Comparison of Cell Cycle Arrest, Transactivation, and Apoptosis Induced by the Simian Immunodeficiency Virus SIVagm and Human Immunodeficiency Virus Type 1 vpr Genes

Abstract

All primate lentiviruses known to date contain one or two open reading frames with homology to the human immunodeficiency virus type 1 (HIV-1) vpr gene. HIV-1 vpr encodes a 96-amino-acid protein with multiple functions in the viral life cycle. These functions include modulation of the viral replication kinetics, transactivation of the long terminal repeat, participation in the nuclear import of preintegration complexes, induction of G₂ arrest, and induction of apoptosis. The simian immunodeficiency virus (SIV) that infects African green monkeys (SIVagm) contains a vpr homologue, which encodes a 118-amino-acid protein. SIVagm vpr is structurally and functionally related to HIV-1 vpr. The present study focuses on how three specific functions (transactivation, induction of G₂ arrest, and induction of apoptosis) are related to one another at a functional level, for HIV-1 and SIVagm vpr. While our study supports previous reports demonstrating a causal relationship between induction of G₂ arrest and transactivation for HIV-1 vpr, we demonstrate that the same is not true for SIVagm vpr. Transactivation by SIVagm vpr is independent of cell cycle perturbation. In addition, we show that induction of G₂ arrest is necessary for the induction of apoptosis by HIV-1 vpr but that the induction of apoptosis by SIVagm vpr is cell cycle independent. Finally, while SIVagm vpr retains its transactivation function in human cells, it is unable to induce G₂ arrest or apoptosis in such cells, suggesting that the cytopathic effects of SIVagm vpr are species specific. Taken together, our results suggest that while the multiple functions of vpr are conserved between HIV-1 and SIVagm, the mechanisms leading to the execution of such functions are divergent.

The vpr gene from human immunodeficiency virus type 1 (HIV-1) encodes a 96-amino-acid protein with multiple functions in the viral life cycle. The first reported role for vpr was a moderate transactivation effect on the viral promoter, the long terminal repeat (LTR) (13). HIV-1 mutants with deletions in vpr replicate with slower kinetics than wild-type viruses do (5, 14). Vpr is encapsidated into virions in significant amounts (1, 12, 57). The presence of vpr in the viral particle facilitates efficient infection of macrophages and other nondividing cells (14, 17, 23) by mediating active nuclear import of preintegration complexes (16, 44). In addition, the presence of vpr enhances the transcriptional activity of the viral LTR in macrophages and T cells, allowing the production of a larger viral progeny (18, 52).

HIV-1 vpr contributes to the multiple cytopathic effects induced by HIV-1 by inducing cell cycle arrest in G₂ (22, 27, 45, 46) and apoptosis (11, 49, 50, 55). The primate immunodeficiency virus strain, SIVagm, encodes an accessory gene, vpr^AGM, which bears sequence (31% amino acid identity) and functional conservation with vpr^HIV-1. At a functional level, vpr^AGM shares with vpr^HIV-1 its abilities to transactivate the viral LTR (39), cause cell cycle arrest in G₂ (42, 51), and become incorporated in the virion particles for subsequent participation in nuclear import of preintegration complexes (1, 8).

Earlier studies have suggested that the HIV-1 and SIVagm vpr genes are functionally conserved in virion encapsidation, nuclear localization, cell cycle arrest, and effect on viral replication kinetics. However, noticeable differences have also been reported. Deletion of SIVagm vpr appears to have a more profound effect on viral replication in primary macrophages and peripheral blood mononuclear cells than does deletion of HIV-1 vpr. SIVagm vpr mutants are incapable of replication in nondividing cells, while HIV-1 vpr mutants are able to replicate at low level. To explain this difference, it was suggested by Campbell and Hirsch that HIV-1 contains redundant nuclear localization signals in addition to Vpr, such as the matrix protein, and perhaps that is not the case for SIVagm (8).

The cause-effect relationships among the different functions of vpr are currently being investigated. The present study focuses on how transactivation, G₂ arrest, and apoptosis are related at a functional level, for both vpr^AGM and vpr^HIV-1.

MATERIALS AND METHODS

Cell lines.

The human breast cancer cell line SKBR3 was cultured in RPMI 1640 medium (BioWhittaker, Walkersville, Md.) containing 10% fetal calf serum (FCS) (Omega Scientific, Inc., Bedford, Ohio). African green monkey kidney Cos-7 cells were grown in Dulbecco's modified Eagle's medium (DMEM; BioWhittaker) supplemented with 10% FCS. The human embryonic 293T cells were propagated in Iscove's modified Dulbecco's medium (BioWhittaker) plus 10% FCS. The human cervical cancer cell line HeLa was maintained in DMEM with 5% FCS.

