT product bands. The levels of Vpr or Vpr fusion proteins normalized to the number of methionines and to the amount of p66 RT protein detected in viral particles were found to be similar. This result indicates that when expressed in cis, both VprCAT and VprIE fusion proteins were packaged with the same efficiency as wild-type Vpr. Interestingly, the production of virus from HxBRU-VprIE- or HxBRU-VprCAT-transfected COS-7 cells was found to be consistently higher than that from HxBRU-transfected cells (Fig. 1). It is likely that the cell cycle arrest resulting from sustained Vpr expression affected the number of transfected cells after 72 h and consequently diminished the amount of virus produced in the culture. In addition, CAT activity was detected in both cell lysates and virions, indicating that the VprCAT fusion protein with amino acids 1 to 88 of Vpr remained enzymatically active even in HIV-1 particles (data not shown). Overall, these data strongly suggest that in the context of the provirus, the Vpr protein with amino acids 1 to 88 is sufficient to specifically target an 18-amino-acid peptide and a 218-amino-acid protein into virions with an efficiency comparable to that of wild-type Vpr.

To further analyze the effects of Vpr fusion proteins on viral replication, we initially tested the ability of virus stocks produced from HxBRU-VprIE- and HxBRU-VprCAT-transfected COS-7 cells to infect CD4⁺ T-cell lines. In Jurkat and MT4 cell lines, replication of HxBRU-VprIE and HxBRU-VprCAT was severely impaired compared to that of the control HxBRU virus (data not shown). Impairment of viral replication was found to result in part from the introduction of additional DNA sequences between the Tat splice acceptor and the Tat initiation codon rather than from the expression of the Vpr fusion proteins only. Indeed, transfection of HxBRU-VprIE or HxBRU-VprCAT in Jurkat cells constitutively expressing Tat revealed that the replication impairment of these viruses could be corrected in part by the expression of Tat (data not shown). Since the effect of VprIE and VprCAT expression on HIV replication could not be precisely evaluated in the context of a replicating virus, Jurkat CD4⁺ T-cell lines expressing VprIE or VprCAT in trans were generated by retroviral vector technology.

Generation and characterization of Jurkat CD4⁺ T-cell lines expressing Vpr fusion proteins.

Amphotropic murine retroviruses carrying pBabepuro, pBabepuro-VprIE, and pBabepuro-VprCAT were used to transduce Jurkat cells. Following puromycin selection, populations of cells consisting of a pool of, respectively, Jurkat-puro, Jurkat-VprIE, and Jurkat-VprCAT puromycin-resistant cell clones were isolated and amplified. To evaluate the level of VprIE or VprCAT gene expression in transduced Jurkat cell lines, we performed Northern blot analysis of purified total mRNAs by using a vpr-specific probe as described in Materials and Methods. Figure 2A shows the expected 2.5- and 2.6-kb vprIE- and vprCAT-specific mRNAs, detected only in Jurkat-VprIE and Jurkat-VprCAT cell lines, respectively (lanes 2 and 3). All attempts to detect VprIE and VprCAT protein expression by immunoprecipitation or Western blot analysis were unsuccessful, indicating that the level of expression of both Vpr fusion proteins in transduced Jurkat cells was low. We also attempted to demonstrate the expression and virion incorporation of the VprCAT fusion protein in HIV-1-infected Jurkat-VprCAT cell lines by CAT activity measurements. Briefly, 3 × 10⁶ parental or Jurkat-VprCAT cells were infected with equivalent amounts (10⁶ cpm of RT activity) of HxBRU and HxBRU-R-. At 6 days postinfection, cells and virus-containing supernatants were collected, and virions were pelleted by ultracentrifugation through a 20% sucrose cushion. CAT activity was undetectable in infected and uninfected Jurkat-VprCAT cells (data not shown). In contrast, CAT activity associated with Vpr⁺ and Vpr⁻ HIV-1 virions was detected (Fig. 2B), indirectly indicating that VprCAT proteins were expressed in Jurkat-VprCAT cells. These results also suggest that trans incorporation concentrated VprCAT in the virions to detectable levels. Importantly, virion incorporation of VprCAT appeared less efficient in HxBRU than in HxBRU-R-, since the level of CAT activity associated with the Vpr⁺ virus was lower than the level of CAT activity associated with the Vpr⁻ virus. This result suggests that competition for virion incorporation occurred between the virus-expressed Vpr and the VprCAT fusion protein.

