ABSTRACT
The human immunodeficiency virus type 1 Vpr protein is both packaged into virions and efficiently localized to the nucleus. In this report, we show that a significant fraction of Vpr also accumulates in the cytoplasm of virus-producing cells. Although Vpr shuttles between the nucleus and the cytoplasm, studies with an export-deficient Vpr mutant reveal that nuclear export is not required for virion incorporation.
The ∼15-kDa human immunodeficiency virus type 1 (HIV-1) accessory protein Vpr is efficiently packaged into progeny virions (4, 25, 32). Although other accessory proteins, namely, Vif (3, 9, 20, 23) and Nef (31,43), are also found in virus particles, the levels of incorporated Vpr appear to be considerably higher. Indeed, recent studies have shown that the ratio of Vpr to Gag in virions is ∼1:7 (29), whereas ratios in the order of ∼1:50 have been reported for Vif to Gag (3, 9, 23). In terms of the sequence determinants of Vpr packaging, a leucine-rich element in the p6 region of the p55Gag polyprotein and a region of Vpr that overlaps a predicted α-helical domain (residues 16 to 33) have each been found to be critical (5, 21, 24, 26).
Vpr is present early in HIV-1 infection as a component of postentry nucleoprotein complexes (frequently called preintegration complexes [PICs]) (12, 18) and is also expressed late in infection as the product of a Rev-dependent transcript (14). Two main functions that directly impact the HIV-1 replicative cycle, one of which is dependent on virion incorporation, have been attributed to Vpr. In the first function, Vpr enhances HIV-1 infectivity by stimulating the nuclear import of PICs (11, 18), an effect that is presumably mediated by the karyophilic properties of Vpr associated with these complexes (7, 42). In the second, Vpr that is expressed in infected cells that are proliferating induces their arrest at the G2/M boundary (17, 19, 35,36). It has been suggested that this helps to maximize virion production because the viral long terminal repeat promoter is most active during the G2 phase of the cell cycle (15,16).
Vpr, in the absence of other viral proteins, has been shown to localize efficiently to the nucleus (5, 7, 25). Given this strong karyophilic potential, it is unclear how Vpr becomes incorporated into progeny virus particles that assemble in the cytoplasm (13). To address this issue, we used two independent methods to examine the subcellular localization of Vpr in the absence or presence of the other viral proteins. Initially, HeLa cell monolayers were transiently transfected with either the Vpr expression vector pCMV/Vpr or the full-length proviral vector pYU-2 (7). At 48 h, the cells were subjected to indirect immunofluorescence using Vpr- or Gag-specific antibodies (Fig.1). As previously demonstrated, transfection with pCMV/Vpr alone resulted in the marked nuclear localization of Vpr (Fig. 1a). In contrast, substantial accumulation of Vpr in the cytoplasm was observed when expression was in the context of the provirus (Fig. 1b). Consistent with previous work, Gag was detected only in the cytoplasm (Fig. 1c).
Subcellular localization of Vpr in the absence or presence of viral proteins. HeLa cell monolayers (35-mm diameter) were transiently transfected with either the pcDNA1-based wild-type Vpr expression vector pCMV/Vpr (a) or an expression vector containing the HIV-1YU-2 provirus (b and c) (8). At 48 h, cells were fixed and analyzed by indirect immunofluorescence using a Vpr-specific monoclonal antibody, followed by a fluorescein isothiocyanate-conjugated secondary antibody. Gag was detected using a p24Gag-specific antiserum raised in rabbits, followed by a Texas red-conjugated secondary antibody (41). Samples were photographed at a magnification of ×400 using a Nikon Microphot-SA microscope attached to a charge-coupled device camera.