Plasmid constructs.

Mammalian vectors directing the expression of thy-1 and a vpr gene, vpr^HIV-1, vpr^HIV-1-FS, or vpr^AGM, were described previously (42). A vector expressing thy-1 and vpr^AGM-FS was constructed by digesting the vpr^AGM expression vector with the restriction endonuclease BsrGI, treating it with Klenow to fill the ends, and religating. This treatment generated a frameshift at codon 85. The vector, pCMV-thy was described earlier (41).

(i) Vectors for luciferase expression.

LTR^HIV-1-Luc was constructed by subcloning the 5′ LTR of HIV-1_NL4-3 (2) into pBluescript II KS(+) (Invitrogen, Carlsbad, Calif.) using AvrII and HindIII. A fragment containing the luciferase gene followed by the simian virus 40 polyadenylation sequences was cloned downstream from the LTR using XhoI and XbaI. LTR^AGM-Luc was constructed by replacing the HIV-1 LTR in the previous construct by that of pSIV_agm677 (26), using HindIII and XbaI.

(ii) Plasmids for production of defective lentivirus vectors.

The envelope construct HCMV-VSVG (7) and the packaging construct pCMVΔR8.2-Δvpr (4) were described previously. D102 is a defective HIV-1-derived lentivirus vector that is described in detail elsewhere (Y. Zhu and V. Planelles, submitted for publication). Briefly, D102-mHSA was derived from the previously described vector HIV-GFP (49) as follows. A 484-bp PflMI-SalI fragment from HIV-GFP was replaced by a 630-bp fragment from pNL-r-HSAS (25) using the same restriction enzymes. This resulted in replacement of the vpr open reading frame with that of the murine heat-stable antigen (mHSA) and also in the inclusion of a unique restriction site, XbaI, for subsequent cloning of additional genes replacing the gene encoding mHSA. D102-vprH and D102-vprA were constructed by replacing mHSA with vpr^HIV-1 or vpr^AGM, respectively, using the newly introduced XbaI site and a natural site, EcoRI, within the vpr remnant.

(iii) Construction of green fluorescent protein (GFP) fusions.

pEGFP-N1 (Clontech, Palo Alto, Calif.) was digested with BglII and BamHI and religated to create pEGFP-bam/bgl-. A ligation was then performed to create pGFP-flex with the following three DNA fragments: an NheI-BsrGI fragment from pEGFP-bam/bgl-, a BamHI-NheI fragment from pcDNA-2, and a flexible linker, which was from pUC-flex (10), digested with BsrGI and BamHI. To construct a GFP-vpr^HIV-1 expression plasmid, an NcoI-XbaI fragment from BSVpr-Thy (42) was subcloned into pGFP-flex digested with XhoI and XbaI. In the previous subcloning, the BsrGI and XhoI sites were filled with Klenow. A GFP-vpr^AGM expression plasmid was made by subcloning the SIVagm vpr cDNA from the vector, BS-677X-Thy (42), using the same restriction sites and recipient plasmid as for its HIV-1 counterpart.

Transfections and luciferase assays.

Transient transfection of cells for luciferase measurement was performed using the TransFast reagent (Promega Corp., Madison, Wis.) as recommended by the manufacturer. Cells were plated at a density of 10⁵/well in 12-well plates. One day later, 0.25 μg of LTR^HIV-1-Luc or LTR^AGM-Luc reporter plasmid and 0.25 μg of a vpr expression plasmid were mixed with 1.5 μl of TransFast reagent in 0.2 ml of serum-free RPMI 1640 (SKBR3 cells) or DMEM (Cos-7 cells), and the mixture was incubated at room temperature for 30 min. The mixture was then added to the cells and incubated at 37°C for 1 h, and then 2 ml of RPMI 1640 with 10% FCS was added. The control plasmid was pCMV-thy. After 72 h, the cells were lysed and assayed for luciferase activity with a commercially available luciferase assay kit (Promega Corp.), using a LumiCount microplate reader (Packard Instrument Corp., Meriden, Conn.). The luciferase assay was performed using the following settings: PMT, 1,100 V; gain level, 5.0; and read length, 0.5 s. The background of the luciferase reading was 25 ± 15 light units for six measurements of lysates from uninfected cells. Each experiment was performed in triplicate, and each measurement was the average of duplicate readings. Luciferase light units were normalized to 1 μg of protein content (Bio-Rad Laboratories, Hercules, Calif.) in cell lysates. Light units in experimental wells were converted to fold increase with respect to control wells.

Viral vector production and titer determination.