FIG. 2.

Characterization of Jurkat cells expressing Vpr fusion proteins. (A) Analysis of vprIE- and vprCAT-specific mRNA expression in transduced Jurkat cells by Northern blotting with a vpr-specific probe. (B) CAT activity associated with HIV-1 virions produced from Jurkat-puro or Jurkat-VprCAT cell lines. CAT assays were performed with 20% sucrose cushion-purified virus isolated from Jurkat-VprCAT cells infected with HxBRU (Vpr⁺) or HxBRU-R- (Vpr⁻). (C) Effect of Vpr fusion protein expression on cell division. 293T cells were transfected with pCEPVpr, pCEPVpr-R-, pCEPVprIE, or pCEPVprCAT expression vectors and cultured in the presence of G-418 for 10 days. The number of G-418-resistant cell clones was then established. Error bars indicate standard deviations of three independent experiments.

The expression of Vpr was shown by many groups to arrest cells at the G₂ phase of the cell cycle (17, 32). To investigate whether the expression of VprIE and VprCAT influences cell physiology, we analyzed the growth of Jurkat-puro, Jurkat-VprIE, and Jurkat-VprCAT cells. All transduced Jurkat cell lines were shown to grow at rates similar to that of the parental Jurkat cell line (data not shown). Moreover, VprIE and VprCAT expression in human fibroblast 293T cells did not affect cell division. Indeed, 293T cells transfected with pCEPVprIE, pCEPVprCAT, or the control pCEPVpr-R- and cultured in the presence of G-418 produced more than 150 cellular clones 12 days posttransfection (Fig. 2C). In contrast, fewer than five clones were obtained when 293T cells transfected with pCEPVpr were cultured under the same conditions. Detection of VprIE and VprCAT by radioimmunoprecipitation with an anti-Vpr antibody confirmed that these proteins were expressed in 293T cells (data not shown). These results strongly indicate that the expression of the fusion proteins does not mediate cell cycle arrest and does not appear to be toxic for the cells.

Effect of Vpr fusion protein expression on HIV-1 replication.

We next evaluated the ability of HIV-1 to replicate in Jurkat-VprIE and Jurkat-VprCAT cell lines. The Jurkat-puro cell line was used as a control. The replication kinetics were analyzed by monitoring viral production in the culture supernatant (determined by RT activity measurements) every 3 days over time. Briefly, each Jurkat cell line was infected with increasing amounts (25,000, 100,000, and 300,000 cpm, as determined by RT activity) of HxBRU and HxBRU-R- viruses produced from transfected MT4 cells. Figure 3A shows the replication kinetics obtained when Jurkat-VprCAT and Jurkat-puro cells were infected with 100,000 cpm of HxBRU or HxBRU-R-. Maximum replication of HxBRU and HxBRU-R- was delayed by 2 to 3 days in Jurkat-VprCAT cells relative to control cells. Interestingly, in the same experiment, when Jurkat-VprIE cells were infected with 100,000 cpm of virus, a more pronounced delay of viral replication was observed (Fig. 3B). Peak viral replication was detected at day 15 in control cells infected with both HxBRU and HxBRU-R-. In contrast, maximum replication in Jurkat-VprIE cells was detected between days 21 and 24 for HxBRU and at approximately day 27 for HxBRU-R- (Fig. 3B). This result reveals that the replication of HxBRU and HxBRU-R- was delayed by 6 and 12 days, respectively, in Jurkat-VprIE cells relative to control cells. This result also indicates that HxBRU-R-, which does not express a native Vpr protein, replicates less efficiently than HxBRU in Jurkat-VprIE cells. The delay in the replication of HxBRU-R- in Jurkat-VprIE cultures also varied according to the input virus. Figure 3C and D show the replication kinetics obtained when Jurkat-VprIE and control cells were infected with 300,000 (Fig. 3C) or 25,000 (Fig. 3D) cpm of HxBRU or HxBRU-R-. When cells were infected with 300,000 cpm of virus, maximum replication of HxBRU and HxBRU-R- was delayed, respectively, by 6 and 9 days (Fig. 3C). However, when cells were infected with 25,000 cpm of virus, delays of approximately 6 and 18 days, respectively, were observed with HxBRU and HxBRU-R- (Fig. 3D). The appearance of virus-associated cytotoxic effects was also delayed in Jurkat cells expressing Vpr fusion proteins relative to control cells (data not shown). These results demonstrate that Jurkat-VprIE cells were less permissive for HIV-1 replication than Jurkat-puro control cells. They also exhibited a higher degree of resistance to cell killing mediated by HxBRU and HxBRU-R- than control cells.

