Human Immunodeficiency Virus Type 1 Nef Functions at the Level of Virus Entry by Enhancing Cytoplasmic Delivery of Virions

Abstract

The Nef protein of the type 1 human immunodeficiency virus (HIV-1) plays a key although poorly understood role in accelerating the progression of clinical disease in vivo. Nef exerts several biological effects in vitro, including enhancement of virion infectivity, downregulation of CD4 and major histocompatibility complex class I receptor expression, and modulation of various intracellular signaling pathways. The positive effect of Nef on virion infectivity requires its expression in the producer cell, although its effect is manifested in the subsequent target cell of infection. Prior studies suggest that Nef does not alter viral entry into target cells; nevertheless, it enhances proviral DNA synthesis, arguing for an action of Nef at the level of viral uncoating or reverse transcription. However, these early studies discounting an effect of Nef on virion entry may be confounded by the recent finding that HIV enters cells by both fusion and endocytosis. Using epifluorescence microscopy to monitor green fluorescent protein-Vpr-labeled HIV virion entry into HeLa cells, we find that endocytosis forms a very active pathway for virus uptake. Virions entering via the endocytic pathway do not support productive infection of the host cell, presumably reflecting their inability to escape from the endosomes. Conversely, our studies now demonstrate that HIV Nef significantly enhances CD4- and chemokine receptor-dependent entry of HIV virions into the cytoplasmic compartment of target cells. Mutations in Nef either impairing its ability to downregulate CD4 or disrupting its polyproline helix compromise virion entry into the cytoplasm. We conclude that Nef acts at least in part as a regulator of cytosolic viral entry and that this action contributes to its positive effects on viral infectivity.

In humans infected with human immunodeficiency virus type 1 (HIV-1) and in adult rhesus monkeys experimentally infected with simian immunodeficiency virus type 1, the absence of the nef gene significantly diminishes plasma viral loads and markedly slows clinical progression to overt disease (11, 26). How Nef accelerates viral pathogenesis in vivo has been the subject of significant recent study (for reviews, see references 15, 21, 40, and 44). In vitro studies clearly demonstrate that Nef enhances virion infectivity, mediates downregulation of surface CD4 and major histocompatibility complex (MHC) class I expression, induces chemokine release from infected macrophages (55), and alters various intracellular signaling pathways (23, 28, 43, 46, 47, 57). The effects of Nef on virion infectivity involve both CD4-dependent and -independent components, as revealed by the use of CD4⁺ and CD4⁻ producer cells (7, 17, 48, 59). Expression of Nef as a transgene in mice recapitulates many of the pathological effects found in patients with AIDS (20), underscoring the multifaceted function of this viral regulatory protein.

A number of studies have shown that Nef increases HIV-1 replication by enhancing the infectivity of virions (8, 16, 25, 34, 36, 52). The fact that Nef-defective viruses can achieve nearly wild-type (wt) levels of infectivity if Nef is provided in trans in the virus-producing cells suggests that Nef either directly or indirectly modifies the virion (2, 37). Of note, approximately 10 to 100 Nef molecules are detectable within the viral particle; Nef itself might directly trigger early events in the subsequent target cells (27, 42, 58). However, a major role for intravirion full-length Nef is drawn into question by the finding that Nef is cleaved by the HIV-1 protease (6, 14, 35, 41). No clear biological activity has yet been ascribed to the resulting Nef fragments (14, 35, 41). Alternatively, the inclusion of Nef in the virion may facilitate the incorporation of Nef-associated cellular kinases that phosphorylate various substrates, including the viral matrix protein (54). Such posttranslational modifications of proteins within the virion could contribute to its ability to enhance viral infectivity (5). Finally, Nef expression in the producer cell may lead to changes in virion structure or composition, producing the observed effects on infectivity.