Lentivirus vectors were produced by transient transfection of 293-T cells. D102-mHSA, D102-vprH, or D102-vprA was cotransfected with HCMV-VSVG and pCMV-ΔR8.2Δvpr using the calcium phosphate-mediated transfection. Virus was collected at 48, 72, 96 h posttransfection. The harvested supernatants were precleared by low-speed centrifugation at 2,000 rpm in a Sorvall RT7 with an RTH-750 rotor (Sorvall) and pelleted by ultracentrifugation at 25,000 rpm in a Discovery 100S centrifuge with a Surespin 630 rotor (Sorvall, Newton, Conn.). Virus pellets were resuspended in fresh tissue culture medium and frozen at −80°C. Vector titers were measured by infection of HeLa cells as described below, followed by flow cytometric analysis of cells positive for the reporter molecule, GFP. Vector titers were calculated as follows: titer = (F × C₀ / V) × D, where F is the frequency of GFP-positive cells by flow cytometry, C₀ is the total number of target cells at the time of infection; V is the volume of inoculum, and D is the virus dilution factor. The virus dilution factor (D) used for titer determinations was 10. The total number of target cells at the time of infection was 10⁶.

Flow cytometry.

Cells were harvested and analyzed by direct immunofluorescence for GFP expression and with propidium iodide to analyze the DNA content. The cells were detached with 2 mM EDTA, washed in phosphate-buffered saline (PBS), fixed with 0.2% paraformaldehyde in PBS for 1 h, and stained with propidium iodide solution (20 μg of propidium iodide per ml, 0.2% Triton X-100, and 11.25 Kunitz units of RNase A per ml in PBS). The percentage of GFP-positive cells in G₂/M was assessed and compared to that of untreated control cells. Flow cytometric analysis was performed in an Epics Elite ESP analyzer (Coulter Corp., Hialeah, Fla.). Gates for detection of GFP were established using mock-infected cells as background. Because electronic settings varied from experiment to experiment, gates were defined such that the percentage of false-positive events was not higher than 0.3% in the mock-infected population. Cell cycle analysis was performed using Multicycle AV software (Phoenix Flow Systems, San Diego, Calif.).

TdT-mediated dUTP-biotin nick end labeling (TUNEL).

To detect in situ apoptosis, the TdT-FragEL kit (Oncogene Research Products, Cambridge, Mass.) was used as specified by the manufacturer. Cells were seeded at 2 × 10⁴ cells/well in chamber slides (Nalge Nunc International, Rochester, N.Y.). At 48 h following infection, cells were inspected for GFP expression to measure the level of infection. At 72 h postinfection, medium was removed and the cells were fixed with 4% formaldehyde (in PBS) for 10 min, washed with 80% ethanol, and stored overnight at −20°C. After being washed with Tris-buffered saline (TBS), the cells were permeabilized with proteinase K at room temperature for 5 min and washed again with TBS. The cells were then treated with 30% H₂O₂ in methanol for 5 min to inactivate endogenous peroxidases. Following application of equilibration buffer, the cells were incubated with a reaction mixture containing terminal deoxynucleotidyltransferase (TdT)- and biotin-labeled dUTP for 1.5 h in a humidified chamber at 37°C. The reaction was stopped by the supplied stop buffer. After the samples were washed with TBS, streptavidin-horseradish peroxidase conjugate was applied in a humidified chamber for 30 min. The cells were then counterstained with methyl green.

Drugs.

Caffeine was purchased from Sigma Chemical Co., and dissolved in water at 50 mM, and kept frozen at −80°C. Taxol (paclitaxel; Sigma Chemical Co.) was dissolved in dimethyl sulfoxide (Fisher Scientific, Fair Lawn, N.J.) at 7 mM and kept frozen at −80°C.

RESULTS

Induction of G₂ arrest by vpr^AGM and vpr^HIV-1 differs in human and African green monkey cells.

We previously described a transient-transfection assay for the study of cell cycle effects by vpr^HIV-1 and related genes from primate lentiviruses (42). This method is based on the use of dual expression vectors which direct the expression of vpr and a reporter gene, murine thy-1 (Fig. 1A). Transfected cells are analyzed by direct immunofluorescence for Thy-1 surface expression and with propidium iodide for DNA content (see Materials and Methods) using flow cytometry. For simplicity, only the cell cycle profiles of Thy-1-positive cells (transfected) cells are shown (Fig. 2).

FIG. 1.

Expression vectors used for analysis of cell cycle distribution and transactivation. (A) Dual expression vectors for vpr variants and the reporter gene, thy-1. Arrows denote promoters. SV40 p(A), simian virus 40 polyadenylation signal; CMV IE, human cytomegalovirus immediate-early promoter; vpr^HIV-1-FS and vpr^AGM-FS, carboxy-terminal truncations of vpr^HIV-1 and vpr^AGM, generated by introducing frameshift mutations at codons 64 and 85, respectively. (B) Luciferase reporter vectors in which the viral LTR serve as a basal promoters.