We next used a membrane flotation assay, in which cell extracts are overlaid with a discontinuous sucrose gradient and the ability of a protein of interest to ascend the gradient during centrifugation is used as a measure of its association with membranes (30). Cytoplasmic extracts were prepared from HeLa cells transfected with either pCMV/Vpr or a full-length provirus expression vector. Samples were then subjected to flotation through a 65% sucrose layer, and the gradient fractions were analyzed for the presence of Vpr or Gag by immunoprecipitation followed by Western blotting (Fig.2). As expected, Vpr alone did not appear to be membrane associated and was detected only in fractions from the bottom of the gradient (Fig. 2A, lanes 8 to 11). However, in cells cotransfected with the provirus, a significant proportion of Vpr was found in membrane-containing fractions at the top of the gradient (Fig.2B, lane 3). A significant percentage of total Gag protein was also detected in this region of the gradient; this percentage presumably represents Gag that is being assembled into progeny virions at the plasma membrane.
Association of Vpr with membranes in HIV-1-infected cells. A total of 4 × 107 HeLa cells were transiently transfected with pCMV/Vpr (A), a wild-type provirus expression vector (B), or a proviral vector containing a disrupted p6Gagregion (the 44LF→PS mutation) (1) (C). At 48 h, cells were fractionated into nuclear and cytoplasmic compartments by treatment with 0.01% digitonin, followed by Dounce homogenization. Samples were centrifuged at 2,000 × g to pellet the nuclei, and the supernatants were concentrated to ∼600 μl using a Centricon Plus 5-kDa nominal molecular weight limit centrifuge filter, mixed with 90% sucrose to yield a final sucrose concentration of 70%, and then processed as described previously (30). Gradient fractions were analyzed for Vpr or Gag by immunoprecipitation using specific antisera raised in rabbits and by Western blotting using a polyclonal Vpr-specific mouse antiserum or a p24Gag-specific monoclonal antibody (41). Bound antibodies were visualized using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence.
These experiments revealed that a significant proportion of HIV-1 Vpr is cytoplasmic in the context of virus-producing cells. This diminution in the efficiency of nuclear accumulation presumably facilitates the packaging of Vpr into virions. Vpr that is cytoplasmic could be derived from two nonexclusive sources: nuclear Vpr could be exported to the cytoplasm, or newly translated Vpr could be precluded from entering the nucleus by retention in the cytoplasm.
Since Vpr is both nuclear and cytoplasmic in provirus-transfected HeLa cells (Fig. 1b), the possibility that nuclear Vpr can undergo export to the cytoplasm was evaluated using transient interspecies heterokaryons (Fig. 3a to c) (28). Here, proteins that have the ability to enter and exit the nucleus (called shuttle proteins) rapidly redistribute between multiple nuclei that have been manipulated to share a common cytoplasm. HeLa cells that had been transfected with pCMV/Vpr (donor cells) were therefore fused to nonexpressing murine L cells (acceptor cells) using polyethylene glycol and cultured for 2 h in the presence of cycloheximide to prevent further translation. Cells were then fixed and double labeled using a Vpr-specific monoclonal antibody (Fig. 3a) and a DNA dye, Hoechst 33258, that stains L nuclei with a signature punctate pattern (Fig. 3b) and thereby distinguishes them from the donor nuclei. As seen in Fig.3a, Vpr rapidly and efficiently accumulated in the L acceptor nuclei, thus defining HIV-1 Vpr as a nucleocytoplasmic shuttle protein.