Since Nef does not appear to alter virion binding or entry but does enhance proviral DNA synthesis, an early postentry action of Nef at the level of viral uncoating or reverse transcription has been proposed (1, 2, 7, 50). These positive effects of Nef are obtained with viruses containing the HIV-1 envelope and also with virions pseudotyped with the amphotropic murine leukemia virus envelope (1, 31, 37). However, HIV-1 virions pseudotyped with the envelope glycoprotein (G) from the vesicular stomatitis virus (VSV-G), which targets virions for entry via endocytosis and fusion within acidified endosomes, do not display Nef-mediated enhancement of infectivity (1). These findings suggest that Nef enhancement of infectivity depends on the route by which the HIV-1 virion enters the cell. Recent studies have further shown that HIV-1 entry occurs both through plasma membrane fusion, leading to productive infection, and through endocytosis, usually leading to nonproductive forms of infection (33). Using subcellular fractionation techniques that segregate the individual contributions of these two pathways of entry, we now demonstrate that Nef significantly enhances the cytosolic entry of viral particles occurring via fusion at the plasma membrane. In previous studies, these effects of Nef at the level of cellular entry were likely masked by a high background of virion entry by endocytosis that was not altered by Nef. These studies also provide a potential explanation for why VSV-G-pseudotyped viruses, which fuse through the endosome rather than at the plasma membrane, fail to display a Nef infectivity phenotype.

MATERIALS AND METHODS

Cells and viruses.

HeLa, HeLa-CD4, and MAGI (HeLa-CD4-LTR-LacZ) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and a mixture of penicillin and streptomycin. Jurkat T cells were grown in RPMI 1640 with the same supplements. NL4-3 molecular clones encoding wt Nef, a deletion mutant of Nef encoding only the N-terminal 35 amino acids (ΔNef), or specific substitution mutants of Nef (WL57AA, E64-68A, R77A, P69P72P75/AAA, LL164AA, and DD174AA) (4, 17) were transfected into 293T cells, and viruses were harvested from the supernatants after 48 h. Viruses lacking the HIV-1 envelope (Δenv) were prepared by transfection of pNL4-3.Luc.R-E- (National Institutes of Health AIDS Research Program, catalog number 3418; donated by N. Landau). In some experiments, these viruses were pseudotyped with VSV-G by cotransfecting 293T cells with the pNL4-3.Luc.R-E- and pHCMV-G (kindly provided by O. Keppler, Gladstone Institutes, San Francisco) expression plasmids. Wt and ΔNef R5-tropic NL4-3 were generated by replacing an EcoRI-BamHI restriction fragment within NL4-3 with an identical fragment derived from clone 81A (56) which contains the V1, V2, and V3 loops of the macrophage-tropic HIV-1 isolate Ba-L. All transfections were performed by using calcium phosphate to precipitate DNA. Viral stocks were normalized by their p24^gag content, measured in an enzyme-linked immunosorbent assay (ELISA; NEN Life Science Products).

Preparation of GFP-Vpr-labeled HIV-1 virions.

Green fluorescent protein (GFP)-expressing virions were produced by cotransfection of 293T cells (plated in a T75 flask) with HIV-1 pNL4-3 proviral DNA (15 μg) and an expression vector (15 μg) encoding a GFP-Vpr fusion protein. After 48 h, the virus-containing supernatant was subjected first to low-speed centrifugation to remove cells and debris and then to ultracentrifugation at 20,000 rpm in an SW41 rotor for 2 h at 4°C to sediment viral particles. The virus-containing pellet was resuspended in complete medium (0.5 ml) and stored in aliquots at −70°C.

Virion binding and entry assays.

Confluent cells cultured in 24-well plates were inoculated in duplicate with viruses containing the wt nef gene or the Δnef deletion mutant (50 ng of p24^gag). These cultures were incubated at 4 or 37°C for 30 min or 4 h, washed with phosphate-buffered saline (PBS), and treated or not with trypsin to remove surface-bound but uninternalized virions. The cells were then washed twice with PBS, and cell lysates were prepared by resuspending the pellet in 200 μl of lysis buffer (PBS containing 1% Triton X-100) and freezing at −20°C. Levels of p24^gag in these lysates were measured by ELISA.

Infectivity analysis.