FIG. 2.

Flow cytometric analysis of vpr-induced G₂ arrest. Cells were transfected with control plasmid (pCMV-thy) or dual expression vectors encoding the murine thy-1 gene and the indicated vpr version (top of each histogram). At 48 h after transfection, the cells were analyzed for thy-1 expression and DNA content. Where indicated, the cells were treated by adding 2 mM caffeine to the medium 1 h posttransfection. Transfected cells were distinguished from untransfected ones by Thy-1 expression and were electronically gated. Histograms depict the cell cycle profile of gated Thy-1-positive cells only. The left peaks constitute cells in G₁, and the right peaks constitute cells in G₂/M; cells in S phase are between the G₁ and G₂ peaks. The frequencies of cells in different stages of the cell cycle were calculated using Multicycle AV software (Phoenix Flow Systems, San Diego, Calif.).

We used parallel constructs encoding vpr mutants, vpr^HIV-1-FS and vpr^AGM-FS (Fig. 1A) containing frameshifts at codons 64 (40, 42) and 85, respectively. The mutant constructs, vpr^HIV-1-FS and vpr^AGM-FS, are unable to cause G₂ arrest (data not shown).

Transfection of vpr^HIV-1 produced detectable cell cycle arrest in G₂ both in human (SKBR3) and African green monkey (Cos-7) cells (Fig. 2). In contrast, vpr^AGM induced cell cycle arrest in Cos-7 cells but did not induce any detectable change in the cell cycle profile of SKBR3 cells (Fig. 2). We have obtained similar results when expressing vpr^AGM and vpr^HIV-1 in a variety of human cell lines, including osteosarcoma, leukemia/lymphoma, fibroblast-like, epidermal, colon carcinoma, rhabdomyosarcoma, and retinoblastoma cells (data not shown).

Relationship between transactivation and G₂ arrest.

The ability of vpr^HIV-1 to induce G₂ arrest has been linked to another property of vpr, transactivation of the viral promoter, the LTR. It was shown that the transcriptional activity of the LTR was increased by 5- to 10-fold (18, 20) during the G₂ phase compared to that during the G₁ phase. Consequently, infected cells that express vpr^HIV-1 are capable of producing 5- to 10-fold-larger amounts of viral progeny than are cells infected with vpr-deleted viruses (18, 20). A direct result of this enhanced transcriptional activity is that HIV-1 strains expressing functional vpr alleles are capable of spreading in culture with faster kinetics than are isogenic strains with mutations in vpr (13, 27, 37).

We wished to investigate the potential relationship between G₂ arrest and transactivation in the context of vpr^AGM. Because vpr^AGM is unable to induce cell cycle arrest in human cells, we hypothesized that vpr^AGM would also be unable to induce transcriptional transactivation of the SIVagm LTR. The transactivation abilities of HIV-1 and SIVagm vpr genes were tested using luciferase reporter vectors (Fig. 1B). Luciferase reporter vectors were cotransfected with a mammalian expression vector expressing vpr or a truncated version of vpr (vpr^HIV-1-FS or vpr^AGM-FS) or with a control plasmid, pCMV-thy (Fig. 1A).

We first tested whether transactivation of the viral LTR by vpr^AGM would behave with species specificity as did induction of G₂ arrest. In African green monkey cells, cotransfection of vpr^AGM with LTR^AGM-Luc produced a significant level of transactivation, as did vpr^HIV-1 when cotransfected with LTR^HIV-1-Luc (Fig. 3A). We then performed a parallel experiment with human cells. As expected, vpr^HIV-1 increased the level of expression of LTR-Luc^HIV-1 in human cells (Fig. 3B). Surprisingly, despite its inability to cause cell cycle arrest in human cells, vpr^AGM produced a significant level of transactivation upon the LTR^AGM-Luc construct. Thus, it appears that vpr^AGM is able to induce transactivation of the SIVagm LTR independently of the induction of G₂ arrest.

FIG. 3.

Transactivation of vpr^HIV-1 and vpr^AGM on the viral LTR. Cos-7 (A) and SKBR3 (B) cells were transiently transfected with a luciferase reporter construct (LTR^AGM-Luc or LTR^HIV-1-Luc) and a vpr expression construct. The structures of all constructs, except the control plasmid, are shown in Fig. 1. The control plasmid was pCMV-thy. At 48 h posttransfection, the cells were lysed in 100 μl of lysis buffer and a 20-μl volume was used to measure luciferase activity. Experiments were normalized to light units measured with control plasmid, to which an arbitrary value of 1 was assigned. All experiments were performed in triplicate, and standard deviations are shown as error bars.