HIV-1 Vpr is a nucleocytoplasmic shuttle protein. HeLa cells were transfected with the wild-type expression vector pCMV/Vpr (a to c and g to i) or its mutated derivative pCMV/VprL68A (d to f). At 24 h, the cells were trypsinized and overlaid onto mouse L cells previously plated on glass coverslips. After the HeLa cells had settled, the coverslips were washed with phosphate-buffered saline (PBS) and inverted onto polyethylene glycol (Sigma P-7181) to initiate cell-to-cell fusion. After ∼100 s, the coverslips were washed gently with PBS, placed into fresh medium containing 50 μg of cycloheximide/ml, and incubated at 37°C for 2 h in the absence (a to f) or presence (g to i) of 5 nM LMB. Samples were fixed and then subjected to indirect immunofluorescence using a Vpr-specific monoclonal antibody, followed by a fluorescein isothiocyanate-conjugated secondary antibody. Before mounting, cell nuclei were stained using Hoechst 33258 (Molecular Probes), followed by a PBS wash. The corresponding phase-contrast images of the heterokaryons are shown in panels c, f, and i. For the Rev controls (j and k), HeLa cells were transfected with a vector that expresses a Rev-green fluorescent protein (GFP) fusion, maintained for 48 h, incubated for 3 h in medium containing 50 μg of cycloheximide/ml and 5 μg of actinomycin D (Act. D)/ml in the absence or presence of 5 nM LMB, and then fixed and mounted. W/T, wild type.
The nuclear export of proteins, like protein nuclear import, is usually mediated by specific peptide sequences that engage nuclear transport receptors (27). In the case of export, these targeting signals are called nuclear export signals (NESs) and are commonly rich in hydrophobic amino acids such as leucine. Among our collection of mutated vpr alleles are a number that encode missense mutants where leucine residues are replaced with other amino acids. To test whether any of these mutant proteins are debilitated for nuclear export, heterokaryon assays were performed using HeLa and L cells as described above. The results for one of the mutants, where the leucine at position 68 has been exchanged for alanine (L68A), are shown in Fig.3d to f. Unlike wild-type Vpr, the L68A protein failed to accumulate in the L nuclei, demonstrating that the leucine at position 68 is essential for the nuclear export function of Vpr. Similar results were obtained using Vpr encoded by a T-cell line-adapted virus, HIV-1IIIB, and with mutant Vpr proteins in which the leucines at positions 64, 67, and 68 were all replaced with alanine (data not shown).
Prototypic leucine-rich NESs, such as that contained in the HIV-1 Rev protein (34), confer nuclear export through binding to the export receptor CRM1; importantly, both binding and export are sensitive to inhibition by the fungal metabolite leptomycin B (LMB) (6, 10). Because leucine 68 is in a region of Vpr that contains several other leucines as well as additional hydrophobic residues, we assessed the sensitivity of Vpr shuttling to inhibition by LMB using the heterokaryon assay. Addition of 5 nM LMB to heterokaryons containing wild-type Vpr had no inhibitory effect on the extent of shuttling (Fig. 3g to i). Control samples expressing a Rev-green fluorescent protein fusion that had been relocalized to the cytoplasm by treatment with 5 μg of actinomycin D/ml (Fig. 3j) (28) confirmed that LMB efficiently inhibited leucine-rich NES-mediated export in our hands (Fig. 3k). The lack of susceptibility of Vpr export to LMB suggests that its export mechanism may differ from that used by classical leucine-rich NESs. For instance, Vpr may utilize CRM1 in an unconventional manner (22) or may achieve export through a CRM1-independent mechanism. Either way, LMB resistance is perhaps not surprising, given the lack of evident sequence alignment between the NES consensus sequence and the region of Vpr that includes residue 68 (Fig. 4).
Organization of 97-amino-acid HIV-1YU-2 Vpr. The amino acid sequence neighboring the leucine at position 68 is shown in the context of a Vpr secondary structure profile predicted by the program PHDsec (37-39); the regions of predicted α-helix are indicated in gray. A consensus leucine-rich NES sequence derived in vivo (2) is shown below the Vpr sequence for comparison. Relatively conserved hydrophobic residues are boxed in gray, X denotes any amino acid, and λ represents amino acid residues with bulky hydrophobic side chains.