HeLa-CD4 or HeLa cells cultured in a 48-well plate were incubated with HIV-1 encoding wt Nef or the various mutants of Nef (50 ng of p24^gag). After 16 h, cells were washed, trypsinized, and replated in 12-well plates; after 2 days, the cells were transferred to six-well plates. Human Jurkat T cells (2 × 10⁵) were similarly infected with HIV-1 pNL4-3 (50 ng of p24^gag) for 16 h, and excess free virus was removed by washing. HIV-1 replication was monitored by measuring p24^gag levels in the culture supernatants during subsequent culture.

Cell fractionation assays.

HeLa-CD4 cells grown to approximately 80% confluence in a 75-cm² culture flask or Jurkat T cells (10⁷ cells) in mid-log phase of growth were incubated with HIV-1 containing wt or mutant nef gene products (500 ng of p24^gag) for 60 min at 37°C. To remove surface-bound virions, the cells were then washed in PBS at 4°C and trypsinized for 5 min at room temperature (HeLa-CD4) or 3 min at 4°C (Jurkat). Cells were then washed once in 5 ml of DMEM supplemented with 10% FBS and twice in ice-cold PBS to deplete the trypsin. The cells were then resuspended in 2 ml of hypotonic buffer (10 mM Tris-HCl [pH 8], 10 mM KCl, 1 mM EDTA) for 15 min (HeLa-CD4) or 1 min (Jurkat) at 4°C and disrupted by Dounce homogenization (15 strokes for HeLa-CD4 and three strokes for Jurkat, 7-ml B pestles). Nuclei and cell debris were pelleted by centrifugation (3,000 rpm for 5 min at 4°C). The postnuclear extracts were centrifuged at 22,000 rpm for 30 min at 4°C in a 28RS Heraeus centrifuge. The supernatant, representing the cytosolic fraction, was adjusted to 0.5% Triton X-100. The pellet representing the vesicular fraction including endosomes was resuspended in lysis buffer (20 mM HEPES [pH 7.4], 0.5% Triton X-100, 150 mM NaCl). p24^gag levels were quantitated in each fraction by ELISA.

Epifluorescence microscopy.

HeLa and HeLa-CD4 cells were grown to 70% confluence on glass coverslips in 24-well plates. Cells were incubated for 20 to 120 min at 37°C with GFP-Vpr-labeled HIV-1 virions (200 ng of p24^gag in 80 μl of culture medium per coverslip) and then fixed in a 3.7% paraformaldehyde–PBS solution for 20 min. After additional washing in PBS, coverslips were secured to microscope slides with Gel Mount (Biomeda Corp., Foster City, Calif.). Samples were imaged with a Nikon Eclipse E600 microscope and a SPOT2 digital camera (Diagnostic Instruments, Inc.). Images were analyzed using Adobe Photoshop.

Colocalization of GFP-Vpr-labeled HIV-1 with plasma membrane, endosomal, or lysosomal markers.

HeLa cells were incubated with GFP-Vpr-labeled HIV-1 for 45 min at 37°C and then washed in PBS. These cells were treated with tetramethylrhodamine-transferrin (TAMRA-Tf; 50 μg/ml final concentration in PBS; Molecular Probes, Eugene, Oreg.) for 10 min at 37°C to label endosomes or with LysoTracker Red (100 nM final concentration in culture medium; Molecular Probes) for 15 min at 37°C to label lysosomes. Cells were washed with PBS, fixed, and analyzed microscopically as described above.

RESULTS

Use of GFP-Vpr-containing HIV virions to monitor virion uptake in HeLa cells expressing or not expressing CD4.

Expression plasmids encoding GFP fused to the N terminus of Vpr and full-length HIV-1 NL4-3 containing wt nef or the Xho deletion of nef (encodes only the N-terminal 35 amino acids of Nef, termed ΔNef) were cotransfected into 293T cells to prepare fluorescently labeled virions. In agreement with previous findings (53), wt viruses containing the GFP-Vpr fusion protein displayed normal single-step infectivity profiles in the MAGI cell assay (HeLa-CD4 cells containing a stably integrated Tat-responsive HIV-1 long terminal repeat-lacZ reporter plasmid). Conversely, viruses containing the Δnef deletion mutant exhibited diminished infectivity (data not shown).