Effect of caffeine on transactivation by vpr^HIV-1 and vpr^AGM.

Caffeine, a methylxanthine, inhibits DNA damage-induced (36) and vpr^HIV-1-induced (43) cell cycle arrest. To further examine whether transactivation by vpr^AGM is independent of cell cycle blockade, we performed an experiment in which caffeine was used to relieve G₂ arrest and the cells were analyzed for transactivation (Fig. 4). Incubation with 2 mM caffeine inhibited vpr^HIV-1-induced G₂ arrest as well as vpr^AGM-induced G₂ arrest (Fig. 2, bottom panels). Treatment with caffeine led to nearly complete (94%) inhibition of vpr^HIV-1 transactivation effect (Fig. 4A). In contrast, treatment with caffeine had no significant effect on the transactivation by vpr^AGM in human cells (Fig. 4B) or in African green monkey kidney cells (Fig. 4C). Thus, it appears that vpr^AGM can induce transactivation of the viral LTR by a mechanism which is independent of the induction of cell cycle arrest, and it represents an important mechanistic difference from the proposed mode of action of vpr^HIV-1 (18).

FIG. 4.

Effect of caffeine on transactivation. SKBR3 (A and B) or Cos-7 (C) cells were transiently transfected with vpr^HIV-1 (A) or vpr^AGM (B and C) vectors and 2 mM caffeine was added to the medium 1 h later; at 72 h posttransfection, the cells were lysed and analyzed for luciferase activity as described in the legend to Fig. 3.

Effect of taxol on transactivation of the HIV-1 and SIVagm LTRs.

It has been suggested that the transactivation effect induced by vpr^HIV-1 is a result of a higher transcriptional level of the viral LTR in the G₂ phase of the cell cycle (18). According to this hypothesis, one would expect that drug-induced G₂ arrest would produce similar transactivation of the LTR. To test this idea, cells were treated with a chemotherapeutic agent, taxol, which specifically induces cell cycle arrest at the G₂/M transition by inhibiting the formation of the mitotic spindle (19, 24, 31). Treatment of SKBR3 cells with taxol induced G₂ arrest (Fig. 5A) and transactivation of the HIV-1 LTR (Fig. 5B) in human cells. Similarly, taxol induced G₂ arrest in Cos-7 cells (data not shown) and increased activity of the SIVagm LTR (Fig. 5B).

FIG. 5.

Induction of G₂/M arrest by taxol leads to LTR transactivation. (A) SKBR3 cells were incubated in medium containing 90 nM taxol [(+) taxol] or medium alone [(−) taxol] and were assayed for cell cycle profile after 48 h. (B) Cells were transfected with the indicated reporter constructs; at 1 h posttransfection they were incubated in medium containing taxol or medium alone, and at 72 h they were lysed and analyzed for luciferase activity. (C) SKBR3 cells were transfected with the indicated plasmids and treated with 90 nM taxol or medium alone; at 72 h posttransfection they were lysed and analyzed for luciferase activity.

Thus, it appears that the SIVagm LTR transcriptional activity can be stimulated by two different mechanisms or pathways, one that is dependent on cell cycle manipulation and one that is independent of it. If this is true, one would expect that treatment with taxol and expression of vpr^AGM may have additive effects on LTR transactivation. To investigate this possibility, we co-transfected LTR^AGM-Luc with vpr^AGM into human cells and treated the cells with taxol (Fig. 5C). If the SIVagm LTR is responsive to cell cycle manipulation, the extent of transactivation by vpr^AGM should be enhanced by taxol. Addition of taxol potentiated the transcriptional activity of the LTR in human cells (Fig. 5C). This indicates that the SIVagm LTR can be transactivated in part by cell cycle manipulation and in part by a cell cycle-independent pathway.

Taken together, the above results suggest that the relationship between transactivation and induction of cell cycle arrest differs in SIVagm and HIV-1 vpr genes. vpr^HIV-1 produces increased activity of the viral promoter through induction of G₂ arrest. Abrogation of G₂ arrest dramatically inhibits such transactivation by vpr^HIV-1. On the other hand, vpr^AGM induces transactivation of the viral LTR by cell cycle-dependent and -independent mechanisms.

Induction of apoptosis by vpr and its relationship to cell cycle perturbation.

Cells that are induced to arrest in G₂ by either vpr^HIV-1 expression or infection with HIV-1 appear to be unable to enter the cell cycle again and, instead, undergo apoptosis (6, 49, 50). Because of the structural and functional similarities between vpr^HIV-1 and vpr^AGM, we wished to investigate the role of vpr^AGM in induction of apoptosis. In particular, we designed experiments to determine (i) whether expression of vpr^AGM leads to apoptosis; (ii) whether apoptosis by vpr^AGM is subject to species specificity, as is G₂ arrest induction; and (iii) whether induction of apoptosis, for both vpr^HIV-1 and vpr^AGM, is linked to the induction of G₂ arrest.