FIG. 3.

HIV replication in transduced Jurkat cells. Jurkat-puro, Jurkat-VprIE, and Jurkat-VprCAT cell lines were challenged with HxBRU (Vpr⁺) and HxBRU-R- (Vpr⁻) (100,000 cpm of RT activity) to assess their susceptibility to viral infection. The infection kinetics were monitored by measuring viral production as RT activity in the supernatant over time for Jurkat-VprCAT cells (A), Jurkat-VprIE cells (B), Jurkat-VprIE cells upon infection with 300,000 cpm (C) or 25,000 cpm (D), and Jurkat-VprIE and Jurkat-RevM10 cells upon infection with 100,000 cpm (E). The experiments presented here were performed two or three times and were found to be highly reproducible.

The antiviral effect of VprIE was also compared to that obtained with RevM10, a trans-dominant mutant form of HIV-1 Rev that has been shown to delay HIV-1 replication in transduced cell lines, primary T cells, and CD34-enriched hematopoietic progenitor stem cells (2, 4, 27). Jurkat cells were transduced with pBabepuro-RevM10, and the resulting puromycin-resistant population, Jurkat-RevM10 cells, was infected with 100,000 cpm of HxBRU or HxBRU-R-. Figure 3E shows that maximum replication of HxBRU was delayed by 6 days in both Jurkat-VprIE and Jurkat-RevM10 cells relative to control Jurkat-puro cells. In contrast, peak HxBRU-R- replication was delayed by 6 and 12 days in Jurkat-RevM10 and Jurkat-VprIE cells, respectively, relative to control cells. This result indicates that in Jurkat cells, VprIE impaired the replication of a vpr⁺ virus (HxBRU) with an efficiency comparable to RevM10. Interestingly, VprIE inhibited the replication of a vpr⁻ virus (HxBRU-R-) more effectively than RevM10.

Impairment of HIV-1 replication by VprIE is dependent on an intact p6^gag domain.

Several studies have shown that Vpr is incorporated into HIV-1 particles through a specific interaction with the p6^gag domain (20, 23). Mutational analysis of the p6^gag domain demonstrated that the integrity of the LXXLF motif at the carboxy terminus of p6 was essential for Vpr incorporation into virions (19). To determine whether the delayed HIV-1 replication in transduced Jurkat cells was due to the specific incorporation of Vpr fusion proteins into virions, Jurkat-VprIE cells were challenged with p6^gag mutants that do not incorporate Vpr. Two mutations targeting the LXXLF motif of p6^gag, L44P and F45S, were introduced into the vpr⁻ HIV-1 infectious molecular clone HxB89-R-, generating HxB89LF-PS-R-, as described in Materials and Methods. Cotransfection of HxB89LF-PS-R- and pCEPVpr into COS-7 cells revealed that Vpr was highly expressed in cells but not incorporated into virions (data not shown). Stocks of HxB89LF-PS-R- and the isogenic p6 wild-type counterpart HxB89-R- were generated by transfection into MT4 cells, and 300,000 cpm was used to infect Jurkat-puro or Jurkat-VprIE cell lines. As shown in Fig. 4, peak viral production appeared at approximately day 17 for HxB89FS-PS-R- in both Jurkat-puro and Jurkat-VprIE cell lines, indicating that the replication of this virus, which had lost its ability to incorporate Vpr, was not affected by VprIE expression. In contrast, the replication of HxB89-R- was delayed by approximately 9 days in Jurkat-VprIE cells relative to Jurkat-puro cells (Fig. 4). These results indicate that the antiviral effect of VprIE is mediated by the incorporation of the fusion protein into HIV-1 virions.