The fluorescent properties of these GFP-Vpr-labeled virions were used to monitor early events of HIV-1 infection in single HeLa-CD4 cells. Equal amounts of these viruses, based on p24^gag content (200 ng of p24^gag, high multiplicity of infection) were added to cells cultured on glass cover slips (Fig. 1A and B). Analysis of the epifluorescence pattern suggested the presence of both surface-bound and internalized virions (see below for results of Z stage analyses of multiple focal planes). No striking differences were obtained with viruses containing wt nef versus Δnef.

FIG. 1.

HIV efficiently enters HeLa cells lacking CD4 and localizes in endosomes. (A to D) Green-fluorescing HIV-1 virions were produced by cotransfection of 293T cells with expression vectors encoding GFP-Vpr and either pNL4-3, containing the wt nef gene, or pNL4-3Δnef, containing a nef deletion mutant. HeLa-CD4 or HeLa cells grown on coverslips were incubated with these fluorescent virions (200 ng of p24^gag) in 80 μl for 20 min at 37°C. The cells were processed for fluorescence microscopy as described in Materials and Methods. Note significant binding and entry of green-fluorescing virions into HeLa cells. (E to J) In addition to exposure to wild-type fluorescent HIV virions, HeLa cells were incubated with TAMRA-Tf to fluorescently label endosomes and with LysoTracker Red to fluorescently label lysosomes. Overlay of the combined fluorescent signals is shown in panels G and J, demonstrating significant colocalization of the GFP-Vpr-labeled viral particles with TAMRA-Tf.

HeLa cells lacking CD4 were examined in parallel as a control for the HeLa-CD4 studies. Surprisingly, significant intracellular staining was obtained using viruses expressing either wt nef or the Δnef mutant. Further, the wt nef virions displayed diffuse staining highlighted by numerous focal accumulations, whereas the HIV Δnef viruses yielded a rather distinct punctate pattern of fluorescence (Fig. 1C and D). Employing a Z stage motor to generate 40 cross-sectional planes through the bodies of these cells, we found that the majority of the punctate fluorescence was intracellular (data not shown). These differences were consistent in five independent virus preparations. Exposure of HeLa-CD4 or HeLa cells to the supernatants of GFP-Vpr-transfected cells did not yield any significant epifluorescence signal, indicating that the HIV genome is required to generate this staining pattern (data not shown). Together, these data indicate that HIV enters HeLa cells lacking CD4 and HeLa cells expressing CD4 a comparable levels.

We next investigated whether these epifluorescent signals were localized in a specific intracellular compartment. Concanavalin A-tetramethylrhodamine, TAMRA-Tf, and LysoTracker Red were used to label the plasma membrane, endosomes, and lysosomes, respectively (Fig. 1E to I). Epifluorescence derived from the GFP-Vpr-labeled virions principally colocalized with TAMRA-Tf (see overlay in Fig. 1G), suggesting significant accumulation of these viral particles in the endosomal compartment. This finding is consistent with a prior report by Marechal et al. (33), who first described significant endocytosis of HIV-1 through a mechanism that is apparently independent of the CD4 receptor.

Analysis of HIV-1 nef⁺ and Δnef virion uptake and replication in HeLa cells expressing or not expressing CD4.

HeLa-CD4 and HeLa cells were used for HIV-1 entry assays (37), employing viruses produced in the presence of the wt nef gene or Δnef deletion mutant. Viruses were incubated with the different cellular hosts for 30 min or 4 h at either 4 or 37°C, followed by incubation of a subset of samples in trypsin to remove virus bound at the cell surface (Fig. 2A). After extensive subsequent washing, the concentration of p24^gag present in the cell lysates was measured by ELISA to assess viral particle entry. The overall level of viral binding and entry in HeLa-CD4 cells varied between 0.5 and 2% of the total p24^gag input. When incubated at 37°C for 4 h 80 to 90% of the binding proved resistant to trypsin treatment, consistent with virion internalization (Fig. 2A, compare columns 3 and 4 with columns 1 and 2). In agreement with previous results (2, 7, 37, 51), no significant difference in entry was observed with viruses containing wt nef versus the Δnef mutant (compare columns 3 and 4). Culturing of cells at 4°C revealed significant albeit lower binding of both viruses (columns 5 and 6). However, as expected under these conditions, both viruses remained at the cell surface, displaying nearly complete sensitivity to trypsin (columns 7 and 8). When HeLa cells lacking CD4 were studied at 37 and 4°C, similar binding and entry results were obtained with the viruses. Furthermore, the presence of wt nef versus Δnef did not affect entry (columns 9 to 16). These findings confirm the microscopic studies indicating that viral entry occurs in the absence of CD4 and furthermore that nef does not significantly enhance or inhibit this response.