For our apoptosis studies, we used a lentivirus transduction method which we described previously (49). This delivery system has important advantages with respect to infection with full-length HIV or transfection procedures (34, 35). First, lentivirus vectors are able to deliver a small subset of viral genes. Second, lentivirus vectors using vesicular stomatitis virus glycoprotein G (VSVG) can be produced at high titers, such that high levels of infection can be achieved. Finally, we have developed lentivirus vectors (49) that express a marker, GFP, for analysis of transduction at the single-cell level (Fig. 6).

FIG. 6.

Viral vectors used for transduction of vpr^AGM, vpr^HIV-1, and mHSA. (A) Lentivirus transfer vectors were constructed by incorporating the indicated genes into the defective vector, D102; the vpr open reading frame was eliminated by mutating the translational start codon, which overlaps with vif; an XbaI restriction endonuclease site was introduced for convenient cloning of desired genes into this region of the genome; an EcoRI site was a present and was used as the downstream cloning site; discontinuous lines depict splicing donors (SD) and acceptors (SA). (B) The lentivirus packaging construct, pCMVΔR8.2-Δvpr, was described previously (4).

A defective HIV-1-derived vector, named D102 (Fig. 6A), was used for construction of three lentivirus vectors expressing vpr^HIV-1, vpr^AGM, or a control gene, encoding mHSA. All the splicing signals were left intact in D102. Only the splicing donors and acceptors involved in the generation of vpr and nef (GFP) mRNAs are shown in Fig. 6.

Infectious vectors were produced by cotransfection of one of the three transfer vectors derived from D102, plus the packaging construct, pCMVΔR8.2-Δvpr (Fig. 6), plus a VSV glycoprotein G expression vector, HCMV-VSVG (3, 7).

Target cells were transduced with lentivirus vectors, and the cells were analyzed at 72 h posttransduction for levels of infection and apoptosis. The levels of infection were assessed by measuring the levels of GFP expression and were between 80 and 95%. The levels of apoptosis were measured by TUNEL (Fig. 7). This method is based on the ability of terminal transferase to extend 3′-hydroxyl (3′-OH) ends of chromosomal breaks. Apoptotic cells, containing nicked DNA, incorporate significant amounts of labeled nucleotides, and these cells stain dark brown. The DNA from nonapoptotic cells is mostly nonfragmented and lacks free 3′-OH ends, and therefore these cells do not stain positive in the TUNEL method.

FIG. 7.

Induction of apoptosis by and subcellular localization of vpr^HIV-1 and vpr^AGM. (A) Apoptosis induction in Cos-7 cells. The cells were either mock infected or infected with one of the indicated lentivirus vectors and assayed for apoptosis at 72 h postinfection, using the TUNEL method; levels of infection were quantitated by measuring the frequency of GFP-positive cells and were between 80 and 95% for all experiments. (B) Apoptosis induction in HeLa cells. Experiments were performed in parallel with those in panel A. (C) Subcellular localization of GFP-vpr fusion proteins. Expression vectors bearing indicated fusion proteins were transiently transfected into HeLa or Cos-7 cells and visualized using fluorescence microscopy after 24 h.

The vector D102-vprH, which drives the expression of vpr^HIV-1 (Fig. 6), induced significant levels of apoptosis in both Cos-7 (Fig. 7A) and HeLa (Fig. 7B) cells. The vector D102-vprA, which directs the expression of vpr^AGM, induced significant levels of apoptosis in Cos-7 cells (Fig. 7A) but did not induce detectable apoptosis in HeLa cells (Fig. 7B). As a negative control, we used another lentivirus vector, D102-mHSA, that was engineered to express mHSA (25, 29). Transduction with D102-mHSA did not induce detectable apoptosis in either cell type (Fig. 7). The levels of apotosis were quantitated as percentages of cells staining positive in the TUNEL method (Fig. 8). Levels of apoptosis were confirmed by nuclear staining with 4′,6-diamidino-2-phenylindole (DAPI) (49) in parallel samples (data not shown). The ability of vpr^HIV-1 and vpr^AGM to induce apoptosis was also tested in the human CD4-positive lymphocyte cell line Sup-T1 (Fig. 8C), and the results further supported the notion that induction of apoptosis by vpr^AGM is species specific.

FIG. 8.