FIG. 4.

Replication of p6^gag HIV mutants in Jurkat-VprIE cells. Jurkat-VprIE or control Jurkat-puro cells were infected with 300,000 cpm of p6^gag mutant HxB89LF-PS-R- or with 300,000 cpm of wild-type p6 isogenic virus HxB89-R-. Viral production was monitored over time by measurement of RT activity in the supernatant. The experiments presented here were performed three times and were found to be highly reproducible.

Incorporation of VprIE and VprCAT into HIV virions impairs viral infectivity but not viral production.

The results presented above indicate that VprIE and, to a lesser extent, VprCAT interfere with viral replication by affecting viral production and/or infectivity. To evaluate whether VprCAT or VprIE expression affected viral production, Jurkat-VprCAT and Jurkat-VprIE cell lines were transfected with HxBRU or HxBRU-R- infectious molecular clones. Measurements of viral production were obtained with a p24 enzyme-linked immunosorbent assay 24 h posttransfection to detect mainly viruses resulting from the first round of replication. The levels of p24 antigen were found to be similar in all samples (data not shown), indicating that there was no impairment at the levels of viral gene expression, assembly, and release of viral particles. This result suggests that the infectivity potential of the released viral particles that incorporated Vpr fusion molecules might be affected. To evaluate the infectious potential of HIV-1 virions containing VprCAT or VprIE fusion proteins, we performed multinuclear activation of galactosidase indicator cell (MAGI) assays (18). Viruses containing Vpr fusion proteins or control virus produced from infected transduced Jurkat cells (Fig. 3A and B) was harvested at the peak of viral production and titrated by RT activity measurements. In three experiments performed with three different virus stocks, MAGI cells were infected with 300,000 cpm of each virus. Figure 5 shows that HIV-1 released from Jurkat-VprCAT cells (HxBRU/VprCAT and HxBRU-R-/VprCAT) was two to three times less infectious than control vpr⁺ or vpr⁻ virus produced from Jurkat-puro cells. However, HIV-1 released from Jurkat-VprIE cells (HxBRU/VprIE and HxBRU-R-/VprIE) was 10 to 30 times less infectious than control virus produced from Jurkat-puro cells (Fig. 5). Furthermore, HxBRU-R-/VprCAT or HxBRU-R-/VprIE viruses exhibited a larger impairment of viral infectivity than HxBRU/VprCAT or HxBRU/VprIE viruses, again suggesting that competition between HIV-1-encoded Vpr and Vpr fusion proteins for virion incorporation occurred. Although less quantitative, the negative effect of Vpr fusion proteins on viral infectivity was also observed in transient experiments. Indeed, virus produced by transient cotransfection of pCEPVprCAT or pCEPVprIE and HxBRU-R- in COS-7 cells exhibited two- and fourfold inhibition of infectivity in a MAGI cell assay, respectively (data not shown). Overall, these results strongly suggest that the impairment of viral infectivity is the consequence of Vpr fusion protein incorporation into HIV-1 virions.

FIG. 5.

Evaluation of viral infectivity. The infectious potential of HIV virions containing Vpr fusion proteins was analyzed by a MAGI cell assay. Briefly, HeLa-CD4-LTR/β-Gal indicator cells were infected with equivalent amounts of virus produced from Jurkat cells or Jurkat-VprCAT and Jurkat-VprIE cells infected with HxBRU (Vpr⁺) and HxBRU-R- (Vpr⁻). At 24 h postinfection, the number of blue colonies was evaluated. The experiment was performed in triplicate with three different virus stocks, and the standard deviations are indicated.