FIG. 2.

Analysis of binding, entry, and replication of CXCR4- and CCR5-tropic HIV containing wt nef or Δnef in HeLa-CD4 and HeLa cells. (A) Similar levels of HIV-1 Nef wt and ΔNef bind to and enter HeLa-CD4 and HeLa cells at 4 h. Confluent cells were inoculated with HIV-1 pNL4-3 wt X4 and ΔNef X4 (50 ng of p24) and incubated at either 4 or 37°C, as indicated. After 4 h, the cells were washed with PBS and either trypsinized (+T) or washed with PBS (−T). Cells were then washed twice with PBS, and cell lysates were prepared. Levels of p24^gag in the cell lysates were measured by ELISA. Data represent averages of three independent infections performed in duplicate. Note comparable entry of both HIV wt nef and HIV Δnef virions into HeLa-CD4 cells and persistent entry of these viruses into HeLa cells lacking CD4. (B) Entry of CXCR4-tropic or CCR5-tropic strains of HIV-1 containing wt nef or Δnef into HeLa-CD4 cells. Assays were performed as described above except that the cells were incubated with virions for 30 min at 37°C. Note effective entry of R5-tropic HIV into HeLa-CD4 cells lacking CCR5. (C) Analysis of HIV replication in HeLa-CD4 and HeLa cells. The viruses and cells described for panels A and B were used to study viral replication. Samples of the culture supernatant were analyzed for p24^gag content on days 2 to 7 of short-term cultures. Data shown are from a representative experiment performed three times with comparable results. Note that effective viral replication occurred only when CXCR4-tropic HIV containing the wt nef gene was cultured with HeLa-CD4 cells.

We next investigated whether altering the chemokine receptor tropism of these viruses would lead to diminished entry (Fig. 2B). Although HeLa-CD4 cells lack CCR5 receptors (these cells display CXCR4), significant entry of a CCR5-tropic version of the NL4-3 virus was observed following incubation for 30 min at 37°C (Fig. 2B, columns 3 and 4 and columns 7 and 8). Of note, a modest nef-dependent difference in entry was observed for the CXCR4-tropic viruses (compare columns 5 and 6); however, when the incubation was extended from 30 min to 4 h, this difference disappeared. These results indicate that HIV can effectively enter HeLa-CD4 cells in the absence of the appropriate chemokine receptor. In view of the endosomal pattern of viral particle localization, these findings are most consistent with significant endocytosis of HIV-1 virions proceeding in a CD4- and chemokine receptor-independent manner.

To compare the biological consequences of virion entry into these various HeLa and HeLa-CD4 cell cultures, each cell type was cultured for 7 days and p24^gag in the supernatant was measured daily. Productive viral replication was detected only in the HeLa-CD4 culture infected with the X4-tropic version of HIV-1 containing the wt nef gene (Fig. 2C). These results confirm a prominent role for Nef in enhancing viral infectivity and further show that viral entry occurring through the endocytic pathway does not lead to productive infection of these target cells. Of note, Fackler and Peterlin have recently reported that endocytosis of HIV-1 SF2 virions leads to viral replication and production. However, such replication of NL4-3-based strains of HIV-1 does not occur (12).

Nef enhances HIV-1 entry mediated through fusion at the plasma membrane.