Effect of caffeine on the induction of apoptosis by vpr^HIV-1 and vpr^AGM. Infections and apoptosis measurements were performed as described in the legend to Fig. 7; TUNEL-positive (apoptotic) cells were visually counted, and the levels of apoptosis were expressed as a percentage of the total cells counted; the percentage of apoptotic cells in mock infections was always lower than 5%. All experiments were done in triplicate, and standard deviations are shown. (A) SKBR3 cells. (B) Cos-7 cells. (C) Sup-T1 cells.

It is not clear whether vpr^HIV-1 and vpr^AGM are able to induce apoptosis directly. In certain instances, apoptosis is thought to be a secondary effect of an abnormal delay in the G₂ phase. In addition, recent observations in various laboratories suggested that the sensitivity of cycling cells to apoptosis is increased during the G₂/M phase and G₂/M arrest (30, 48, 56). If vpr induces apoptosis directly, we predict that inhibition of G₂ arrest will have no effect on the level of apoptosis. Conversely, if the induction of apoptosis by vpr is dependent on the induction of G₂ arrest, we would expect that alleviation of G₂ arrest (i.e., by incubation with caffeine) would diminish the level of apoptosis. In a study on the effects of caffeine and radiation (36), radiation-induced G₂ arrest and apoptosis decreased when the cells were treated with caffeine.

Treatment with caffeine led to a 60% decrease in the level of apoptosis induced by vpr^HIV-1 in human cells (Fig. 8A). This degree of inhibition was statistically significant (P < 0.05; Student's t test). Because abrogation of G₂ arrest led to a dramatic decrease in the amount of apoptosis, we conclude that G₂ arrest is a necessary step in the full induction of apoptosis by vpr^HIV-1.

Treatment of vpr^AGM-expressing cells with caffeine showed a different response from the one we observed with vpr^HIV-1. While caffeine was able to inhibit the cell cycle arrest in Cos-7 cells (Fig. 2), it produced only a moderate degree of inhibition of apoptosis (Fig. 8B), which was not statistically significant (P > 0.05; Student's t test).

Nuclear localization of vpr^HIV-1 and vpr^AGM in human versus African green monkey cells.

In view of the above results and since vpr^HIV-1 and vpr^AGM naturally localize to the cell nucleus (8, 15, 33, 51, 53), we hypothesized that species-specific differences in the functions of these proteins may be a consequence of potential differences in subcellular localization.

In an effort to compare the subcellular localization properties of vpr^HIV-1 and vpr^AGM and how localization may relate to their function in human versus African green monkey kidney cells, we constructed GFP-vpr fusion proteins and studied their localization by transient transfection followed by fluorescence microscopy (Fig. 7C). The study of subcellular localization of proteins via fusion with GFP offers the significant advantage that fluorescence may be studied in living cells, in the absence of fixation or staining procedures which may lead to artifactual patterns of distribution. In the fusion proteins, we incorporated a flexible linker (10) between the fusion partners to prevent steric hindrance. GFP displayed a diffuse pattern of distribution and did not selectively accumulate in the nucleus or the cytoplasm in human HeLa and simian Cos-7 cells. As previously reported, GFP-vpr^HIV-1 selectively localized to the nucleus and only very faint fluorescence could be detected in the cytoplasm. GFP-vpr^AGM appeared to localize in the nucleus, with a pattern that was indistinguishable from that of GFP-vpr^HIV-1. A previous report indicated that while both vpr^HIV-1 and vpr^AGM selectively localized to the nucleus of human cells, vpr^HIV-1 was evenly distributed whereas vpr^AGM displayed a punctate distribution (51). Our observations with GFP-vpr fusion proteins suggest that both HIV-1 and African green monkey kidney Vpr proteins display a punctate nuclear distribution. Thus, subcellular localization appears not to be the determinant behind the failure of vpr^AGM to induce cell cycle arrest and apotosis in human cells.

DISCUSSION

All known primate lentiviruses contain one or two genes (vpr and/or vpx) that are considered vpr homologs. The various genes in the vpr family are structurally related because they have predicted amino acid sequence homology and encode products that are virion-encapsidated (54). These genes are involved in multiple aspects of the biology of primate lentiviruses, although their precise roles in viral replication and modulation of host functions are still unclear.

We and others previously reported that species-specific factors modulate the induction of G₂ arrest by members of the vpr family (42, 51). SIVagm and SIVsyk vpr genes are capable of arresting African green monkey kidney cells but are unable to do so in human cells. In contrast, HIV-1, HIV-2, and SIVsm vpr genes function in both simian and human cell types, although SIVsm Vpr functions more efficiently in simian cells than it does in human cells. These differences could not be explained on the basis of differential protein stability or subcellular localization (51). The species-specific cell cycle arrest differences between SIVagm and HIV-1 Vpr proteins indicate that these proteins may signal through cellular pathways that are divergent among primates.