Having found that HIV-1 Vpr shuttles and that the L68A mutation blocks export, we were in a position to address whether virion packaging is dependent on nuclear export. 293T monolayers were therefore cotransfected with pYU-2/Δvpr and hemagglutinin (HA)-tagged versions of pCMV/Vpr or pCMV/VprL68A. At 24 h, virions were pelleted from culture supernatants by centrifugation at 100,000 × g for 90 min and analyzed by Western blotting for Gag and Vpr content in parallel with the corresponding whole-cell lysates. As seen in Fig. 5, wild-type Vpr and the L68A mutant were incorporated into virions with comparable efficiencies (lanes 3 and 4). The lower level of packaged L68A is most likely a consequence of reduced accumulation in the 293T producer cells (compare lane 1 to lane 2), a phenomenon that is frequently observed for mutant Vpr proteins (5). In sum, HIV-1 Vpr can be incorporated into progeny virions irrespective of nuclear export.
Nuclear export of Vpr is not required for virion packaging. A total of 107 293T cells were cotransfected with pYU-2/Δvpr and vectors containing HA-tagged versions of wild-type Vpr, pCMV/HA-Vpr, or the L68A mutant pCMV/HA-VprL68A. At 24 h, viral supernatants were clarified by centrifugation and filtered through a 0.45-μm-pore-size membrane and virions were pelleted through a 20% sucrose cushion by ultracentrifugation at 100,000 × g for 90 min. The transfected cells (lanes 1 and 2) and pelleted virions (lanes 3 and 4) were resuspended in loading buffer and resolved in parallel by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Vpr and Gag were detected by Western blotting using the HA-specific monoclonal antibody HA.11 (Covance) and a p24Gag-specific monoclonal antibody, respectively, and by enhanced chemiluminescence. W/T, wild type.
In order for Vpr packaging to occur, newly translated Vpr must be retargeted away from the nucleus and towards the cytoplasmic site(s) of virus assembly and budding. In light of Vpr's ability to interact with p6Gag (1, 40) and the essential role of the leucine-rich region of p6 in Vpr packaging (21, 24), we wished to determine if this Gag sequence also mediated the cytoplasmic retention of Vpr in HIV-1-infected cells. To address this, a p6 mutant provirus expression vector was constructed such that the critical leucine and phenylalanine residues at positions 44 and 45 (1,21) were exchanged for proline and serine. Using transfected cells and the same membrane flotation assay employed earlier, we found that Vpr accumulated in cytoplasmic membrane-containing fractions with an efficiency that closely matched that seen for the wild-type virus (Fig. 2, compare panels B and C). Thus, the retention of Vpr in the cytoplasm of virus-producing cells is not mediated solely through an interaction with the leucine-rich element of Gag. Whether cytoplasmic accumulation is dictated by other regions of Gag, other viral proteins, or as an indirect consequence of HIV-1 gene expression in host cells remains to be determined.
Finally, our observations also raise the question, what is the biological significance of Vpr shuttling? Although resolution of this issue awaits further experimentation, it is tempting to speculate that the nucleocytoplasmic movement of Vpr, like that of many cellular regulators of the cell cycle (33, 44), may relate to the ability of Vpr to influence normal cell cycle progression.
ACKNOWLEDGMENTS
We thank Minoru Yoshida for the generous gift of LMB, David Rekosh and Marie-Louise Hammarskjöld for important reagents, and Laurie Zimmerman for excellent secretarial support.
This work was supported by U.S. Public Health Service grants AI46942 (M.H.M.), AI09996 (Y.J.), and GM18907 (P.V.S.).
FOOTNOTES
- Received 18 December 2000.
- Accepted 30 May 2001.
↵* Corresponding author. Mailing address: Department of Microbiology, University of Pennsylvania School of Medicine, 347 B Clinical Research Bldg., 415 Curie Blvd., Philadelphia, PA 19104-6148. Phone: (215) 573-3493. Fax: (215) 573-2172. E-mail:malim{at}mail.med.upenn.edu.
REFERENCES
- Copyright © 2001 American Society for Microbiology