Since viral entry by endocytosis occurred at high levels and could potentially mask an effect of Nef on fusion-mediated entry at the plasma membrane, it was important to segregate and quantitate virion entry occurring through these two pathways. For this purpose, we prepared cytosolic and membrane-bound fractions enriched in endosomes from HeLa-CD4 and human Jurkat T cells infected with viruses containing either wt nef or Δnef or, as a control, viruses lacking the HIV-1 envelope gene (Δenv). This last virus is internalized exclusively within the vesicular endosome fraction (33). Cytosolic and membrane-bound fractions were prepared 1 h after infection and analyzed for p24^gag content, which is expressed as a percentage of the whole-cell p24^gag detected (Fig. 3). Following infection of HeLa-CD4 cells with HIV-1 containing wt nef, approximately 13% of the p24^gag appeared in the cytosolic fraction. In sharp contrast, infection of the same cells with HIV-1 containing Δnef yielded less than 1% cytosolic p24^gag. As expected, infection with the Δenv HIV-1 produced virtually no detectable cytosolic p24^gag (Fig. 3A). These results indicate that Nef significantly enhances the entry of viral particles into the cytosol occurring as a result of fusion at the plasma membrane.

FIG. 3.

Nef enhances HIV-1 entry into the cytosol of HeLa-CD4 and Jurkat T cells. HeLa-CD4 cells or Jurkat T cells were infected with HIV viruses containing the wt nef gene or the Δnef mutation or viruses lacking the HIV envelope gene (Δenv). Additionally, the wt nef and or Δnef viruses were tested after pseudotyping with the VSV-G envelope (VSV env) (500 ng of p24^gag). After a 1-h incubation of cells and viruses at 37°C, cytosolic and endosomal fractions were prepared as described in Materials and Methods. The level of p24^gag present in each of these fractions was measured by ELISA. Values correspond to the percentages of p24^gag in the cytosol relative to the total intracellular p24 Gag after substraction of the background values measured with the Δenv mutant. Data are representative of three to six experiments performed with three independent virus preparations. Cytosolic delivery averaged 0.1% of total p24^gag input for HIV-1 and 0.6% for HIV pseudotyped with the VSV-G envelope.

To extend this analysis to lymphoid cells, Jurkat T cells were infected with HIV-1 containing wt nef or Δnef. HIV-1 nef⁺ viruses yielded approximately 38% cytosolic p24^gag, whereas infection with the Δnef virus produced 19% cytosolic p24^gag. In agreement with prior studies (1), pseudotyping of HIV-1 with the VSV-G envelope, which mediates fusion after endocytosis and acidification of the endosome, led to a loss of the nef-induced differences in cytosolic p24^gag expression (Fig. 3B). These results indicate that Nef also enhances HIV-1 entry into the cytosolic compartment of Jurkat CD4⁺ T cells and that viral entry into the cytosol is higher in Jurkat T cells (38%) than in HeLa-CD4 cells (13%). This result is in agreement with the prior results of Maréchal et al. (33) and argues for cell type-specific differences in fusion efficiency at the plasma membrane. We suspect that the smaller difference between cytosolic p24^gag levels obtained with wt nef and Δnef viruses in Jurkat T cells compared with HeLa-CD4 cells may derive in part from increased friability of the Jurkat endosomes in the fractionation assay employed. Rupture of these endosomes during isolation could artifactually increase p24^gag levels in the cytosolic function.

Viruses containing various alanine substitution mutations in nef were next examined for differences in cytosolic entry into HeLa-CD4 cells (Fig. 4). Mutation of the polyproline helix implicated in MHC class I downregulation and p21-activated kinase (PAK) binding (P69-P72-P75, designated PPP) (39, 48, 49, 59) or of various residues involved in CD4 downregulation (WL57 and LL164) (4, 9, 18) proved much less effective than wt nef in enhancing cytosolic p24^gag delivery. Conversely, alanine substitution at E64-68 (19) or R77 (10), implicated in MHC class I downregulation and Hck binding, respectively, did not alter the ability of the resulting Nef analogues to enhance cytosolic p24^gag expression. Alanine substitution of the two aspartic acid residues at positions 174 and 175, involved in vacuolar ATPase binding (30), gave an intermediate phenotype. These results suggest that Nef mutants exhibiting either compromised downregulation of CD4 or defective binding to select SH3 target domains or PAKs fail to support enhanced HIV entry into the cytoplasm.