DISCUSSION

Recent progress in our understanding of the structural and functional domains of Vpr makes the design of very efficient Vpr-based antiviral genes possible. To address the question of whether Vpr could be used as a shuttle to target foreign amino acid sequences of different sizes into HIV-1 virions, an XbaI restriction site was inserted at the 3′ end of vpr in the infectious molecular clone HxBRU (22). The resulting construct, which replicates as efficiently as wild-type HxBRU (unpublished data), had the advantage of providing a simple means to screen several Vpr fusion proteins for their ability to be packaged into HIV-1 virions. Moreover, HxBRUXbaI provided a quantitative and physiological assay for Vpr fusion protein incorporation into HIV particles. Indeed, our results revealed that the first 88 amino acids of Vpr were sufficient to incorporate the 18-amino-acid Vpu C-terminal immunodominant epitope and the 218-amino-acid CAT enzyme into HIV-1 virions. Importantly, the packaging efficiency of the VprCAT and VprIE fusion proteins was comparable to that obtained with wild-type Vpr (Fig. 1).

Previous studies reported that Vpr fusion proteins could severely impair HIV replication in CD4⁺ T cells when expressed in cis from the provirus (28, 33). All of our attempts to express Vpr fusion proteins in the context of the HxBRUXbaI proviral construct revealed that this biological system was not reliable for the evaluation of the viral replication potential. Indeed, we found that the introduction of foreign DNA sequences between the Tat splice acceptor site and the Tat initiation codon had a cis-acting negative effect on HIV protein expression (data not shown). On the other hand, transient expression experiments in which a vpr⁺ or a vpr⁻ provirus was cotransfected with a Vpr fusion protein expressor did not allow for the sensitive measurement of antiviral activity, since the different transfection efficiencies of the two plasmids gave highly variable results. For quantitative measurement of antiviral activity, T-cell lines that express Vpr fusion proteins constitute a more reliable system. In addition, the stable expression of therapeutic genes in cell lines is essential to evaluate their possible effect on cell homeostasis as well as their effectiveness in a gene therapy approach.

Consequently, we generated CD4⁺ Jurkat cell populations constitutively expressing Vpr fusion proteins by using retroviral vectors. This system provided a reliable method to evaluate the effect of Vpr fusion proteins on viral replication and spread in vitro, since the ideal situation, in which 100% of HIV-1 target cells carried the therapeutic gene, was achieved. We first investigated the expression of VprIE and VprCAT in selected transduced Jurkat cell populations. The expression of vprIE- and vprCAT-specific mRNAs was demonstrated by Northern blot analysis (Fig. 2B). Moreover, CAT enzymatic activity associated with the VprCAT fusion protein was efficiently incorporated in trans into HIV-1 virions (Fig. 2C). However, we failed to detect the expression of the transgenes by immunoprecipitation and Western blot analysis. It has been shown that proteins are poorly expressed from the Moloney murine leukemia virus LTR in human-derived CD4⁺ T-cell lines (29).