We first compared vpr^AGM and vpr^HIV-1 at the level of transactivation. We wished to ascertain whether transactivation by vpr^AGM is species specific, as is its ability to induce G₂ arrest. Surprisingly, vpr^AGM was able to induce transactivation in human cells, despite its inability to induce cell cycle alteration in these cells. This observation suggests that vpr^AGM may exert transactivation by a mechanism that is, at least in part, independent of cell cycle manipulation. To corroborate this observation, we resorted to the use of caffeine, a known inhibitor of vpr-induced (43) and DNA damage-induced (36) G₂ arrest. Caffeine potently inhibited the cell cycle arrest by vpr^AGM in African green monkey kidney cells but produced little or no inhibition of its transactivation effect in either monkey or human cells.

The transactivation effects reported here and elsewhere for the various primate vpr genes are modest compared with the transactivation effects exerted by other, more classical viral transactivators (e.g., HIV-1 tat or human T-cell leukemia virus tax). The importance of vpr-induced transactivation, however, is underscored by two observations. First, HIV-1, HIV-2, SIVmac, and SIVagm vpr genes are able to transactivate their respective LTRs (39). This observation establishes the conservation of this function through evolution. Second, experiments in which the abilities of tat and vpr to transactivate were tested in parallel indicated that they were not mutually exclusive but were synergistic (28, 47). Thus, coexpression of the two genes provided a multiplicative effect on the promoter activity of the LTR.

We also wished to examine whether the ability of vpr^HIV-1 to cause apoptosis would be paralleled by vpr^AGM. Indeed, vpr^AGM is capable of inducing apoptosis in African green monkey kidney cells. This induction of apoptosis was not observed when vpr^AGM was expressed in human cells, and therefore apoptosis appears be determined by species-specific factors.

Abrogation of the cell cycle effect by caffeine treatment suppressed the levels of apoptosis by vpr^AGM, as it did for vpr^HIV-1. The potential link between the induction of G₂ arrest and the onset of apoptosis is likely to be complex, since multiple signaling connections have been proposed. Perhaps the area of radiation-induced DNA damage has produced the clearest results to date. DNA damage is a natural signal that may induce cell cycle arrest and apoptosis. Cell cycle arrest following cellular insult allows for the repair of damaged DNA to protect the organism from the repercussions of mutation (21). Cell cycle block can occur before DNA replication, at the G₁/S checkpoint, or before chromosome segregation, at the G₂/M checkpoint (21, 38). Many DNA-damaging agents, including certain antineoplastic drugs, exert their effect at the G₂/M phase (9) and, ultimately, commit the cell to death (32). The effects of vpr, specifically the G₂ block, may be mediated along similar pathways to the effects of genotoxic agents (43).

It was initially thought that the cell cycle perturbation function of vpr^HIV-1 would require that the protein be localized in the nucleus. This was supported by experiments by Di Marzio et al. (15), who failed to find vpr^HIV-1 mutants that could cause G₂ arrest in the absence of nuclear localization. Later mutagenesis experiments demonstrated that nuclear localization was not a requisite for vpr^HIV-1 to function as a cell cycle inhibitor (33, 53). We tested whether the subcellular localization of vpr^AGM might explain the species specificity of the effects of this protein. We found that in both human and African green monkey kidney cells a GFP-vpr^AGM fusion protein is also found in the nucleus. Thus, subcellular localization does not explain the inability of vpr^AGM to cause G₂ arrest and apoptosis in human cells.

Our results point out important functional differences between the HIV-1 and SIVagm vpr genes (Fig. 9). HIV-1 vpr appears to induce G₂ arrest as a primary effect, whereas transactivation and apoptosis appear to be downstream effects of G₂ arrest. Thus, inhibition of G₂ arrest will abrogate such downstream effects. SIVagm vpr induces these three effects independently. Thus, abrogation of vpr^AGM-induced G₂ arrest has no effect on the onset of apoptosis or transactivation of the LTR.

FIG. 9.

Proposed model to summarize the relationships among various functions of vpr^HIV-1 and vpr^AGM. The observed effects of drugs are depicted in grey. Caffeine inhibits vpr^HIV-1 G₂ arrest and also its downstream effects, apoptosis and transactivation. For vpr^AGM, caffeine inhibits G₂ arrest but not apoptosis or transactivation. Transactivation by vpr^AGM is due, in part, to a cell cycle-independent mechanism, and therefore, caffeine has little or no effect on transactivation. However, the SIVagm LTR is also responsive to cell cycle arrest since taxol can induce transactivation. Because treatment with caffeine does not significantly affect transactivation, the significance of G₂ arrest for vpr^AGM is uncertain (question mark).

Taken together, our results suggest that while the multiple functions of vpr are conserved among various primate lentiviruses, the mechanisms leading to the execution of such functions are divergent.