FIG. 4.

Effect of nef mutations on cytosolic p24^gag entry into HeLa-CD4 cells. HeLa-CD4 cells were incubated with viruses containing wt nef, a nef deletion (Δnef), or various alanine substitution mutants of Nef for 1 h at 37°C and fractionated as described in the legend to Fig. 3. Histograms present the percentage of p24^gag present in the cytosolic fraction of these cultures. Note that mutation of nef either in its central proline helix or at two sites implicated in CD4 downregulation significantly impairs the cytosolic appearance of p24^gag.

To examine whether Nef-mediated enhancement of cytosolic entry correlates with viral infectivity, these different viruses containing the wild-type or mutant nef alleles were examined in Jurkat T cells after 3 days of culture (Fig. 5). Results are expressed as a percentage of supernatant p24^gag generated in Jurkat T cells following HIV wt nef infection. The levels of replication correlated with the levels of p24^gag measured in the cytosol. Similar results were obtained when primary peripheral blood mononuclear cells were used as targets for infection (results not shown). Thus, for the Δnef, PPP, WL57, and LL164 nef mutant viruses, diminished cytosolic entry observed 1 h after infection closely correlated with decreases in viral infectivity measured 3 days later. Conversely, viruses containing nef mutations that did not diminish virion entry displayed full infectivity.

FIG. 5.

Analysis of infectivity of HIV viruses containing wild-type or mutant nef genes. Viral replication was measured in Jurkat T cells infected with the indicated viruses. Levels of p24 Gag released into the culture medium were measured 3 days after infection. Values are representative of three independent experiments performed in duplicate. Standard deviation was less than 15% for all measurements.

DISCUSSION

The aim of these studies was to further define how the HIV-1 Nef protein enhances HIV-1 infectivity. It is important to note that enhancement of virus infectivity by Nef involves both CD4-dependent and CD4-independent components. The expression of Nef serves to counteract the inhibitory effect of surface CD4 receptor expression, leading to either more efficient release of budding virions or augmented infectivity of the released virions (3, 28, 47; for a review, see reference 22). Our studies specifically address the CD4-independent effects of Nef on viral infectivity, since all of the viruses used in this study were produced in 293T cells lacking CD4.

Our data demonstrate for the first time that Nef functions as an entry factor that enhances delivery of HIV-1 particles into the cytoplasm of target cells (see model provided in Fig. 6). Prior studies suggested that Nef exerted no effects at the level of viral entry but enhanced proviral DNA synthesis. These findings argued for an early effect of Nef at the level of viral uncoating or reverse transcription (2, 7, 37). However, it is now clear that these early-entry assays were confounded by a high background of unappreciated virion endocytosis (33) that likely masked the positive effects of Nef on cytoplasmic entry of HIV virions occurring via fusion at the plasma membrane. Endocytosis of virus, although not generally leading to productive forms of viral infection, can far exceed fusion-mediated entry in many cells. For example, in HeLa-CD4 cells, 85 to 90% of total viral entry is mediated through endocytosis occuring in a CD4- and chemokine receptor-independent manner. In contrast, in Jurkat T cells, approximately 40% of entry occurs through fusion at the plasma membrane and 60% through endocytosis. In contrast to HeLa-CD4 cells, we also find that endocytosis of HIV in primary T lymphocytes is significantly blocked by anti-CD4 antibodies, arguing for a role of CD4, perhaps as a viral tether, in this endocytic process (data not shown). Thus, endocytosis of HIV in HeLa cells proceeds independently of CD4, while this surface receptor may commonly participate in both fusion- and endocytosis-mediated entry of HIV in human T cells.

FIG. 6.

A schematic model summarizing the two pathways of HIV-1 entry into cells, including CD4- and chemokine-receptor-dependent fusion at the plasma membrane and receptor-independent endocytosis. HIV-1 virions entering by fusion at the plasma membrane support subsequent steps in the retroviral life cycle leading to HIV-1 replication and viral spread, while virions entering by endocytosis in general do not lead to productive infection. The nef gene product, acting either directly or indirectly, significantly enhances cytoplasmic virion entry occurring by fusion.