The effect of VprCAT and VprIE fusion proteins on cell division was quantitatively evaluated with 293T cells. The results indicated that cells expressing VprCAT or VprIE fusion proteins proliferated and formed colonies with the same efficiency as the pCEPVpr-R- positive control, which did not express Vpr (Fig. 2C). The first 88 amino acids of Vpr therefore appeared sufficient to achieve Vpr fusion protein virion incorporation without any negative effect on cellular growth. This finding allowed us to isolate Jurkat-puromycin-resistant clones that expressed Vpr fusion proteins and that were later pooled to obtain a stable cell population representative of the parental Jurkat cell line. Infection of transduced Jurkat-VprCAT or Jurkat-VprIE cell lines resulted in partial inhibition of HIV-1 replication compared to that for control cells (Fig. 3A and B). To confirm that the delays in viral replication observed when we challenged the different Jurkat cell lines were not an artifact arising from cell line selection, we analyzed CD4 expression at the cell surface. Fluorescence-activated cell sorter analysis demonstrated that CD4 expression at the plasma membrane of both Jurkat-VprCAT and Jurkat-VprIE cells was similar to that of parental Jurkat cells (data not shown). Moreover, a p6^gag mutant of HIV-1 that did not incorporate Vpr replicated as efficiently in Jurkat-VprIE cells as in Jurkat-puro control cells (Fig. 4), indicating that the impairment of viral replication in Jurkat-VprIE cells was the consequence of VprIE fusion protein incorporation into virions. Overall, these results argue strongly against a clonal effect resulting from cell population selection as a major contributor to the impairment of viral replication.

Virion infectivity has been shown to correlate with proper viral assembly and maturation (12, 34). The presence of a nonrelevant amino acid sequence fused to Vpr during assembly and budding may affect viral morphogenesis, resulting in the production of virus with a drastically reduced infectious potential. Indeed, as shown in Fig. 3B, the introduction of 18 amino acids derived from HIV-1 Vpu at the C-terminal end of Vpr affected considerably the replication of HIV-1 in tissue cultures. This antiviral effect of VprIE was shown to be as efficient as that exerted by a transdominant mutant of Rev, the RevM10 protein, which inhibits HIV-1 replication through a distinct mechanism (Fig. 3E). In contrast, the 35-kDa VprCAT fusion protein delayed HIV-1 replication less efficiently (Fig. 3A). Analysis of virus produced from transduced Jurkat cells revealed that virus released from Jurkat-VprCAT cells was two to three times less infectious than the control virus, whereas virus produced from Jurkat-VprIE cells exhibited a 10- to 30-fold decrease in viral infectivity (Fig. 5). We currently assume that the highly hydrophilic Vpu sequence fused to Vpr may provoke an important steric hindrance, which in turn affects normal viral morphogenesis. Indeed, electron microscopy revealed a strong morphological alteration of virions containing VprIE fusion proteins (unpublished data). However, at this point we cannot rule out the possibility that early events following the entry of wild-type virus into VprIE-expressing cells may be affected.

From a therapeutical point of view, an efficient Vpr fusion protein should impair the replication of a large range of HIV isolates, which could be vpr⁺ or vpr⁻. Figure 2B shows that the vpr⁺ HxBRU virus incorporated less VprCAT fusion protein than the vpr⁻ HxBRU-R- virus. Moreover, HxBRU-R- was less infectious than HxBRU when produced in Jurkat-VprIE or Jurkat-VprCAT cell lines (Fig. 5). These results suggest competition for virion incorporation between wild-type Vpr expressed from HxBRU and Vpr fusion proteins provided in trans by the transduced cell lines. Consequently, it was not surprising to observe a greater delay in viral replication in Jurkat-VprIE cells infected with HxBRU-R- than in Jurkat-VprIE cells infected with HxBRU (Fig. 3B, C, and D). No such difference was observed in Jurkat-VprCAT cells. This finding likely reflects the weak antiviral effect of VprCAT on HIV replication (Fig. 3A). The competition for virion incorporation between Vpr fusion proteins and wild-type Vpr expressed by the virus remains a critical point. Indeed, an increased delay in viral replication was observed with decreasing amounts of HxBRU-R- virus in Jurkat-VprIE cells, an effect that was not observed with the HxBRU virus (compare Fig. 3B, C, and D). Therefore, the development of Vpr fusion proteins that could target both HIV-1 production and infectivity and that could efficiently compete with wild-type Vpr for virion incorporation remains an essential and promising goal.

This report presents, for the first time, proof of principal experiments showing that virion-targeted viral inactivation with Vpr fusion proteins can delay HIV-1 replication in transduced CD4⁺ T cells. Therefore, the development of more efficient Vpr-based fusion proteins represents a promising approach that may contribute to efficient strategies against HIV infection.