In terms of structure-function relationships for Nef-mediated enhancement of HIV cytosolic entry, we find that various Nef mutants exhibiting compromised downregulation of CD4 (LL164, WL57, and DD174) (4, 9, 19) and/or diminished binding to SH3 domains (P69, P72, and P75) (39, 48, 49, 59) all display reduced activity compared to wt Nef. In contrast, alanine substitutions at R77 or E64-68, which inhibit binding to the Hck tyrosine kinase and MHC class I receptor downregulation (10), respectively, do not impair cytosolic entry. These data reveal the surprising finding that, despite the absence of CD4 in the producer cells, Nef mutants lacking the ability to downregulate CD4 continue to influence virion entry into the cytosol. It is of course possible that these mutations commonly impair an unrecognized overlapping domain in Nef required for enhancement of virion entry.

How does Nef produce these effects on viral entry into the cytoplasmic compartment of target cells? Nef expression in the producer cell is sufficient to enhance virion infectivity in the subsequent target cell. While a direct effect mediated through the inclusion of small quantities of Nef within the virion cannot be discounted, intravirion Nef is efficiently cleaved by the viral protease. Thus far, no biological functions have been attributed to the resultant Nef fragments. It is also possible that Nef acts indirectly in other ways. For example, the presence of Nef in the producer cells may influence the subsequent function of the viral envelope protein. While previous studies indicate that Nef does not alter the efficiency of gp120 incorporation into virions that do not express CD4 (37), an indirect effect of Nef on the viral envelope could occur as a consequence of modifications of the viral Gag matrix protein (MA). Specifically, HIV-1 Nef enhances serine phosphorylation of MA involving an action of a Nef-associated kinase, presumably a member of the PAK or mitogen activated protein kinase (MAPK) families (54). ERK/MAPK has also been implicated in the phosphorylation of MA (24). Phosphorylated forms of MA may in turn alter or influence the function of the HIV envelope protein. The proline motif of Nef binds tyrosine kinases of the Src family, including Lck, Hck, Vav, and serine-threonine kinases such as the Nef-associated kinase and PAK2 (32; for a review, see reference 46). Interestingly, binding to PAK correlates with the enhancement of viral infectivity (39, 45, 49, 59) and also mediates phosphorylation of the HIV-1 MA protein (54). In this regard, MA plays an essential role in the targeting of the Gag polyprotein precursor to the plasma membrane and in the incorporation of envelope glycoproteins into budding virions. Moreover, the MA protein interacts with the cytoplasmic tail of the HIV-1 envelope protein gp41 in infected T cells (38). Finally, the importance of MA for the production of fully infectious viral particles is well established (13). Such a cascade of interactions involving Nef-facilitated phosphorylation of MA with secondary effects on HIV envelope function manifested by more efficient fusion clearly merits further investigation. However, such an effect cannot be restricted solely to the HIV envelope glycoprotein, since Nef is able to mediate enhanced infectivity of HIV-1 virions pseudotyped with the amphotropic murine leukemia virus envelope as well. Alternatively, through its binding to a thioesterase enzyme (29), Nef may indirectly control lipid modification and thus affect the rigidity or fluidity of the viral membrane, which is reflected by enhanced fusion-mediated entry. Additional study is required to test these and other possibilities.

The function of Nef as a cytoplasmic entry factor is also consistent with the recent finding that pseudotyping HIV-1 with the envelope protein from the VSVs leads to a loss of the effects of Nef on infectivity (1, 31). In view of our current results, we suggest that targeting HIV-1 for entry via fusion within the acidic endosome with VSV-G likely bypasses the function of Nef occurring on entry at the plasma membrane. The Nef phenotype is thus lost in such infections.

Our findings raise the possibility that the previously described effects of Nef on proviral DNA synthesis likely result from an earlier action at the level of viral entry. However, additional effects of Nef on proviral DNA synthesis cannot be completely excluded, yet the recognition that HIV-1 Nef functions as an entry factor provides new insights into the molecular basis of its function as an accelerator of HIV-1 pathogenesis.